MyArxiv
Robotics 52
☆ EfficientFlow: Efficient Equivariant Flow Policy Learning for Embodied AI AAAI 2026
Generative modeling has recently shown remarkable promise for visuomotor policy learning, enabling flexible and expressive control across diverse embodied AI tasks. However, existing generative policies often struggle with data inefficiency, requiring large-scale demonstrations, and sampling inefficiency, incurring slow action generation during inference. We introduce EfficientFlow, a unified framework for efficient embodied AI with flow-based policy learning. To enhance data efficiency, we bring equivariance into flow matching. We theoretically prove that when using an isotropic Gaussian prior and an equivariant velocity prediction network, the resulting action distribution remains equivariant, leading to improved generalization and substantially reduced data demands. To accelerate sampling, we propose a novel acceleration regularization strategy. As direct computation of acceleration is intractable for marginal flow trajectories, we derive a novel surrogate loss that enables stable and scalable training using only conditional trajectories. Across a wide range of robotic manipulation benchmarks, the proposed algorithm achieves competitive or superior performance under limited data while offering dramatically faster inference. These results highlight EfficientFlow as a powerful and efficient paradigm for high-performance embodied AI.
comment: Accepted by AAAI 2026. Project Page: https://efficientflow.github.io/
☆ Data-Centric Visual Development for Self-Driving Labs
Self-driving laboratories offer a promising path toward reducing the labor-intensive, time-consuming, and often irreproducible workflows in the biological sciences. Yet their stringent precision requirements demand highly robust models whose training relies on large amounts of annotated data. However, this kind of data is difficult to obtain in routine practice, especially negative samples. In this work, we focus on pipetting, the most critical and precision sensitive action in SDLs. To overcome the scarcity of training data, we build a hybrid pipeline that fuses real and virtual data generation. The real track adopts a human-in-the-loop scheme that couples automated acquisition with selective human verification to maximize accuracy with minimal effort. The virtual track augments the real data using reference-conditioned, prompt-guided image generation, which is further screened and validated for reliability. Together, these two tracks yield a class-balanced dataset that enables robust bubble detection training. On a held-out real test set, a model trained entirely on automatically acquired real images reaches 99.6% accuracy, and mixing real and generated data during training sustains 99.4% accuracy while reducing collection and review load. Our approach offers a scalable and cost-effective strategy for supplying visual feedback data to SDL workflows and provides a practical solution to data scarcity in rare event detection and broader vision tasks.
comment: 11 pages, 4 figures
☆ Visual Sync: Multi-Camera Synchronization via Cross-View Object Motion NeurIPS 2025
Today, people can easily record memorable moments, ranging from concerts, sports events, lectures, family gatherings, and birthday parties with multiple consumer cameras. However, synchronizing these cross-camera streams remains challenging. Existing methods assume controlled settings, specific targets, manual correction, or costly hardware. We present VisualSync, an optimization framework based on multi-view dynamics that aligns unposed, unsynchronized videos at millisecond accuracy. Our key insight is that any moving 3D point, when co-visible in two cameras, obeys epipolar constraints once properly synchronized. To exploit this, VisualSync leverages off-the-shelf 3D reconstruction, feature matching, and dense tracking to extract tracklets, relative poses, and cross-view correspondences. It then jointly minimizes the epipolar error to estimate each camera's time offset. Experiments on four diverse, challenging datasets show that VisualSync outperforms baseline methods, achieving an median synchronization error below 50 ms.
comment: Accepted to NeurIPS 2025. Project page: https://stevenlsw.github.io/visualsync/
☆ ManualVLA: A Unified VLA Model for Chain-of-Thought Manual Generation and Robotic Manipulation
Vision-Language-Action (VLA) models have recently emerged, demonstrating strong generalization in robotic scene understanding and manipulation. However, when confronted with long-horizon tasks that require defined goal states, such as LEGO assembly or object rearrangement, existing VLA models still face challenges in coordinating high-level planning with precise manipulation. Therefore, we aim to endow a VLA model with the capability to infer the "how" process from the "what" outcomes, transforming goal states into executable procedures. In this paper, we introduce ManualVLA, a unified VLA framework built upon a Mixture-of-Transformers (MoT) architecture, enabling coherent collaboration between multimodal manual generation and action execution. Unlike prior VLA models that directly map sensory inputs to actions, we first equip ManualVLA with a planning expert that generates intermediate manuals consisting of images, position prompts, and textual instructions. Building upon these multimodal manuals, we design a Manual Chain-of-Thought (ManualCoT) reasoning process that feeds them into the action expert, where each manual step provides explicit control conditions, while its latent representation offers implicit guidance for accurate manipulation. To alleviate the burden of data collection, we develop a high-fidelity digital-twin toolkit based on 3D Gaussian Splatting, which automatically generates manual data for planning expert training. ManualVLA demonstrates strong real-world performance, achieving an average success rate 32% higher than the previous hierarchical SOTA baseline on LEGO assembly and object rearrangement tasks.
☆ Learning Dexterous Manipulation Skills from Imperfect Simulations
Reinforcement learning and sim-to-real transfer have made significant progress in dexterous manipulation. However, progress remains limited by the difficulty of simulating complex contact dynamics and multisensory signals, especially tactile feedback. In this work, we propose \ours, a sim-to-real framework that addresses these limitations and demonstrates its effectiveness on nut-bolt fastening and screwdriving with multi-fingered hands. The framework has three stages. First, we train reinforcement learning policies in simulation using simplified object models that lead to the emergence of correct finger gaits. We then use the learned policy as a skill primitive within a teleoperation system to collect real-world demonstrations that contain tactile and proprioceptive information. Finally, we train a behavior cloning policy that incorporates tactile sensing and show that it generalizes to nuts and screwdrivers with diverse geometries. Experiments across both tasks show high task progress ratios compared to direct sim-to-real transfer and robust performance even on unseen object shapes and under external perturbations. Videos and code are available on https://dexscrew.github.io.
☆ LLM-Driven Corrective Robot Operation Code Generation with Static Text-Based Simulation
Recent advances in Large language models (LLMs) have demonstrated their promising capabilities of generating robot operation code to enable LLM-driven robots. To enhance the reliability of operation code generated by LLMs, corrective designs with feedback from the observation of executing code have been increasingly adopted in existing research. However, the code execution in these designs relies on either a physical experiment or a customized simulation environment, which limits their deployment due to the high configuration effort of the environment and the potential long execution time. In this paper, we explore the possibility of directly leveraging LLM to enable static simulation of robot operation code, and then leverage it to design a new reliable LLM-driven corrective robot operation code generation framework. Our framework configures the LLM as a static simulator with enhanced capabilities that reliably simulate robot code execution by interpreting actions, reasoning over state transitions, analyzing execution outcomes, and generating se- mantic observations that accurately capture trajectory dynamics. To validate the performance of our framework, we performed experiments on various operation tasks for different robots, including UAVs and small ground vehicles. The experiment results not only demonstrated the high accuracy of our static text-based simulation but also the reliable code generation of our LLM-driven corrective framework, which achieves a comparable performance with state-of-the-art research while does not rely on dynamic code execution using physical experiments or simulators.
comment: 8 pages, 2 figures
☆ Learning Sim-to-Real Humanoid Locomotion in 15 Minutes
Massively parallel simulation has reduced reinforcement learning (RL) training time for robots from days to minutes. However, achieving fast and reliable sim-to-real RL for humanoid control remains difficult due to the challenges introduced by factors such as high dimensionality and domain randomization. In this work, we introduce a simple and practical recipe based on off-policy RL algorithms, i.e., FastSAC and FastTD3, that enables rapid training of humanoid locomotion policies in just 15 minutes with a single RTX 4090 GPU. Our simple recipe stabilizes off-policy RL algorithms at massive scale with thousands of parallel environments through carefully tuned design choices and minimalist reward functions. We demonstrate rapid end-to-end learning of humanoid locomotion controllers on Unitree G1 and Booster T1 robots under strong domain randomization, e.g., randomized dynamics, rough terrain, and push perturbations, as well as fast training of whole-body human-motion tracking policies. We provide videos and open-source implementation at: https://younggyo.me/fastsac-humanoid.
comment: Project website: https://younggyo.me/fastsac-humanoid
☆ RoaD: Rollouts as Demonstrations for Closed-Loop Supervised Fine-Tuning of Autonomous Driving Policies
Autonomous driving policies are typically trained via open-loop behavior cloning of human demonstrations. However, such policies suffer from covariate shift when deployed in closed loop, leading to compounding errors. We introduce Rollouts as Demonstrations (RoaD), a simple and efficient method to mitigate covariate shift by leveraging the policy's own closed-loop rollouts as additional training data. During rollout generation, RoaD incorporates expert guidance to bias trajectories toward high-quality behavior, producing informative yet realistic demonstrations for fine-tuning. This approach enables robust closed-loop adaptation with orders of magnitude less data than reinforcement learning, and avoids restrictive assumptions of prior closed-loop supervised fine-tuning (CL-SFT) methods, allowing broader applications domains including end-to-end driving. We demonstrate the effectiveness of RoaD on WOSAC, a large-scale traffic simulation benchmark, where it performs similar or better than the prior CL-SFT method; and in AlpaSim, a high-fidelity neural reconstruction-based simulator for end-to-end driving, where it improves driving score by 41\% and reduces collisions by 54\%.
comment: Preprint
☆ Forecasting in Offline Reinforcement Learning for Non-stationary Environments NeurIPS 2025
Offline Reinforcement Learning (RL) provides a promising avenue for training policies from pre-collected datasets when gathering additional interaction data is infeasible. However, existing offline RL methods often assume stationarity or only consider synthetic perturbations at test time, assumptions that often fail in real-world scenarios characterized by abrupt, time-varying offsets. These offsets can lead to partial observability, causing agents to misperceive their true state and degrade performance. To overcome this challenge, we introduce Forecasting in Non-stationary Offline RL (FORL), a framework that unifies (i) conditional diffusion-based candidate state generation, trained without presupposing any specific pattern of future non-stationarity, and (ii) zero-shot time-series foundation models. FORL targets environments prone to unexpected, potentially non-Markovian offsets, requiring robust agent performance from the onset of each episode. Empirical evaluations on offline RL benchmarks, augmented with real-world time-series data to simulate realistic non-stationarity, demonstrate that FORL consistently improves performance compared to competitive baselines. By integrating zero-shot forecasting with the agent's experience, we aim to bridge the gap between offline RL and the complexities of real-world, non-stationary environments.
comment: The Thirty-Ninth Annual Conference on Neural Information Processing Systems, NeurIPS 2025
☆ GrndCtrl: Grounding World Models via Self-Supervised Reward Alignment
Recent advances in video world modeling have enabled large-scale generative models to simulate embodied environments with high visual fidelity, providing strong priors for prediction, planning, and control. Yet, despite their realism, these models often lack geometric grounding, limiting their use in navigation tasks that require spatial coherence and long-horizon stability. We introduce Reinforcement Learning with World Grounding (RLWG), a self-supervised post-training framework that aligns pretrained world models with a physically verifiable structure through geometric and perceptual rewards. Analogous to reinforcement learning from verifiable feedback (RLVR) in language models, RLWG can use multiple rewards that measure pose cycle-consistency, depth reprojection, and temporal coherence. We instantiate this framework with GrndCtrl, a reward-aligned adaptation method based on Group Relative Policy Optimization (GRPO), yielding world models that maintain stable trajectories, consistent geometry, and reliable rollouts for embodied navigation. Like post-training alignment in large language models, GrndCtrl leverages verifiable rewards to bridge generative pretraining and grounded behavior, achieving superior spatial coherence and navigation stability over supervised fine-tuning in outdoor environments.
☆ Guardian: Detecting Robotic Planning and Execution Errors with Vision-Language Models
Robust robotic manipulation requires reliable failure detection and recovery. Although current Vision-Language Models (VLMs) show promise, their accuracy and generalization are limited by the scarcity of failure data. To address this data gap, we propose an automatic robot failure synthesis approach that procedurally perturbs successful trajectories to generate diverse planning and execution failures. This method produces not only binary classification labels but also fine-grained failure categories and step-by-step reasoning traces in both simulation and the real world. With it, we construct three new failure detection benchmarks: RLBench-Fail, BridgeDataV2-Fail, and UR5-Fail, substantially expanding the diversity and scale of existing failure datasets. We then train Guardian, a VLM with multi-view images for detailed failure reasoning and detection. Guardian achieves state-of-the-art performance on both existing and newly introduced benchmarks. It also effectively improves task success rates when integrated into a state-of-the-art manipulation system in simulation and real robots, demonstrating the impact of our generated failure data.
comment: 9 pages, 9 figures, 6 tables
☆ Real-World Robot Control by Deep Active Inference With a Temporally Hierarchical World Model
Robots in uncertain real-world environments must perform both goal-directed and exploratory actions. However, most deep learning-based control methods neglect exploration and struggle under uncertainty. To address this, we adopt deep active inference, a framework that accounts for human goal-directed and exploratory actions. Yet, conventional deep active inference approaches face challenges due to limited environmental representation capacity and high computational cost in action selection. We propose a novel deep active inference framework that consists of a world model, an action model, and an abstract world model. The world model encodes environmental dynamics into hidden state representations at slow and fast timescales. The action model compresses action sequences into abstract actions using vector quantization, and the abstract world model predicts future slow states conditioned on the abstract action, enabling low-cost action selection. We evaluate the framework on object-manipulation tasks with a real-world robot. Results show that it achieves high success rates across diverse manipulation tasks and switches between goal-directed and exploratory actions in uncertain settings, while making action selection computationally tractable. These findings highlight the importance of modeling multiple timescale dynamics and abstracting actions and state transitions.
comment: Accepted for publication in IEEE Robotics and Automation Letters (RA-L)
☆ NeuroHJR: Hamilton-Jacobi Reachability-based Obstacle Avoidance in Complex Environments with Physics-Informed Neural Networks
Autonomous ground vehicles (AGVs) must navigate safely in cluttered environments while accounting for complex dynamics and environmental uncertainty. Hamilton-Jacobi Reachability (HJR) offers formal safety guarantees through the computation of forward and backward reachable sets, but its application is hindered by poor scalability in environments with numerous obstacles. In this paper, we present a novel framework called NeuroHJR that leverages Physics-Informed Neural Networks (PINNs) to approximate the HJR solution for real-time obstacle avoidance. By embedding system dynamics and safety constraints directly into the neural network loss function, our method bypasses the need for grid-based discretization and enables efficient estimation of reachable sets in continuous state spaces. We demonstrate the effectiveness of our approach through simulation results in densely cluttered scenarios, showing that it achieves safety performance comparable to that of classical HJR solvers while significantly reducing the computational cost. This work provides a new step toward real-time, scalable deployment of reachability-based obstacle avoidance in robotics.
comment: Author-accepted version. Accepted at IEEE 11th Indian Control Conference (ICC), 2025
☆ Is Image-based Object Pose Estimation Ready to Support Grasping?
We present a framework for evaluating 6-DoF instance-level object pose estimators, focusing on those that require a single RGB (not RGB-D) image as input. Besides gaining intuition about how accurate these estimators are, we are interested in the degree to which they can serve as the sole perception mechanism for robotic grasping. To assess this, we perform grasping trials in a physics-based simulator, using image-based pose estimates to guide a parallel gripper and an underactuated robotic hand in picking up 3D models of objects. Our experiments on a subset of the BOP (Benchmark for 6D Object Pose Estimation) dataset compare five open-source object pose estimators and provide insights that were missing from the literature.
☆ Register Any Point: Scaling 3D Point Cloud Registration by Flow Matching
Point cloud registration aligns multiple unposed point clouds into a common frame, and is a core step for 3D reconstruction and robot localization. In this work, we cast registration as conditional generation: a learned continuous, point-wise velocity field transports noisy points to a registered scene, from which the pose of each view is recovered. Unlike previous methods that conduct correspondence matching to estimate the transformation between a pair of point clouds and then optimize the pairwise transformations to realize multi-view registration, our model directly generates the registered point cloud. With a lightweight local feature extractor and test-time rigidity enforcement, our approach achieves state-of-the-art results on pairwise and multi-view registration benchmarks, particularly with low overlap, and generalizes across scales and sensor modalities. It further supports downstream tasks including relocalization, multi-robot SLAM, and multi-session map merging. Source code available at: https://github.com/PRBonn/RAP.
comment: 22 pages
☆ Much Ado About Noising: Dispelling the Myths of Generative Robotic Control
Generative models, like flows and diffusions, have recently emerged as popular and efficacious policy parameterizations in robotics. There has been much speculation as to the factors underlying their successes, ranging from capturing multi-modal action distribution to expressing more complex behaviors. In this work, we perform a comprehensive evaluation of popular generative control policies (GCPs) on common behavior cloning (BC) benchmarks. We find that GCPs do not owe their success to their ability to capture multi-modality or to express more complex observation-to-action mappings. Instead, we find that their advantage stems from iterative computation, as long as intermediate steps are supervised during training and this supervision is paired with a suitable level of stochasticity. As a validation of our findings, we show that a minimum iterative policy (MIP), a lightweight two-step regression-based policy, essentially matches the performance of flow GCPs, and often outperforms distilled shortcut models. Our results suggest that the distribution-fitting component of GCPs is less salient than commonly believed, and point toward new design spaces focusing solely on control performance. Project page: https://simchowitzlabpublic.github.io/much-ado-about-noising-project/
☆ GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation
We present GR-RL, a robotic learning framework that turns a generalist vision-language-action (VLA) policy into a highly capable specialist for long-horizon dexterous manipulation. Assuming the optimality of human demonstrations is core to existing VLA policies. However, we claim that in highly dexterous and precise manipulation tasks, human demonstrations are noisy and suboptimal. GR-RL proposes a multi-stage training pipeline that filters, augments, and reinforces the demonstrations by reinforcement learning. First, GR-RL learns a vision-language-conditioned task progress, filters the demonstration trajectories, and only keeps the transitions that contribute positively to the progress. Specifically, we show that by directly applying offline RL with sparse reward, the resulting $Q$-values can be treated as a robust progress function. Next, we introduce morphological symmetry augmentation that greatly improves the generalization and performance of GR-RL. Lastly, to better align the VLA policy with its deployment behaviors for high-precision control, we perform online RL by learning a latent space noise predictor. With this pipeline, GR-RL is, to our knowledge, the first learning-based policy that can autonomously lace up a shoe by threading shoelaces through multiple eyelets with an 83.3% success rate, a task requiring long-horizon reasoning, millimeter-level precision, and compliant soft-body interaction. We hope GR-RL provides a step toward enabling generalist robot foundations models to specialize into reliable real-world experts.
☆ IGen: Scalable Data Generation for Robot Learning from Open-World Images
The rise of generalist robotic policies has created an exponential demand for large-scale training data. However, on-robot data collection is labor-intensive and often limited to specific environments. In contrast, open-world images capture a vast diversity of real-world scenes that naturally align with robotic manipulation tasks, offering a promising avenue for low-cost, large-scale robot data acquisition. Despite this potential, the lack of associated robot actions hinders the practical use of open-world images for robot learning, leaving this rich visual resource largely unexploited. To bridge this gap, we propose IGen, a framework that scalably generates realistic visual observations and executable actions from open-world images. IGen first converts unstructured 2D pixels into structured 3D scene representations suitable for scene understanding and manipulation. It then leverages the reasoning capabilities of vision-language models to transform scene-specific task instructions into high-level plans and generate low-level actions as SE(3) end-effector pose sequences. From these poses, it synthesizes dynamic scene evolution and renders temporally coherent visual observations. Experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data. This highlights the potential of IGen to support scalable data generation from open-world images for generalist robotic policy training.
comment: 8 pages, 8 figures
☆ AgriLiRa4D: A Multi-Sensor UAV Dataset for Robust SLAM in Challenging Agricultural Fields
Multi-sensor Simultaneous Localization and Mapping (SLAM) is essential for Unmanned Aerial Vehicles (UAVs) performing agricultural tasks such as spraying, surveying, and inspection. However, real-world, multi-modal agricultural UAV datasets that enable research on robust operation remain scarce. To address this gap, we present AgriLiRa4D, a multi-modal UAV dataset designed for challenging outdoor agricultural environments. AgriLiRa4D spans three representative farmland types-flat, hilly, and terraced-and includes both boundary and coverage operation modes, resulting in six flight sequence groups. The dataset provides high-accuracy ground-truth trajectories from a Fiber Optic Inertial Navigation System with Real-Time Kinematic capability (FINS_RTK), along with synchronized measurements from a 3D LiDAR, a 4D Radar, and an Inertial Measurement Unit (IMU), accompanied by complete intrinsic and extrinsic calibrations. Leveraging its comprehensive sensor suite and diverse real-world scenarios, AgriLiRa4D supports diverse SLAM and localization studies and enables rigorous robustness evaluation against low-texture crops, repetitive patterns, dynamic vegetation, and other challenges of real agricultural environments. To further demonstrate its utility, we benchmark four state-of-the-art multi-sensor SLAM algorithms across different sensor combinations, highlighting the difficulty of the proposed sequences and the necessity of multi-modal approaches for reliable UAV localization. By filling a critical gap in agricultural SLAM datasets, AgriLiRa4D provides a valuable benchmark for the research community and contributes to advancing autonomous navigation technologies for agricultural UAVs. The dataset can be downloaded from: https://zhan994.github.io/AgriLiRa4D.
☆ DiG-Flow: Discrepancy-Guided Flow Matching for Robust VLA Models
Vision-Language-Action (VLA) models trained with flow matching have demonstrated impressive capabilities on robotic manipulation tasks. However, their performance often degrades under distribution shift and on complex multi-step tasks, suggesting that the learned representations may not robustly capture task-relevant semantics. We introduce DiG-Flow, a principled framework that enhances VLA robustness through geometric regularization. Our key insight is that the distributional discrepancy between observation and action embeddings provides a meaningful geometric signal: lower transport cost indicates compatible representations, while higher cost suggests potential misalignment. DiG-Flow computes a discrepancy measure between empirical distributions of observation and action embeddings, maps it to a modulation weight via a monotone function, and applies residual updates to the observation embeddings before flow matching. Crucially, this intervention operates at the representation level without modifying the flow matching path or target vector field. We provide theoretical guarantees showing that discrepancy-guided training provably decreases the training objective, and that guided inference refinement converges with contraction. Empirically, DiG-Flow integrates into existing VLA architectures with negligible overhead and consistently improves performance, with particularly pronounced gains on complex multi-step tasks and under limited training data.
☆ Dynamic Log-Gaussian Process Control Barrier Function for Safe Robotic Navigation in Dynamic Environments
Control Barrier Functions (CBFs) have emerged as efficient tools to address the safe navigation problem for robot applications. However, synthesizing informative and obstacle motion-aware CBFs online using real-time sensor data remains challenging, particularly in unknown and dynamic scenarios. Motived by this challenge, this paper aims to propose a novel Gaussian Process-based formulation of CBF, termed the Dynamic Log Gaussian Process Control Barrier Function (DLGP-CBF), to enable real-time construction of CBF which are both spatially informative and responsive to obstacle motion. Firstly, the DLGP-CBF leverages a logarithmic transformation of GP regression to generate smooth and informative barrier values and gradients, even in sparse-data regions. Secondly, by explicitly modeling the DLGP-CBF as a function of obstacle positions, the derived safety constraint integrates predicted obstacle velocities, allowing the controller to proactively respond to dynamic obstacles' motion. Simulation results demonstrate significant improvements in obstacle avoidance performance, including increased safety margins, smoother trajectories, and enhanced responsiveness compared to baseline methods.
comment: To be presented in the 64th IEEE Conference on Decision and Control (CDC 2025)
☆ SPARK: Sim-ready Part-level Articulated Reconstruction with VLM Knowledge
Articulated 3D objects are critical for embodied AI, robotics, and interactive scene understanding, yet creating simulation-ready assets remains labor-intensive and requires expert modeling of part hierarchies and motion structures. We introduce SPARK, a framework for reconstructing physically consistent, kinematic part-level articulated objects from a single RGB image. Given an input image, we first leverage VLMs to extract coarse URDF parameters and generate part-level reference images. We then integrate the part-image guidance and the inferred structure graph into a generative diffusion transformer to synthesize consistent part and complete shapes of articulated objects. To further refine the URDF parameters, we incorporate differentiable forward kinematics and differentiable rendering to optimize joint types, axes, and origins under VLM-generated open-state supervision. Extensive experiments show that SPARK produces high-quality, simulation-ready articulated assets across diverse categories, enabling downstream applications such as robotic manipulation and interaction modeling.
☆ Integrated YOLOP Perception and Lyapunov-based Control for Autonomous Mobile Robot Navigation on Track
This work presents a real-time autonomous track navigation framework for nonholonomic differential-drive mobile robots by jointly integrating multi-task visual perception and a provably stable tracking controller. The perception pipeline reconstructs lane centerlines using 2D-to-3D camera projection, arc-length based uniform point resampling, and cubic polynomial fitting solved via robust QR least-squares optimization. The controller regulates robot linear and angular velocities through a Lyapunov-stability grounded design, ensuring bounded error dynamics and asymptotic convergence of position and heading deviations even in dynamic and partially perceived lane scenarios, without relying on HD prior maps or global satellite localization. Real-world experiments on embedded platforms verify system fidelity, real-time execution, trajectory smoothness, and closed-loop stability for reliable autonomous navigation.
comment: This is a master's graduation thesis that has not been formally published. Uploaded with the author's copyright permission. No confidential content involved
☆ A Cross-Embodiment Gripper Benchmark for Rigid-Object Manipulation in Aerial and Industrial Robotics
Robotic grippers are increasingly deployed across industrial, collaborative, and aerial platforms, where each embodiment imposes distinct mechanical, energetic, and operational constraints. Established YCB and NIST benchmarks quantify grasp success, force, or timing on a single platform, but do not evaluate cross-embodiment transferability or energy-aware performance, capabilities essential for modern mobile and aerial manipulation. This letter introduces the Cross-Embodiment Gripper Benchmark (CEGB), a compact and reproducible benchmarking suite extending YCB and selected NIST metrics with three additional components: a transfer-time benchmark measuring the practical effort required to exchange embodiments, an energy-consumption benchmark evaluating grasping and holding efficiency, and an intent-specific ideal payload assessment reflecting design-dependent operational capability. Together, these metrics characterize both grasp performance and the suitability of reusing a single gripper across heterogeneous robotic systems. A lightweight self-locking gripper prototype is implemented as a reference case. Experiments demonstrate rapid embodiment transfer (median ~= 17.6 s across user groups), low holding energy for gripper prototype (~= 1.5 J per 10 s), and consistent grasp performance with cycle times of 3.2 - 3.9 s and success rates exceeding 90%. CEGB thus provides a reproducible foundation for cross-platform, energy-aware evaluation of grippers in aerial and manipulators domains.
☆ L2M-Calib: One-key Calibration Method for LiDAR and Multiple Magnetic Sensors
Multimodal sensor fusion enables robust environmental perception by leveraging complementary information from heterogeneous sensing modalities. However, accurate calibration is a critical prerequisite for effective fusion. This paper proposes a novel one-key calibration framework named L2M-Calib for a fused magnetic-LiDAR system, jointly estimating the extrinsic transformation between the two kinds of sensors and the intrinsic distortion parameters of the magnetic sensors. Magnetic sensors capture ambient magnetic field (AMF) patterns, which are invariant to geometry, texture, illumination, and weather, making them suitable for challenging environments. Nonetheless, the integration of magnetic sensing into multimodal systems remains underexplored due to the absence of effective calibration techniques. To address this, we optimize extrinsic parameters using an iterative Gauss-Newton scheme, coupled with the intrinsic calibration as a weighted ridge-regularized total least squares (w-RRTLS) problem, ensuring robustness against measurement noise and ill-conditioned data. Extensive evaluations on both simulated datasets and real-world experiments, including AGV-mounted sensor configurations, demonstrate that our method achieves high calibration accuracy and robustness under various environmental and operational conditions.
☆ NavForesee: A Unified Vision-Language World Model for Hierarchical Planning and Dual-Horizon Navigation Prediction
Embodied navigation for long-horizon tasks, guided by complex natural language instructions, remains a formidable challenge in artificial intelligence. Existing agents often struggle with robust long-term planning about unseen environments, leading to high failure rates. To address these limitations, we introduce NavForesee, a novel Vision-Language Model (VLM) that unifies high-level language planning and predictive world model imagination within a single, unified framework. Our approach empowers a single VLM to concurrently perform planning and predictive foresight. Conditioned on the full instruction and historical observations, the model is trained to understand the navigation instructions by decomposing the task, tracking its progress, and formulating the subsequent sub-goal. Simultaneously, it functions as a generative world model, providing crucial foresight by predicting short-term environmental dynamics and long-term navigation milestones. The VLM's structured plan guides its targeted prediction, while the imagined future provides rich context to inform the navigation actions, creating a powerful internal feedback loop of perception-planning/prediction-action. We demonstrate through extensive experiments on the R2R-CE and RxR-CE benchmark that NavForesee achieves highly competitive performance in complex scenarios. Our work highlights the immense potential of fusing explicit language planning with implicit spatiotemporal prediction, paving the way for more intelligent and capable embodied agents.
☆ $\mathbf{M^3A}$ Policy: Mutable Material Manipulation Augmentation Policy through Photometric Re-rendering
Material generalization is essential for real-world robotic manipulation, where robots must interact with objects exhibiting diverse visual and physical properties. This challenge is particularly pronounced for objects made of glass, metal, or other materials whose transparent or reflective surfaces introduce severe out-of-distribution variations. Existing approaches either rely on simulated materials in simulators and perform sim-to-real transfer, which is hindered by substantial visual domain gaps, or depend on collecting extensive real-world demonstrations, which is costly, time-consuming, and still insufficient to cover various materials. To overcome these limitations, we resort to computational photography and introduce Mutable Material Manipulation Augmentation (M$^3$A), a unified framework that leverages the physical characteristics of materials as captured by light transport for photometric re-rendering. The core idea is simple yet powerful: given a single real-world demonstration, we photometrically re-render the scene to generate a diverse set of highly realistic demonstrations with different material properties. This augmentation effectively decouples task-specific manipulation skills from surface appearance, enabling policies to generalize across materials without additional data collection. To systematically evaluate this capability, we construct the first comprehensive multi-material manipulation benchmark spanning both simulation and real-world environments. Extensive experiments show that the M$^3$A policy significantly enhances cross-material generalization, improving the average success rate across three real-world tasks by 58.03\%, and demonstrating robust performance on previously unseen materials.
comment: under submission
☆ Accelerating Probabilistic Response-Time Analysis: Revised Critical Instant and Optimized Convolution
Accurate estimation of the Worst-Case Deadline Failure Probability (WCDFP) has attracted growing attention as a means to provide safety assurances in complex systems such as robotic platforms and autonomous vehicles. WCDFP quantifies the likelihood of deadline misses under the most pessimistic operating conditions, and safe estimation is essential for dependable real-time applications. However, achieving high accuracy in WCDFP estimation often incurs significant computational cost. Recent studies have revealed that the classical assumption of the critical instant, the activation pattern traditionally considered to trigger the worst-case behavior, can lead to underestimation of WCDFP in probabilistic settings. This observation motivates the use of a revised critical instant formulation that more faithfully captures the true worst-case scenario. This paper investigates convolution-based methods for WCDFP estimation under this revised setting and proposes an optimization technique that accelerates convolution by improving the merge order. Extensive experiments with diverse execution-time distributions demonstrate that the proposed optimized Aggregate Convolution reduces computation time by up to an order of magnitude compared to Sequential Convolution, while retaining accurate and safe-sided WCDFP estimates. These results highlight the potential of the approach to provide both efficiency and reliability in probabilistic timing analysis for safety-critical real-time applications.
comment: 8 pages, 5 figures. Proceedings of APRIS2025
☆ Modality-Augmented Fine-Tuning of Foundation Robot Policies for Cross-Embodiment Manipulation on GR1 and G1
This paper presents a modality-augmented fine-tuning framework designed to adapt foundation robot policies to diverse humanoid embodiments. We validate our approach across two distinct settings: (i) the GR1 embodiment, utilizing public datasets where we introduce post-processed modalities, including binary contact signals and ZoeDepth-generated metric depth; and (ii) the Unitree G1 embodiment, for which we contribute a novel multi-modal dataset incorporating cuRobo motion planning, inverse kinematics, and ground-truth contact-force measurements. Our experiments demonstrate that modality augmentation consistently enhances policy performance across different embodiments. Specifically, for the GR1, integrating contact-state cues and RGB-D fusion improves online success rates from 51% to 63%. Furthermore, in the G1 "Pick Apple to Bowl" task, our contact-augmented model achieves a success rate of 94%, significantly outperforming the 48% achieved by standard fine-tuning and the 0% baseline of zero-shot transfer. These results highlight that lightweight post-processing effectively strengthens policies for GR1, while high-quality multi-modal data is crucial for reliable transfer to the Unitree G1. Consequently, this work establishes a unified, data-centric pathway for extending foundation robot policies through targeted modality design and multi-modal fine-tuning.
comment: 8 pages, 10 figures
☆ Discovering Self-Protective Falling Policy for Humanoid Robot via Deep Reinforcement Learning
Humanoid robots have received significant research interests and advancements in recent years. Despite many successes, due to their morphology, dynamics and limitation of control policy, humanoid robots are prone to fall as compared to other embodiments like quadruped or wheeled robots. And its large weight, tall Center of Mass, high Degree-of-Freedom would cause serious hardware damages when falling uncontrolled, to both itself and surrounding objects. Existing researches in this field mostly focus on using control based methods that struggle to cater diverse falling scenarios and may introduce unsuitable human prior. On the other hand, large-scale Deep Reinforcement Learning and Curriculum Learning could be employed to incentivize humanoid agent discovering falling protection policy that fits its own nature and property. In this work, with carefully designed reward functions and domain diversification curriculum, we successfully train humanoid agent to explore falling protection behaviors and discover that by forming a `triangle' structure, the falling damages could be significantly reduced with its rigid-material body. With comprehensive metrics and experiments, we quantify its performance with comparison to other methods, visualize its falling behaviors and successfully transfer it to real world platform.
☆ Visibility-aware Cooperative Aerial Tracking with Decentralized LiDAR-based Swarms
Autonomous aerial tracking with drones offers vast potential for surveillance, cinematography, and industrial inspection applications. While single-drone tracking systems have been extensively studied, swarm-based target tracking remains underexplored, despite its unique advantages of distributed perception, fault-tolerant redundancy, and multidirectional target coverage. To bridge this gap, we propose a novel decentralized LiDAR-based swarm tracking framework that enables visibility-aware, cooperative target tracking in complex environments, while fully harnessing the unique capabilities of swarm systems. To address visibility, we introduce a novel Spherical Signed Distance Field (SSDF)-based metric for 3-D environmental occlusion representation, coupled with an efficient algorithm that enables real-time onboard SSDF updating. A general Field-of-View (FOV) alignment cost supporting heterogeneous LiDAR configurations is proposed for consistent target observation. Swarm coordination is enhanced through cooperative costs that enforce inter-robot safe clearance, prevent mutual occlusions, and notably facilitate 3-D multidirectional target encirclement via a novel electrostatic-potential-inspired distribution metric. These innovations are integrated into a hierarchical planner, combining a kinodynamic front-end searcher with a spatiotemporal $SE(3)$ back-end optimizer to generate collision-free, visibility-optimized trajectories.Deployed on heterogeneous LiDAR swarms, our fully decentralized implementation features collaborative perception, distributed planning, and dynamic swarm reconfigurability. Validated through rigorous real-world experiments in cluttered outdoor environments, the proposed system demonstrates robust cooperative tracking of agile targets (drones, humans) while achieving superior visibility maintenance.
☆ COMET: A Dual Swashplate Autonomous Coaxial Bi-copter AAV with High-Maneuverability and Long-Endurance
Coaxial bi-copter autonomous aerial vehicles (AAVs) have garnered attention due to their potential for improved rotor system efficiency and compact form factor. However, balancing efficiency, maneuverability, and compactness in coaxial bi-copter systems remains a key design challenge, limiting their practical deployment. This letter introduces COMET, a coaxial bi-copter AAV platform featuring a dual swashplate mechanism. The coaxial bi-copter system's efficiency and compactness are optimized through bench tests, and the whole prototype's efficiency and robustness under varying payload conditions are verified through flight endurance experiments. The maneuverability performance of the system is evaluated in comprehensive trajectory tracking tests. The results indicate that the dual swashplate configuration enhances tracking performance and improves flight efficiency compared to the single swashplate alternative. Successful autonomous flight trials across various scenarios verify COMET's potential for real-world applications.
comment: 8 pages, 8 figures, accepted at IEEE RA-L
☆ How do trout regulate patterns of muscle contraction to optimize propulsive efficiency during steady swimming
Understanding efficient fish locomotion offers insights for biomechanics, fluid dynamics, and engineering. Traditional studies often miss the link between neuromuscular control and whole-body movement. To explore energy transfer in carangiform swimming, we created a bio-inspired digital trout. This model combined multibody dynamics, Hill-type muscle modeling, and a high-fidelity fluid-structure interaction algorithm, accurately replicating a real trout's form and properties. Using deep reinforcement learning, the trout's neural system achieved hierarchical spatiotemporal control of muscle activation. We systematically examined how activation strategies affect speed and energy use. Results show that axial myomere coupling-with activation spanning over 0.5 body lengths-is crucial for stable body wave propagation. Moderate muscle contraction duration ([0.1,0.3] of a tail-beat cycle) lets the body and fluid act as a passive damping system, cutting energy use. Additionally, the activation phase lag of myomeres shapes the body wave; if too large, it causes antagonistic contractions that hinder thrust. These findings advance bio-inspired locomotion understanding and aid energy-efficient underwater system design.
☆ RoboLoc: A Benchmark Dataset for Point Place Recognition and Localization in Indoor-Outdoor Integrated Environments
Robust place recognition is essential for reliable localization in robotics, particularly in complex environments with fre- quent indoor-outdoor transitions. However, existing LiDAR-based datasets often focus on outdoor scenarios and lack seamless domain shifts. In this paper, we propose RoboLoc, a benchmark dataset designed for GPS-free place recognition in indoor-outdoor environments with floor transitions. RoboLoc features real-world robot trajectories, diverse elevation profiles, and transitions between structured indoor and unstructured outdoor domains. We benchmark a variety of state-of-the-art models, point-based, voxel-based, and BEV-based architectures, highlighting their generalizability domain shifts. RoboLoc provides a realistic testbed for developing multi-domain localization systems in robotics and autonomous navigation
☆ Real-World Reinforcement Learning of Active Perception Behaviors NeurIPS 2025
A robot's instantaneous sensory observations do not always reveal task-relevant state information. Under such partial observability, optimal behavior typically involves explicitly acting to gain the missing information. Today's standard robot learning techniques struggle to produce such active perception behaviors. We propose a simple real-world robot learning recipe to efficiently train active perception policies. Our approach, asymmetric advantage weighted regression (AAWR), exploits access to "privileged" extra sensors at training time. The privileged sensors enable training high-quality privileged value functions that aid in estimating the advantage of the target policy. Bootstrapping from a small number of potentially suboptimal demonstrations and an easy-to-obtain coarse policy initialization, AAWR quickly acquires active perception behaviors and boosts task performance. In evaluations on 8 manipulation tasks on 3 robots spanning varying degrees of partial observability, AAWR synthesizes reliable active perception behaviors that outperform all prior approaches. When initialized with a "generalist" robot policy that struggles with active perception tasks, AAWR efficiently generates information-gathering behaviors that allow it to operate under severe partial observability for manipulation tasks. Website: https://penn-pal-lab.github.io/aawr/
comment: NeurIPS 2025 camera ready
☆ Real-Time On-the-Go Annotation Framework Using YOLO for Automated Dataset Generation CEC 2025
Efficient and accurate annotation of datasets remains a significant challenge for deploying object detection models such as You Only Look Once (YOLO) in real-world applications, particularly in agriculture where rapid decision-making is critical. Traditional annotation techniques are labor-intensive, requiring extensive manual labeling post data collection. This paper presents a novel real-time annotation approach leveraging YOLO models deployed on edge devices, enabling immediate labeling during image capture. To comprehensively evaluate the efficiency and accuracy of our proposed system, we conducted an extensive comparative analysis using three prominent YOLO architectures (YOLOv5, YOLOv8, YOLOv12) under various configurations: single-class versus multi-class annotation and pretrained versus scratch-based training. Our analysis includes detailed statistical tests and learning dynamics, demonstrating significant advantages of pretrained and single-class configurations in terms of model convergence, performance, and robustness. Results strongly validate the feasibility and effectiveness of our real-time annotation framework, highlighting its capability to drastically reduce dataset preparation time while maintaining high annotation quality.
comment: Copyright 2025 IEEE. This is the author's version of the work that has been accepted for publication in Proceedings of the 5. Interdisciplinary Conference on Electrics and Computer (INTCEC 2025) 15-16 September 2025, Chicago-USA. The final version of record is available at: https://doi.org/10.1109/INTCEC65580.2025.11256048
♻ ☆ Securing the Skies: A Comprehensive Survey on Anti-UAV Methods, Benchmarking, and Future Directions CVPR
Unmanned Aerial Vehicles (UAVs) are indispensable for infrastructure inspection, surveillance, and related tasks, yet they also introduce critical security challenges. This survey provides a wide-ranging examination of the anti-UAV domain, centering on three core objectives-classification, detection, and tracking-while detailing emerging methodologies such as diffusion-based data synthesis, multi-modal fusion, vision-language modeling, self-supervised learning, and reinforcement learning. We systematically evaluate state-of-the-art solutions across both single-modality and multi-sensor pipelines (spanning RGB, infrared, audio, radar, and RF) and discuss large-scale as well as adversarially oriented benchmarks. Our analysis reveals persistent gaps in real-time performance, stealth detection, and swarm-based scenarios, underscoring pressing needs for robust, adaptive anti-UAV systems. By highlighting open research directions, we aim to foster innovation and guide the development of next-generation defense strategies in an era marked by the extensive use of UAVs.
comment: Best Paper, Accepted at CVPR Workshop Anti-UAV 2025. 16 pages
♻ ☆ PRIMT: Preference-based Reinforcement Learning with Multimodal Feedback and Trajectory Synthesis from Foundation Models
Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for teaching robots complex behaviors without reward engineering. However, its effectiveness is often limited by two critical challenges: the reliance on extensive human input and the inherent difficulties in resolving query ambiguity and credit assignment during reward learning. In this paper, we introduce PRIMT, a PbRL framework designed to overcome these challenges by leveraging foundation models (FMs) for multimodal synthetic feedback and trajectory synthesis. Unlike prior approaches that rely on single-modality FM evaluations, PRIMT employs a hierarchical neuro-symbolic fusion strategy, integrating the complementary strengths of large language models and vision-language models in evaluating robot behaviors for more reliable and comprehensive feedback. PRIMT also incorporates foresight trajectory generation, which reduces early-stage query ambiguity by warm-starting the trajectory buffer with bootstrapped samples, and hindsight trajectory augmentation, which enables counterfactual reasoning with a causal auxiliary loss to improve credit assignment. We evaluate PRIMT on 2 locomotion and 6 manipulation tasks on various benchmarks, demonstrating superior performance over FM-based and scripted baselines.
♻ ☆ Multimodal "Puppeteer": Exploring Robot Teleoperation Via Virtual Counterpart with LLM-Driven Voice and Gesture Interaction in Augmented Reality
The integration of robotics and augmented reality (AR) offers promising opportunities to enhance human-robot interaction (HRI) by making teleoperation more transparent, spatially grounded, and intuitive. We present a head-mounted AR "puppeteer" framework in which users control a physical robot via interacting with its virtual counterpart robot using large language model (LLM)-driven voice commands and hand-gesture interaction on the Meta Quest 3. In a within-subject user study with 42 participants performing an AR-based robotic pick-and-place pattern-matching task, we compare two interaction conditions: gesture-only (GO) and combined voice+gesture (VG). Our results show that GO currently provides more reliable and efficient control for this time-critical task, while VG introduces additional flexibility but also latency and recognition issues that can increase workload. We further explore how prior robotics experience shapes participants' perceptions of each modality. Based on these findings, we distill a set of evidence-based design guidelines for AR puppeteer metaphoric robot teleoperation, implicating multimodality as an adaptive strategy that must balance efficiency, robustness, and user expertise rather than assuming that additional modalities are universally beneficial. Our work contributes empirical insights into how multimodal (voice+gesture) interaction influences task efficiency, usability, and user experience in AR-based HRI.
comment: This work is under peer review
♻ ☆ How to Adapt Control Barrier Functions? A Learning-Based Approach with Applications to a VTOL Quadplane
In this paper, we present a novel theoretical framework for online adaptation of Control Barrier Function (CBF) parameters, i.e., of the class K functions included in the CBF condition, under input constraints. We introduce the concept of locally validated CBF parameters, which are adapted online to guarantee finite-horizon safety, based on conditions derived from Nagumo's theorem and tangent cone analysis. To identify these parameters online, we integrate a learning-based approach with an uncertainty-aware verification process that account for both epistemic and aleatoric uncertainties inherent in neural network predictions. Our method is demonstrated on a VTOL quadplane model during challenging transition and landing maneuvers, showcasing enhanced performance while maintaining safety.
comment: 2025 IEEE Conference on Decision and Control (CDC). Project page: https://www.taekyung.me/how-to-adapt-cbf
♻ ☆ High-Speed Event Vision-Based Tactile Roller Sensor for Large Surface Measurements
Inspecting large-scale industrial surfaces like aircraft fuselages for quality control requires capturing their precise 3D surface geometry at high resolution. Vision-based tactile sensors (VBTSs) offer high local resolution but require slow 'press-and-lift' measurements stitched for large areas. Approaches with sliding or roller/belt VBTS designs provide measurements continuity. However, they face significant challenges respectively: sliding struggles with friction/wear and both approaches are speed-limited by conventional camera frame rates and motion blur, making large-area scanning time consuming. Thus, a rapid, continuous, high-resolution method is needed. We introduce a novel tactile sensor integrating a neuromorphic camera in a rolling mechanism to achieve this. Leveraging its high temporal resolution and robustness to motion blur, our system uses a modified event-based multi-view stereo approach for 3D reconstruction. We demonstrate state-of-the-art scanning speeds up to 0.5 m/s, achieving Mean Absolute Error below 100 microns -- 11 times faster than prior continuous tactile sensing methods. A multi-reference Bayesian fusion strategy enhances accuracy (reducing MAE by 25.2\% compared to EMVS) and mitigates curvature errors. We also validate high-speed feature recognition via Braille reading 2.6 times faster than previous approaches.
comment: Under Review - Project Page: https://akramekhairi.github.io/TheySeeMeRolling/. 14 pages, 11 figures
♻ ☆ A Unified Framework for Probabilistic Dynamic-, Trajectory- and Vision-based Virtual Fixtures
Probabilistic Virtual Fixtures (VFs) enable the adaptive selection of the most suitable haptic feedback for each phase of a task, based on learned or perceived uncertainty. While keeping the human in the loop remains essential, for instance, to ensure high precision, partial automation of certain task phases is critical for productivity. We present a unified framework for probabilistic VFs that seamlessly switches between manual fixtures, semi-automated fixtures (with the human handling precise tasks), and full autonomy. We introduce a novel probabilistic Dynamical System-based VF for coarse guidance, enabling the robot to autonomously complete certain task phases while keeping the human operator in the loop. For tasks requiring precise guidance, we extend probabilistic position-based trajectory fixtures with automation allowing for seamless human interaction as well as geometry-awareness and optimal impedance gains. For manual tasks requiring very precise guidance, we also extend visual servoing fixtures with the same geometry-awareness and impedance behavior. We validate our approach experimentally on different robots, showcasing multiple operation modes and the ease of programming fixtures.
comment: for the supplementary video, see https://www.youtube.com/watch?v=vUXzcpMbMnY
♻ ☆ VITA: Vision-to-Action Flow Matching Policy
Conventional flow matching and diffusion-based policies sample through iterative denoising from standard noise distributions (e.g., Gaussian), and require conditioning modules to repeatedly incorporate visual information during the generative process, incurring substantial time and memory overhead. To reduce the complexity, we develop VITA(VIsion-To-Action policy), a noise-free and conditioning-free flow matching policy learning framework that directly flows from visual representations to latent actions. Since the source of the flow is visually grounded, VITA eliminates the need of visual conditioning during generation. As expected, bridging vision and action is challenging, because actions are lower-dimensional, less structured, and sparser than visual representations; moreover, flow matching requires the source and target to have the same dimensionality. To overcome this, we introduce an action autoencoder that maps raw actions into a structured latent space aligned with visual latents, trained jointly with flow matching. To further prevent latent space collapse, we propose flow latent decoding, which anchors the latent generation process by backpropagating the action reconstruction loss through the flow matching ODE (ordinary differential equation) solving steps. We evaluate VITA on 9 simulation and 5 real-world tasks from ALOHA and Robomimic. VITA achieves 1.5x-2x faster inference compared to conventional methods with conditioning modules, while outperforming or matching state-of-the-art policies. Codes, datasets, and demos are available at our project page: https://ucd-dare.github.io/VITA/.
comment: Project page: https://ucd-dare.github.io/VITA/ Code: https://github.com/ucd-dare/VITA
♻ ☆ RobustVLA: Robustness-Aware Reinforcement Post-Training for Vision-Language-Action Models
Vision-Language-Action (VLA) models have recently emerged as powerful general-purpose policies for robotic manipulation, benefiting from large-scale multi-modal pre-training. However, they often fail to generalize reliably in out-of-distribution deployments, where unavoidable disturbances such as observation noise, sensor errors, or actuation perturbations become prevalent. While recent Reinforcement Learning (RL)-based post-training provides a practical means to adapt pre-trained VLA models, existing methods mainly emphasize reward maximization and overlook robustness to environmental uncertainty. In this work, we introduce RobustVLA, a lightweight online RL post-training method designed to explicitly enhance the resilience of VLA models. Through a systematic robustness analysis, we identify two key regularizations: Jacobian regularization, which mitigates sensitivity to observation noise, and smoothness regularization, which stabilizes policies under action perturbations. Extensive experiments across diverse robotic environments demonstrate that RobustVLA significantly outperforms prior state-of-the-art methods in robustness and reliability. Our results highlight the importance of principled robustness-aware RL post-training as a key step toward improving the reliability and robustness of VLA models.
♻ ☆ Compliant Residual DAgger: Improving Real-World Contact-Rich Manipulation with Human Corrections
We address key challenges in Dataset Aggregation (DAgger) for real-world contact-rich manipulation: how to collect informative human correction data and how to effectively update policies with this new data. We introduce Compliant Residual DAgger (CR-DAgger), which contains two novel components: 1) a Compliant Intervention Interface that leverages compliance control, allowing humans to provide gentle, accurate delta action corrections without interrupting the ongoing robot policy execution; and 2) a Compliant Residual Policy formulation that learns from human corrections while incorporating force feedback and force control. Our system significantly enhances performance on precise contact-rich manipulation tasks using minimal correction data, improving base policy success rates by over 50\% on two challenging tasks (book flipping and belt assembly) while outperforming both retraining-from-scratch and finetuning approaches. Through extensive real-world experiments, we provide practical guidance for implementing effective DAgger in real-world robot learning tasks. Result videos are available at: https://compliant-residual-dagger.github.io/
♻ ☆ 3EED: Ground Everything Everywhere in 3D NeurIPS 2025
Visual grounding in 3D is the key for embodied agents to localize language-referred objects in open-world environments. However, existing benchmarks are limited to indoor focus, single-platform constraints, and small scale. We introduce 3EED, a multi-platform, multi-modal 3D grounding benchmark featuring RGB and LiDAR data from vehicle, drone, and quadruped platforms. We provide over 128,000 objects and 22,000 validated referring expressions across diverse outdoor scenes -- 10x larger than existing datasets. We develop a scalable annotation pipeline combining vision-language model prompting with human verification to ensure high-quality spatial grounding. To support cross-platform learning, we propose platform-aware normalization and cross-modal alignment techniques, and establish benchmark protocols for in-domain and cross-platform evaluations. Our findings reveal significant performance gaps, highlighting the challenges and opportunities of generalizable 3D grounding. The 3EED dataset and benchmark toolkit are released to advance future research in language-driven 3D embodied perception.
comment: NeurIPS 2025 DB Track; 38 pages, 17 figures, 10 tables; Project Page at https://project-3eed.github.io/
♻ ☆ AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios AAAI 2026
Referring Multi-Object Tracking (RMOT) aims to achieve precise object detection and tracking through natural language instructions, representing a fundamental capability for intelligent robotic systems. However, current RMOT research remains mostly confined to ground-level scenarios, which constrains their ability to capture broad-scale scene contexts and perform comprehensive tracking and path planning. In contrast, Unmanned Aerial Vehicles (UAVs) leverage their expansive aerial perspectives and superior maneuverability to enable wide-area surveillance. Moreover, UAVs have emerged as critical platforms for Embodied Intelligence, which has given rise to an unprecedented demand for intelligent aerial systems capable of natural language interaction. To this end, we introduce AerialMind, the first large-scale RMOT benchmark in UAV scenarios, which aims to bridge this research gap. To facilitate its construction, we develop an innovative semi-automated collaborative agent-based labeling assistant (COALA) framework that significantly reduces labor costs while maintaining annotation quality. Furthermore, we propose HawkEyeTrack (HETrack), a novel method that collaboratively enhances vision-language representation learning and improves the perception of UAV scenarios. Comprehensive experiments validated the challenging nature of our dataset and the effectiveness of our method.
comment: AAAI 2026
♻ ☆ UniFucGrasp: Human-Hand-Inspired Unified Functional Grasp Annotation Strategy and Dataset for Diverse Dexterous Hands
Dexterous grasp datasets are vital for embodied intelligence, but mostly emphasize grasp stability, ignoring functional grasps needed for tasks like opening bottle caps or holding cup handles. Most rely on bulky, costly, and hard-to-control high-DOF Shadow Hands. Inspired by the human hand's underactuated mechanism, we establish UniFucGrasp, a universal functional grasp annotation strategy and dataset for multiple dexterous hand types. Based on biomimicry, it maps natural human motions to diverse hand structures and uses geometry-based force closure to ensure functional, stable, human-like grasps. This method supports low-cost, efficient collection of diverse, high-quality functional grasps. Finally, we establish the first multi-hand functional grasp dataset and provide a synthesis model to validate its effectiveness. Experiments on the UFG dataset, IsaacSim, and complex robotic tasks show that our method improves functional manipulation accuracy and grasp stability, demonstrates improved adaptability across multiple robotic hands, helping to alleviate annotation cost and generalization challenges in dexterous grasping. The project page is at https://haochen611.github.io/UFG.
comment: Accepted to IEEE Robotics and Automation Letters (RA-L). The project page is at https://haochen611.github.io/UFG
♻ ☆ HybridWorldSim: A Scalable and Controllable High-fidelity Simulator for Autonomous Driving
Realistic and controllable simulation is critical for advancing end-to-end autonomous driving, yet existing approaches often struggle to support novel view synthesis under large viewpoint changes or to ensure geometric consistency. We introduce HybridWorldSim, a hybrid simulation framework that integrates multi-traversal neural reconstruction for static backgrounds with generative modeling for dynamic agents. This unified design addresses key limitations of previous methods, enabling the creation of diverse and high-fidelity driving scenarios with reliable visual and spatial consistency. To facilitate robust benchmarking, we further release a new multi-traversal dataset MIRROR that captures a wide range of routes and environmental conditions across different cities. Extensive experiments demonstrate that HybridWorldSim surpasses previous state-of-the-art methods, providing a practical and scalable solution for high-fidelity simulation and a valuable resource for research and development in autonomous driving.
♻ ☆ Curvature-Constrained Vector Field for Motion Planning of Nonholonomic Robots
Vector fields are advantageous in handling nonholonomic motion planning as they provide reference orientation for robots. However, additionally incorporating curvature constraints becomes challenging, due to the interconnection between the design of the curvature-bounded vector field and the tracking controller under underactuation. In this paper, we present a novel framework to co-develop the vector field and the control laws, guiding the nonholonomic robot to the target configuration with curvature-bounded trajectory. First, we formulate the problem by introducing the target positive limit set, which allows the robot to converge to or pass through the target configuration, depending on different dynamics and tasks. Next, we construct a curvature-constrained vector field (CVF) via blending and distributing basic flow fields in workspace and propose the saturated control laws with a dynamic gain, under which the tracking error's magnitude decreases even when saturation occurs. Under the control laws, kinematically constrained nonholonomic robots are guaranteed to track the reference CVF and converge to the target positive limit set with bounded trajectory curvature. Numerical simulations show that the proposed CVF method outperforms other vector-field-based algorithms. Experiments on Ackermann UGVs and semi-physical fixed-wing UAVs demonstrate that the method can be effectively implemented in real-world scenarios.
comment: IEEE T-RO accepted, 20 pages, 22 figures
♻ ☆ A $1000\times$ Faster LLM-enhanced Algorithm For Path Planning in Large-scale Grid Maps
Path planning in grid maps, arising from various applications, has garnered significant attention. Existing methods, such as A*, Dijkstra, and their variants, work well for small-scale maps but fail to address large-scale ones due to high search time and memory consumption. Recently, Large Language Models (LLMs) have shown remarkable performance in path planning but still suffer from spatial illusion and poor planning performance. Among all the works, LLM-A* \cite{meng2024llm} leverages LLM to generate a series of waypoints and then uses A* to plan the paths between the neighboring waypoints. In this way, the complete path is constructed. However, LLM-A* still suffers from high computational time for large-scale maps. To fill this gap, we conducted a deep investigation into LLM-A* and found its bottleneck, resulting in limited performance. Accordingly, we design an innovative LLM-enhanced algorithm, abbr. as iLLM-A*. iLLM-A* includes 3 carefully designed mechanisms, including the optimization of A*, an incremental learning method for LLM to generate high-quality waypoints, and the selection of the appropriate waypoints for A* for path planning. Finally, a comprehensive evaluation on various grid maps shows that, compared with LLM-A*, iLLM-A* \textbf{1) achieves more than $1000\times$ speedup on average, and up to $2349.5\times$ speedup in the extreme case, 2) saves up to $58.6\%$ of the memory cost, 3) achieves both obviously shorter path length and lower path length standard deviation.}
♻ ☆ AutoDrive-R$^2$: Incentivizing Reasoning and Self-Reflection Capacity for VLA Model in Autonomous Driving
Vision-Language-Action (VLA) models in autonomous driving systems have recently demonstrated transformative potential by integrating multimodal perception with decision-making capabilities. However, the interpretability and coherence of the decision process and the plausibility of action sequences remain largely underexplored. To address these issues, we propose AutoDrive-R$^2$, a novel VLA framework that enhances both reasoning and self-reflection capabilities of autonomous driving systems through chain-of-thought (CoT) processing and reinforcement learning (RL). Specifically, we first propose an innovative CoT dataset named nuScenesR$^2$-6K for supervised fine-tuning, which effectively builds cognitive bridges between input information and output trajectories through a four-step logical chain with self-reflection for validation. Moreover, to maximize both reasoning and self-reflection during the RL stage, we further employ the Group Relative Policy Optimization (GRPO) algorithm within a physics-grounded reward framework that incorporates spatial alignment, vehicle dynamic, and temporal smoothness criteria to ensure reliable and realistic trajectory planning. Extensive evaluation results across both nuScenes and Waymo datasets demonstrates the state-of-the-art performance and robust generalization capacity of our proposed method.
Robotics 35
☆ Ethically-Aware Participatory Design of a Productivity Social Robot for College Students
College students often face academic and life stressors affecting productivity, especially students with Attention Deficit Hyperactivity Disorder (ADHD) who experience executive functioning challenges. Conventional productivity tools typically demand sustained self-discipline and consistent use, which many students struggle with, leading to disruptive app-switching behaviors. Socially Assistive Robots (SARs), known for their intuitive and interactive nature, offer promising potential to support productivity in academic environments, having been successfully utilized in domains like education, cognitive development, and mental health. To leverage SARs effectively in addressing student productivity, this study employed a Participatory Design (PD) approach, directly involving college students and a Student Success and Well-Being Coach in the design process. Through interviews and a collaborative workshop, we gathered detailed insights on productivity challenges and identified desirable features for a productivity-focused SAR. Importantly, ethical considerations were integrated from the onset, facilitating responsible and user-aligned design choices. Our contributions include comprehensive insights into student productivity challenges, SAR design preferences, and actionable recommendations for effective robot characteristics. Additionally, we present stakeholder-derived ethical guidelines to inform responsible future implementations of productivity-focused SARs in higher education.
☆ Think Fast: Real-Time Kinodynamic Belief-Space Planning for Projectile Interception
Intercepting fast moving objects, by its very nature, is challenging because of its tight time constraints. This problem becomes further complicated in the presence of sensor noise because noisy sensors provide, at best, incomplete information, which results in a distribution over target states to be intercepted. Since time is of the essence, to hit the target, the planner must begin directing the interceptor, in this case a robot arm, while still receiving information. We introduce an tree-like structure, which is grown using kinodynamic motion primitives in state-time space. This tree-like structure encodes reachability to multiple goals from a single origin, while enabling real-time value updates as the target belief evolves and seamless transitions between goals. We evaluate our framework on an interception task on a 6 DOF industrial arm (ABB IRB-1600) with an onboard stereo camera (ZED 2i). A robust Innovation-based Adaptive Estimation Adaptive Kalman Filter (RIAE-AKF) is used to track the target and perform belief updates.
☆ Tactile Robotics: Past and Future
What is the future of tactile robotics? To help define that future, this article provides a historical perspective on tactile sensing in robotics from the wealth of knowledge and expert opinion in nearly 150 reviews over almost half a century. This history is characterized by a succession of generations: 1965-79 (origins), 1980-94 (foundations and growth), 1995-2009 (tactile winter) and 2010-2024 (expansion and diversification). Recent expansion has led to diverse themes emerging of e-skins, tactile robotic hands, vision-based tactile sensing, soft/biomimetic touch, and the tactile Internet. In the next generation from 2025, tactile robotics could mature to widespread commercial use, with applications in human-like dexterity, understanding human intelligence, and telepresence impacting all robotics and AI. By linking past expert insights to present themes, this article highlights recurring challenges in tactile robotics, showing how the field has evolved, why progress has often stalled, and which opportunities are most likely to define its future.
comment: Accepted in International Journal of Robotics Research (IJRR)
☆ Supporting Productivity Skill Development in College Students through Social Robot Coaching: A Proof-of-Concept
College students often face academic challenges that hamper their productivity and well-being. Although self-help books and productivity apps are popular, they often fall short. Books provide generalized, non-interactive guidance, and apps are not inherently educational and can hinder the development of key organizational skills. Traditional productivity coaching offers personalized support, but is resource-intensive and difficult to scale. In this study, we present a proof-of-concept for a socially assistive robot (SAR) as an educational coach and a potential solution to the limitations of existing productivity tools and coaching approaches. The SAR delivers six different lessons on time management and task prioritization. Users interact via a chat interface, while the SAR responds through speech (with a toggle option). An integrated dashboard monitors progress, mood, engagement, confidence per lesson, and time spent per lesson. It also offers personalized productivity insights to foster reflection and self-awareness. We evaluated the system with 15 college students, achieving a System Usability Score of 79.2 and high ratings for overall experience and engagement. Our findings suggest that SAR-based productivity coaching can offer an effective and scalable solution to improve productivity among college students.
☆ Estimation of Kinematic Motion from Dashcam Footage
The goal of this paper is to explore the accuracy of dashcam footage to predict the actual kinematic motion of a car-like vehicle. Our approach uses ground truth information from the vehicle's on-board data stream, through the controller area network, and a time-synchronized dashboard camera, mounted to a consumer-grade vehicle, for 18 hours of footage and driving. The contributions of the paper include neural network models that allow us to quantify the accuracy of predicting the vehicle speed and yaw, as well as the presence of a lead vehicle, and its relative distance and speed. In addition, the paper describes how other researchers can gather their own data to perform similar experiments, using open-source tools and off-the-shelf technology.
comment: 8 pages, 10 figures
☆ Reinforcement Learning for Gliding Projectile Guidance and Control
This paper presents the development of a control law, which is intended to be implemented on an optical guided glider. This guiding law follows an innovative approach, the reinforcement learning. This control law is used to make navigation more flexible and autonomous in a dynamic environment. The final objective is to track a target detected with the camera and then guide the glider to this point with high precision. Already applied on quad-copter drones, we wish by this study to demonstrate the applicability of reinforcement learning for fixed-wing aircraft on all of its axis.
comment: 6 pages
☆ Opening the Sim-to-Real Door for Humanoid Pixel-to-Action Policy Transfer
Recent progress in GPU-accelerated, photorealistic simulation has opened a scalable data-generation path for robot learning, where massive physics and visual randomization allow policies to generalize beyond curated environments. Building on these advances, we develop a teacher-student-bootstrap learning framework for vision-based humanoid loco-manipulation, using articulated-object interaction as a representative high-difficulty benchmark. Our approach introduces a staged-reset exploration strategy that stabilizes long-horizon privileged-policy training, and a GRPO-based fine-tuning procedure that mitigates partial observability and improves closed-loop consistency in sim-to-real RL. Trained entirely on simulation data, the resulting policy achieves robust zero-shot performance across diverse door types and outperforms human teleoperators by up to 31.7% in task completion time under the same whole-body control stack. This represents the first humanoid sim-to-real policy capable of diverse articulated loco-manipulation using pure RGB perception.
comment: https://doorman-humanoid.github.io/
☆ Autonomous Grasping On Quadruped Robot With Task Level Interaction
Quadruped robots are increasingly used in various applications due to their high mobility and ability to operate in diverse terrains. However, most available quadruped robots are primarily focused on mobility without object manipulation capabilities. Equipping a quadruped robot with a robotic arm and gripper introduces a challenge in manual control, especially in remote scenarios that require complex commands. This research aims to develop an autonomous grasping system on a quadruped robot using a task-level interaction approach. The system includes hardware integration of a robotic arm and gripper onto the quadruped robot's body, a layered control system designed using ROS, and a web-based interface for human-robot interaction. The robot is capable of autonomously performing tasks such as navigation, object detection, and grasping using GraspNet. Testing was conducted through real-world scenarios to evaluate navigation, object selection and grasping, and user experience. The results show that the robot can perform tasks accurately and consistently, achieving a grasping success rate of 75 % from 12 trials. Therefore, the system demonstrates significant potential in enhancing the capabilities of quadruped robots as service robots in real-world environments.
comment: This work has been submitted to the IEEE for possible publication
☆ VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inference
Vision-Language-Action models (VLAs) are becoming increasingly capable across diverse robotic tasks. However, their real-world deployment remains slow and inefficient: demonstration videos are often sped up by 5-10x to appear smooth, with noticeable action stalls and delayed reactions to environmental changes. Asynchronous inference offers a promising solution to achieve continuous and low-latency control by enabling robots to execute actions and perform inference simultaneously. However, because the robot and environment continue to evolve during inference, a temporal misalignment arises between the prediction and execution intervals. This leads to significant action instability, while existing methods either degrade accuracy or introduce runtime overhead to mitigate it. We propose VLASH, a general asynchronous inference framework for VLAs that delivers smooth, accurate, and fast reaction control without additional overhead or architectural changes. VLASH estimates the future execution-time state by rolling the robot state forward with the previously generated action chunk, thereby bridging the gap between prediction and execution. Experiments show that VLASH achieves up to 2.03x speedup and reduces reaction latency by up to 17.4x compared to synchronous inference while fully preserving the original accuracy. Moreover, it empowers VLAs to handle fast-reaction, high-precision tasks such as playing ping-pong and playing whack-a-mole, where traditional synchronous inference fails. Code is available at https://github.com/mit-han-lab/vlash
☆ CycleManip: Enabling Cyclic Task Manipulation via Effective Historical Perception and Understanding
In this paper, we explore an important yet underexplored task in robot manipulation: cycle-based manipulation, where robots need to perform cyclic or repetitive actions with an expected terminal time. These tasks are crucial in daily life, such as shaking a bottle or knocking a nail. However, few prior works have explored this task, leading to two main challenges: 1) the imitation methods often fail to complete these tasks within the expected terminal time due to the ineffective utilization of history; 2) the absence of a benchmark with sufficient data and automatic evaluation tools hinders development of effective solutions in this area. To address these challenges, we first propose the CycleManip framework to achieve cycle-based task manipulation in an end-to-end imitation manner without requiring any extra models, hierarchical structure or significant computational overhead. The core insight is to enhance effective history perception by a cost-aware sampling strategy and to improve historical understanding by multi-task learning. Second, we introduce a cycle-based task manipulation benchmark, which provides diverse cycle-based tasks, and an automatic evaluation method. Extensive experiments conducted in both simulation and real-world settings demonstrate that our method achieves high success rates in cycle-based task manipulation. The results further show strong adaptability performance in general manipulation, and the plug-and-play ability on imitation policies such as Vision-Language-Action (VLA) models. Moreover, the results show that our approach can be applied across diverse robotic platforms, including bi-arm grippers, dexterous hands, and humanoid robots.
comment: Project page: https://isee-laboratory.github.io/OmniDexGrasp/
☆ Integration of UWB Radar on Mobile Robots for Continuous Obstacle and Environment Mapping
This paper presents an infrastructure-free approach for obstacle detection and environmental mapping using ultra-wideband (UWB) radar mounted on a mobile robotic platform. Traditional sensing modalities such as visual cameras and Light Detection and Ranging (LiDAR) fail in environments with poor visibility due to darkness, smoke, or reflective surfaces. In these visioned-impaired conditions, UWB radar offers a promising alternative. To this end, this work explores the suitability of robot-mounted UWB radar for environmental mapping in dynamic, anchor-free scenarios. The study investigates how different materials (metal, concrete and plywood) and UWB radio channels (5 and 9) influence the Channel Impulse Response (CIR). Furthermore, a processing pipeline is proposed to achieve reliable mapping of detected obstacles, consisting of 3 steps: (i) target identification (based on CIR peak detection), (ii) filtering (based on peak properties, signal-to-noise score, and phase-difference of arrival), and (iii) clustering (based on distance estimation and angle-of-arrival estimation). The proposed approach successfully reduces noise and multipath effects, resulting in an obstacle detection precision of at least 82.36% and a recall of 89.46% on channel 9 even when detecting low-reflective materials such as plywood. This work offers a foundation for further development of UWB-based localisation and mapping (SLAM) systems that do not rely on visual features and, unlike conventional UWB localisation systems, do not require on fixed anchor nodes for triangulation.
comment: This paper has been submitted to IEEE Access Journal and is currently undergoing review
☆ FOM-Nav: Frontier-Object Maps for Object Goal Navigation
This paper addresses the Object Goal Navigation problem, where a robot must efficiently find a target object in an unknown environment. Existing implicit memory-based methods struggle with long-term memory retention and planning, while explicit map-based approaches lack rich semantic information. To address these challenges, we propose FOM-Nav, a modular framework that enhances exploration efficiency through Frontier-Object Maps and vision-language models. Our Frontier-Object Maps are built online and jointly encode spatial frontiers and fine-grained object information. Using this representation, a vision-language model performs multimodal scene understanding and high-level goal prediction, which is executed by a low-level planner for efficient trajectory generation. To train FOM-Nav, we automatically construct large-scale navigation datasets from real-world scanned environments. Extensive experiments validate the effectiveness of our model design and constructed dataset. FOM-Nav achieves state-of-the-art performance on the MP3D and HM3D benchmarks, particularly in navigation efficiency metric SPL, and yields promising results on a real robot.
comment: Project page: https://www.di.ens.fr/willow/research/fom-nav/
☆ MM-ACT: Learn from Multimodal Parallel Generation to Act
A generalist robotic policy needs both semantic understanding for task planning and the ability to interact with the environment through predictive capabilities. To tackle this, we present MM-ACT, a unified Vision-Language-Action (VLA) model that integrates text, image, and action in shared token space and performs generation across all three modalities. MM-ACT adopts a re-mask parallel decoding strategy for text and image generation, and employs a one-step parallel decoding strategy for action generation to improve efficiency. We introduce Context-Shared Multimodal Learning, a unified training paradigm that supervises generation in all three modalities from a shared context, enhancing action generation through cross-modal learning. Experiments were conducted on the LIBERO simulation and Franka real-robot setups as well as RoboTwin2.0 to assess in-domain and out-of-domain performances respectively. Our approach achieves a success rate of 96.3% on LIBERO, 72.0% across three tasks of real Franka, and 52.38% across eight bimanual tasks of RoboTwin2.0 with an additional gain of 9.25% from cross-modal learning. We release our codes, models and data at https://github.com/HHYHRHY/MM-ACT.
comment: 17 pages
☆ H-Zero: Cross-Humanoid Locomotion Pretraining Enables Few-shot Novel Embodiment Transfer
The rapid advancement of humanoid robotics has intensified the need for robust and adaptable controllers to enable stable and efficient locomotion across diverse platforms. However, developing such controllers remains a significant challenge because existing solutions are tailored to specific robot designs, requiring extensive tuning of reward functions, physical parameters, and training hyperparameters for each embodiment. To address this challenge, we introduce H-Zero, a cross-humanoid locomotion pretraining pipeline that learns a generalizable humanoid base policy. We show that pretraining on a limited set of embodiments enables zero-shot and few-shot transfer to novel humanoid robots with minimal fine-tuning. Evaluations show that the pretrained policy maintains up to 81% of the full episode duration on unseen robots in simulation while enabling few-shot transfer to unseen humanoids and upright quadrupeds within 30 minutes of fine-tuning.
comment: in submission, under review
☆ Constant-Time Motion Planning with Manipulation Behaviors
Recent progress in contact-rich robotic manipulation has been striking, yet most deployed systems remain confined to simple, scripted routines. One of the key barriers is the lack of motion planning algorithms that can provide verifiable guarantees for safety, efficiency and reliability. To address this, a family of algorithms called Constant-Time Motion Planning (CTMP) was introduced, which leverages a preprocessing phase to enable collision-free motion queries in a fixed, user-specified time budget (e.g., 10 milliseconds). However, existing CTMP methods do not explicitly incorporate the manipulation behaviors essential for object handling. To bridge this gap, we introduce the \textit{Behavioral Constant-Time Motion Planner} (B-CTMP), an algorithm that extends CTMP to solve a broad class of two-step manipulation tasks: (1) a collision-free motion to a behavior initiation state, followed by (2) execution of a manipulation behavior (such as grasping or insertion) to reach the goal. By precomputing compact data structures, B-CTMP guarantees constant-time query in mere milliseconds while ensuring completeness and successful task execution over a specified set of states. We evaluate B-CTMP on two canonical manipulation tasks in simulation, shelf picking and plug insertion,and demonstrate its effectiveness on a real robot. Our results show that B-CTMP unifies collision-free planning and object manipulation within a single constant-time framework, providing provable guarantees of speed and success for manipulation in semi-structured environments.
comment: In submission
☆ Partially Equivariant Reinforcement Learning in Symmetry-Breaking Environments
Group symmetries provide a powerful inductive bias for reinforcement learning (RL), enabling efficient generalization across symmetric states and actions via group-invariant Markov Decision Processes (MDPs). However, real-world environments almost never realize fully group-invariant MDPs; dynamics, actuation limits, and reward design usually break symmetries, often only locally. Under group-invariant Bellman backups for such cases, local symmetry-breaking introduces errors that propagate across the entire state-action space, resulting in global value estimation errors. To address this, we introduce Partially group-Invariant MDP (PI-MDP), which selectively applies group-invariant or standard Bellman backups depending on where symmetry holds. This framework mitigates error propagation from locally broken symmetries while maintaining the benefits of equivariance, thereby enhancing sample efficiency and generalizability. Building on this framework, we present practical RL algorithms -- Partially Equivariant (PE)-DQN for discrete control and PE-SAC for continuous control -- that combine the benefits of equivariance with robustness to symmetry-breaking. Experiments across Grid-World, locomotion, and manipulation benchmarks demonstrate that PE-DQN and PE-SAC significantly outperform baselines, highlighting the importance of selective symmetry exploitation for robust and sample-efficient RL.
comment: 27 pages, 10 figures
☆ Magnetic Tactile-Driven Soft Actuator for Intelligent Grasping and Firmness Evaluation
Soft robots are powerful tools for manipulating delicate objects, yet their adoption is hindered by two gaps: the lack of integrated tactile sensing and sensor signal distortion caused by actuator deformations. This paper addresses these challenges by introducing the SoftMag actuator: a magnetic tactile-sensorized soft actuator. Unlike systems relying on attached sensors or treating sensing and actuation separately, SoftMag unifies them through a shared architecture while confronting the mechanical parasitic effect, where deformations corrupt tactile signals. A multiphysics simulation framework models this coupling, and a neural-network-based decoupling strategy removes the parasitic component, restoring sensing fidelity. Experiments including indentation, quasi-static and step actuation, and fatigue tests validate the actuator's performance and decoupling effectiveness. Building upon this foundation, the system is extended into a two-finger SoftMag gripper, where a multi-task neural network enables real-time prediction of tri-axial contact forces and position. Furthermore, a probing-based strategy estimates object firmness during grasping. Validation on apricots shows a strong correlation (Pearson r over 0.8) between gripper-estimated firmness and reference measurements, confirming the system's capability for non-destructive quality assessment. Results demonstrate that combining integrated magnetic sensing, learning-based correction, and real-time inference enables a soft robotic platform that adapts its grasp and quantifies material properties. The framework offers an approach for advancing sensorized soft actuators toward intelligent, material-aware robotics.
comment: 25 pages, 24 figures
☆ SwiftVLA: Unlocking Spatiotemporal Dynamics for Lightweight VLA Models at Minimal Overhead
Vision-Language-Action (VLA) models built on pretrained Vision-Language Models (VLMs) show strong potential but are limited in practicality due to their large parameter counts. To mitigate this issue, using a lightweight VLM has been explored, but it compromises spatiotemporal reasoning. Although some methods suggest that incorporating additional 3D inputs can help, they usually rely on large VLMs to fuse 3D and 2D inputs and still lack temporal understanding. Therefore, we propose SwiftVLA, an architecture that enhances a compact model with 4D understanding while preserving design efficiency. Specifically, our approach features a pretrained 4D visual geometry transformer with a temporal cache that extracts 4D features from 2D images. Then, to enhance the VLM's ability to exploit both 2D images and 4D features, we introduce Fusion Tokens, a set of learnable tokens trained with a future prediction objective to generate unified representations for action generation. Finally, we introduce a mask-and-reconstruct strategy that masks 4D inputs to the VLM and trains the VLA to reconstruct them, enabling the VLM to learn effective 4D representations and allowing the 4D branch to be dropped at inference with minimal performance loss. Experiments in real and simulated environments show that SwiftVLA outperforms lightweight baselines and rivals VLAs up to 7 times larger, achieving comparable performance on edge devices while being 18 times faster and reducing memory footprint by 12 times.
☆ A Novel MDP Decomposition Framework for Scalable UAV Mission Planning in Complex and Uncertain Environments
This paper presents a scalable and fault-tolerant framework for unmanned aerial vehicle (UAV) mission management in complex and uncertain environments. The proposed approach addresses the computational bottleneck inherent in solving large-scale Markov Decision Processes (MDPs) by introducing a two-stage decomposition strategy. In the first stage, a factor-based algorithm partitions the global MDP into smaller, goal-specific sub-MDPs by leveraging domain-specific features such as goal priority, fault states, spatial layout, and energy constraints. In the second stage, a priority-based recombination algorithm solves each sub-MDP independently and integrates the results into a unified global policy using a meta-policy for conflict resolution. Importantly, we present a theoretical analysis showing that, under mild probabilistic independence assumptions, the combined policy is provably equivalent to the optimal global MDP policy. Our work advances artificial intelligence (AI) decision scalability by decomposing large MDPs into tractable subproblems with provable global equivalence. The proposed decomposition framework enhances the scalability of Markov Decision Processes, a cornerstone of sequential decision-making in artificial intelligence, enabling real-time policy updates for complex mission environments. Extensive simulations validate the effectiveness of our method, demonstrating orders-of-magnitude reduction in computation time without sacrificing mission reliability or policy optimality. The proposed framework establishes a practical and robust foundation for scalable decision-making in real-time UAV mission execution.
☆ Transforming Monolithic Foundation Models into Embodied Multi-Agent Architectures for Human-Robot Collaboration
Foundation models have become central to unifying perception and planning in robotics, yet real-world deployment exposes a mismatch between their monolithic assumption that a single model can handle all cognitive functions and the distributed, dynamic nature of practical service workflows. Vision-language models offer strong semantic understanding but lack embodiment-aware action capabilities while relying on hand-crafted skills. Vision-Language-Action policies enable reactive manipulation but remain brittle across embodiments, weak in geometric grounding, and devoid of proactive collaboration mechanisms. These limitations indicate that scaling a single model alone cannot deliver reliable autonomy for service robots operating in human-populated settings. To address this gap, we present InteractGen, an LLM-powered multi-agent framework that decomposes robot intelligence into specialized agents for continuous perception, dependency-aware planning, decision and verification, failure reflection, and dynamic human delegation, treating foundation models as regulated components within a closed-loop collective. Deployed on a heterogeneous robot team and evaluated in a three-month open-use study, InteractGen improves task success, adaptability, and human-robot collaboration, providing evidence that multi-agent orchestration offers a more feasible path toward socially grounded service autonomy than further scaling standalone models.
comment: 21 pages, 16 figures, 4 tables
☆ Sigma: The Key for Vision-Language-Action Models toward Telepathic Alignment
To address the gap in humanoid robot cognitive systems regarding the lack of a time-updable mediating thought space between semantics and continuous control, this study constructs and trains a VLA model named "Sigma" that runs on a single RTX 4090. It uses the open-source pi05_base model as a foundation and preprocesses svla_so101_pickplace into a training dataset. The researcher independently designed an architecture for a vision-language-action model that combines deep semantic understanding and association to achieve telepathic communication. The training process involved repeated optimizations of data preprocessing, LoRA fine-tuning, and the inference-stage adapter. The experiment employed offline closed-loop replay, comparing Sigma with the untuned pure pi05_base_base model under data conditions. Results showed that Sigma exhibited a stable decrease in control MSE across vector, fragment, and entire trajectory timescales, while maintaining the telepathy norm and semantic-text alignment quality unchanged. It demonstrates that mind-responsive alignment control is quantified through an architecture that combines deep understanding of semantics and association without retraining the base model, which provides reproducible experience for semantic alignment and intention-driven behavior in humanoid robots.
comment: The Sigma model has been open-sourced on Hugging Face. Weights, dataset, some scripts, and logs are all available. The link is: https://huggingface.co/Veltraxor/Sigma
☆ Sign Language Recognition using Bidirectional Reservoir Computing
Sign language recognition (SLR) facilitates communication between deaf and hearing individuals. Deep learning is widely used to develop SLR-based systems; however, it is computationally intensive and requires substantial computational resources, making it unsuitable for resource-constrained devices. To address this, we propose an efficient sign language recognition system using MediaPipe and an echo state network (ESN)-based bidirectional reservoir computing (BRC) architecture. MediaPipe extracts hand joint coordinates, which serve as inputs to the ESN-based BRC architecture. The BRC processes these features in both forward and backward directions, efficiently capturing temporal dependencies. The resulting states of BRC are concatenated to form a robust representation for classification. We evaluated our method on the Word-Level American Sign Language (WLASL) video dataset, achieving a competitive accuracy of 57.71% and a significantly lower training time of only 9 seconds, in contrast to the 55 minutes and $38$ seconds required by the deep learning-based Bi-GRU approach. Consequently, the BRC-based SLR system is well-suited for edge devices.
☆ SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
Linear Temporal Logic (LTL) provides a rigorous framework for complex robotic tasks, yet existing methods often rely on accurate dynamics models or expensive online interactions. In this work, we address LTL-constrained control in a challenging offline, model-free setting, utilizing only fixed, task-agnostic datasets of fragmented trajectories. We propose SAGAS, a novel framework combining graph-assisted trajectory stitching with automata-guided planning. First, we construct a latent reachability graph from a learned temporal-distance representation. To bridge the semantic gap, we augment this graph with certified anchor nodes and probabilistic soft labels. We then translate the specification into a Büchi automaton and search the implicit product space to derive a cost-minimal prefix-suffix plan. Finally, a subgoal-conditioned low-level policy is deployed to execute these latent waypoints. Experiments on OGBench locomotion domains demonstrate that SAGAS successfully synthesizes efficient trajectories for diverse LTL tasks, effectively bridging the gap between fragmented offline data and complex logical constraints.
☆ MS-PPO: Morphological-Symmetry-Equivariant Policy for Legged Robot Locomotion
Reinforcement learning has recently enabled impressive locomotion capabilities on legged robots; however, most policy architectures remain morphology- and symmetry-agnostic, leading to inefficient training and limited generalization. This work introduces MS-PPO, a morphological-symmetry-equivariant policy learning framework that encodes robot kinematic structure and morphological symmetries directly into the policy network. We construct a morphology-informed graph neural architecture that is provably equivariant with respect to the robot's morphological symmetry group actions, ensuring consistent policy responses under symmetric states while maintaining invariance in value estimation. This design eliminates the need for tedious reward shaping or costly data augmentation, which are typically required to enforce symmetry. We evaluate MS-PPO in simulation on Unitree Go2 and Xiaomi CyberDog2 robots across diverse locomotion tasks, including trotting, pronking, slope walking, and bipedal turning, and further deploy the learned policies on hardware. Extensive experiments show that MS-PPO achieves superior training stability, symmetry generalization ability, and sample efficiency in challenging locomotion tasks, compared to state-of-the-art baselines. These findings demonstrate that embedding both kinematic structure and morphological symmetry into policy learning provides a powerful inductive bias for legged robot locomotion control. Our code will be made publicly available at https://lunarlab-gatech.github.io/MS-PPO/.
☆ TrajDiff: End-to-end Autonomous Driving without Perception Annotation
End-to-end autonomous driving systems directly generate driving policies from raw sensor inputs. While these systems can extract effective environmental features for planning, relying on auxiliary perception tasks, developing perception annotation-free planning paradigms has become increasingly critical due to the high cost of manual perception annotation. In this work, we propose TrajDiff, a Trajectory-oriented BEV Conditioned Diffusion framework that establishes a fully perception annotation-free generative method for end-to-end autonomous driving. TrajDiff requires only raw sensor inputs and future trajectory, constructing Gaussian BEV heatmap targets that inherently capture driving modalities. We design a simple yet effective trajectory-oriented BEV encoder to extract the TrajBEV feature without perceptual supervision. Furthermore, we introduce Trajectory-oriented BEV Diffusion Transformer (TB-DiT), which leverages ego-state information and the predicted TrajBEV features to directly generate diverse yet plausible trajectories, eliminating the need for handcrafted motion priors. Beyond architectural innovations, TrajDiff enables exploration of data scaling benefits in the annotation-free setting. Evaluated on the NAVSIM benchmark, TrajDiff achieves 87.5 PDMS, establishing state-of-the-art performance among all annotation-free methods. With data scaling, it further improves to 88.5 PDMS, which is comparable to advanced perception-based approaches. Our code and model will be made publicly available.
♻ ☆ Gentle Object Retraction in Dense Clutter Using Multimodal Force Sensing and Imitation Learning
Dense collections of movable objects are common in everyday spaces-from cabinets in a home to shelves in a warehouse. Safely retracting objects from such collections is difficult for robots, yet people do it frequently, leveraging learned experience in tandem with vision and non-prehensile tactile sensing on the sides and backs of their hands and arms. We investigate the role of contact force sensing for training robots to gently reach into constrained clutter and extract objects. The available sensing modalities are (1) "eye-in-hand" vision, (2) proprioception, (3) non-prehensile triaxial tactile sensing, (4) contact wrenches estimated from joint torques, and (5) a measure of object acquisition obtained by monitoring the vacuum line of a suction cup. We use imitation learning to train policies from a set of demonstrations on randomly generated scenes, then conduct an ablation study of wrench and tactile information. We evaluate each policy's performance across 40 unseen environment configurations. Policies employing any force sensing show fewer excessive force failures, an increased overall success rate, and faster completion times. The best performance is achieved using both tactile and wrench information, producing an 80% improvement above the baseline without force information.
comment: Accepted in IEEE Robotics and Automation Letters (RA-L)
♻ ☆ HiMo: High-Speed Objects Motion Compensation in Point Clouds
LiDAR point cloud is essential for autonomous vehicles, but motion distortions from dynamic objects degrade the data quality. While previous work has considered distortions caused by ego motion, distortions caused by other moving objects remain largely overlooked, leading to errors in object shape and position. This distortion is particularly pronounced in high-speed environments such as highways and in multi-LiDAR configurations, a common setup for heavy vehicles. To address this challenge, we introduce HiMo, a pipeline that repurposes scene flow estimation for non-ego motion compensation, correcting the representation of dynamic objects in point clouds. During the development of HiMo, we observed that existing self-supervised scene flow estimators often produce degenerate or inconsistent estimates under high-speed distortion. We further propose SeFlow++, a real-time scene flow estimator that achieves state-of-the-art performance on both scene flow and motion compensation. Since well-established motion distortion metrics are absent in the literature, we introduce two evaluation metrics: compensation accuracy at a point level and shape similarity of objects. We validate HiMo through extensive experiments on Argoverse 2, ZOD, and a newly collected real-world dataset featuring highway driving and multi-LiDAR-equipped heavy vehicles. Our findings show that HiMo improves the geometric consistency and visual fidelity of dynamic objects in LiDAR point clouds, benefiting downstream tasks such as semantic segmentation and 3D detection. See https://kin-zhang.github.io/HiMo for more details.
comment: 15 pages, 13 figures, Published in Transactions on Robotics (Volume 41)
♻ ☆ SRPO: Self-Referential Policy Optimization for Vision-Language-Action Models
Vision-Language-Action (VLA) models excel in robotic manipulation but are constrained by their heavy reliance on expert demonstrations, leading to demonstration bias and limiting performance. Reinforcement learning (RL) is a vital post-training strategy to overcome these limits, yet current VLA-RL methods, including group-based optimization approaches, are crippled by severe reward sparsity. Relying on binary success indicators wastes valuable information in failed trajectories, resulting in low training efficiency. To solve this, we propose Self-Referential Policy Optimization (SRPO), a novel VLA-RL framework. SRPO eliminates the need for external demonstrations or manual reward engineering by leveraging the model's own successful trajectories, generated within the current training batch, as a self-reference. This allows us to assign a progress-wise reward to failed attempts. A core innovation is the use of latent world representations to measure behavioral progress robustly. Instead of relying on raw pixels or requiring domain-specific fine-tuning, we utilize the compressed, transferable encodings from a world model's latent space. These representations naturally capture progress patterns across environments, enabling accurate, generalized trajectory comparison. Empirical evaluations on the LIBERO benchmark demonstrate SRPO's efficiency and effectiveness. Starting from a supervised baseline with 48.9% success, SRPO achieves a new state-of-the-art success rate of 99.2% in just 200 RL steps, representing a 103% relative improvement without any extra supervision. Furthermore, SRPO shows substantial robustness, achieving a 167% performance improvement on the LIBERO-Plus benchmark.
♻ ☆ Differentiable Contact Dynamics for Stable Object Placement Under Geometric Uncertainties
From serving a cup of coffee to positioning mechanical parts during assembly, stable object placement is a crucial skill for future robots. It becomes particularly challenging under geometric uncertainties, e.g., when the object pose or shape is not known accurately. This work leverages a differentiable simulation model of contact dynamics to tackle this challenge. We derive a novel gradient that relates force-torque sensor readings to geometric uncertainties, thus enabling uncertainty estimation by minimizing discrepancies between sensor data and model predictions via gradient descent. Gradient-based methods are sensitive to initialization. To mitigate this effect, we maintain a belief over multiple estimates and choose the robot action based on the current belief at each timestep. In experiments on a Franka robot arm, our method achieved promising results on multiple objects under various geometric uncertainties, including the in-hand pose uncertainty of a grasped object, the object shape uncertainty, and the environment uncertainty.
♻ ☆ GigaWorld-0: World Models as Data Engine to Empower Embodied AI
World models are emerging as a foundational paradigm for scalable, data-efficient embodied AI. In this work, we present GigaWorld-0, a unified world model framework designed explicitly as a data engine for Vision-Language-Action (VLA) learning. GigaWorld-0 integrates two synergistic components: GigaWorld-0-Video, which leverages large-scale video generation to produce diverse, texture-rich, and temporally coherent embodied sequences under fine-grained control of appearance, camera viewpoint, and action semantics; and GigaWorld-0-3D, which combines 3D generative modeling, 3D Gaussian Splatting reconstruction, physically differentiable system identification, and executable motion planning to ensure geometric consistency and physical realism. Their joint optimization enables the scalable synthesis of embodied interaction data that is visually compelling, spatially coherent, physically plausible, and instruction-aligned. Training at scale is made feasible through our efficient GigaTrain framework, which exploits FP8-precision and sparse attention to drastically reduce memory and compute requirements. We conduct comprehensive evaluations showing that GigaWorld-0 generates high-quality, diverse, and controllable data across multiple dimensions. Critically, VLA model (e.g., GigaBrain-0) trained on GigaWorld-0-generated data achieve strong real-world performance, significantly improving generalization and task success on physical robots without any real-world interaction during training.
comment: Project Page: https://giga-world-0.github.io/
♻ ☆ Automaton Constrained Q-Learning NeurIPS 2025
Real-world robotic tasks often require agents to achieve sequences of goals while respecting time-varying safety constraints. However, standard Reinforcement Learning (RL) paradigms are fundamentally limited in these settings. A natural approach to these problems is to combine RL with Linear-time Temporal Logic (LTL), a formal language for specifying complex, temporally extended tasks and safety constraints. Yet, existing RL methods for LTL objectives exhibit poor empirical performance in complex and continuous environments. As a result, no scalable methods support both temporally ordered goals and safety simultaneously, making them ill-suited for realistic robotics scenarios. We propose Automaton Constrained Q-Learning (ACQL), an algorithm that addresses this gap by combining goal-conditioned value learning with automaton-guided reinforcement. ACQL supports most LTL task specifications and leverages their automaton representation to explicitly encode stage-wise goal progression and both stationary and non-stationary safety constraints. We show that ACQL outperforms existing methods across a range of continuous control tasks, including cases where prior methods fail to satisfy either goal-reaching or safety constraints. We further validate its real-world applicability by deploying ACQL on a 6-DOF robotic arm performing a goal-reaching task in a cluttered, cabinet-like space with safety constraints. Our results demonstrate that ACQL is a robust and scalable solution for learning robotic behaviors according to rich temporal specifications.
comment: 10 main content pages, 4 main content figures, 11 appendix pages, 5 appendix figures, camera ready version submitted to 39th Conference on Neural Information Processing Systems (NeurIPS 2025)
♻ ☆ SkillWrapper: Generative Predicate Invention for Skill Abstraction
Generalizing from individual skill executions to solving long-horizon tasks remains a core challenge in building autonomous agents. A promising direction is learning high-level, symbolic abstractions of the low-level skills of the agents, enabling reasoning and planning independent of the low-level state space. Among possible high-level representations, object-centric skill abstraction with symbolic predicates has been proven to be efficient because of its compatibility with domain-independent planners. Recent advances in foundation models have made it possible to generate symbolic predicates that operate on raw sensory inputs, a process we call generative predicate invention, to facilitate downstream abstraction learning. However, it remains unclear which formal properties the learned representations must satisfy, and how they can be learned to guarantee these properties. In this paper, we address both questions by presenting a formal theory of generative predicate invention for skill abstraction, resulting in symbolic operators that can be used for provably sound and complete planning. Within this framework, we propose SkillWrapper, a method that leverages foundation models to actively collect robot data and learn human-interpretable, plannable representations of black-box skills, using only RGB image observations. Our extensive empirical evaluation in simulation and on real robots shows that SkillWrapper learns abstract representations that enable solving unseen, long-horizon tasks in the real world with black-box skills.
♻ ☆ Adaptive Legged Locomotion via Online Learning for Model Predictive Control
We provide an algorithm for adaptive legged locomotion via online learning and model predictive control. The algorithm is composed of two interacting modules: model predictive control (MPC) and online learning of residual dynamics. The residual dynamics can represent modeling errors and external disturbances. We are motivated by the future of autonomy where quadrupeds will autonomously perform complex tasks despite real-world unknown uncertainty, such as unknown payload and uneven terrains. The algorithm uses random Fourier features to approximate the residual dynamics in reproducing kernel Hilbert spaces. Then, it employs MPC based on the current learned model of the residual dynamics. The model is updated online in a self-supervised manner using least squares based on the data collected while controlling the quadruped. The algorithm enjoys sublinear \textit{dynamic regret}, defined as the suboptimality against an optimal clairvoyant controller that knows how the residual dynamics. We validate our algorithm in Gazebo and MuJoCo simulations, where the quadruped aims to track reference trajectories. The Gazebo simulations include constant unknown external forces up to $12\boldsymbol{g}$, where $\boldsymbol{g}$ is the gravity vector, in flat terrain, slope terrain with $20\degree$ inclination, and rough terrain with $0.25m$ height variation. The MuJoCo simulations include time-varying unknown disturbances with payload up to $8~kg$ and time-varying ground friction coefficients in flat terrain.
comment: IEEE Robotics and Automation Letters
♻ ☆ APULSE: A Scalable Hybrid Algorithm for the RCSPP on Large-Scale Dense Graphs
The resource-constrained shortest path problem (RCSPP) is a fundamental NP-hard optimization challenge with broad applications, from network routing to autonomous navigation. This problem involves finding a path that minimizes a primary cost subject to a budget on a secondary resource. While various RCSPP solvers exist, they often face critical scalability limitations when applied to the large, dense graphs characteristic of complex, real-world scenarios, making them impractical for time-critical planning. This challenge is particularly acute in domains like mission planning for unmanned ground vehicles (UGVs), which demand solutions on large-scale terrain graphs. This paper introduces APULSE, a hybrid label-setting algorithm designed to efficiently solve the RCSPP on such challenging graphs. APULSE integrates a best-first search guided by an A* heuristic with aggressive, Pulse-style pruning mechanisms and a time-bucketing strategy for effective state-space reduction. A computational study, using a large-scale UGV planning scenario, benchmarks APULSE against state-of-the-art algorithms. The results demonstrate that APULSE consistently finds near-optimal solutions while being orders of magnitude faster and more robust, particularly on large problem instances where competing methods fail. This superior scalability establishes APULSE as an effective solution for RCSPP in complex, large-scale environments, enabling capabilities such as interactive decision support and dynamic replanning.
comment: This version corrects keywords and reference [9]. 9 pages
♻ ☆ SAD-Flower: Flow Matching for Safe, Admissible, and Dynamically Consistent Planning
Flow matching (FM) has shown promising results in data-driven planning. However, it inherently lacks formal guarantees for ensuring state and action constraints, whose satisfaction is a fundamental and crucial requirement for the safety and admissibility of planned trajectories on various systems. Moreover, existing FM planners do not ensure the dynamical consistency, which potentially renders trajectories inexecutable. We address these shortcomings by proposing SAD-Flower, a novel framework for generating Safe, Admissible, and Dynamically consistent trajectories. Our approach relies on an augmentation of the flow with a virtual control input. Thereby, principled guidance can be derived using techniques from nonlinear control theory, providing formal guarantees for state constraints, action constraints, and dynamic consistency. Crucially, SAD-Flower operates without retraining, enabling test-time satisfaction of unseen constraints. Through extensive experiments across several tasks, we demonstrate that SAD-Flower outperforms various generative-model-based baselines in ensuring constraint satisfaction.
Robotics 23
☆ Active Learning of Fractional-Order Viscoelastic Model Parameters for Realistic Haptic Rendering
Effective medical simulators necessitate realistic haptic rendering of biological tissues that display viscoelastic material properties, such as creep and stress relaxation. Fractional-order models provide an effective means of describing intrinsically time-dependent viscoelastic dynamics with few parameters, as these models can naturally capture memory effects. However, due to the unintuitive frequency-dependent coupling between the order of the fractional element and the other parameters, determining appropriate parameters for fractional-order models that yield high perceived realism remains a significant challenge. In this study, we propose a systematic means of determining the parameters of fractional-order viscoelastic models that optimizes the perceived realism of haptic rendering across general populations. First, we demonstrate that the parameters of fractional-order models can be effectively optimized through active learning, via qualitative feedback-based human-in-the-loop~(HiL) optimizations, to ensure consistently high realism ratings for each individual. Second, we propose a rigorous method to combine HiL optimization results to form an aggregate perceptual map trained on the entire dataset and demonstrate the selection of population-level optimal parameters from this representation that are broadly perceived as realistic across general populations. Finally, we provide evidence of the effectiveness of the generalized fractional-order viscoelastic model parameters by characterizing their perceived realism through human-subject experiments. Overall, generalized fractional-order viscoelastic models established through the proposed HiL optimization and aggregation approach possess the potential to significantly improve the sim-to-real transition performance of medical training simulators.
comment: This work has been submitted to the IEEE for possible publication. 14 pages, 9 figures
☆ Fast, Robust, Permutation-and-Sign Invariant SO(3) Pattern Alignment
We address the correspondence-free alignment of two rotation sets on \(SO(3)\), a core task in calibration and registration that is often impeded by missing time alignment, outliers, and unknown axis conventions. Our key idea is to decompose each rotation into its \emph{Transformed Basis Vectors} (TBVs)-three unit vectors on \(S^2\)-and align the resulting spherical point sets per axis using fast, robust matchers (SPMC, FRS, and a hybrid). To handle axis relabels and sign flips, we introduce a \emph{Permutation-and-Sign Invariant} (PASI) wrapper that enumerates the 24 proper signed permutations, scores them via summed correlations, and fuses the per-axis estimates into a single rotation by projection/Karcher mean. The overall complexity remains linear in the number of rotations (\(\mathcal{O}(n)\)), contrasting with \(\mathcal{O}(N_r^3\log N_r)\) for spherical/\(SO(3)\) correlation. Experiments on EuRoC Machine Hall simulations (axis-consistent) and the ETH Hand-Eye benchmark (\texttt{robot\_arm\_real}) (axis-ambiguous) show that our methods are accurate, 6-60x faster than traditional methods, and robust under extreme outlier ratios (up to 90\%), all without correspondence search.
☆ HAVEN: Hierarchical Adversary-aware Visibility-Enabled Navigation with Cover Utilization using Deep Transformer Q-Networks
Autonomous navigation in partially observable environments requires agents to reason beyond immediate sensor input, exploit occlusion, and ensure safety while progressing toward a goal. These challenges arise in many robotics domains, from urban driving and warehouse automation to defense and surveillance. Classical path planning approaches and memoryless reinforcement learning often fail under limited fields of view (FoVs) and occlusions, committing to unsafe or inefficient maneuvers. We propose a hierarchical navigation framework that integrates a Deep Transformer Q-Network (DTQN) as a high-level subgoal selector with a modular low-level controller for waypoint execution. The DTQN consumes short histories of task-aware features, encoding odometry, goal direction, obstacle proximity, and visibility cues, and outputs Q-values to rank candidate subgoals. Visibility-aware candidate generation introduces masking and exposure penalties, rewarding the use of cover and anticipatory safety. A low-level potential field controller then tracks the selected subgoal, ensuring smooth short-horizon obstacle avoidance. We validate our approach in 2D simulation and extend it directly to a 3D Unity-ROS environment by projecting point-cloud perception into the same feature schema, enabling transfer without architectural changes. Results show consistent improvements over classical planners and RL baselines in success rate, safety margins, and time to goal, with ablations confirming the value of temporal memory and visibility-aware candidate design. These findings highlight a generalizable framework for safe navigation under uncertainty, with broad relevance across robotic platforms.
☆ Describe Anything Anywhere At Any Moment
Computer vision and robotics applications ranging from augmented reality to robot autonomy in large-scale environments require spatio-temporal memory frameworks that capture both geometric structure for accurate language-grounding as well as semantic detail. Existing methods face a tradeoff, where producing rich open-vocabulary descriptions comes at the expense of real-time performance when these descriptions have to be grounded in 3D. To address these challenges, we propose Describe Anything, Anywhere, at Any Moment (DAAAM), a novel spatio-temporal memory framework for large-scale and real-time 4D scene understanding. DAAAM introduces a novel optimization-based frontend to infer detailed semantic descriptions from localized captioning models, such as the Describe Anything Model (DAM), leveraging batch processing to speed up inference by an order of magnitude for online processing. It leverages such semantic understanding to build a hierarchical 4D scene graph (SG), which acts as an effective globally spatially and temporally consistent memory representation. DAAAM constructs 4D SGs with detailed, geometrically grounded descriptions while maintaining real-time performance. We show that DAAAM's 4D SG interfaces well with a tool-calling agent for inference and reasoning. We thoroughly evaluate DAAAM in the complex task of spatio-temporal question answering on the NaVQA benchmark and show its generalization capabilities for sequential task grounding on the SG3D benchmark. We further curate an extended OC-NaVQA benchmark for large-scale and long-time evaluations. DAAAM achieves state-of-the-art results in both tasks, improving OC-NaVQA question accuracy by 53.6%, position errors by 21.9%, temporal errors by 21.6%, and SG3D task grounding accuracy by 27.8% over the most competitive baselines, respectively. We release our data and code open-source.
comment: 14 pages, 5 figures, 6 tables
☆ Image Generation as a Visual Planner for Robotic Manipulation CVPR 2026
Generating realistic robotic manipulation videos is an important step toward unifying perception, planning, and action in embodied agents. While existing video diffusion models require large domain-specific datasets and struggle to generalize, recent image generation models trained on language-image corpora exhibit strong compositionality, including the ability to synthesize temporally coherent grid images. This suggests a latent capacity for video-like generation even without explicit temporal modeling. We explore whether such models can serve as visual planners for robots when lightly adapted using LoRA finetuning. We propose a two-part framework that includes: (1) text-conditioned generation, which uses a language instruction and the first frame, and (2) trajectory-conditioned generation, which uses a 2D trajectory overlay and the same initial frame. Experiments on the Jaco Play dataset, Bridge V2, and the RT1 dataset show that both modes produce smooth, coherent robot videos aligned with their respective conditions. Our findings indicate that pretrained image generators encode transferable temporal priors and can function as video-like robotic planners under minimal supervision. Code is released at \href{https://github.com/pangye202264690373/Image-Generation-as-a-Visual-Planner-for-Robotic-Manipulation}{https://github.com/pangye202264690373/Image-Generation-as-a-Visual-Planner-for-Robotic-Manipulation}.
comment: 11 pages 9 figures Under review at CVPR 2026
☆ LAP: Fast LAtent Diffusion Planner with Fine-Grained Feature Distillation for Autonomous Driving
Diffusion models have demonstrated strong capabilities for modeling human-like driving behaviors in autonomous driving, but their iterative sampling process induces substantial latency, and operating directly on raw trajectory points forces the model to spend capacity on low-level kinematics, rather than high-level multi-modal semantics. To address these limitations, we propose LAtent Planner (LAP), a framework that plans in a VAE-learned latent space that disentangles high-level intents from low-level kinematics, enabling our planner to capture rich, multi-modal driving strategies. We further introduce a fine-grained feature distillation mechanism to guide a better interaction and fusion between the high-level semantic planning space and the vectorized scene context. Notably, LAP can produce high-quality plans in one single denoising step, substantially reducing computational overhead. Through extensive evaluations on the large-scale nuPlan benchmark, LAP achieves state-of-the-art closed-loop performance among learning-based planning methods, while demonstrating an inference speed-up of at most 10 times over previous SOTA approaches.
☆ Sample-Efficient Expert Query Control in Active Imitation Learning via Conformal Prediction
Active imitation learning (AIL) combats covariate shift by querying an expert during training. However, expert action labeling often dominates the cost, especially in GPU-intensive simulators, human-in-the-loop settings, and robot fleets that revisit near-duplicate states. We present Conformalized Rejection Sampling for Active Imitation Learning (CRSAIL), a querying rule that requests an expert action only when the visited state is under-represented in the expert-labeled dataset. CRSAIL scores state novelty by the distance to the $K$-th nearest expert state and sets a single global threshold via conformal prediction. This threshold is the empirical $(1-α)$ quantile of on-policy calibration scores, providing a distribution-free calibration rule that links $α$ to the expected query rate and makes $α$ a task-agnostic tuning knob. This state-space querying strategy is robust to outliers and, unlike safety-gate-based AIL, can be run without real-time expert takeovers: we roll out full trajectories (episodes) with the learner and only afterward query the expert on a subset of visited states. Evaluated on MuJoCo robotics tasks, CRSAIL matches or exceeds expert-level reward while reducing total expert queries by up to 96% vs. DAgger and up to 65% vs. prior AIL methods, with empirical robustness to $α$ and $K$, easing deployment on novel systems with unknown dynamics.
☆ Hardware-Software Collaborative Computing of Photonic Spiking Reinforcement Learning for Robotic Continuous Control
Robotic continuous control tasks impose stringent demands on the energy efficiency and latency of computing architectures due to their high-dimensional state spaces and real-time interaction requirements. Conventional electronic computing platforms face computational bottlenecks, whereas the fusion of photonic computing and spiking reinforcement learning (RL) offers a promising alternative. Here, we propose a novel computing architecture based on photonic spiking RL, which integrates the Twin Delayed Deep Deterministic policy gradient (TD3) algorithm with spiking neural network (SNN). The proposed architecture employs an optical-electronic hybrid computing paradigm wherein a silicon photonic Mach-Zehnder interferometer (MZI) chip executes linear matrix computations, while nonlinear spiking activations are performed in the electronic domain. Experimental validation on the Pendulum-v1 and HalfCheetah-v2 benchmarks demonstrates the system capability for software-hardware co-inference, achieving a control policy reward of 5831 on HalfCheetah-v2, a 23.33% reduction in convergence steps, and an action deviation below 2.2%. Notably, this work represents the first application of a programmable MZI photonic computing chip to robotic continuous control tasks, attaining an energy efficiency of 1.39 TOPS/W and an ultralow computational latency of 120 ps. Such performance underscores the promise of photonic spiking RL for real-time decision-making in autonomous and industrial robotic systems.
☆ Balancing Efficiency and Fairness: An Iterative Exchange Framework for Multi-UAV Cooperative Path Planning
Multi-UAV cooperative path planning (MUCPP) is a fundamental problem in multi-agent systems, aiming to generate collision-free trajectories for a team of unmanned aerial vehicles (UAVs) to complete distributed tasks efficiently. A key challenge lies in achieving both efficiency, by minimizing total mission cost, and fairness, by balancing the workload among UAVs to avoid overburdening individual agents. This paper presents a novel Iterative Exchange Framework for MUCPP, balancing efficiency and fairness through iterative task exchanges and path refinements. The proposed framework formulates a composite objective that combines the total mission distance and the makespan, and iteratively improves the solution via local exchanges under feasibility and safety constraints. For each UAV, collision-free trajectories are generated using A* search over a terrain-aware configuration space. Comprehensive experiments on multiple terrain datasets demonstrate that the proposed method consistently achieves superior trade-offs between total distance and makespan compared to existing baselines.
☆ DPNet: Doppler LiDAR Motion Planning for Highly-Dynamic Environments
Existing motion planning methods often struggle with rapid-motion obstacles due to an insufficient understanding of environmental changes. To address this limitation, we propose integrating motion planners with Doppler LiDARs which provide not only ranging measurements but also instantaneous point velocities. However, this integration is nontrivial due to the dual requirements of high accuracy and high frequency. To this end, we introduce Doppler Planning Network (DPNet), which tracks and reacts to rapid obstacles using Doppler model-based learning. Particularly, we first propose a Doppler Kalman neural network (D-KalmanNet) to track the future states of obstacles under partially observable Gaussian state space model. We then leverage the estimated motions to construct a Doppler-tuned model predictive control (DT-MPC) framework for ego-motion planning, enabling runtime auto-tuning of the controller parameters. These two model-based learners allow DPNet to maintain lightweight while learning fast environmental changes using minimum data, and achieve both high frequency and high accuracy in tracking and planning. Experiments on both high-fidelity simulator and real-world datasets demonstrate the superiority of DPNet over extensive benchmark schemes.
☆ MILE: A Mechanically Isomorphic Exoskeleton Data Collection System with Fingertip Visuotactile Sensing for Dexterous Manipulation
Imitation learning provides a promising approach to dexterous hand manipulation, but its effectiveness is limited by the lack of large-scale, high-fidelity data. Existing data-collection pipelines suffer from inaccurate motion retargeting, low data-collection efficiency, and missing high-resolution fingertip tactile sensing. We address this gap with MILE, a mechanically isomorphic teleoperation and data-collection system co-designed from human hand to exoskeleton to robotic hand. The exoskeleton is anthropometrically derived from the human hand, and the robotic hand preserves one-to-one joint-position isomorphism, eliminating nonlinear retargeting and enabling precise, natural control. The exoskeleton achieves a multi-joint mean absolute angular error below one degree, while the robotic hand integrates compact fingertip visuotactile modules that provide high-resolution tactile observations. Built on this retargeting-free interface, we teleoperate complex, contact-rich in-hand manipulation and efficiently collect a multimodal dataset comprising high-resolution fingertip visuotactile signals, RGB-D images, and joint positions. The teleoperation pipeline achieves a mean success rate improvement of 64%. Incorporating fingertip tactile observations further increases the success rate by an average of 25% over the vision-only baseline, validating the fidelity and utility of the dataset. Further details are available at: https://sites.google.com/view/mile-system.
☆ Data-Driven Modeling and Correction of Vehicle Dynamics
We develop a data-driven framework for learning and correcting non-autonomous vehicle dynamics. Physics-based vehicle models are often simplified for tractability and therefore exhibit inherent model-form uncertainty, motivating the need for data-driven correction. Moreover, non-autonomous dynamics are governed by time-dependent control inputs, which pose challenges in learning predictive models directly from temporal snapshot data. To address these, we reformulate the vehicle dynamics via a local parameterization of the time-dependent inputs, yielding a modified system composed of a sequence of local parametric dynamical systems. We approximate these parametric systems using two complementary approaches. First, we employ the DRIPS (dimension reduction and interpolation in parameter space) methodology to construct efficient linear surrogate models, equipped with lifted observable spaces and manifold-based operator interpolation. This enables data-efficient learning of vehicle models whose dynamics admit accurate linear representations in the lifted spaces. Second, for more strongly nonlinear systems, we employ FML (Flow Map Learning), a deep neural network approach that approximates the parametric evolution map without requiring special treatment of nonlinearities. We further extend FML with a transfer-learning-based model correction procedure, enabling the correction of misspecified prior models using only a sparse set of high-fidelity or experimental measurements, without assuming a prescribed form for the correction term. Through a suite of numerical experiments on unicycle, simplified bicycle, and slip-based bicycle models, we demonstrate that DRIPS offers robust and highly data-efficient learning of non-autonomous vehicle dynamics, while FML provides expressive nonlinear modeling and effective correction of model-form errors under severe data scarcity.
☆ RealAppliance: Let High-fidelity Appliance Assets Controllable and Workable as Aligned Real Manuals
Existing appliance assets suffer from poor rendering, incomplete mechanisms, and misalignment with manuals, leading to simulation-reality gaps that hinder appliance manipulation development. In this work, we introduce the RealAppliance dataset, comprising 100 high-fidelity appliances with complete physical, electronic mechanisms, and program logic aligned with their manuals. Based on these assets, we propose the RealAppliance-Bench benchmark, which evaluates multimodal large language models and embodied manipulation planning models across key tasks in appliance manipulation planning: manual page retrieval, appliance part grounding, open-loop manipulation planning, and closed-loop planning adjustment. Our analysis of model performances on RealAppliance-Bench provides insights for advancing appliance manipulation research
☆ Dependent Reachable Sets for the Constant Bearing Pursuit Strategy
This paper introduces a novel reachability problem for the scenario where one agent follows another agent using the constant bearing pursuit strategy, and analyzes the geometry of the reachable set of the follower. Key theoretical results are derived, providing bounds for the associated dependent reachable set. Simulation results are presented to empirically establish the shape of the dependent reachable set. In the process, an original optimization problem for the constant bearing strategy is formulated and analyzed.
comment: This work has been submitted to a journal for possible publication
☆ "Why the face?": Exploring Robot Error Detection Using Instrumented Bystander Reactions
How do humans recognize and rectify social missteps? We achieve social competence by looking around at our peers, decoding subtle cues from bystanders - a raised eyebrow, a laugh - to evaluate the environment and our actions. Robots, however, struggle to perceive and make use of these nuanced reactions. By employing a novel neck-mounted device that records facial expressions from the chin region, we explore the potential of previously untapped data to capture and interpret human responses to robot error. First, we develop NeckNet-18, a 3D facial reconstruction model to map the reactions captured through the chin camera onto facial points and head motion. We then use these facial responses to develop a robot error detection model which outperforms standard methodologies such as using OpenFace or video data, generalizing well especially for within-participant data. Through this work, we argue for expanding human-in-the-loop robot sensing, fostering more seamless integration of robots into diverse human environments, pushing the boundaries of social cue detection and opening new avenues for adaptable robotics.
♻ ☆ RoboArena: Distributed Real-World Evaluation of Generalist Robot Policies
Comprehensive, unbiased, and comparable evaluation of modern generalist policies is uniquely challenging: existing approaches for robot benchmarking typically rely on heavy standardization, either by specifying fixed evaluation tasks and environments, or by hosting centralized ''robot challenges'', and do not readily scale to evaluating generalist policies across a broad range of tasks and environments. In this work, we propose RoboArena, a new approach for scalable evaluation of generalist robot policies in the real world. Instead of standardizing evaluations around fixed tasks, environments, or locations, we propose to crowd-source evaluations across a distributed network of evaluators. Importantly, evaluators can freely choose the tasks and environments they evaluate on, enabling easy scaling of diversity, but they are required to perform double-blind evaluations over pairs of policies. Then, by aggregating preference feedback from pairwise comparisons across diverse tasks and environments, we can derive a ranking of policies. We instantiate our approach across a network of evaluators at seven academic institutions using the DROID robot platform. Through more than 600 pairwise real-robot evaluation episodes across seven generalist policies, we demonstrate that our crowd-sourced approach can more accurately rank the performance of existing generalist policies than conventional, centralized evaluation approaches, while being more scalable, resilient, and trustworthy. We open our evaluation network to the community and hope that it can enable more accessible comparisons of generalist robot policies.
comment: Website: https://robo-arena.github.io/
♻ ☆ Ensuring Force Safety in Vision-Guided Robotic Manipulation via Implicit Tactile Calibration
In dynamic environments, robots often encounter constrained movement trajectories when manipulating objects with specific properties, such as doors. Therefore, applying the appropriate force is crucial to prevent damage to both the robots and the objects. However, current vision-guided robot state generation methods often falter in this regard, as they lack the integration of tactile perception. To tackle this issue, this paper introduces a novel state diffusion framework termed SafeDiff. It generates a prospective state sequence from the current robot state and visual context observation while incorporating real-time tactile feedback to refine the sequence. As far as we know, this is the first study specifically focused on ensuring force safety in robotic manipulation. It significantly enhances the rationality of state planning, and the safe action trajectory is derived from inverse dynamics based on this refined planning. In practice, unlike previous approaches that concatenate visual and tactile data to generate future robot state sequences, our method employs tactile data as a calibration signal to adjust the robot's state within the state space implicitly. Additionally, we've developed a large-scale simulation dataset called SafeDoorManip50k, offering extensive multimodal data to train and evaluate the proposed method. Extensive experiments show that our visual-tactile model substantially mitigates the risk of harmful forces in the door opening, across both simulated and real-world settings.
comment: Website URL: see https://i-am-future.github.io/safediff/
♻ ☆ A Minimal Subset Approach for Informed Keyframe Sampling in Large-Scale SLAM
Typical LiDAR SLAM architectures feature a front-end for odometry estimation and a back-end for refining and optimizing the trajectory and map, commonly through loop closures. However, loop closure detection in large-scale missions presents significant computational challenges due to the need to identify, verify, and process numerous candidate pairs for pose graph optimization. Keyframe sampling bridges the front-end and back-end by selecting frames for storing and processing during global optimization. This article proposes an online keyframe sampling approach that constructs the pose graph using the most impactful keyframes for loop closure. We introduce the Minimal Subset Approach (MSA), which optimizes two key objectives: redundancy minimization and information preservation, implemented within a sliding window framework. By operating in the feature space rather than 3-D space, MSA efficiently reduces redundant keyframes while retaining essential information. Evaluations on diverse public datasets show that the proposed approach outperforms naive methods in reducing false positive rates in place recognition, while delivering superior ATE and RPE in metric localization, without the need for manual parameter tuning. Additionally, MSA demonstrates efficiency and scalability by reducing memory usage and computational overhead during loop closure detection and pose graph optimization.
comment: Please cite the published version. 8 pages, 9 figures
♻ ☆ Have We Scene It All? Scene Graph-Aware Deep Point Cloud Compression
Efficient transmission of 3D point cloud data is critical for advanced perception in centralized and decentralized multi-agent robotic systems, especially nowadays with the growing reliance on edge and cloud-based processing. However, the large and complex nature of point clouds creates challenges under bandwidth constraints and intermittent connectivity, often degrading system performance. We propose a deep compression framework based on semantic scene graphs. The method decomposes point clouds into semantically coherent patches and encodes them into compact latent representations with semantic-aware encoders conditioned by Feature-wise Linear Modulation (FiLM). A folding-based decoder, guided by latent features and graph node attributes, enables structurally accurate reconstruction. Experiments on the SemanticKITTI and nuScenes datasets show that the framework achieves state-of-the-art compression rates, reducing data size by up to 98% while preserving both structural and semantic fidelity. In addition, it supports downstream applications such as multi-robot pose graph optimization and map merging, achieving trajectory accuracy and map alignment comparable to those obtained with raw LiDAR scans.
comment: Please cite published version. 8 pages, 6 figures
♻ ☆ Adversarial Exploitation of Data Diversity Improves Visual Localization
Visual localization, which estimates a camera's pose within a known scene, is a fundamental capability for autonomous systems. While absolute pose regression (APR) methods have shown promise for efficient inference, they often struggle with generalization. Recent approaches attempt to address this through data augmentation with varied viewpoints, yet they overlook a critical factor: appearance diversity. In this work, we identify appearance variation as the key to robust localization. Specifically, we first lift real 2D images into 3D Gaussian Splats with varying appearance and deblurring ability, enabling the synthesis of diverse training data that varies not just in poses but also in environmental conditions such as lighting and weather. To fully unleash the potential of the appearance-diverse data, we build a two-branch joint training pipeline with an adversarial discriminator to bridge the syn-to-real gap. Extensive experiments demonstrate that our approach significantly outperforms state-of-the-art methods, reducing translation and rotation errors by 50\% and 41\% on indoor datasets, and 38\% and 44\% on outdoor datasets. Most notably, our method shows remarkable robustness in dynamic driving scenarios under varying weather conditions and in day-to-night scenarios, where previous APR methods fail. Project Page: https://ai4ce.github.io/RAP/
comment: 24 pages, 22 figures
♻ ☆ Towards Fully Onboard State Estimation and Trajectory Tracking for UAVs with Suspended Payloads
This paper addresses the problem of tracking the position of a cable-suspended payload carried by an unmanned aerial vehicle, with a focus on real-world deployment and minimal hardware requirements. In contrast to many existing approaches that rely on motion-capture systems, additional onboard cameras, or instrumented payloads, we propose a framework that uses only standard onboard sensors--specifically, real-time kinematic global navigation satellite system measurements and data from the onboard inertial measurement unit--to estimate and control the payload's position. The system models the full coupled dynamics of the aerial vehicle and payload, and integrates a linear Kalman filter for state estimation, a model predictive contouring control planner, and an incremental model predictive controller. The control architecture is designed to remain effective despite sensing limitations and estimation uncertainty. Extensive simulations demonstrate that the proposed system achieves performance comparable to control based on ground-truth measurements, with only minor degradation (< 6%). The system also shows strong robustness to variations in payload parameters. Field experiments further validate the framework, confirming its practical applicability and reliable performance in outdoor environments using only off-the-shelf aerial vehicle hardware.
comment: Updated to match the published version. Added journal reference and DOI
♻ ☆ Delta-Triplane Transformers as Occupancy World Models
Occupancy World Models (OWMs) aim to predict future scenes via 3D voxelized representations of the environment to support intelligent motion planning. Existing approaches typically generate full future occupancy states from VAE-style latent encodings, which can be computationally expensive and redundant. We propose Delta-Triplane Transformers (DTT), a novel 4D OWM for autonomous driving, that introduces two key innovations: (1) a triplane based representation that encodes 3D occupancy more compactly than previous approaches, and (2) an incremental prediction strategy for OWM that models {\em changes} in occupancy rather than dealing with full states. The core insight is that changes in the compact 3D latent space are naturally sparser and easier to model, enabling higher accuracy with a lighter-weight architecture. Building on this representation, DTT extracts multi-scale motion features from historical data and iteratively predict future triplane deltas. These deltas are combined with past states to decode future occupancy and ego-motion trajectories. Extensive experiments demonstrate that DTT delivers a 1.44$\times$ speedup (26 FPS) over the state of the art, improves mean IoU to 30.85, and reduces the mean absolute planning error to 1.0 meters. Demo videos are provided in the supplementary material.
♻ ☆ A Neuro-inspired Theory of Joint Human-Swarm Interaction ICRA
Human-swarm interaction (HSI) is an active research challenge in the realms of swarm robotics and human-factors engineering. Here we apply a cognitive systems engineering perspective and introduce a neuro-inspired joint systems theory of HSI. The mindset defines predictions for adaptive, robust and scalable HSI dynamics and therefore has the potential to inform human-swarm loop design.
comment: ICRA Workshop on Human-Swarm Interaction 2020
Robotics 40
☆ From CAD to POMDP: Probabilistic Planning for Robotic Disassembly of End-of-Life Products
To support the circular economy, robotic systems must not only assemble new products but also disassemble end-of-life (EOL) ones for reuse, recycling, or safe disposal. Existing approaches to disassembly sequence planning often assume deterministic and fully observable product models, yet real EOL products frequently deviate from their initial designs due to wear, corrosion, or undocumented repairs. We argue that disassembly should therefore be formulated as a Partially Observable Markov Decision Process (POMDP), which naturally captures uncertainty about the product's internal state. We present a mathematical formulation of disassembly as a POMDP, in which hidden variables represent uncertain structural or physical properties. Building on this formulation, we propose a task and motion planning framework that automatically derives specific POMDP models from CAD data, robot capabilities, and inspection results. To obtain tractable policies, we approximate this formulation with a reinforcement-learning approach that operates on stochastic action outcomes informed by inspection priors, while a Bayesian filter continuously maintains beliefs over latent EOL conditions during execution. Using three products on two robotic systems, we demonstrate that this probabilistic planning framework outperforms deterministic baselines in terms of average disassembly time and variance, generalizes across different robot setups, and successfully adapts to deviations from the CAD model, such as missing or stuck parts.
☆ Design, modelling and experimental validation of bipenniform shape memory alloy-based linear actuator integrable with hydraulic stroke amplification mechanism
The increasing industrial demand for alternative actuators over conventional electromagnetism-based systems having limited efficiency, bulky size, complex design due to in-built gear-train mechanisms, and high production and amortization costs necessitates the innovation in new actuator development. Integrating bio-inspired design principles into linear actuators could bring forth the next generation of adaptive and energy efficient smart material-based actuation systems. The present study amalgamates the advantages of bipenniform architecture, which generates high force in the given physiological region and a high power-to-weight ratio of shape memory alloy (SMA), into a novel bio-inspired SMA-based linear actuator. A mathematical model of a multi-layered bipenniform configuration-based SMA actuator was developed and validated experimentally. The current research also caters to the incorporation of failure mitigation strategies using design failure mode and effects analysis along with the experimental assessment of the performance of the developed actuator. The system has been benchmarked against an industry-developed stepper motor-driven actuator. It has shown promising results generating an actuation force of 257 N with 15 V input voltage, meeting the acceptable range for actuation operation. It further exhibits about 67% reduction in the weight of the drive mechanism, with 80% lesser component, 32% cost reduction, and 19% energy savings and similar envelope dimensions for assembly compatibility with dampers and louvers for easy onsite deployment. The study introduces SMA coil-based actuator as an advanced design that can be deployed for high force-high stroke applications. The bio-inspired SMA-based linear actuator has applications ranging from building automation controls to lightweight actuation systems for space robotics and medical prosthesis.
☆ SimScale: Learning to Drive via Real-World Simulation at Scale
Achieving fully autonomous driving systems requires learning rational decisions in a wide span of scenarios, including safety-critical and out-of-distribution ones. However, such cases are underrepresented in real-world corpus collected by human experts. To complement for the lack of data diversity, we introduce a novel and scalable simulation framework capable of synthesizing massive unseen states upon existing driving logs. Our pipeline utilizes advanced neural rendering with a reactive environment to generate high-fidelity multi-view observations controlled by the perturbed ego trajectory. Furthermore, we develop a pseudo-expert trajectory generation mechanism for these newly simulated states to provide action supervision. Upon the synthesized data, we find that a simple co-training strategy on both real-world and simulated samples can lead to significant improvements in both robustness and generalization for various planning methods on challenging real-world benchmarks, up to +6.8 EPDMS on navhard and +2.9 on navtest. More importantly, such policy improvement scales smoothly by increasing simulation data only, even without extra real-world data streaming in. We further reveal several crucial findings of such a sim-real learning system, which we term SimScale, including the design of pseudo-experts and the scaling properties for different policy architectures. Our simulation data and code would be released.
comment: Project page: https://opendrivelab.com/SimScale
☆ SafeHumanoid: VLM-RAG-driven Control of Upper Body Impedance for Humanoid Robot
Safe and trustworthy Human Robot Interaction (HRI) requires robots not only to complete tasks but also to regulate impedance and speed according to scene context and human proximity. We present SafeHumanoid, an egocentric vision pipeline that links Vision Language Models (VLMs) with Retrieval-Augmented Generation (RAG) to schedule impedance and velocity parameters for a humanoid robot. Egocentric frames are processed by a structured VLM prompt, embedded and matched against a curated database of validated scenarios, and mapped to joint-level impedance commands via inverse kinematics. We evaluate the system on tabletop manipulation tasks with and without human presence, including wiping, object handovers, and liquid pouring. The results show that the pipeline adapts stiffness, damping, and speed profiles in a context-aware manner, maintaining task success while improving safety. Although current inference latency (up to 1.4 s) limits responsiveness in highly dynamic settings, SafeHumanoid demonstrates that semantic grounding of impedance control is a viable path toward safer, standard-compliant humanoid collaboration.
☆ Incorporating Ephemeral Traffic Waves in A Data-Driven Framework for Microsimulation in CARLA
This paper introduces a data-driven traffic microsimulation framework in CARLA that reconstructs real-world wave dynamics using high-fidelity time-space data from the I-24 MOTION testbed. Calibration of road networks in microsimulators to reproduce ephemeral phenomena such as traffic waves for large-scale simulation is a process that is fraught with challenges. This work reconsiders the existence of the traffic state data as boundary conditions on an ego vehicle moving through previously recorded traffic data, rather than reproducing those traffic phenomena in a calibrated microsim. Our approach is to autogenerate a 1 mile highway segment corresponding to I-24, and use the I-24 data to power a cosimulation module that injects traffic information into the simulation. The CARLA and cosimulation simulations are centered around an ego vehicle sampled from the empirical data, with autogeneration of "visible" traffic within the longitudinal range of the ego vehicle. Boundary control beyond these visible ranges is achieved using ghost cells behind (upstream) and ahead (downstream) of the ego vehicle. Unlike prior simulation work that focuses on local car-following behavior or abstract geometries, our framework targets full time-space diagram fidelity as the validation objective. Leveraging CARLA's rich sensor suite and configurable vehicle dynamics, we simulate wave formation and dissipation in both low-congestion and high-congestion scenarios for qualitative analysis. The resulting emergent behavior closely mirrors that of real traffic, providing a novel cosimulation framework for evaluating traffic control strategies, perception-driven autonomy, and future deployment of wave mitigation solutions. Our work bridges microscopic modeling with physical experimental data, enabling the first perceptually realistic, boundary-driven simulation of empirical traffic wave phenomena in CARLA.
comment: Submitted to IEEE IV 2026
☆ Field-programmable dynamics in a soft magnetic actuator enabling true random number generation and reservoir computing
Complex and even chaotic dynamics, though prevalent in many natural and engineered systems, has been largely avoided in the design of electromechanical systems due to concerns about wear and controlability. Here, we demonstrate that complex dynamics might be particularly advantageous in soft robotics, offering new functionalities beyond motion not easily achievable with traditional actuation methods. We designed and realized resilient magnetic soft actuators capable of operating in a tunable dynamic regime for tens of thousands cycles without fatigue. We experimentally demonstrated the application of these actuators for true random number generation and stochastic computing. {W}e validate soft robots as physical reservoirs capable of performing Mackey--Glass time series prediction. These findings show that exploring the complex dynamics in soft robotics would extend the application scenarios in soft computing, human-robot interaction and collaborative robots as we demonstrate with biomimetic blinking and randomized voice modulation.
☆ Fault-Tolerant MARL for CAVs under Observation Perturbations for Highway On-Ramp Merging
Multi-Agent Reinforcement Learning (MARL) holds significant promise for enabling cooperative driving among Connected and Automated Vehicles (CAVs). However, its practical application is hindered by a critical limitation, i.e., insufficient fault tolerance against observational faults. Such faults, which appear as perturbations in the vehicles' perceived data, can substantially compromise the performance of MARL-based driving systems. Addressing this problem presents two primary challenges. One is to generate adversarial perturbations that effectively stress the policy during training, and the other is to equip vehicles with the capability to mitigate the impact of corrupted observations. To overcome the challenges, we propose a fault-tolerant MARL method for cooperative on-ramp vehicles incorporating two key agents. First, an adversarial fault injection agent is co-trained to generate perturbations that actively challenge and harden the vehicle policies. Second, we design a novel fault-tolerant vehicle agent equipped with a self-diagnosis capability, which leverages the inherent spatio-temporal correlations in vehicle state sequences to detect faults and reconstruct credible observations, thereby shielding the policy from misleading inputs. Experiments in a simulated highway merging scenario demonstrate that our method significantly outperforms baseline MARL approaches, achieving near-fault-free levels of safety and efficiency under various observation fault patterns.
☆ Obstruction reasoning for robotic grasping
Successful robotic grasping in cluttered environments not only requires a model to visually ground a target object but also to reason about obstructions that must be cleared beforehand. While current vision-language embodied reasoning models show emergent spatial understanding, they remain limited in terms of obstruction reasoning and accessibility planning. To bridge this gap, we present UNOGrasp, a learning-based vision-language model capable of performing visually-grounded obstruction reasoning to infer the sequence of actions needed to unobstruct the path and grasp the target object. We devise a novel multi-step reasoning process based on obstruction paths originated by the target object. We anchor each reasoning step with obstruction-aware visual cues to incentivize reasoning capability. UNOGrasp combines supervised and reinforcement finetuning through verifiable reasoning rewards. Moreover, we construct UNOBench, a large-scale dataset for both training and benchmarking, based on MetaGraspNetV2, with over 100k obstruction paths annotated by humans with obstruction ratios, contact points, and natural-language instructions. Extensive experiments and real-robot evaluations show that UNOGrasp significantly improves obstruction reasoning and grasp success across both synthetic and real-world environments, outperforming generalist and proprietary alternatives. Project website: https://tev-fbk.github.io/UnoGrasp/.
☆ Automated Generation of MDPs Using Logic Programming and LLMs for Robotic Applications
We present a novel framework that integrates Large Language Models (LLMs) with automated planning and formal verification to streamline the creation and use of Markov Decision Processes (MDP). Our system leverages LLMs to extract structured knowledge in the form of a Prolog knowledge base from natural language (NL) descriptions. It then automatically constructs an MDP through reachability analysis, and synthesises optimal policies using the Storm model checker. The resulting policy is exported as a state-action table for execution. We validate the framework in three human-robot interaction scenarios, demonstrating its ability to produce executable policies with minimal manual effort. This work highlights the potential of combining language models with formal methods to enable more accessible and scalable probabilistic planning in robotics.
comment: 9 pages, 11 figures, 2 tables, 2 algorithms, accepted for publication in IEEE Robotics and Automation Letters
☆ LatBot: Distilling Universal Latent Actions for Vision-Language-Action Models
Learning transferable latent actions from large-scale object manipulation videos can significantly enhance generalization in downstream robotics tasks, as such representations are agnostic to different robot embodiments. Existing approaches primarily rely on visual reconstruction objectives while neglecting physical priors, leading to sub-optimal performance in learning universal representations. To address these challenges, we propose a Universal Latent Action Learning framework that takes task instructions and multiple frames as inputs, and optimizes both future frame reconstruction and action sequence prediction. Unlike prior works, incorporating action predictions (e.g., gripper or hand trajectories and orientations) allows the model to capture richer physical priors such as real-world distances and orientations, thereby enabling seamless transferability to downstream tasks. We further decompose the latent actions into learnable motion and scene tokens to distinguish the robot's active movements from environmental changes, thus filtering out irrelevant dynamics. By distilling the learned latent actions into the latest VLA models, we achieve strong performance across both simulated (SIMPLER and LIBERO) and real-world robot settings. Notably, with only 10 real-world trajectories per task collected on a Franka robot, our approach successfully completes all five challenging tasks, demonstrating strong few-shot transferability in robotic manipulation.
comment: Project Page: https://mm-robot.github.io/distill_latent_action/
☆ DiskChunGS: Large-Scale 3D Gaussian SLAM Through Chunk-Based Memory Management
Recent advances in 3D Gaussian Splatting (3DGS) have demonstrated impressive results for novel view synthesis with real-time rendering capabilities. However, integrating 3DGS with SLAM systems faces a fundamental scalability limitation: methods are constrained by GPU memory capacity, restricting reconstruction to small-scale environments. We present DiskChunGS, a scalable 3DGS SLAM system that overcomes this bottleneck through an out-of-core approach that partitions scenes into spatial chunks and maintains only active regions in GPU memory while storing inactive areas on disk. Our architecture integrates seamlessly with existing SLAM frameworks for pose estimation and loop closure, enabling globally consistent reconstruction at scale. We validate DiskChunGS on indoor scenes (Replica, TUM-RGBD), urban driving scenarios (KITTI), and resource-constrained Nvidia Jetson platforms. Our method uniquely completes all 11 KITTI sequences without memory failures while achieving superior visual quality, demonstrating that algorithmic innovation can overcome the memory constraints that have limited previous 3DGS SLAM methods.
☆ Control Barrier Function for Unknown Systems: An Approximation-free Approach
We study the prescribed-time reach-avoid (PT-RA) control problem for nonlinear systems with unknown dynamics operating in environments with moving obstacles. Unlike robust or learning based Control Barrier Function (CBF) methods, the proposed framework requires neither online model learning nor uncertainty bound estimation. A CBF-based Quadratic Program (CBF-QP) is solved on a simple virtual system to generate a safe reference satisfying PT-RA conditions with respect to time-varying, tightened obstacle and goal sets. The true system is confined to a Virtual Confinement Zone (VCZ) around this reference using an approximation-free feedback law. This construction guarantees real-time safety and prescribed-time target reachability under unknown dynamics and dynamic constraints without explicit model identification or offline precomputation. Simulation results illustrate reliable dynamic obstacle avoidance and timely convergence to the target set.
☆ Adaptive Factor Graph-Based Tightly Coupled GNSS/IMU Fusion for Robust Positionin
Reliable positioning in GNSS-challenged environments remains a critical challenge for navigation systems. Tightly coupled GNSS/IMU fusion improves robustness but remains vulnerable to non-Gaussian noise and outliers. We present a robust and adaptive factor graph-based fusion framework that directly integrates GNSS pseudorange measurements with IMU preintegration factors and incorporates the Barron loss, a general robust loss function that unifies several m-estimators through a single tunable parameter. By adaptively down weighting unreliable GNSS measurements, our approach improves resilience positioning. The method is implemented in an extended GTSAM framework and evaluated on the UrbanNav dataset. The proposed solution reduces positioning errors by up to 41% relative to standard FGO, and achieves even larger improvements over extended Kalman filter (EKF) baselines in urban canyon environments. These results highlight the benefits of Barron loss in enhancing the resilience of GNSS/IMU-based navigation in urban and signal-compromised environments.
☆ MrGS: Multi-modal Radiance Fields with 3D Gaussian Splatting for RGB-Thermal Novel View Synthesis ICRA 2025
Recent advances in Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS) have achieved considerable performance in RGB scene reconstruction. However, multi-modal rendering that incorporates thermal infrared imagery remains largely underexplored. Existing approaches tend to neglect distinctive thermal characteristics, such as heat conduction and the Lambertian property. In this study, we introduce MrGS, a multi-modal radiance field based on 3DGS that simultaneously reconstructs both RGB and thermal 3D scenes. Specifically, MrGS derives RGB- and thermal-related information from a single appearance feature through orthogonal feature extraction and employs view-dependent or view-independent embedding strategies depending on the degree of Lambertian reflectance exhibited by each modality. Furthermore, we leverage two physics-based principles to effectively model thermal-domain phenomena. First, we integrate Fourier's law of heat conduction prior to alpha blending to model intensity interpolation caused by thermal conduction between neighboring Gaussians. Second, we apply the Stefan-Boltzmann law and the inverse-square law to formulate a depth-aware thermal radiation map that imposes additional geometric constraints on thermal rendering. Experimental results demonstrate that the proposed MrGS achieves high-fidelity RGB-T scene reconstruction while reducing the number of Gaussians.
comment: Accepted at Thermal Infrared in Robotics (TIRO) Workshop, ICRA 2025 (Best Poster Award)
☆ Analytical Inverse Kinematic Solution for "Moz1" NonSRS 7-DOF Robot arm with novel arm angle
This paper presents an analytical solution to the inverse kinematic problem(IKP) for the seven degree-of-freedom (7-DOF) Moz1 Robot Arm with offsets on wrist. We provide closed-form solutions with the novel arm angle . it allow fully self-motion and solve the problem of algorithmic singularities within the workspace. It also provides information on how the redundancy is resolved in a new arm angle representation where traditional SEW angle faied to be defined and how singularities are handled. The solution is simple, fast and exact, providing full solution space (i.e. all 16 solutions) per pose.
☆ Commanding Humanoid by Free-form Language: A Large Language Action Model with Unified Motion Vocabulary
Enabling humanoid robots to follow free-form language commands is critical for seamless human-robot interaction, collaborative task execution, and general-purpose embodied intelligence. While recent advances have improved low-level humanoid locomotion and robot manipulation, language-conditioned whole-body control remains a significant challenge. Existing methods are often limited to simple instructions and sacrifice either motion diversity or physical plausibility. To address this, we introduce Humanoid-LLA, a Large Language Action Model that maps expressive language commands to physically executable whole-body actions for humanoid robots. Our approach integrates three core components: a unified motion vocabulary that aligns human and humanoid motion primitives into a shared discrete space; a vocabulary-directed controller distilled from a privileged policy to ensure physical feasibility; and a physics-informed fine-tuning stage using reinforcement learning with dynamics-aware rewards to enhance robustness and stability. Extensive evaluations in simulation and on a real-world Unitree G1 humanoid show that Humanoid-LLA delivers strong language generalization while maintaining high physical fidelity, outperforming existing language-conditioned controllers in motion naturalness, stability, and execution success rate.
comment: Project page: https://humanoidlla.github.io/
☆ RobotSeg: A Model and Dataset for Segmenting Robots in Image and Video
Accurate robot segmentation is a fundamental capability for robotic perception. It enables precise visual servoing for VLA systems, scalable robot-centric data augmentation, accurate real-to-sim transfer, and reliable safety monitoring in dynamic human-robot environments. Despite the strong capabilities of modern segmentation models, surprisingly it remains challenging to segment robots. This is due to robot embodiment diversity, appearance ambiguity, structural complexity, and rapid shape changes. Embracing these challenges, we introduce RobotSeg, a foundation model for robot segmentation in image and video. RobotSeg is built upon the versatile SAM 2 foundation model but addresses its three limitations for robot segmentation, namely the lack of adaptation to articulated robots, reliance on manual prompts, and the need for per-frame training mask annotations, by introducing a structure-enhanced memory associator, a robot prompt generator, and a label-efficient training strategy. These innovations collectively enable a structure-aware, automatic, and label-efficient solution. We further construct the video robot segmentation (VRS) dataset comprising over 2.8k videos (138k frames) with diverse robot embodiments and environments. Extensive experiments demonstrate that RobotSeg achieves state-of-the-art performance on both images and videos, establishing a strong foundation for future advances in robot perception.
comment: Project page: https://github.com/showlab/RobotSeg
☆ Seeing before Observable: Potential Risk Reasoning in Autonomous Driving via Vision Language Models
Ensuring safety remains a key challenge for autonomous vehicles (AVs), especially in rare and complex scenarios. One critical but understudied aspect is the \textbf{potential risk} situations, where the risk is \textbf{not yet observable} but can be inferred from subtle precursors, such as anomalous behaviors or commonsense violations. Recognizing these precursors requires strong semantic understanding and reasoning capabilities, which are often absent in current AV systems due to the scarcity of such cases in existing driving or risk-centric datasets. Moreover, current autonomous driving accident datasets often lack annotations of the causal reasoning chains behind incidents, which are essential for identifying potential risks before they become observable. To address these gaps, we introduce PotentialRiskQA, a novel vision-language dataset designed for reasoning about potential risks prior to observation. Each sample is annotated with structured scene descriptions, semantic precursors, and inferred risk outcomes. Based on this dataset, we further propose PR-Reasoner, a vision-language-model-based framework tailored for onboard potential risk reasoning. Experimental results show that fine-tuning on PotentialRiskQA enables PR-Reasoner to significantly enhance its performance on the potential risk reasoning task compared to baseline VLMs. Together, our dataset and model provide a foundation for developing autonomous systems with improved foresight and proactive safety capabilities, moving toward more intelligent and resilient AVs.
☆ SUPER-AD: Semantic Uncertainty-aware Planning for End-to-End Robust Autonomous Driving
End-to-End (E2E) planning has become a powerful paradigm for autonomous driving, yet current systems remain fundamentally uncertainty-blind. They assume perception outputs are fully reliable, even in ambiguous or poorly observed scenes, leaving the planner without an explicit measure of uncertainty. To address this limitation, we propose a camera-only E2E framework that estimates aleatoric uncertainty directly in BEV space and incorporates it into planning. Our method produces a dense, uncertainty-aware drivability map that captures both semantic structure and geometric layout at pixel-level resolution. To further promote safe and rule-compliant behavior, we introduce a lane-following regularization that encodes lane structure and traffic norms. This prior stabilizes trajectory planning under normal conditions while preserving the flexibility needed for maneuvers such as overtaking or lane changes. Together, these components enable robust and interpretable trajectory planning, even under challenging uncertainty conditions. Evaluated on the NAVSIM benchmark, our method achieves state-of-the-art performance, delivering substantial gains on both the challenging NAVHARD and NAVSAFE subsets. These results demonstrate that our principled aleatoric uncertainty modeling combined with driving priors significantly advances the safety and reliability of camera-only E2E autonomous driving.
☆ MARVO: Marine-Adaptive Radiance-aware Visual Odometry CVPR2026
Underwater visual localization remains challenging due to wavelength-dependent attenuation, poor texture, and non-Gaussian sensor noise. We introduce MARVO, a physics-aware, learning-integrated odometry framework that fuses underwater image formation modeling, differentiable matching, and reinforcement-learning optimization. At the front-end, we extend transformer-based feature matcher with a Physics Aware Radiance Adapter that compensates for color channel attenuation and contrast loss, yielding geometrically consistent feature correspondences under turbidity. These semi dense matches are combined with inertial and pressure measurements inside a factor-graph backend, where we formulate a keyframe-based visual-inertial-barometric estimator using GTSAM library. Each keyframe introduces (i) Pre-integrated IMU motion factors, (ii) MARVO-derived visual pose factors, and (iii) barometric depth priors, giving a full-state MAP estimate in real time. Lastly, we introduce a Reinforcement-Learningbased Pose-Graph Optimizer that refines global trajectories beyond local minima of classical least-squares solvers by learning optimal retraction actions on SE(2).
comment: 10 pages, 5 figures, 3 tables, Submitted to CVPR2026
☆ Threat-Aware UAV Dodging of Human-Thrown Projectiles with an RGB-D Camera
Uncrewed aerial vehicles (UAVs) performing tasks such as transportation and aerial photography are vulnerable to intentional projectile attacks from humans. Dodging such a sudden and fast projectile poses a significant challenge for UAVs, requiring ultra-low latency responses and agile maneuvers. Drawing inspiration from baseball, in which pitchers' body movements are analyzed to predict the ball's trajectory, we propose a novel real-time dodging system that leverages an RGB-D camera. Our approach integrates human pose estimation with depth information to predict the attacker's motion trajectory and the subsequent projectile trajectory. Additionally, we introduce an uncertainty-aware dodging strategy to enable the UAV to dodge incoming projectiles efficiently. Our perception system achieves high prediction accuracy and outperforms the baseline in effective distance and latency. The dodging strategy addresses temporal and spatial uncertainties to ensure UAV safety. Extensive real-world experiments demonstrate the framework's reliable dodging capabilities against sudden attacks and its outstanding robustness across diverse scenarios.
☆ Safe Autonomous Lane Changing: Planning with Dynamic Risk Fields and Time-Varying Convex Space Generation
This paper presents a novel trajectory planning pipeline for complex driving scenarios like autonomous lane changing, by integrating risk-aware planning with guaranteed collision avoidance into a unified optimization framework. We first construct a dynamic risk fields (DRF) that captures both the static and dynamic collision risks from surrounding vehicles. Then, we develop a rigorous strategy for generating time-varying convex feasible spaces that ensure kinematic feasibility and safety requirements. The trajectory planning problem is formulated as a finite-horizon optimal control problem and solved using a constrained iterative Linear Quadratic Regulator (iLQR) algorithm that jointly optimizes trajectory smoothness, control effort, and risk exposure while maintaining strict feasibility. Extensive simulations demonstrate that our method outperforms traditional approaches in terms of safety and efficiency, achieving collision-free trajectories with shorter lane-changing distances (28.59 m) and times (2.84 s) while maintaining smooth and comfortable acceleration patterns. In dense roundabout environments the planner further demonstrates robust adaptability, producing larger safety margins, lower jerk, and superior curvature smoothness compared with APF, MPC, and RRT based baselines. These results confirm that the integrated DRF with convex feasible space and constrained iLQR solver provides a balanced solution for safe, efficient, and comfortable trajectory generation in dynamic and interactive traffic scenarios.
♻ ☆ Underactuated Robotic Hand with Grasp State Estimation Using Tendon-Based Proprioception
Anthropomorphic underactuated hands are valued for their structural simplicity and inherent adaptability. However, the uncertainty arising from interdependent joint motions makes it challenging to capture various grasp states during hand-object interaction without increasing structural complexity through multiple embedded sensors. This motivates the need for an approach that can extract rich grasp-state information from a single sensing source while preserving the simplicity of underactuation. This study proposes an anthropomorphic underactuated hand that achieves comprehensive grasp state estimation, using only tendon-based proprioception provided by series elastic actuators (SEAs). Our approach is enabled by the design of a compact SEA with high accuracy and reliability that can be seamlessly integrated into sensorless fingers. By coupling accurate proprioceptive measurements with potential energy-based modeling, the system estimates multiple key grasp state variables, including contact timing, joint angles, relative object stiffness, and external disturbances. Finger-level experimental validations and extensive hand-level grasp functionality demonstrations confirmed the effectiveness of the proposed approach. These results highlight tendon-based proprioception as a compact and robust sensing modality for practical manipulation without reliance on vision or tactile feedback.
comment: 11 pages, 15 figures, 3 tables, Supplementary video
♻ ☆ Material-informed Gaussian Splatting for 3D World Reconstruction in a Digital Twin
3D reconstruction for Digital Twins often relies on LiDAR-based methods, which provide accurate geometry but lack the semantics and textures naturally captured by cameras. Traditional LiDAR-camera fusion approaches require complex calibration and still struggle with certain materials like glass, which are visible in images but poorly represented in point clouds. We propose a camera-only pipeline that reconstructs scenes using 3D Gaussian Splatting from multi-view images, extracts semantic material masks via vision models, converts Gaussian representations to mesh surfaces with projected material labels, and assigns physics-based material properties for accurate sensor simulation in modern graphics engines and simulators. This approach combines photorealistic reconstruction with physics-based material assignment, providing sensor simulation fidelity comparable to LiDAR-camera fusion while eliminating hardware complexity and calibration requirements. We validate our camera-only method using an internal dataset from an instrumented test vehicle, leveraging LiDAR as ground truth for reflectivity validation alongside image similarity metrics.
comment: 8 pages, 5 figures. Submitted to IEEE Intelligent Vehicles Symposium (IV) 2026 for possible publication. Revised version (v2) to correct author order
♻ ☆ Design and Measurements of mmWave FMCW Radar Based Non-Contact Multi-Patient Heart Rate and Breath Rate Monitoring System
Recent developments in mmWave radar technologies have enabled the truly non-contact heart-rate (HR) and breath-rate (BR) measurement approaches, which provides a great ease in patient monitoring. Additionally, these technologies also provide opportunities to simultaneously detect HR and BR of multiple patients, which has become increasingly important for efficient mass monitoring scenarios. In this work, a frequency modulated continuous wave (FMCW) mmWave radar based truly non-contact multiple patient HR and BR monitoring system has been presented. Furthermore, a novel approach is also proposed, which combines multiple processing methods using a least squares solution to improve measurement accuracy, generalization, and handle measurement error. The proposed system has been developed using Texas Instruments' FMCW radar and experimental results with multiple subjects are also presented, which show >97% and >93% accuracy in the measured BR and HR values, respectively.
comment: Presented at BioCAS 2023
♻ ☆ Quality-guided UAV Surface Exploration for 3D Reconstruction
Reasons for mapping an unknown environment with autonomous robots are wide-ranging, but in practice, they are often overlooked when developing planning strategies. Rapid information gathering and comprehensive structural assessment of buildings have different requirements and therefore necessitate distinct methodologies. In this paper, we propose a novel modular Next-Best-View (NBV) planning framework for aerial robots that explicitly uses a reconstruction quality objective to guide the exploration planning. In particular, our approach introduces new and efficient methods for view generation and selection of viewpoint candidates that are adaptive to the user-defined quality requirements, fully exploiting the uncertainty encoded in a Truncated Signed Distance field (TSDF) representation of the environment. This results in informed and efficient exploration decisions tailored towards the predetermined objective. Finally, we validate our method via extensive simulations in realistic environments. We demonstrate that it successfully adjusts its behavior to the user goal while consistently outperforming conventional NBV strategies in terms of coverage, quality of the final 3D map and path efficiency.
♻ ☆ Scalable Multisubject Vital Sign Monitoring With mmWave FMCW Radar and FPGA Prototyping
In this work, we introduce an innovative approach to estimate the vital signs of multiple human subjects simultaneously in a non-contact way using a Frequency Modulated Continuous Wave (FMCW) radar-based system. Traditional vital sign monitoring methods often face significant limitations, including subject discomfort with wearable devices, challenges in calibration, and the risk of infection transmission through contact measurement devices. To address these issues, this research is motivated by the need for versatile, non-contact vital monitoring solutions applicable in various critical scenarios. This work also explores the challenges of extending this capability to an arbitrary number of subjects, including hardware and theoretical limitations. Supported by rigorous experimental results and discussions, the paper illustrates the system's potential to redefine vital sign monitoring. An FPGA-based implementation is also presented as proof of concept for a hardware-based and portable solution, improving upon previous works by offering 2.7x faster execution and 18.4% less Look-Up Table (LUT) utilization, as well as providing over 7400x acceleration compared to its software counterpart.
comment: Published in IEEE Sensors Journal
♻ ☆ Real-Time Obstacle Avoidance for a Mobile Robot Using CNN-Based Sensor Fusion
Obstacle avoidance is a critical component of the navigation stack required for mobile robots to operate effectively in complex and unknown environments. In this research, three end-to-end Convolutional Neural Networks (CNNs) were trained and evaluated offline and deployed on a differential-drive mobile robot for real-time obstacle avoidance to generate low-level steering commands from synchronized color and depth images acquired by an Intel RealSense D415 RGB-D camera in diverse environments. Offline evaluation showed that the NetConEmb model achieved the best performance with a notably low MedAE of $0.58 \times 10^{-3}$ rad/s. In comparison, the lighter NetEmb architecture, which reduces the number of trainable parameters by approximately 25\% and converges faster, produced comparable results with an RMSE of $21.68 \times 10^{-3}$ rad/s, close to the $21.42 \times 10^{-3}$ rad/s obtained by NetConEmb. Real-time navigation further confirmed NetConEmb's robustness, achieving a 100\% success rate in both known and unknown environments, while NetEmb and NetGated succeeded only in navigating the known environment.
♻ ☆ Efficient Path Planning and Task Allocation Algorithm for Boolean Specifications
This paper addresses path planning and task allocation in multi-robot systems subject to global Boolean specifications defined on the final state. The main contribution is the exploitation of the structural properties of a Petri net model: we prove that the associated constraint matrix is totally unimodular (TU). This property allows relaxing the original Integer Linear Programming (ILP) formulation to a Mixed Integer Linear Programming (MILP) in which all variables are continuous except for those that are corresponding to the atomic propositions in the Boolean specification. This yields a substantial reduction in complexity. In the special case where the specification is a conjunction of atomic propositions of cardinality equal to the team size, i.e., the standard Task-Assignment and Path Finding (TAPF) problem, the formulation reduces to a Linear Programming (LP). Collision-free paths are ensured by introducing intermediate synchronization points only when necessary, while robots move in parallel between them. These structural insights enable a computationally efficient and scalable solution, achieving tractability and safety for large-scale systems with up to 2500 robots.
♻ ☆ Steady-State Drifting Equilibrium Analysis of Single-Track Two-Wheeled Robots for Controller Design IROS
Drifting is an advanced driving technique where the wheeled robot's tire-ground interaction breaks the common non-holonomic pure rolling constraint. This allows high-maneuverability tasks like quick cornering, and steady-state drifting control enhances motion stability under lateral slip conditions. While drifting has been successfully achieved in four-wheeled robot systems, its application to single-track two-wheeled (STTW) robots, such as unmanned motorcycles or bicycles, has not been thoroughly studied. To bridge this gap, this paper extends the drifting equilibrium theory to STTW robots and reveals the mechanism behind the steady-state drifting maneuver. Notably, the counter-steering drifting technique used by skilled motorcyclists is explained through this theory. In addition, an analytical algorithm based on intrinsic geometry and kinematics relationships is proposed, reducing the computation time by four orders of magnitude while maintaining less than 6% error compared to numerical methods. Based on equilibrium analysis, a model predictive controller (MPC) is designed to achieve steady-state drifting and equilibrium points transition, with its effectiveness and robustness validated through simulations.
comment: Submitted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025
♻ ☆ Trajectory Optimization for In-Hand Manipulation with Tactile Force Control IROS 2025
The strength of the human hand lies in its ability to manipulate small objects precisely and robustly. In contrast, simple robotic grippers have low dexterity and fail to handle small objects effectively. This is why many automation tasks remain unsolved by robots. This paper presents an optimization-based framework for in-hand manipulation with a robotic hand equipped with compact Magnetic Tactile Sensors (MTSs). The small form factor of the robotic hand from Shadow Robot introduces challenges in estimating the state of the object while satisfying contact constraints. To address this, we formulate a trajectory optimization problem using Nonlinear Programming (NLP) for finger movements while ensuring contact points to change along the geometry of the fingers. Using the optimized trajectory from the solver, we implement and test an open-loop controller for rolling motion. To further enhance robustness and accuracy, we introduce a force controller for the fingers and a state estimator for the object utilizing MTSs. The proposed framework is validated through comparative experiments, showing that incorporating the force control with compliance consideration improves the accuracy and robustness of the rolling motion. Rolling an object with the force controller is 30\% more likely to succeed than running an open-loop controller. The demonstration video is available at https://youtu.be/6J_muL_AyE8.
comment: This paper has been accepted to IROS 2025
♻ ☆ Enhancing Kinematic Performances of Soft Continuum Robots for Magnetic Actuation
Soft continuum robots achieve complex deformation through elastic equilibrium, making their reachable motions governed jointly by structural design and actuation-induced mechanics. This work develops a general formulation that integrates equilibrium computation with kinematic performances by evaluating Riemannian Jacobian spectra on the equilibrium manifold shaped by internal/external loading. The resulting framework yields a global performance functional that directly links structural parameters, actuation inputs, and the induced configuration space geometry. We apply this general framework to magnetic actuation. Analytical characterization is obtained under weak uniform fields, revealing optimal placement and orientation of the embedded magnet with invariant scale properties. To address nonlinear deformation and spatially varying fields, a two-level optimization algorithm is developed that alternates between energy based equilibrium search and gradient based structural updates. Simulations and physical experiments across uniform field, dipole field, and multi-magnet configurations demonstrate consistent structural tendencies: aligned moments favor distal or mid-distal solutions through constructive torque amplification, whereas opposing moments compress optimal designs toward proximal regions due to intrinsic cancellation zones.
♻ ☆ Vectorized Online POMDP Planning ICRA 2026
Planning under partial observability is an essential capability of autonomous robots. The Partially Observable Markov Decision Process (POMDP) provides a powerful framework for planning under partial observability problems, capturing the stochastic effects of actions and the limited information available through noisy observations. POMDP solving could benefit tremendously from massive parallelization of today's hardware, but parallelizing POMDP solvers has been challenging. They rely on interleaving numerical optimization over actions with the estimation of their values, which creates dependencies and synchronization bottlenecks between parallel processes that can quickly offset the benefits of parallelization. In this paper, we propose Vectorized Online POMDP Planner (VOPP), a novel parallel online solver that leverages a recent POMDP formulation that analytically solves part of the optimization component, leaving only the estimation of expectations for numerical computation. VOPP represents all data structures related to planning as a collection of tensors and implements all planning steps as fully vectorized computations over this representation. The result is a massively parallel solver with no dependencies and synchronization bottlenecks between parallel computations. Experimental results indicate that VOPP is at least 20X more efficient in computing near-optimal solutions compared to an existing state-of-the-art parallel online solver.
comment: 9 pages, 3 figures. Submitted to ICRA 2026
♻ ☆ Rethinking Progression of Memory State in Robotic Manipulation: An Object-Centric Perspective AAAI 2026
As embodied agents operate in increasingly complex environments, the ability to perceive, track, and reason about individual object instances over time becomes essential, especially in tasks requiring sequenced interactions with visually similar objects. In these non-Markovian settings, key decision cues are often hidden in object-specific histories rather than the current scene. Without persistent memory of prior interactions (what has been interacted with, where it has been, or how it has changed) visuomotor policies may fail, repeat past actions, or overlook completed ones. To surface this challenge, we introduce LIBERO-Mem, a non-Markovian task suite for stress-testing robotic manipulation under object-level partial observability. It combines short- and long-horizon object tracking with temporally sequenced subgoals, requiring reasoning beyond the current frame. However, vision-language-action (VLA) models often struggle in such settings, with token scaling quickly becoming intractable even for tasks spanning just a few hundred frames. We propose Embodied-SlotSSM, a slot-centric VLA framework built for temporal scalability. It maintains spatio-temporally consistent slot identities and leverages them through two mechanisms: (1) slot-state-space modeling for reconstructing short-term history, and (2) a relational encoder to align the input tokens with action decoding. Together, these components enable temporally grounded, context-aware action prediction. Experiments show Embodied-SlotSSM's baseline performance on LIBERO-Mem and general tasks, offering a scalable solution for non-Markovian reasoning in object-centric robotic policies.
comment: Accepted at AAAI 2026
♻ ☆ SlotVLA: Towards Modeling of Object-Relation Representations in Robotic Manipulation
Inspired by how humans reason over discrete objects and their relationships, we explore whether compact object-centric and object-relation representations can form a foundation for multitask robotic manipulation. Most existing robotic multitask models rely on dense embeddings that entangle both object and background cues, raising concerns about both efficiency and interpretability. In contrast, we study object-relation-centric representations as a pathway to more structured, efficient, and explainable visuomotor control. Our contributions are two-fold. First, we introduce LIBERO+, a fine-grained benchmark dataset designed to enable and evaluate object-relation reasoning in robotic manipulation. Unlike prior datasets, LIBERO+ provides object-centric annotations that enrich demonstrations with box- and mask-level labels as well as instance-level temporal tracking, supporting compact and interpretable visuomotor representations. Second, we propose SlotVLA, a slot-attention-based framework that captures both objects and their relations for action decoding. It uses a slot-based visual tokenizer to maintain consistent temporal object representations, a relation-centric decoder to produce task-relevant embeddings, and an LLM-driven module that translates these embeddings into executable actions. Experiments on LIBERO+ demonstrate that object-centric slot and object-relation slot representations drastically reduce the number of required visual tokens, while providing competitive generalization. Together, LIBERO+ and SlotVLA provide a compact, interpretable, and effective foundation for advancing object-relation-centric robotic manipulation.
comment: under review
♻ ☆ BactoBot: A Low-Cost, Bacteria-Inspired Soft Underwater Robot for Marine Exploration
Traditional rigid underwater vehicles pose risks to delicate marine ecosystems due to high-speed propellers and rigid hulls. This paper presents BactoBot, a low-cost, soft underwater robot designed for safe and gentle marine exploration. Inspired by the efficient flagellar propulsion of bacteria, BactoBot features 12 flexible, silicone-based arms arranged on a dodecahedral frame. Unlike high-cost research platforms, this prototype was fabricated using accessible DIY methods, including food-grade silicone molding, FDM 3D printing, and off-the-shelf DC motors. A novel multi-stage waterproofing protocol was developed to seal rotating shafts using a grease-filled chamber system, ensuring reliability at low cost. The robot was successfully tested in a controlled aquatic environment, demonstrating stable forward propulsion and turning maneuvers. With a total fabrication cost of approximately $355 USD, this project validates the feasibility of democratizing soft robotics for marine science in resource-constrained settings.
comment: Revised version. Updated literature review, improved figures, and added clarification on waterproofing methodology
♻ ☆ HAFO: Humanoid Force-Adaptive Control for Intense External Force Interaction Environments
Reinforcement learning controllers have made impressive progress in humanoid locomotion and light load manipulation. However, achieving robust and precise motion with strong force interaction remains a significant challenge. Based on the above limitations, this paper proposes HAFO, a dual-agent reinforcement learning control framework that simultaneously optimizes both a robust locomotion strategy and a precise upper-body manipulation strategy through coupled training under external force interaction environments. Simultaneously, we explicitly model the external pulling disturbances through a spring-damper system and achieve fine-grained force control by manipulating the virtual spring. During this process, the reinforcement-learning policy spontaneously generates disturbance-rejection response by exploiting environmental feedback. Moreover, HAFO employs an asymmetric Actor-Critic framework in which the Critic-network access to privileged spring-damping forces guides the actor-network to learn a generalizable, robust policy for resisting external disturbances. The experimental results demonstrate that HAFO achieves stable control of humanoid robot under various strong force interactions, showing remarkable performance in load tasks and ensuring stable robot operation under rope tension disturbances. Project website: hafo-robot.github.io.
♻ ☆ Spectral Signature Mapping from RGB Imagery for Terrain-Aware Navigation
Successful navigation in outdoor environments requires accurate prediction of the physical interactions between the robot and the terrain. Many prior methods rely on geometric or semantic labels to classify traversable surfaces. However, such labels cannot distinguish visually similar surfaces that differ in material properties. Spectral sensors enable inference of material composition from surface reflectance measured across multiple wavelength bands. Although spectral sensing is gaining traction in robotics, widespread deployment remains constrained by the need for custom hardware integration, high sensor costs, and compute-intensive processing pipelines. In this paper, we present the RGB Image to Spectral Signature Neural Network (RS-Net), a deep neural network designed to bridge the gap between the accessibility of RGB sensing and the rich material information provided by spectral data. RS-Net predicts spectral signatures from RGB patches, which we map to terrain labels and friction coefficients. The resulting terrain classifications are integrated into a sampling-based motion planner for a wheeled robot operating in outdoor environments. Likewise, the friction estimates are incorporated into a contact-force-based MPC for a quadruped robot navigating slippery surfaces. Overall, our framework learns the task-relevant physical properties offline during training and thereafter relies solely on RGB sensing at run time.
comment: 8 pages, 11 figures, accepted to Robotic Computing & Communication
♻ ☆ Coordinating Spinal and Limb Dynamics for Enhanced Sprawling Robot Mobility ICRA 2025
Sprawling locomotion in vertebrates, particularly salamanders, demonstrates how body undulation and spinal mobility enhance stability, maneuverability, and adaptability across complex terrains. While prior work has separately explored biologically inspired gait design or deep reinforcement learning (DRL), these approaches face inherent limitations: open-loop gait designs often lack adaptability to unforeseen terrain variations, whereas end-to-end DRL methods are data-hungry and prone to unstable behaviors when transferring from simulation to real robots. We propose a hybrid control framework that integrates Hildebrand's biologically grounded gait design with DRL, enabling a salamander-inspired quadruped robot to exploit active spinal joints for robust crawling motion. Our evaluation across multiple robot configurations in target-directed navigation tasks reveals that this hybrid approach systematically improves robustness under environmental uncertainties such as surface irregularities. By bridging structured gait design with learning-based methodology, our work highlights the promise of interdisciplinary control strategies for developing efficient, resilient, and biologically informed spinal actuation in robotic systems.
comment: Initial version of the work has been accepted for presentation at the Mechanical Intelligence in Robotics workshop at ICRA 2025
♻ ☆ Gemini Robotics 1.5: Pushing the Frontier of Generalist Robots with Advanced Embodied Reasoning, Thinking, and Motion Transfer
General-purpose robots need a deep understanding of the physical world, advanced reasoning, and general and dexterous control. This report introduces the latest generation of the Gemini Robotics model family: Gemini Robotics 1.5, a multi-embodiment Vision-Language-Action (VLA) model, and Gemini Robotics-ER 1.5, a state-of-the-art Embodied Reasoning (ER) model. We are bringing together three major innovations. First, Gemini Robotics 1.5 features a novel architecture and a Motion Transfer (MT) mechanism, which enables it to learn from heterogeneous, multi-embodiment robot data and makes the VLA more general. Second, Gemini Robotics 1.5 interleaves actions with a multi-level internal reasoning process in natural language. This enables the robot to "think before acting" and notably improves its ability to decompose and execute complex, multi-step tasks, and also makes the robot's behavior more interpretable to the user. Third, Gemini Robotics-ER 1.5 establishes a new state-of-the-art for embodied reasoning, i.e., for reasoning capabilities that are critical for robots, such as visual and spatial understanding, task planning, and progress estimation. Together, this family of models takes us a step towards an era of physical agents-enabling robots to perceive, think and then act so they can solve complex multi-step tasks.
Robotics 46
☆ Switching control of underactuated multi-channel systems with input constraints for cooperative manipulation
This work presents an event-triggered switching control framework for a class of nonlinear underactuated multi-channel systems with input constraints. These systems are inspired by cooperative manipulation tasks involving underactuation, where multiple underactuated agents collaboratively push or pull an object to a target pose. Unlike existing approaches for multi-channel systems, our method addresses underactuation and the potential loss of controllability by additionally addressing channel assignment of agents. To simultaneously account for channel assignment, input constraints, and stabilization, we formulate the control problem as a Mixed Integer Linear Programming and derive sufficient conditions for its feasibility. To improve real-time computation efficiency, we introduce an event-triggered control scheme that maintains stability even between switching events through a quadratic programming-based stabilizing controller. We theoretically establish the semi-global exponential stability of the proposed method and the asymptotic stability of its extension to nonprehensile cooperative manipulation under noninstantaneous switching. The proposed framework is further validated through numerical simulations on 2D and 3D free-flyer systems and multi-robot nonprehensile pushing tasks.
comment: 14 pages
☆ Distracted Robot: How Visual Clutter Undermine Robotic Manipulation
In this work, we propose an evaluation protocol for examining the performance of robotic manipulation policies in cluttered scenes. Contrary to prior works, we approach evaluation from a psychophysical perspective, therefore we use a unified measure of clutter that accounts for environmental factors as well as the distractors quantity, characteristics, and arrangement. Using this measure, we systematically construct evaluation scenarios in both hyper-realistic simulation and real-world and conduct extensive experimentation on manipulation policies, in particular vision-language-action (VLA) models. Our experiments highlight the significant impact of scene clutter, lowering the performance of the policies, by as much as 34% and show that despite achieving similar average performance across the tasks, different VLA policies have unique vulnerabilities and a relatively low agreement on success scenarios. We further show that our clutter measure is an effective indicator of performance degradation and analyze the impact of distractors in terms of their quantity and occluding influence. At the end, we show that finetuning on enhanced data, although effective, does not equally remedy all negative impacts of clutter on performance.
comment: 12 figures, 2 tables
☆ Improving Robotic Manipulation Robustness via NICE Scene Surgery
Learning robust visuomotor policies for robotic manipulation remains a challenge in real-world settings, where visual distractors can significantly degrade performance and safety. In this work, we propose an effective and scalable framework, Naturalistic Inpainting for Context Enhancement (NICE). Our method minimizes out-of-distribution (OOD) gap in imitation learning by increasing visual diversity through construction of new experiences using existing demonstrations. By utilizing image generative frameworks and large language models, NICE performs three editing operations, object replacement, restyling, and removal of distracting (non-target) objects. These changes preserve spatial relationships without obstructing target objects and maintain action-label consistency. Unlike previous approaches, NICE requires no additional robot data collection, simulator access, or custom model training, making it readily applicable to existing robotic datasets. Using real-world scenes, we showcase the capability of our framework in producing photo-realistic scene enhancement. For downstream tasks, we use NICE data to finetune a vision-language model (VLM) for spatial affordance prediction and a vision-language-action (VLA) policy for object manipulation. Our evaluations show that NICE successfully minimizes OOD gaps, resulting in over 20% improvement in accuracy for affordance prediction in highly cluttered scenes. For manipulation tasks, success rate increases on average by 11% when testing in environments populated with distractors in different quantities. Furthermore, we show that our method improves visual robustness, lowering target confusion by 6%, and enhances safety by reducing collision rate by 7%.
comment: 11 figures, 3 tables
☆ CAPE: Context-Aware Diffusion Policy Via Proximal Mode Expansion for Collision Avoidance
In robotics, diffusion models can capture multi-modal trajectories from demonstrations, making them a transformative approach in imitation learning. However, achieving optimal performance following this regiment requires a large-scale dataset, which is costly to obtain, especially for challenging tasks, such as collision avoidance. In those tasks, generalization at test time demands coverage of many obstacles types and their spatial configurations, which are impractical to acquire purely via data. To remedy this problem, we propose Context-Aware diffusion policy via Proximal mode Expansion (CAPE), a framework that expands trajectory distribution modes with context-aware prior and guidance at inference via a novel prior-seeded iterative guided refinement procedure. The framework generates an initial trajectory plan and executes a short prefix trajectory, and then the remaining trajectory segment is perturbed to an intermediate noise level, forming a trajectory prior. Such a prior is context-aware and preserves task intent. Repeating the process with context-aware guided denoising iteratively expands mode support to allow finding smoother, less collision-prone trajectories. For collision avoidance, CAPE expands trajectory distribution modes with collision-aware context, enabling the sampling of collision-free trajectories in previously unseen environments while maintaining goal consistency. We evaluate CAPE on diverse manipulation tasks in cluttered unseen simulated and real-world settings and show up to 26% and 80% higher success rates respectively compared to SOTA methods, demonstrating better generalization to unseen environments.
comment: 4 tables, 9 figures
☆ Beyond Egocentric Limits: Multi-View Depth-Based Learning for Robust Quadrupedal Locomotion
Recent progress in legged locomotion has allowed highly dynamic and parkour-like behaviors for robots, similar to their biological counterparts. Yet, these methods mostly rely on egocentric (first-person) perception, limiting their performance, especially when the viewpoint of the robot is occluded. A promising solution would be to enhance the robot's environmental awareness by using complementary viewpoints, such as multiple actors exchanging perceptual information. Inspired by this idea, this work proposes a multi-view depth-based locomotion framework that combines egocentric and exocentric observations to provide richer environmental context during agile locomotion. Using a teacher-student distillation approach, the student policy learns to fuse proprioception with dual depth streams while remaining robust to real-world sensing imperfections. To further improve robustness, we introduce extensive domain randomization, including stochastic remote-camera dropouts and 3D positional perturbations that emulate aerial-ground cooperative sensing. Simulation results show that multi-viewpoints policies outperform single-viewpoint baseline in gap crossing, step descent, and other dynamic maneuvers, while maintaining stability when the exocentric camera is partially or completely unavailable. Additional experiments show that moderate viewpoint misalignment is well tolerated when incorporated during training. This study demonstrates that heterogeneous visual feedback improves robustness and agility in quadrupedal locomotion. Furthermore, to support reproducibility, the implementation accompanying this work is publicly available at https://anonymous.4open.science/r/multiview-parkour-6FB8
comment: 12 pages, 6 figures, code available at https://anonymous.4open.science/r/multiview-parkour-6FB8
☆ A Two Degrees-of-Freedom Floor-Based Robot for Transfer and Rehabilitation Applications
The ability to accomplish a sit-to-stand (STS) motion is key to increase functional mobility and reduce rehospitalization risks. While raising aid (transfer) devices and partial bodyweight support (rehabilitation) devices exist, both are unable to adjust the STS training to different mobility levels. Therefore, We have developed an STS training device that allows various configurations of impedance and vertical/forward forces to adapt to many training needs while maintaining commercial raising aid transfer capabilities. Experiments with healthy adults (both men and women) of various heights and weights show that the device 1) has a low impact on the natural STS kinematics, 2) can provide precise weight unloading at the patient's center of mass and 3) can add a forward virtual spring to assist the transfer of the bodyweight to the feet for seat-off, at the start of the STS motion.
comment: 13 pages, 16 figures
☆ Mechanistic Finetuning of Vision-Language-Action Models via Few-Shot Demonstrations
Vision-Language Action (VLAs) models promise to extend the remarkable success of vision-language models (VLMs) to robotics. Yet, unlike VLMs in the vision-language domain, VLAs for robotics require finetuning to contend with varying physical factors like robot embodiment, environment characteristics, and spatial relationships of each task. Existing fine-tuning methods lack specificity, adapting the same set of parameters regardless of a task's visual, linguistic, and physical characteristics. Inspired by functional specificity in neuroscience, we hypothesize that it is more effective to finetune sparse model representations specific to a given task. In this work, we introduce Robotic Steering, a finetuning approach grounded in mechanistic interpretability that leverages few-shot demonstrations to identify and selectively finetune task-specific attention heads aligned with the physical, visual, and linguistic requirements of robotic tasks. Through comprehensive on-robot evaluations with a Franka Emika robot arm, we demonstrate that Robotic Steering outperforms LoRA while achieving superior robustness under task variation, reduced computational cost, and enhanced interpretability for adapting VLAs to diverse robotic tasks.
☆ Deadlock-Free Hybrid RL-MAPF Framework for Zero-Shot Multi-Robot Navigation
Multi-robot navigation in cluttered environments presents fundamental challenges in balancing reactive collision avoidance with long-range goal achievement. When navigating through narrow passages or confined spaces, deadlocks frequently emerge that prevent agents from reaching their destinations, particularly when Reinforcement Learning (RL) control policies encounter novel configurations out of learning distribution. Existing RL-based approaches suffer from limited generalization capability in unseen environments. We propose a hybrid framework that seamlessly integrates RL-based reactive navigation with on-demand Multi-Agent Path Finding (MAPF) to explicitly resolve topological deadlocks. Our approach integrates a safety layer that monitors agent progress to detect deadlocks and, when detected, triggers a coordination controller for affected agents. The framework constructs globally feasible trajectories via MAPF and regulates waypoint progression to reduce inter-agent conflicts during navigation. Extensive evaluation on dense multi-agent benchmarks shows that our method boosts task completion from marginal to near-universal success, markedly reducing deadlocks and collisions. When integrated with hierarchical task planning, it enables coordinated navigation for heterogeneous robots, demonstrating that coupling reactive RL navigation with selective MAPF intervention yields a robust, zero-shot performance.
☆ MG-Nav: Dual-Scale Visual Navigation via Sparse Spatial Memory
We present MG-Nav (Memory-Guided Navigation), a dual-scale framework for zero-shot visual navigation that unifies global memory-guided planning with local geometry-enhanced control. At its core is the Sparse Spatial Memory Graph (SMG), a compact, region-centric memory where each node aggregates multi-view keyframe and object semantics, capturing both appearance and spatial structure while preserving viewpoint diversity. At the global level, the agent is localized on SMG and a goal-conditioned node path is planned via an image-to-instance hybrid retrieval, producing a sequence of reachable waypoints for long-horizon guidance. At the local level, a navigation foundation policy executes these waypoints in point-goal mode with obstacle-aware control, and switches to image-goal mode when navigating from the final node towards the visual target. To further enhance viewpoint alignment and goal recognition, we introduce VGGT-adapter, a lightweight geometric module built on the pre-trained VGGT model, which aligns observation and goal features in a shared 3D-aware space. MG-Nav operates global planning and local control at different frequencies, using periodic re-localization to correct errors. Experiments on HM3D Instance-Image-Goal and MP3D Image-Goal benchmarks demonstrate that MG-Nav achieves state-of-the-art zero-shot performance and remains robust under dynamic rearrangements and unseen scene conditions.
comment: 10pages, 5 figures
☆ Beyond Success: Refining Elegant Robot Manipulation from Mixed-Quality Data via Just-in-Time Intervention
Vision-Language-Action (VLA) models have enabled notable progress in general-purpose robotic manipulation, yet their learned policies often exhibit variable execution quality. We attribute this variability to the mixed-quality nature of human demonstrations, where the implicit principles that govern how actions should be carried out are only partially satisfied. To address this challenge, we introduce the LIBERO-Elegant benchmark with explicit criteria for evaluating execution quality. Using these criteria, we develop a decoupled refinement framework that improves execution quality without modifying or retraining the base VLA policy. We formalize Elegant Execution as the satisfaction of Implicit Task Constraints (ITCs) and train an Elegance Critic via offline Calibrated Q-Learning to estimate the expected quality of candidate actions. At inference time, a Just-in-Time Intervention (JITI) mechanism monitors critic confidence and intervenes only at decision-critical moments, providing selective, on-demand refinement. Experiments on LIBERO-Elegant and real-world manipulation tasks show that the learned Elegance Critic substantially improves execution quality, even on unseen tasks. The proposed model enables robotic control that values not only whether tasks succeed, but also how they are performed.
☆ BUDD-e: an autonomous robotic guide for visually impaired users
This paper describes the design and the realization of a prototype of the novel guide robot BUDD-e for visually impaired users. The robot has been tested in a real scenario with the help of visually disabled volunteers at ASST Grande Ospedale Metropolitano Niguarda, in Milan. The results of the experimental campaign are throughly described in the paper, displaying its remarkable performance and user-acceptance.
comment: 14 pages
☆ RealD$^2$iff: Bridging Real-World Gap in Robot Manipulation via Depth Diffusion
Robot manipulation in the real world is fundamentally constrained by the visual sim2real gap, where depth observations collected in simulation fail to reflect the complex noise patterns inherent to real sensors. In this work, inspired by the denoising capability of diffusion models, we invert the conventional perspective and propose a clean-to-noisy paradigm that learns to synthesize noisy depth, thereby bridging the visual sim2real gap through purely simulation-driven robotic learning. Building on this idea, we introduce RealD$^2$iff, a hierarchical coarse-to-fine diffusion framework that decomposes depth noise into global structural distortions and fine-grained local perturbations. To enable progressive learning of these components, we further develop two complementary strategies: Frequency-Guided Supervision (FGS) for global structure modeling and Discrepancy-Guided Optimization (DGO) for localized refinement. To integrate RealD$^2$iff seamlessly into imitation learning, we construct a pipeline that spans six stages. We provide comprehensive empirical and experimental validation demonstrating the effectiveness of this paradigm. RealD$^2$iff enables two key applications: (1) generating real-world-like depth to construct clean-noisy paired datasets without manual sensor data collection. (2) Achieving zero-shot sim2real robot manipulation, substantially improving real-world performance without additional fine-tuning.
☆ Motion-to-Motion Latency Measurement Framework for Connected and Autonomous Vehicle Teleoperation
Latency is a key performance factor for the teleoperation of Connected and Autonomous Vehicles (CAVs). It affects how quickly an operator can perceive changes in the driving environment and apply corrective actions. Most existing work focuses on Glass-to-Glass (G2G) latency, which captures delays only in the video pipeline. However, there is no standard method for measuring Motion-to-Motion (M2M) latency, defined as the delay between the physical steering movement of the remote operator and the corresponding steering motion in the vehicle. This paper presents an M2M latency measurement framework that uses Hall-effect sensors and two synchronized Raspberry Pi~5 devices. The system records interrupt-based timestamps on both sides to estimate M2M latency, independently of the underlying teleoperation architecture. Precision tests show an accuracy of 10--15~ms, while field results indicate that actuator delays dominate M2M latency, with median values above 750~ms.
☆ Visual-Geometry Diffusion Policy: Robust Generalization via Complementarity-Aware Multimodal Fusion
Imitation learning has emerged as a crucial ap proach for acquiring visuomotor skills from demonstrations, where designing effective observation encoders is essential for policy generalization. However, existing methods often struggle to generalize under spatial and visual randomizations, instead tending to overfit. To address this challenge, we propose Visual Geometry Diffusion Policy (VGDP), a multimodal imitation learning framework built around a Complementarity-Aware Fusion Module where modality-wise dropout enforces balanced use of RGB and point-cloud cues, with cross-attention serving only as a lightweight interaction layer. Our experiments show that the expressiveness of the fused latent space is largely induced by the enforced complementarity from modality-wise dropout, with cross-attention serving primarily as a lightweight interaction mechanism rather than the main source of robustness. Across a benchmark of 18 simulated tasks and 4 real-world tasks, VGDP outperforms seven baseline policies with an average performance improvement of 39.1%. More importantly, VGDP demonstrates strong robustness under visual and spatial per turbations, surpassing baselines with an average improvement of 41.5% in different visual conditions and 15.2% in different spatial settings.
☆ BINDER: Instantly Adaptive Mobile Manipulation with Open-Vocabulary Commands
Open-vocabulary mobile manipulation (OVMM) requires robots to follow language instructions, navigate, and manipulate while updating their world representation under dynamic environmental changes. However, most prior approaches update their world representation only at discrete update points such as navigation targets, waypoints, or the end of an action step, leaving robots blind between updates and causing cascading failures: overlooked objects, late error detection, and delayed replanning. To address this limitation, we propose BINDER (Bridging INstant and DEliberative Reasoning), a dual process framework that decouples strategic planning from continuous environment monitoring. Specifically, BINDER integrates a Deliberative Response Module (DRM, a multimodal LLM for task planning) with an Instant Response Module (IRM, a VideoLLM for continuous monitoring). The two modules play complementary roles: the DRM performs strategic planning with structured 3D scene updates and guides what the IRM attends to, while the IRM analyzes video streams to update memory, correct ongoing actions, and trigger replanning when necessary. Through this bidirectional coordination, the modules address the trade off between maintaining awareness and avoiding costly updates, enabling robust adaptation under dynamic conditions. Evaluated in three real world environments with dynamic object placement, BINDER achieves substantially higher success and efficiency than SoTA baselines, demonstrating its effectiveness for real world deployment.
comment: 12 pages, 8 figures
☆ LLM-Based Generalizable Hierarchical Task Planning and Execution for Heterogeneous Robot Teams with Event-Driven Replanning ICRA 2026
This paper introduces CoMuRoS (Collaborative Multi-Robot System), a generalizable hierarchical architecture for heterogeneous robot teams that unifies centralized deliberation with decentralized execution, and supports event-driven replanning. A Task Manager LLM interprets natural-language goals, classifies tasks, and allocates subtasks using static rules plus dynamic contexts (task, history, robot and task status, and events).Each robot runs a local LLM that composes executable Python code from primitive skills (ROS2 nodes, policies), while onboard perception (VLMs/image processing) continuously monitors events and classifies them into relevant or irrelevant to the task. Task failures or user intent changes trigger replanning, allowing robots to assist teammates, resume tasks, or request human help. Hardware studies demonstrate autonomous recovery from disruptive events, filtering of irrelevant distractions, and tightly coordinated transport with emergent human-robot cooperation (e.g., multirobot collaborative object recovery success rate: 9/10, coordinated transport: 8/8, human-assisted recovery: 5/5).Simulation studies show intention-aware replanning. A curated textual benchmark spanning 22 scenarios (3 tasks each, around 20 robots) evaluates task allocation, classification, IoU, executability, and correctness, with high average scores (e.g., correctness up to 0.91) across multiple LLMs, a separate replanning set (5 scenarios) achieves 1.0 correctness. Compared with prior LLM-based systems, CoMuRoS uniquely demonstrates runtime, event-driven replanning on physical robots, delivering robust, flexible multi-robot and human-robot collaboration.
comment: submitted to ICRA 2026
☆ Nonholonomic Narrow Dead-End Escape with Deep Reinforcement Learning
Nonholonomic constraints restrict feasible velocities without reducing configuration-space dimension, which makes collision-free geometric paths generally non-executable for car-like robots. Ackermann steering further imposes curvature bounds and forbids in-place rotation, so escaping from narrow dead ends typically requires tightly sequenced forward and reverse maneuvers. Classical planners that decouple global search and local steering struggle in these settings because narrow passages occupy low-measure regions and nonholonomic reachability shrinks the set of valid connections, which degrades sampling efficiency and increases sensitivity to clearances. We study nonholonomic narrow dead-end escape for Ackermann vehicles and contribute three components. First, we construct a generator that samples multi-phase forward-reverse trajectories compatible with Ackermann kinematics and inflates their envelopes to synthesize families of narrow dead ends that are guaranteed to admit at least one feasible escape. Second, we construct a training environment that enforces kinematic constraints and train a policy using the soft actor-critic algorithm. Third, we evaluate against representative classical planners that combine global search with nonholonomic steering. Across parameterized dead-end families, the learned policy solves a larger fraction of instances, reduces maneuver count, and maintains comparable path length and planning time while under the same sensing and control limits. We provide our project as open source at https://github.com/gitagitty/cisDRL-RobotNav.git
comment: 14 pages, 5 figures, 1 table, submitted to arXiv
☆ Soft Fluidic Sheet Transistor for Soft Robotic System Enabling Fluid Logic Operations
Aiming to achieve both high functionality and flexibility in soft robot system, this paper presents a soft urethane sheet-like valve with an amplifier that can perform logical operations using only pneumatic signals. When the control chamber in the valve is pressurized, the main path is compressed along its central axis, buckling and being pressed,resulting in blockage. This allows control by a pressure signal smaller than that within the main channel. Furthermore, similar to transistors in electrical circuits, when combined, the proposed valve can perform a variety of logical operations. The basic type operates as a NOT logic element, which is named the fluidic sheet transistor (FST). By integrating multiple FSTs, logical operations such as positive logic, NAND, and NOR can be performed on a single sheet. This paper describes the operating principle, fabrication method, and characteristics of the FST,followed by a method for configuring logical operations.Moreover, we demonstrate the construction of a latch circuit(self-holding logic circuit) using FST, introducing a prototype of a fluid robot system that combines a tactile tube as a fluidic detector and fluid actuators. This demonstrates that it is possible to generate behavior that actively changes posture when hitting an obstacle using only air pressure from a single pipe, which verifies the effectiveness of the proposed methods.
comment: 7 pages, 16 figures
☆ MLATC: Fast Hierarchical Topological Mapping from 3D LiDAR Point Clouds Based on Adaptive Resonance Theory
This paper addresses the problem of building global topological maps from 3D LiDAR point clouds for autonomous mobile robots operating in large-scale, dynamic, and unknown environments. Adaptive Resonance Theory-based Topological Clustering with Different Topologies (ATC-DT) builds global topological maps represented as graphs while mitigating catastrophic forgetting during sequential processing. However, its winner selection mechanism relies on an exhaustive nearest-neighbor search over all existing nodes, leading to scalability limitations as the map grows. To address this challenge, we propose a hierarchical extension called Multi-Layer ATC (MLATC). MLATC organizes nodes into a hierarchy, enabling the nearest-neighbor search to proceed from coarse to fine resolutions, thereby drastically reducing the number of distance evaluations per query. The number of layers is not fixed in advance. MLATC employs an adaptive layer addition mechanism that automatically deepens the hierarchy when lower layers become saturated, keeping the number of user-defined hyperparameters low. Simulation experiments on synthetic large-scale environments show that MLATC accelerates topological map building compared to the original ATC-DT and exhibits a sublinear, approximately logarithmic scaling of search time with respect to the number of nodes. Experiments on campus-scale real-world LiDAR datasets confirm that MLATC maintains a millisecond-level per-frame runtime and enables real-time global topological map building in large-scale environments, significantly outperforming the original ATC-DT in terms of computational efficiency.
☆ Bayesian Decentralized Decision-making for Multi-Robot Systems: Sample-efficient Estimation of Event Rates
Effective collective decision-making in swarm robotics often requires balancing exploration, communication and individual uncertainty estimation, especially in hazardous environments where direct measurements are limited or costly. We propose a decentralized Bayesian framework that enables a swarm of simple robots to identify the safer of two areas, each characterized by an unknown rate of hazardous events governed by a Poisson process. Robots employ a conjugate prior to gradually predict the times between events and derive confidence estimates to adapt their behavior. Our simulation results show that the robot swarm consistently chooses the correct area while reducing exposure to hazardous events by being sample-efficient. Compared to baseline heuristics, our proposed approach shows better performance in terms of safety and speed of convergence. The proposed scenario has potential to extend the current set of benchmarks in collective decision-making and our method has applications in adaptive risk-aware sampling and exploration in hazardous, dynamic environments.
comment: 7 pages, 3 figures, submitted to IEEE MRS 2025
☆ BiCQL-ML: A Bi-Level Conservative Q-Learning Framework for Maximum Likelihood Inverse Reinforcement Learning
Offline inverse reinforcement learning (IRL) aims to recover a reward function that explains expert behavior using only fixed demonstration data, without any additional online interaction. We propose BiCQL-ML, a policy-free offline IRL algorithm that jointly optimizes a reward function and a conservative Q-function in a bi-level framework, thereby avoiding explicit policy learning. The method alternates between (i) learning a conservative Q-function via Conservative Q-Learning (CQL) under the current reward, and (ii) updating the reward parameters to maximize the expected Q-values of expert actions while suppressing over-generalization to out-of-distribution actions. This procedure can be viewed as maximum likelihood estimation under a soft value matching principle. We provide theoretical guarantees that BiCQL-ML converges to a reward function under which the expert policy is soft-optimal. Empirically, we show on standard offline RL benchmarks that BiCQL-ML improves both reward recovery and downstream policy performance compared to existing offline IRL baselines.
comment: 8 pages, 3 figures
☆ 3D Affordance Keypoint Detection for Robotic Manipulation IROS 2024
This paper presents a novel approach for affordance-informed robotic manipulation by introducing 3D keypoints to enhance the understanding of object parts' functionality. The proposed approach provides direct information about what the potential use of objects is, as well as guidance on where and how a manipulator should engage, whereas conventional methods treat affordance detection as a semantic segmentation task, focusing solely on answering the what question. To address this gap, we propose a Fusion-based Affordance Keypoint Network (FAKP-Net) by introducing 3D keypoint quadruplet that harnesses the synergistic potential of RGB and Depth image to provide information on execution position, direction, and extent. Benchmark testing demonstrates that FAKP-Net outperforms existing models by significant margins in affordance segmentation task and keypoint detection task. Real-world experiments also showcase the reliability of our method in accomplishing manipulation tasks with previously unseen objects.
comment: Accepted to IROS 2024
☆ HybridWorldSim: A Scalable and Controllable High-fidelity Simulator for Autonomous Driving
Realistic and controllable simulation is critical for advancing end-to-end autonomous driving, yet existing approaches often struggle to support novel view synthesis under large viewpoint changes or to ensure geometric consistency. We introduce HybridWorldSim, a hybrid simulation framework that integrates multi-traversal neural reconstruction for static backgrounds with generative modeling for dynamic agents. This unified design addresses key limitations of previous methods, enabling the creation of diverse and high-fidelity driving scenarios with reliable visual and spatial consistency. To facilitate robust benchmarking, we further release a new multi-traversal dataset MIRROR that captures a wide range of routes and environmental conditions across different cities. Extensive experiments demonstrate that HybridWorldSim surpasses previous state-of-the-art methods, providing a practical and scalable solution for high-fidelity simulation and a valuable resource for research and development in autonomous driving.
☆ MTR-VP: Towards End-to-End Trajectory Planning through Context-Driven Image Encoding and Multiple Trajectory Prediction
We present a method for trajectory planning for autonomous driving, learning image-based context embeddings that align with motion prediction frameworks and planning-based intention input. Within our method, a ViT encoder takes raw images and past kinematic state as input and is trained to produce context embeddings, drawing inspiration from those generated by the recent MTR (Motion Transformer) encoder, effectively substituting map-based features with learned visual representations. MTR provides a strong foundation for multimodal trajectory prediction by localizing agent intent and refining motion iteratively via motion query pairs; we name our approach MTR-VP (Motion Transformer for Vision-based Planning), and instead of the learnable intention queries used in the MTR decoder, we use cross attention on the intent and the context embeddings, which reflect a combination of information encoded from the driving scene and past vehicle states. We evaluate our methods on the Waymo End-to-End Driving Dataset, which requires predicting the agent's future 5-second trajectory in bird's-eye-view coordinates using prior camera images, agent pose history, and routing goals. We analyze our architecture using ablation studies, removing input images and multiple trajectory output. Our results suggest that transformer-based methods that are used to combine the visual features along with the kinetic features such as the past trajectory features are not effective at combining both modes to produce useful scene context embeddings, even when intention embeddings are augmented with foundation-model representations of scene context from CLIP and DINOv2, but that predicting a distribution over multiple futures instead of a single future trajectory boosts planning performance.
comment: 8 pages, 3 figures, 4 tables
☆ Lips-Jaw and Tongue-Jaw Articulatory Tradeoff in DYNARTmo
This paper investigates how the dynamic articulatory model DYNARTmo accounts for articulatory tradeoffs between primary and secondary articulators, with a focus on lips-jaw and tongue-jaw coordination. While DYNARTmo does not implement full task-dynamic second-order biomechanics, it adopts first-order task-space gesture specifications comparable to those used in articulatory phonology and integrates a simplified mechanism for distributing articulatory effort across multiple articulators. We first outline the conceptual relationship between task dynamics and DYNARTmo, emphasizing the distinction between high-level task-space trajectories and their low-level articulatory execution. We then present simulation results for a set of CV syllables that illustrate how jaw displacement varies as a function of both place of articulation (labial, apical, dorsal) and vowel context (/a/, /i/, /u/). The model reproduces empirically attested patterns of articulatory synergy, including jaw-supported apical closures, lower-lip elevation in bilabial stops, tongue-jaw co-movement, and saturation effects in labial constrictions. These results demonstrate that even with computationally simplified assumptions, DYNARTmo can generate realistic spatio-temporal movement patterns that capture key aspects of articulatory tradeoff and synergy across a range of consonant-vowel combinations.
comment: 12 pages, 3 figures, supplementary material: python code
☆ DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action
To build a generalizable Vision-Language-Action (VLA) model with strong reasoning ability, a common strategy is to first train a specialist VLA on robot demonstrations to acquire reliable manipulation skills, and then incorporate mixed annotated robot data together with multimodal data to restore broader reasoning capabilities. However, we observe that the resulting reasoning VLA often suffers from degraded action performance compared to the specialist model before fine-tuning, a phenomenon we refer to as action degeneration. To address this issue, we propose DualVLA, which enhances action performance through carefully designed post-training while still preserving reasoning capability. We first introduce a dual-layer data pruning method that removes redundant embodied reasoning, preventing it from adversely influencing action learning. To further strengthen action generation, we design a dual-teacher adaptive distillation strategy that assigns different supervision signals to different data domains while maintaining reasoning ability. To fill the evaluation gap for generalist VLAs, we also propose VLA Score, which decouples VLA capability into reasoning, intention, action, and alignment dimensions for a more fine-grained assessment. Experiments show that DualVLA achieves an average success rate of 61.0 in SimplerEnv and an average score of 65.4 across eight competitive multimodal benchmarks, demonstrating a stronger balance between precise action execution and multimodal understanding. Project Website: https://costaliya.github.io/DualVLA/.
☆ Design of an Adaptive Modular Anthropomorphic Dexterous Hand for Human-like Manipulation
Biological synergies have emerged as a widely adopted paradigm for dexterous hand design, enabling human-like manipulation with a small number of actuators. Nonetheless, excessive coupling tends to diminish the dexterity of hands. This paper tackles the trade-off between actuation complexity and dexterity by proposing an anthropomorphic finger topology with 4 DoFs driven by 2 actuators, and by developing an adaptive, modular dexterous hand based on this finger topology. We explore the biological basis of hand synergies and human gesture analysis, translating joint-level coordination and structural attributes into a modular finger architecture. Leveraging these biomimetic mappings, we design a five-finger modular hand and establish its kinematic model to analyze adaptive grasping and in-hand manipulation. Finally, we construct a physical prototype and conduct preliminary experiments, which validate the effectiveness of the proposed design and analysis.
comment: 7 pages, 8 figures
☆ SoftNash: Entropy-Regularized Nash Games for Non-Fighting Virtual Fixtures
Virtual fixtures (VFs) improve precision in teleoperation but often ``fight'' the user, inflating mental workload and eroding the sense of agency. We propose Soft-Nash Virtual Fixtures, a game-theoretic shared-control policy that softens the classic two-player linear-quadratic (LQ) Nash solution by inflating the fixture's effort weight with a single, interpretable scalar parameter $τ$. This yields a continuous dial on controller assertiveness: $τ=0$ recovers a hard, performance-focused Nash / virtual fixture controller, while larger $τ$ reduce gains and pushback, yet preserve the equilibrium structure and continuity of closed-loop stability. We derive Soft-Nash from both a KL-regularized trust-region and a maximum-entropy viewpoint, obtaining a closed-form robot best response that shrinks authority and aligns the fixture with the operator's input as $τ$ grows. We implement Soft-Nash on a 6-DoF haptic device in 3D tracking task ($n=12$). Moderate softness ($τ\approx 1-3$, especially $τ=2$) maintains tracking error statistically indistinguishable from a tuned classic VF while sharply reducing controller-user conflict, lowering NASA-TLX workload, and increasing Sense of Agency (SoAS). A composite BalancedScore that combines normalized accuracy and non-fighting behavior peaks near $τ=2-3$. These results show that a one-parameter Soft-Nash policy can preserve accuracy while improving comfort and perceived agency, providing a practical and interpretable pathway to personalized shared control in haptics and teleoperation.
☆ SwordRiding: A Unified Navigation Framework for Quadrotors in Unknown Complex Environments via Online Guiding Vector Fields
Although quadrotor navigation has achieved high performance in trajectory planning and control, real-time adaptability in unknown complex environments remains a core challenge. This difficulty mainly arises because most existing planning frameworks operate in an open-loop manner, making it hard to cope with environmental uncertainties such as wind disturbances or external perturbations. This paper presents a unified real-time navigation framework for quadrotors in unknown complex environments, based on the online construction of guiding vector fields (GVFs) from discrete reference path points. In the framework, onboard perception modules build a Euclidean Signed Distance Field (ESDF) representation of the environment, which enables obstacle awareness and path distance evaluation. The system first generates discrete, collision-free path points using a global planner, and then parameterizes them via uniform B-splines to produce a smooth and physically feasible reference trajectory. An adaptive GVF is then synthesized from the ESDF and the optimized B-spline trajectory. Unlike conventional approaches, the method adopts a closed-loop navigation paradigm, which significantly enhances robustness under external disturbances. Compared with conventional GVF methods, the proposed approach directly accommodates discretized paths and maintains compatibility with standard planning algorithms. Extensive simulations and real-world experiments demonstrate improved robustness against external disturbances and superior real-time performance.
comment: For an experimental demo, see https://www.youtube.com/watch?v=tKYCg266c4o. For the lemma proof, see https://github.com/SmartGroupSystems/GVF_close_loop_planning/blob/main/proofs.md
☆ Constant-Volume Deformation Manufacturing for Material-Efficient Shaping
Additive and subtractive manufacturing enable complex geometries but rely on discrete stacking or local removal, limiting continuous and controllable deformation and causing volume loss and shape deviations. We present a volumepreserving digital-mold paradigm that integrates real-time volume-consistency modeling with geometry-informed deformation prediction and an error-compensation strategy to achieve highly predictable shaping of plastic materials. By analyzing deformation patterns and error trends from post-formed point clouds, our method corrects elastic rebound and accumulation errors, maintaining volume consistency and surface continuity. Experiments on five representative geometries demonstrate that the system reproduces target shapes with high fidelity while achieving over 98% material utilization. This approach establishes a digitally driven, reproducible pathway for sustainable, zero-waste shaping of user-defined designs, bridging digital modeling, real-time sensing, and adaptive forming, and advancing next-generation sustainable and customizable manufacturing.
comment: 46 pages, 27 figures
♻ ☆ VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation
A key barrier to the real-world deployment of humanoid robots is the lack of autonomous loco-manipulation skills. We introduce VIRAL, a visual sim-to-real framework that learns humanoid loco-manipulation entirely in simulation and deploys it zero-shot to real hardware. VIRAL follows a teacher-student design: a privileged RL teacher, operating on full state, learns long-horizon loco-manipulation using a delta action space and reference state initialization. A vision-based student policy is then distilled from the teacher via large-scale simulation with tiled rendering, trained with a mixture of online DAgger and behavior cloning. We find that compute scale is critical: scaling simulation to tens of GPUs (up to 64) makes both teacher and student training reliable, while low-compute regimes often fail. To bridge the sim-to-real gap, VIRAL combines large-scale visual domain randomization over lighting, materials, camera parameters, image quality, and sensor delays--with real-to-sim alignment of the dexterous hands and cameras. Deployed on a Unitree G1 humanoid, the resulting RGB-based policy performs continuous loco-manipulation for up to 54 cycles, generalizing to diverse spatial and appearance variations without any real-world fine-tuning, and approaching expert-level teleoperation performance. Extensive ablations dissect the key design choices required to make RGB-based humanoid loco-manipulation work in practice.
comment: Project website: https://viral-humanoid.github.io/
♻ ☆ Robot-mediated physical Human-Human Interaction in Neurorehabilitation: a position paper
Neurorehabilitation conventionally relies on the interaction between a patient and a physical therapist. Robotic systems can improve and enrich the physical feedback provided to patients after neurological injury, but they under-utilize the adaptability and clinical expertise of trained therapists. In this position paper, we advocate for a novel approach that integrates the therapist's clinical expertise and nuanced decision-making with the strength, accuracy, and repeatability of robotics: Robot-mediated physical Human-Human Interaction. This framework, which enables two individuals to physically interact through robotic devices, has been studied across diverse research groups and has recently emerged as a promising link between conventional manual therapy and rehabilitation robotics, harmonizing the strengths of both approaches. This paper presents the rationale of a multidisciplinary team-including engineers, doctors, and physical therapists-for conducting research that utilizes: a unified taxonomy to describe robot-mediated rehabilitation, a framework of interaction based on social psychology, and a technological approach that makes robotic systems seamless facilitators of natural human-human interaction.
comment: Accepted in IEEE Reviews in Biomedical Engineering
♻ ☆ ArtiBench and ArtiBrain: Benchmarking Generalizable Vision-Language Articulated Object Manipulation
Interactive articulated manipulation requires long-horizon, multi-step interactions with appliances while maintaining physical consistency. Existing vision-language and diffusion-based policies struggle to generalize across parts, instances, and categories. We first introduce ArtiBench, a five-level benchmark covering kitchen, storage, office, and tool environments. ArtiBench enables structured evaluation from cross-part and cross-instance variation to long-horizon multi-object tasks, revealing the core generalization challenges of articulated object manipulation. Building on this benchmark, we propose ArtiBrain, a modular framework that unifies high-level reasoning with adaptive low-level control. ArtiBrain uses a VLM-based Task Reasoner (GPT-4.1) to decompose and validate subgoals, and employs a Hybrid Controller that combines geometry-aware keyframe execution with affordance-guided diffusion for precise and interpretable manipulation. An Affordance Memory Bank continually accumulates successful execution episodes and propagates part-level actionable affordances to unseen articulated parts and configurations. Extensive experiments on ArtiBench show that our ArtiBrain significantly outperforms state-of-the-art multimodal and diffusion-based methods in robustness and generalization. Code and dataset will be released upon acceptance.
♻ ☆ Safe Multi-Robotic Arm Interaction via 3D Convex Shapes
Inter-robot collisions pose a significant safety risk when multiple robotic arms operate in close proximity. We present an online collision avoidance methodology leveraging High-Order Control Barrier Functions (HOCBFs) constructed for safe interactions among 3D convex shapes to address this issue. While prior works focused on using Control Barrier Functions (CBFs) for human-robotic arm and single-arm collision avoidance, we explore the problem of collision avoidance between multiple robotic arms operating in a shared space. In our methodology, we utilize the proposed HOCBFs as centralized and decentralized safety filters. These safety filters are compatible with many nominal controllers and ensure safety without significantly restricting the robots' workspace. A key challenge in implementing these filters is the computational overhead caused by the large number of safety constraints and the computation of a Hessian matrix per constraint. We address this challenge by employing numerical differentiation methods to approximate computationally intensive terms. The effectiveness of our method is demonstrated through extensive simulation studies and real-world experiments with Franka Research 3 robotic arms. The project video is available at this link.
comment: The formal publication is available at DOI: https://doi.org/10.1016/j.robot.2025.105263
♻ ☆ D-LIO: 6DoF Direct LiDAR-Inertial Odometry based on Simultaneous Truncated Distance Field Mapping
This paper presents a new approach for 6DoF Direct LiDAR-Inertial Odometry (D-LIO) based on the simultaneous mapping of truncated distance fields on CPU. Such continuous representation (in the vicinity of the points) enables working with raw 3D LiDAR data online, avoiding the need of LiDAR feature selection and tracking, simplifying the odometry pipeline and easily generalizing to many scenarios. The method is based on the proposed Fast Truncated Distance Field (Fast-TDF) method as a convenient tool to represent the environment. Such representation enables i) solving the LiDAR point-cloud registration as a nonlinear optimization process without the need of selecting/tracking LiDAR features in the input data, ii) simultaneously producing an accurate truncated distance field map of the environment, and iii) updating such map at constant time independently of its size. The approach is tested using open datasets, aerial and ground. It is also benchmarked against other state-of-the-art odometry approaches, demonstrating the same or better level of accuracy with the added value of an online-generated TDF representation of the environment, that can be used for other robotics tasks as planning or collision avoidance. The source code is publicly available at https://anonymous.4open.science/r/D-LIO
comment: 9 pages, 3 figures and 43 references
♻ ☆ EfficientNav: Towards On-Device Object-Goal Navigation with Navigation Map Caching and Retrieval NeurIPS 2025
Object-goal navigation (ObjNav) tasks an agent with navigating to the location of a specific object in an unseen environment. Embodied agents equipped with large language models (LLMs) and online constructed navigation maps can perform ObjNav in a zero-shot manner. However, existing agents heavily rely on giant LLMs on the cloud, e.g., GPT-4, while directly switching to small LLMs, e.g., LLaMA3.2-11b, suffer from significant success rate drops due to limited model capacity for understanding complex navigation maps, which prevents deploying ObjNav on local devices. At the same time, the long prompt introduced by the navigation map description will cause high planning latency on local devices. In this paper, we propose EfficientNav to enable on-device efficient LLM-based zero-shot ObjNav. To help the smaller LLMs better understand the environment, we propose semantics-aware memory retrieval to prune redundant information in navigation maps. To reduce planning latency, we propose discrete memory caching and attention-based memory clustering to efficiently save and re-use the KV cache. Extensive experimental results demonstrate that EfficientNav achieves 11.1% improvement in success rate on HM3D benchmark over GPT-4-based baselines, and demonstrates 6.7x real-time latency reduction and 4.7x end-to-end latency reduction over GPT-4 planner. Our code is available on https://github.com/PKU-SEC-Lab/EfficientNav.
comment: NeurIPS 2025
♻ ☆ A K-means Inspired Solution Framework for Large-Scale Multi-Traveling Salesman Problems
The Multi-Traveling Salesman Problem (MTSP) is a commonly used mathematical model for multi-agent task allocation. However, as the number of agents and task targets increases, existing optimization-based methods often incur prohibitive computational costs, posing significant challenges to large-scale coordination in unmanned systems. To address this issue, this paper proposes a K-means-inspired task allocation framework that reformulates the MTSP as a spatially constrained classification process. By leveraging spatial coherence, the proposed method enables fast estimation of path costs and efficient task grouping, thereby fundamentally reducing overall computational complexity. Extensive simulation results demonstrate that the framework can maintain high solution quality even in extremely large-scale scenarios-for instance, in tasks involving 1000 agents and 5000 targets. The findings indicate that this "cluster-then-route" decomposition strategy offers an efficient and reliable solution for large-scale multi-agent task allocation.
♻ ☆ MonoDream: Monocular Vision-Language Navigation with Panoramic Dreaming
Vision-Language Navigation (VLN) tasks often leverage panoramic RGB and depth inputs to provide rich spatial cues for action planning, but these sensors can be costly or less accessible in real-world deployments. Recent approaches based on Vision-Language Action (VLA) models achieve strong results with monocular input, yet they still lag behind methods using panoramic RGB-D information. We present MonoDream, a lightweight VLA framework that enables monocular agents to learn a Unified Navigation Representation (UNR). This shared feature representation jointly aligns navigation-relevant visual semantics (e.g., global layout, depth, and future cues) and language-grounded action intent, enabling more reliable action prediction. MonoDream further introduces Latent Panoramic Dreaming (LPD) tasks to supervise the UNR, which train the model to predict latent features of panoramic RGB and depth observations at both current and future steps based on only monocular input. Experiments on multiple VLN benchmarks show that MonoDream consistently improves monocular navigation performance and significantly narrows the gap with panoramic-based agents.
♻ ☆ Foundation Models in Autonomous Driving: A Survey on Scenario Generation and Scenario Analysis
For autonomous vehicles, safe navigation in complex environments depends on handling a broad range of diverse and rare driving scenarios. Simulation- and scenario-based testing have emerged as key approaches to development and validation of autonomous driving systems. Traditional scenario generation relies on rule-based systems, knowledge-driven models, and data-driven synthesis, often producing limited diversity and unrealistic safety-critical cases. With the emergence of foundation models, which represent a new generation of pre-trained, general-purpose AI models, developers can process heterogeneous inputs (e.g., natural language, sensor data, HD maps, and control actions), enabling the synthesis and interpretation of complex driving scenarios. In this paper, we conduct a survey about the application of foundation models for scenario generation and scenario analysis in autonomous driving (as of May 2025). Our survey presents a unified taxonomy that includes large language models, vision-language models, multimodal large language models, diffusion models, and world models for the generation and analysis of autonomous driving scenarios. In addition, we review the methodologies, open-source datasets, simulation platforms, and benchmark challenges, and we examine the evaluation metrics tailored explicitly to scenario generation and analysis. Finally, the survey concludes by highlighting the open challenges and research questions, and outlining promising future research directions. All reviewed papers are listed in a continuously maintained repository, which contains supplementary materials and is available at https://github.com/TUM-AVS/FM-for-Scenario-Generation-Analysis.
comment: Revised manuscript with separate evaluation metrics table
♻ ☆ Scaling Spatial Intelligence with Multimodal Foundation Models
Despite remarkable progress, multimodal foundation models still exhibit surprising deficiencies in spatial intelligence. In this work, we explore scaling up multimodal foundation models to cultivate spatial intelligence within the SenseNova-SI family, built upon established multimodal foundations including visual understanding models (i.e., Qwen3-VL and InternVL3) and unified understanding and generation models (i.e., Bagel). We take a principled approach to constructing high-performing and robust spatial intelligence by systematically curating SenseNova-SI-8M: eight million diverse data samples under a rigorous taxonomy of spatial capabilities. SenseNova-SI demonstrates unprecedented performance across a broad range of spatial intelligence benchmarks: 68.7% on VSI-Bench, 43.3% on MMSI, 85.6% on MindCube, 54.6% on ViewSpatial, and 50.1% on SITE, while maintaining strong general multimodal understanding (e.g., 84.9% on MMBench-En). More importantly, we analyze the impact of data scaling, discuss early signs of emergent generalization capabilities enabled by diverse data training, analyze the risk of overfitting and language shortcuts, present a preliminary study on spatial chain-of-thought reasoning, and validate the potential downstream application. SenseNova-SI is an ongoing project, and this report will be updated continuously. All newly trained multimodal foundation models are publicly released to facilitate further research in this direction.
comment: Codebase: https://github.com/OpenSenseNova/SenseNova-SI; Models: https://huggingface.co/collections/sensenova/sensenova-si
♻ ☆ Holistic Evaluation of Multimodal LLMs on Spatial Intelligence
Multimodal models have achieved remarkable progress in recent years. Nevertheless, they continue to exhibit notable limitations in spatial understanding and reasoning, the very capability that anchors artificial general intelligence in the physical world. With the recent release of GPT-5, allegedly the most powerful AI model to date, it is timely to examine where the leading models (GPT, Gemini, Grok, Seed, Qwen, and Intern) stand on the path toward spatial intelligence (SI). We thus propose EASI for holistic Evaluation of multimodAl LLMs on Spatial Intelligence. EASI conceptualizes a comprehensive taxonomy of spatial tasks that unifies existing benchmarks and a growing collection of newly curated ones, enabling systematic evaluation of state-of-the-art models. In this report, we conduct the study across eight key benchmarks, at a cost exceeding ten billion total tokens. Our empirical study then reveals that (1) GPT-5 demonstrates unprecedented strength in SI, yet (2) still falls short of human performance significantly across a broad spectrum of SI-tasks. Moreover, we (3) show that SI-tasks expose greater model capability deficiency than non-SI tasks, to the extent that (4) proprietary models do not exhibit a decisive advantage when facing the most difficult ones. In addition, we conduct a qualitative evaluation across a diverse set of scenarios that are intuitive for humans, yet fail the most advanced multimodal models. EASI is an ongoing community effort: we have open-sourced the EASI codebase that provides a one-stop and reproducible solution with standardized interfaces, integrated protocols and prompts that significantly reduce the friction of configuring and running multiple benchmarks; we have also launched an accompanying EASI leaderboard to provide a continually updated snapshot of model performance across the full SI spectrum, accelerating collective progress toward robust SI.
comment: Codebase: https://github.com/EvolvingLMMs-Lab/EASI/; Leaderboard: https://huggingface.co/spaces/lmms-lab-si/EASI-Leaderboard
♻ ☆ Heuristic Step Planning for Learning Dynamic Bipedal Locomotion: A Comparative Study of Model-Based and Model-Free Approaches
This work presents an extended framework for learning-based bipedal locomotion that incorporates a heuristic step-planning strategy guided by desired torso velocity tracking. The framework enables precise interaction between a humanoid robot and its environment, supporting tasks such as crossing gaps and accurately approaching target objects. Unlike approaches based on full or simplified dynamics, the proposed method avoids complex step planners and analytical models. Step planning is primarily driven by heuristic commands, while a Raibert-type controller modulates the foot placement length based on the error between desired and actual torso velocity. We compare our method with a model-based step-planning approach -- the Linear Inverted Pendulum Model (LIPM) controller. Experimental results demonstrate that our approach attains comparable or superior accuracy in maintaining target velocity (up to 80%), significantly greater robustness on uneven terrain (over 50% improvement), and improved energy efficiency. These results suggest that incorporating complex analytical, model-based components into the training architecture may be unnecessary for achieving stable and robust bipedal walking, even in unstructured environments.
♻ ☆ Periodic Skill Discovery NeurIPS 2025
Unsupervised skill discovery in reinforcement learning (RL) aims to learn diverse behaviors without relying on external rewards. However, current methods often overlook the periodic nature of learned skills, focusing instead on increasing the mutual dependence between states and skills or maximizing the distance traveled in latent space. Considering that many robotic tasks - particularly those involving locomotion - require periodic behaviors across varying timescales, the ability to discover diverse periodic skills is essential. Motivated by this, we propose Periodic Skill Discovery (PSD), a framework that discovers periodic behaviors in an unsupervised manner. The key idea of PSD is to train an encoder that maps states to a circular latent space, thereby naturally encoding periodicity in the latent representation. By capturing temporal distance, PSD can effectively learn skills with diverse periods in complex robotic tasks, even with pixel-based observations. We further show that these learned skills achieve high performance on downstream tasks such as hurdling. Moreover, integrating PSD with an existing skill discovery method offers more diverse behaviors, thus broadening the agent's repertoire. Our code and demos are available at https://jonghaepark.github.io/psd/
comment: NeurIPS 2025
♻ ☆ ROVER: Recursive Reasoning Over Videos with Vision-Language Models for Embodied Tasks
Vision-language models (VLMs) have exhibited impressive capabilities across diverse image understanding tasks, but still struggle in settings that require reasoning over extended sequences of camera frames from a video. This limits their utility in embodied settings, which require reasoning over long frame sequences from a continuous stream of visual input at each moment of a task attempt. To address this limitation, we propose ROVER (Reasoning Over VidEo Recursively), a framework that enables the model to recursively decompose long-horizon video trajectories into segments corresponding to shorter subtasks within the trajectory. In doing so, ROVER facilitates more focused and accurate reasoning over temporally localized frame sequences without losing global context. We evaluate ROVER, implemented using an in-context learning approach, on diverse OpenX Embodiment videos and on a new dataset derived from RoboCasa that consists of 543 videos showing both expert and perturbed non-expert trajectories across 27 robotic manipulation tasks. ROVER outperforms strong baselines across three video reasoning tasks: task progress estimation, frame-level natural language reasoning, and video question answering. We observe that, by reducing the number of frames the model reasons over at each timestep, ROVER mitigates hallucinations, especially during unexpected or non-optimal moments of a trajectory. In addition, by enabling the implementation of a subtask-specific sliding context window, ROVER's time complexity scales linearly with video length, an asymptotic improvement over baselines. Demos, code, and data available at: https://rover-vlm.github.io
♻ ☆ Advancing Embodied Intelligence in Robotic-Assisted Endovascular Procedures: A Systematic Review of AI Solutions
Endovascular procedures have revolutionized vascular disease treatment, yet their manual execution is challenged by the demands for high precision, operator fatigue, and radiation exposure. Robotic systems have emerged as transformative solutions to mitigate these inherent limitations. A pivotal moment has arrived, where a confluence of pressing clinical needs and breakthroughs in AI creates an opportunity for a paradigm shift toward Embodied Intelligence (EI), enabling robots to navigate complex vascular networks and adapt to dynamic physiological conditions. Data-driven approaches, leveraging advanced computer vision, medical image analysis, and machine learning, drive this evolution by enabling real-time vessel segmentation, device tracking, and anatomical landmark detection. Reinforcement learning and imitation learning further enhance navigation strategies and replicate expert techniques. This review systematically analyzes the integration of EI into endovascular robotics, identifying profound systemic challenges such as the heterogeneity in validation standards and the gap between human mimicry and machine-native capabilities. Based on this analysis, a conceptual roadmap is proposed that reframes the ultimate objective away from systems that supplant clinical decision-making. This vision of augmented intelligence, where the clinician's role evolves into that of a high-level supervisor, provides a principled foundation for the future of the field.
comment: 20 pages, 6 figures
♻ ☆ Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning NeurIPS 2025
To enable embodied agents to operate effectively over extended timeframes, it is crucial to develop models that form and access memories to stay contextualized in their environment. In the current paradigm of training transformer-based policies for embodied sequential decision-making tasks, visual inputs often overwhelm the context limits of transformers, while humans can maintain and utilize a lifetime of experience compressed as memories. Significant compression is possible in principle, as much of the input is irrelevant and can be abstracted. However, existing approaches predominantly focus on either recurrent models with fixed-size memory or transformers with full-context reliance. In this work, we propose Memo, a transformer-based architecture and training recipe for reinforcement learning (RL) on memory-intensive, long-horizon tasks. Memo incorporates the creation and retrieval of memory by interleaving periodic summarization tokens with the inputs of a model during training. We demonstrate Memo's effectiveness on a gridworld meta-RL benchmark and a multi-object navigation task in photo-realistic indoor settings. Memo outperforms naive long-context transformer baselines while being more compute and storage efficient. Additionally, Memo generalizes better to longer contexts at inference time and remains robust in streaming settings, where historical context must be truncated to fit inference constraints. Our code is available at: https://github.com/gunshi/memo.
comment: Accepted for Spotlight Presentation at NeurIPS 2025
Robotics 47
☆ TraceGen: World Modeling in 3D Trace Space Enables Learning from Cross-Embodiment Videos
Learning new robot tasks on new platforms and in new scenes from only a handful of demonstrations remains challenging. While videos of other embodiments - humans and different robots - are abundant, differences in embodiment, camera, and environment hinder their direct use. We address the small-data problem by introducing a unifying, symbolic representation - a compact 3D "trace-space" of scene-level trajectories - that enables learning from cross-embodiment, cross-environment, and cross-task videos. We present TraceGen, a world model that predicts future motion in trace-space rather than pixel space, abstracting away appearance while retaining the geometric structure needed for manipulation. To train TraceGen at scale, we develop TraceForge, a data pipeline that transforms heterogeneous human and robot videos into consistent 3D traces, yielding a corpus of 123K videos and 1.8M observation-trace-language triplets. Pretraining on this corpus produces a transferable 3D motion prior that adapts efficiently: with just five target robot videos, TraceGen attains 80% success across four tasks while offering 50-600x faster inference than state-of-the-art video-based world models. In the more challenging case where only five uncalibrated human demonstration videos captured on a handheld phone are available, it still reaches 67.5% success on a real robot, highlighting TraceGen's ability to adapt across embodiments without relying on object detectors or heavy pixel-space generation.
☆ Uncertainty Quantification for Visual Object Pose Estimation SP
Quantifying the uncertainty of an object's pose estimate is essential for robust control and planning. Although pose estimation is a well-studied robotics problem, attaching statistically rigorous uncertainty is not well understood without strict distributional assumptions. We develop distribution-free pose uncertainty bounds about a given pose estimate in the monocular setting. Our pose uncertainty only requires high probability noise bounds on pixel detections of 2D semantic keypoints on a known object. This noise model induces an implicit, non-convex set of pose uncertainty constraints. Our key contribution is SLUE (S-Lemma Uncertainty Estimation), a convex program to reduce this set to a single ellipsoidal uncertainty bound that is guaranteed to contain the true object pose with high probability. SLUE solves a relaxation of the minimum volume bounding ellipsoid problem inspired by the celebrated S-lemma. It requires no initial guess of the bound's shape or size and is guaranteed to contain the true object pose with high probability. For tighter uncertainty bounds at the same confidence, we extend SLUE to a sum-of-squares relaxation hierarchy which is guaranteed to converge to the minimum volume ellipsoidal uncertainty bound for a given set of keypoint constraints. We show this pose uncertainty bound can easily be projected to independent translation and axis-angle orientation bounds. We evaluate SLUE on two pose estimation datasets and a real-world drone tracking scenario. Compared to prior work, SLUE generates substantially smaller translation bounds and competitive orientation bounds. We release code at https://github.com/MIT-SPARK/PoseUncertaintySets.
comment: 18 pages, 9 figures. Code available: https://github.com/MIT-SPARK/PoseUncertaintySets
☆ Model-Based Policy Adaptation for Closed-Loop End-to-End Autonomous Driving NeurIPS 2025
End-to-end (E2E) autonomous driving models have demonstrated strong performance in open-loop evaluations but often suffer from cascading errors and poor generalization in closed-loop settings. To address this gap, we propose Model-based Policy Adaptation (MPA), a general framework that enhances the robustness and safety of pretrained E2E driving agents during deployment. MPA first generates diverse counterfactual trajectories using a geometry-consistent simulation engine, exposing the agent to scenarios beyond the original dataset. Based on this generated data, MPA trains a diffusion-based policy adapter to refine the base policy's predictions and a multi-step Q value model to evaluate long-term outcomes. At inference time, the adapter proposes multiple trajectory candidates, and the Q value model selects the one with the highest expected utility. Experiments on the nuScenes benchmark using a photorealistic closed-loop simulator demonstrate that MPA significantly improves performance across in-domain, out-of-domain, and safety-critical scenarios. We further investigate how the scale of counterfactual data and inference-time guidance strategies affect overall effectiveness.
comment: Published at NeurIPS 2025: https://openreview.net/forum?id=4OLbpaTKJe
☆ VacuumVLA: Boosting VLA Capabilities via a Unified Suction and Gripping Tool for Complex Robotic Manipulation
Vision Language Action models have significantly advanced general purpose robotic manipulation by harnessing large scale pretrained vision and language representations. Among existing approaches, a majority of current VLA systems employ parallel two finger grippers as their default end effectors. However, such grippers face inherent limitations in handling certain real world tasks such as wiping glass surfaces or opening drawers without handles due to insufficient contact area or lack of adhesion. To overcome these challenges, we present a low cost, integrated hardware design that combines a mechanical two finger gripper with a vacuum suction unit, enabling dual mode manipulation within a single end effector. Our system supports flexible switching or synergistic use of both modalities, expanding the range of feasible tasks. We validate the efficiency and practicality of our design within two state of the art VLA frameworks: DexVLA and Pi0. Experimental results demonstrate that with the proposed hybrid end effector, robots can successfully perform multiple complex tasks that are infeasible for conventional two finger grippers alone. All hardware designs and controlling systems will be released.
comment: 8 pages
☆ $\mathcal{E}_0$: Enhancing Generalization and Fine-Grained Control in VLA Models via Continuized Discrete Diffusion
Vision-Language-Action (VLA) models offer a unified framework for robotic manipulation by integrating visual perception, language understanding, and control generation. Yet existing VLA models still struggle to generalize across diverse tasks, scenes, and camera viewpoints, and often produce coarse or unstable actions. We introduce E0, a continuized discrete diffusion framework that formulates action generation as iterative denoising over quantized action tokens. Compared with continuous diffusion policies, E0 offers two key advantages: (1) discrete action tokens align naturally with the symbolic structure of pretrained VLM/VLA backbones, enabling stronger semantic conditioning; and 2. discrete diffusion matches the true quantized nature of real-world robot control-whose hardware constraints (e.g., encoder resolution, control frequency, actuation latency) inherently discretize continuous signals-and therefore benefits from a Bayes-optimal denoiser that models the correct discrete action distribution, leading to stronger generalization. Compared with discrete autoregressive and mask-based discrete diffusion models, E0 supports a significantly larger and finer-grained action vocabulary and avoids the distributional mismatch introduced by masking-based corruptions-yielding more accurate fine-grained action control. We further introduce a spherical viewpoint perturbation augmentation method to improve robustness to camera shifts without additional data. Experiments on LIBERO, VLABench, and ManiSkill show that E0 achieves state-of-the-art performance across 14 diverse environments, outperforming strong baselines by 10.7% on average. Real-world evaluation on a Franka arm confirms that E0 delivers precise, robust, and transferable manipulation, establishing discrete diffusion as a promising direction for generalizable VLA policy learning.
☆ Predictive Safety Shield for Dyna-Q Reinforcement Learning
Obtaining safety guarantees for reinforcement learning is a major challenge to achieve applicability for real-world tasks. Safety shields extend standard reinforcement learning and achieve hard safety guarantees. However, existing safety shields commonly use random sampling of safe actions or a fixed fallback controller, therefore disregarding future performance implications of different safe actions. In this work, we propose a predictive safety shield for model-based reinforcement learning agents in discrete space. Our safety shield updates the Q-function locally based on safe predictions, which originate from a safe simulation of the environment model. This shielding approach improves performance while maintaining hard safety guarantees. Our experiments on gridworld environments demonstrate that even short prediction horizons can be sufficient to identify the optimal path. We observe that our approach is robust to distribution shifts, e.g., between simulation and reality, without requiring additional training.
☆ Hybrid Control for Robotic Nut Tightening Task
An autonomous robotic nut tightening system for a serial manipulator equipped with a parallel gripper is proposed. The system features a hierarchical motion-primitive-based planner and a control-switching scheme that alternates between force and position control. Extensive simulations demonstrate the system's robustness to variance in initial conditions. Additionally, the proposed controller tightens threaded screws 14% faster than the baseline while applying 40 times less contact force on manipulands. For the benefit of the research community, the system's implementation is open-sourced.
☆ Scalable Multisubject Vital Sign Monitoring With mmWave FMCW Radar and FPGA Prototyping
In this work, we introduce an innovative approach to estimate the vital signs of multiple human subjects simultaneously in a non-contact way using a Frequency Modulated Continuous Wave (FMCW) radar-based system. Traditional vital sign monitoring methods often face significant limitations, including subject discomfort with wearable devices, challenges in calibration, and the risk of infection transmission through contact measurement devices. To address these issues, this research is motivated by the need for versatile, non-contact vital monitoring solutions applicable in various critical scenarios. This work also explores the challenges of extending this capability to an arbitrary number of subjects, including hardware and theoretical limitations. Supported by rigorous experimental results and discussions, the paper illustrates the system's potential to redefine vital sign monitoring. An FPGA-based implementation is also presented as proof of concept for a hardware-based and portable solution, improving upon previous works by offering 2.7x faster execution and 18.4% less Look-Up Table (LUT) utilization, as well as providing over 7400x acceleration compared to its software counterpart.
comment: Published in IEEE Sensors Journal
☆ Neural NMPC through Signed Distance Field Encoding for Collision Avoidance
This paper introduces a neural Nonlinear Model Predictive Control (NMPC) framework for mapless, collision-free navigation in unknown environments with Aerial Robots, using onboard range sensing. We leverage deep neural networks to encode a single range image, capturing all the available information about the environment, into a Signed Distance Function (SDF). The proposed neural architecture consists of two cascaded networks: a convolutional encoder that compresses the input image into a low-dimensional latent vector, and a Multi-Layer Perceptron that approximates the corresponding spatial SDF. This latter network parametrizes an explicit position constraint used for collision avoidance, which is embedded in a velocity-tracking NMPC that outputs thrust and attitude commands to the robot. First, a theoretical analysis of the contributed NMPC is conducted, verifying recursive feasibility and stability properties under fixed observations. Subsequently, we evaluate the open-loop performance of the learning-based components as well as the closed-loop performance of the controller in simulations and experiments. The simulation study includes an ablation study, comparisons with two state-of-the-art local navigation methods, and an assessment of the resilience to drifting odometry. The real-world experiments are conducted in forest environments, demonstrating that the neural NMPC effectively performs collision avoidance in cluttered settings against an adversarial reference velocity input and drifting position estimates.
comment: accepted for publication in IJRR
☆ Improvement of Collision Avoidance in Cut-In Maneuvers Using Time-to-Collision Metrics
This paper proposes a new strategy for collision avoidance system leveraging Time-to-Collision (TTC) metrics for handling cut-in scenarios, which are particularly challenging for autonomous vehicles (AVs). By integrating a deep learning with TTC calculations, the system predicts potential collisions and determines appropriate evasive actions compared to traditional TTC -based approaches.
☆ Sampling-Based Optimization with Parallelized Physics Simulator for Bimanual Manipulation
In recent years, dual-arm manipulation has become an area of strong interest in robotics, with end-to-end learning emerging as the predominant strategy for solving bimanual tasks. A critical limitation of such learning-based approaches, however, is their difficulty in generalizing to novel scenarios, especially within cluttered environments. This paper presents an alternative paradigm: a sampling-based optimization framework that utilizes a GPU-accelerated physics simulator as its world model. We demonstrate that this approach can solve complex bimanual manipulation tasks in the presence of static obstacles. Our contribution is a customized Model Predictive Path Integral Control (MPPI) algorithm, \textbf{guided by carefully designed task-specific cost functions,} that uses GPU-accelerated MuJoCo for efficiently evaluating robot-object interaction. We apply this method to solve significantly more challenging versions of tasks from the PerAct$^{2}$ benchmark, such as requiring the point-to-point transfer of a ball through an obstacle course. Furthermore, we establish that our method achieves real-time performance on commodity GPUs and facilitates successful sim-to-real transfer by leveraging unique features within MuJoCo. The paper concludes with a statistical analysis of the sample complexity and robustness, quantifying the performance of our approach. The project website is available at: https://sites.google.com/view/bimanualakslabunitartu .
comment: 9 pages, 5 figures
☆ Design and Measurements of mmWave FMCW Radar Based Non-Contact Multi-Patient Heart Rate and Breath Rate Monitoring System
Recent developments in mmWave radar technologies have enabled the truly non-contact heart-rate (HR) and breath-rate (BR) measurement approaches, which provides a great ease in patient monitoring. Additionally, these technologies also provide opportunities to simultaneously detect HR and BR of multiple patients, which has become increasingly important for efficient mass monitoring scenarios. In this work, a frequency modulated continuous wave (FMCW) mmWave radar based truly non-contact multiple patient HR and BR monitoring system has been presented. Furthermore, a novel approach is also proposed, which combines multiple processing methods using a least squares solution to improve measurement accuracy, generalization, and handle measurement error. The proposed system has been developed using Texas Instruments' FMCW radar and experimental results with multiple subjects are also presented, which show >97% and >93% accuracy in the measured BR and HR values, respectively.
comment: Presented at BioCAS 2023
Transformer Driven Visual Servoing and Dual Arm Impedance Control for Fabric Texture Matching
In this paper, we propose a method to align and place a fabric piece on top of another using a dual-arm manipulator and a grayscale camera, so that their surface textures are accurately matched. We propose a novel control scheme that combines Transformer-driven visual servoing with dualarm impedance control. This approach enables the system to simultaneously control the pose of the fabric piece and place it onto the underlying one while applying tension to keep the fabric piece flat. Our transformer-based network incorporates pretrained backbones and a newly introduced Difference Extraction Attention Module (DEAM), which significantly enhances pose difference prediction accuracy. Trained entirely on synthetic images generated using rendering software, the network enables zero-shot deployment in real-world scenarios without requiring prior training on specific fabric textures. Real-world experiments demonstrate that the proposed system accurately aligns fabric pieces with different textures.
comment: 8 pages, 11 figures. Accepted to IEEE Robotics and Automation Letters (RA-L)
☆ Dual Preintegration for Relative State Estimation
Relative State Estimation perform mutually localization between two mobile agents undergoing six-degree-of-freedom motion. Based on the principle of circular motion, the estimation accuracy is sensitive to nonlinear rotations of the reference platform, particularly under large inter-platform distances. This phenomenon is even obvious for linearized kinematics, because cumulative linearization errors significantly degrade precision. In virtual reality (VR) applications, this manifests as substantial positional errors in 6-DoF controller tracking during rapid rotations of the head-mounted display. The linearization errors introduce drift in the estimate and render the estimator inconsistent. In the field of odometry, IMU preintegration is proposed as a kinematic observation to enable efficient relinearization, thus mitigate linearized error. Building on this theory, we propose dual preintegration, a novel observation integrating IMU preintegration from both platforms. This method serves as kinematic constraints for consecutive relative state and supports efficient relinearization. We also perform observability analysis of the state and analytically formulate the accordingly null space. Algorithm evaluation encompasses both simulations and real-world experiments. Multiple nonlinear rotations on the reference platform are simulated to compare the precision of the proposed method with that of other state-of-the-art (SOTA) algorithms. The field test compares the proposed method and SOTA algorithms in the application of VR controller tracking from the perspectives of bias observability, nonlinear rotation, and background texture. The results demonstrate that the proposed method is more precise and robust than the SOTA algorithms.
☆ Kinematics-Aware Multi-Policy Reinforcement Learning for Force-Capable Humanoid Loco-Manipulation
Humanoid robots, with their human-like morphology, hold great potential for industrial applications. However, existing loco-manipulation methods primarily focus on dexterous manipulation, falling short of the combined requirements for dexterity and proactive force interaction in high-load industrial scenarios. To bridge this gap, we propose a reinforcement learning-based framework with a decoupled three-stage training pipeline, consisting of an upper-body policy, a lower-body policy, and a delta-command policy. To accelerate upper-body training, a heuristic reward function is designed. By implicitly embedding forward kinematics priors, it enables the policy to converge faster and achieve superior performance. For the lower body, a force-based curriculum learning strategy is developed, enabling the robot to actively exert and regulate interaction forces with the environment.
☆ MarketGen: A Scalable Simulation Platform with Auto-Generated Embodied Supermarket Environments
The development of embodied agents for complex commercial environments is hindered by a critical gap in existing robotics datasets and benchmarks, which primarily focus on household or tabletop settings with short-horizon tasks. To address this limitation, we introduce MarketGen, a scalable simulation platform with automatic scene generation for complex supermarket environments. MarketGen features a novel agent-based Procedural Content Generation (PCG) framework. It uniquely supports multi-modal inputs (text and reference images) and integrates real-world design principles to automatically generate complete, structured, and realistic supermarkets. We also provide an extensive and diverse 3D asset library with a total of 1100+ supermarket goods and parameterized facilities assets. Building on this generative foundation, we propose a novel benchmark for assessing supermarket agents, featuring two daily tasks in a supermarket: (1) Checkout Unloading: long-horizon tabletop tasks for cashier agents, and (2) In-Aisle Item Collection: complex mobile manipulation tasks for salesperson agents. We validate our platform and benchmark through extensive experiments, including the deployment of a modular agent system and successful sim-to-real transfer. MarketGen provides a comprehensive framework to accelerate research in embodied AI for complex commercial applications.
comment: Project Page: https://xuhu0529.github.io/MarketGen
☆ Maglev-Pentabot: Magnetic Levitation System for Non-Contact Manipulation using Deep Reinforcement Learning
Non-contact manipulation has emerged as a transformative approach across various industrial fields. However, current flexible 2D and 3D non-contact manipulation techniques are often limited to microscopic scales, typically controlling objects in the milligram range. In this paper, we present a magnetic levitation system, termed Maglev-Pentabot, designed to address this limitation. The Maglev-Pentabot leverages deep reinforcement learning (DRL) to develop complex control strategies for manipulating objects in the gram range. Specifically, we propose an electromagnet arrangement optimized through numerical analysis to maximize controllable space. Additionally, an action remapping method is introduced to address sample sparsity issues caused by the strong nonlinearity in magnetic field intensity, hence allowing the DRL controller to converge. Experimental results demonstrate flexible manipulation capabilities, and notably, our system can generalize to transport tasks it has not been explicitly trained for. Furthermore, our approach can be scaled to manipulate heavier objects using larger electromagnets, offering a reference framework for industrial-scale robotic applications.
☆ SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied Navigation
Embodied navigation that adheres to social norms remains an open research challenge. Our \textbf{SocialNav} is a foundational model for socially-aware navigation with a hierarchical "brain-action" architecture, capable of understanding high-level social norms and generating low-level, socially compliant trajectories. To enable such dual capabilities, we construct the SocNav Dataset, a large-scale collection of 7 million samples, comprising (1) a Cognitive Activation Dataset providing social reasoning signals such as chain-of-thought explanations and social traversability prediction, and (2) an Expert Trajectories Pyramid aggregating diverse navigation demonstrations from internet videos, simulated environments, and real-world robots. A multi-stage training pipeline is proposed to gradually inject and refine navigation intelligence: we first inject general navigation skills and social norms understanding into the model via imitation learning, and then refine such skills through a deliberately designed Socially-Aware Flow Exploration GRPO (SAFE-GRPO), the first flow-based reinforcement learning framework for embodied navigation that explicitly rewards socially compliant behaviors. SocialNav achieves +38% success rate and +46% social compliance rate compared to the state-of-the-art method, demonstrating strong gains in both navigation performance and social compliance. Our project page: https://amap-eai.github.io/SocialNav/
☆ Dual-Agent Reinforcement Learning for Adaptive and Cost-Aware Visual-Inertial Odometry
Visual-Inertial Odometry (VIO) is a critical component for robust ego-motion estimation, enabling foundational capabilities such as autonomous navigation in robotics and real-time 6-DoF tracking for augmented reality. Existing methods face a well-known trade-off: filter-based approaches are efficient but prone to drift, while optimization-based methods, though accurate, rely on computationally prohibitive Visual-Inertial Bundle Adjustment (VIBA) that is difficult to run on resource-constrained platforms. Rather than removing VIBA altogether, we aim to reduce how often and how heavily it must be invoked. To this end, we cast two key design choices in modern VIO, when to run the visual frontend and how strongly to trust its output, as sequential decision problems, and solve them with lightweight reinforcement learning (RL) agents. Our framework introduces a lightweight, dual-pronged RL policy that serves as our core contribution: (1) a Select Agent intelligently gates the entire VO pipeline based only on high-frequency IMU data; and (2) a composite Fusion Agent that first estimates a robust velocity state via a supervised network, before an RL policy adaptively fuses the full (p, v, q) state. Experiments on the EuRoC MAV and TUM-VI datasets show that, in our unified evaluation, the proposed method achieves a more favorable accuracy-efficiency-memory trade-off than prior GPU-based VO/VIO systems: it attains the best average ATE while running up to 1.77 times faster and using less GPU memory. Compared to classical optimization-based VIO systems, our approach maintains competitive trajectory accuracy while substantially reducing computational load.
☆ AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios AAAI 2026
Referring Multi-Object Tracking (RMOT) aims to achieve precise object detection and tracking through natural language instructions, representing a fundamental capability for intelligent robotic systems. However, current RMOT research remains mostly confined to ground-level scenarios, which constrains their ability to capture broad-scale scene contexts and perform comprehensive tracking and path planning. In contrast, Unmanned Aerial Vehicles (UAVs) leverage their expansive aerial perspectives and superior maneuverability to enable wide-area surveillance. Moreover, UAVs have emerged as critical platforms for Embodied Intelligence, which has given rise to an unprecedented demand for intelligent aerial systems capable of natural language interaction. To this end, we introduce AerialMind, the first large-scale RMOT benchmark in UAV scenarios, which aims to bridge this research gap. To facilitate its construction, we develop an innovative semi-automated collaborative agent-based labeling assistant (COALA) framework that significantly reduces labor costs while maintaining annotation quality. Furthermore, we propose HawkEyeTrack (HETrack), a novel method that collaboratively enhances vision-language representation learning and improves the perception of UAV scenarios. Comprehensive experiments validated the challenging nature of our dataset and the effectiveness of our method.
comment: AAAI 2026
Dataset Poisoning Attacks on Behavioral Cloning Policies SP 2025
Behavior Cloning (BC) is a popular framework for training sequential decision policies from expert demonstrations via supervised learning. As these policies are increasingly being deployed in the real world, their robustness and potential vulnerabilities are an important concern. In this work, we perform the first analysis of the efficacy of clean-label backdoor attacks on BC policies. Our backdoor attacks poison a dataset of demonstrations by injecting a visual trigger to create a spurious correlation that can be exploited at test time. We evaluate how policy vulnerability scales with the fraction of poisoned data, the strength of the trigger, and the trigger type. We also introduce a novel entropy-based test-time trigger attack that substantially degrades policy performance by identifying critical states where test-time triggering of the backdoor is expected to be most effective at degrading performance. We empirically demonstrate that BC policies trained on even minimally poisoned datasets exhibit deceptively high, near-baseline task performance despite being highly vulnerable to backdoor trigger attacks during deployment. Our results underscore the urgent need for more research into the robustness of BC policies, particularly as large-scale datasets are increasingly used to train policies for real-world cyber-physical systems. Videos and code are available at https://sites.google.com/view/dataset-poisoning-in-bc.
comment: Accepted at EAI SmartSP 2025
☆ ENACT: Evaluating Embodied Cognition with World Modeling of Egocentric Interaction
Embodied cognition argues that intelligence arises from sensorimotor interaction rather than passive observation. It raises an intriguing question: do modern vision-language models (VLMs), trained largely in a disembodied manner, exhibit signs of embodied cognition? We introduce ENACT, a benchmark that casts evaluation of embodied cognition as world modeling from egocentric interaction in a visual question answering (VQA) format. Framed as a partially observable Markov decision process (POMDP) whose actions are scene graph changes, ENACT comprises two complementary sequence reordering tasks: forward world modeling (reorder shuffled observations given actions) and inverse world modeling (reorder shuffled actions given observations). While conceptually simple, solving these tasks implicitly demands capabilities central to embodied cognition-affordance recognition, action-effect reasoning, embodied awareness, and interactive, long-horizon memory from partially observable egocentric input, while avoiding low-level image synthesis that could confound the evaluation. We provide a scalable pipeline that synthesizes QA pairs from robotics simulation (BEHAVIOR) and evaluates models on 8,972 QA pairs spanning long-horizon home-scale activities. Experiments reveal a performance gap between frontier VLMs and humans that widens with interaction horizon. Models consistently perform better on the inverse task than the forward one and exhibit anthropocentric biases, including a preference for right-handed actions and degradation when camera intrinsics or viewpoints deviate from human vision. Website at https://enact-embodied-cognition.github.io/.
comment: Preprint version
☆ RSPECT: Robust and Scalable Planner for Energy-Aware Coordination of UAV-UGV Teams in Aerial Monitoring
We consider the robust planning of energy-constrained unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs), which act as mobile charging stations, to perform long-horizon aerial monitoring missions. More specifically, given a set of points to be visited by the UAVs and desired final positions of the UAV-UGV teams, the objective is to find a robust plan (the vehicle trajectories) that can be realized without a major revision in the face of uncertainty (e.g., unknown obstacles/terrain, wind) to complete this mission in minimum time. We provide a formal description of this problem as a mixed-integer program (MIP), which is NP-hard. Since exact solution methods are computationally intractable for such problems, we propose RSPECT, a scalable and efficient heuristic. We provide theoretical results on the complexity of our algorithm and the feasibility and robustness of resulting plans. We also demonstrate the performance of our method via simulations and experiments.
☆ OpenTwinMap: An Open-Source Digital Twin Generator for Urban Autonomous Driving
Digital twins of urban environments play a critical role in advancing autonomous vehicle (AV) research by enabling simulation, validation, and integration with emerging generative world models. While existing tools have demonstrated value, many publicly available solutions are tightly coupled to specific simulators, difficult to extend, or introduce significant technical overhead. For example, CARLA-the most widely used open-source AV simulator-provides a digital twin framework implemented entirely as an Unreal Engine C++ plugin, limiting flexibility and rapid prototyping. In this work, we propose OpenTwinMap, an open-source, Python-based framework for generating high-fidelity 3D urban digital twins. The completed framework will ingest LiDAR scans and OpenStreetMap (OSM) data to produce semantically segmented static environment assets, including road networks, terrain, and urban structures, which can be exported into Unreal Engine for AV simulation. OpenTwinMap emphasizes extensibility and parallelization, lowering the barrier for researchers to adapt and scale the pipeline to diverse urban contexts. We describe the current capabilities of the OpenTwinMap, which includes preprocessing of OSM and LiDAR data, basic road mesh and terrain generation, and preliminary support for CARLA integration.
☆ Bridging Planning and Execution: Multi-Agent Path Finding Under Real-World Deadlines
The Multi-Agent Path Finding (MAPF) problem aims to find collision-free paths for multiple agents while optimizing objectives such as the sum of costs or makespan. MAPF has wide applications in domains like automated warehouses, manufacturing systems, and airport logistics. However, most MAPF formulations assume a simplified robot model for planning, which overlooks execution-time factors such as kinodynamic constraints, communication latency, and controller variability. This gap between planning and execution is problematic for time-sensitive applications. To bridge this gap, we propose REMAP, an execution-informed MAPF planning framework that can be combined with leading search-based MAPF planners with minor changes. Our framework integrates the proposed ExecTimeNet to accurately estimate execution time based on planned paths. We demonstrate our method for solving MAPF with Real-world Deadlines (MAPF-RD) problem, where agents must reach their goals before a predefined wall-clock time. We integrate our framework with two popular MAPF methods, MAPF-LNS and CBS. Experiments show that REMAP achieves up to 20% improvement in solution quality over baseline methods (e.g., constant execution speed estimators) on benchmark maps with up to 300 agents.
☆ Massively Parallel Imitation Learning of Mouse Forelimb Musculoskeletal Reaching Dynamics NeurIPS 2025
The brain has evolved to effectively control the body, and in order to understand the relationship we need to model the sensorimotor transformations underlying embodied control. As part of a coordinated effort, we are developing a general-purpose platform for behavior-driven simulation modeling high fidelity behavioral dynamics, biomechanics, and neural circuit architectures underlying embodied control. We present a pipeline for taking kinematics data from the neuroscience lab and creating a pipeline for recapitulating those natural movements in a biomechanical model. We implement a imitation learning framework to perform a dexterous forelimb reaching task with a musculoskeletal model in a simulated physics environment. The mouse arm model is currently training at faster than 1 million training steps per second due to GPU acceleration with JAX and Mujoco-MJX. We present results that indicate that adding naturalistic constraints on energy and velocity lead to simulated musculoskeletal activity that better predict real EMG signals. This work provides evidence to suggest that energy and control constraints are critical to modeling musculoskeletal motor control.
comment: Accepted at NeurIPS 2025 Workshop Data on the Brain & Mind: Concrete Applications of AI to Neuroscience and Cognitive Science. 12 pages, 4 figures
♻ ☆ X-Nav: Learning End-to-End Cross-Embodiment Navigation for Mobile Robots
Existing navigation methods are primarily designed for specific robot embodiments, limiting their generalizability across diverse robot platforms. In this paper, we introduce X-Nav, a novel framework for end-to-end cross-embodiment navigation where a single unified policy can be deployed across various embodiments for both wheeled and quadrupedal robots. X-Nav consists of two learning stages: 1) multiple expert policies are trained using deep reinforcement learning with privileged observations on a wide range of randomly generated robot embodiments; and 2) a single general policy is distilled from the expert policies via navigation action chunking with transformer (Nav-ACT). The general policy directly maps visual and proprioceptive observations to low-level control commands, enabling generalization to novel robot embodiments. Simulated experiments demonstrated that X-Nav achieved zero-shot transfer to both unseen embodiments and photorealistic environments. A scalability study showed that the performance of X-Nav improves when trained with an increasing number of randomly generated embodiments. An ablation study confirmed the design choices of X-Nav. Furthermore, real-world experiments were conducted to validate the generalizability of X-Nav in real-world environments.
♻ ☆ Safety Control of Service Robots with LLMs and Embodied Knowledge Graphs
Safety limitations in service robotics across various industries have raised significant concerns about the need for robust mechanisms ensuring that robots adhere to safe practices, thereby preventing actions that might harm humans or cause property damage. Despite advances, including the integration of Knowledge Graphs (KGs) with Large Language Models (LLMs), challenges in ensuring consistent safety in autonomous robot actions persist. In this paper, we propose a novel integration of Large Language Models with Embodied Robotic Control Prompts (ERCPs) and Embodied Knowledge Graphs (EKGs) to enhance the safety framework for service robots. ERCPs are designed as predefined instructions that ensure LLMs generate safe and precise responses. These responses are subsequently validated by EKGs, which provide a comprehensive knowledge base ensuring that the actions of the robot are continuously aligned with safety protocols, thereby promoting safer operational practices in varied contexts. Our experimental setup involved diverse real-world tasks, where robots equipped with our framework demonstrated significantly higher compliance with safety standards compared to traditional methods. This integration fosters secure human-robot interactions and positions our methodology at the forefront of AI-driven safety innovations in service robotics.
♻ ☆ MonoMPC: Monocular Vision Based Navigation with Learned Collision Model and Risk-Aware Model Predictive Control
Navigating unknown environments with a single RGB camera is challenging, as the lack of depth information prevents reliable collision-checking. While some methods use estimated depth to build collision maps, we found that depth estimates from vision foundation models are too noisy for zero-shot navigation in cluttered environments. We propose an alternative approach: instead of using noisy estimated depth for direct collision-checking, we use it as a rich context input to a learned collision model. This model predicts the distribution of minimum obstacle clearance that the robot can expect for a given control sequence. At inference, these predictions inform a risk-aware MPC planner that minimizes estimated collision risk. We proposed a joint learning pipeline that co-trains the collision model and risk metric using both safe and unsafe trajectories. Crucially, our joint-training ensures well calibrated uncertainty in our collision model that improves navigation in highly cluttered environments. Consequently, real-world experiments show reductions in collision-rate and improvements in goal reaching and speed over several strong baselines.
♻ ☆ Floor Plan-Guided Visual Navigation Incorporating Depth and Directional Cues
Guiding an agent to a specific target in indoor environments based solely on RGB inputs and a floor plan is a promising yet challenging problem. Although existing methods have made significant progress, two challenges remain unresolved. First, the modality gap between egocentric RGB observations and the floor plan hinders the integration of visual and spatial information for both local obstacle avoidance and global planning. Second, accurate localization is critical for navigation performance, but remains challenging at deployment in unseen environments due to the lack of explicit geometric alignment between RGB inputs and floor plans. We propose a novel diffusion-based policy, denoted as GlocDiff, which integrates global path planning from the floor plan with local depth-aware features derived from RGB observations. The floor plan offers explicit global guidance, while the depth features provide implicit geometric cues, collectively enabling precise prediction of optimal navigation directions and robust obstacle avoidance. Moreover, GlocDiff introduces noise perturbation during training to enhance robustness against pose estimation errors, and we find that combining this with a relatively stable VO module during inference results in significantly improved navigation performance. Extensive experiments on the FloNa benchmark demonstrate GlocDiff's efficiency and effectiveness in achieving superior navigation performance, and the success of real-world deployments also highlights its potential for widespread practical applications.
♻ ☆ scipy.spatial.transform: Differentiable Framework-Agnostic 3D Transformations in Python
Three-dimensional rigid-body transforms, i.e. rotations and translations, are central to modern differentiable machine learning pipelines in robotics, vision, and simulation. However, numerically robust and mathematically correct implementations, particularly on SO(3), are error-prone due to issues such as axis conventions, normalizations, composition consistency and subtle errors that only appear in edge cases. SciPy's spatial$.$transform module is a rigorously tested Python implementation. However, it historically only supported NumPy, limiting adoption in GPU-accelerated and autodiff-based workflows. We present a complete overhaul of SciPy's spatial$.$transform functionality that makes it compatible with any array library implementing the Python array API, including JAX, PyTorch, and CuPy. The revised implementation preserves the established SciPy interface while enabling GPU/TPU execution, JIT compilation, vectorized batching, and differentiation via native autodiff of the chosen backend. We demonstrate how this foundation supports differentiable scientific computing through two case studies: (i) scalability of 3D transforms and rotations and (ii) a JAX drone simulation that leverages SciPy's Rotation for accurate integration of rotational dynamics. Our contributions have been merged into SciPy main and will ship in the next release, providing a framework-agnostic, production-grade basis for 3D spatial math in differentiable systems and ML.
comment: Accepted as oral at the 1st Workshop on Differentiable Systems and Scientific Machine Learning @ EurIPS 2025
♻ ☆ Heterogeneous Multi-robot Task Allocation for Long-Endurance Missions in Dynamic Scenarios
We present a framework for Multi-Robot Task Allocation (MRTA) in heterogeneous teams performing long-endurance missions in dynamic scenarios. Given the limited battery of robots, especially for aerial vehicles, we allow for robot recharges and the possibility of fragmenting and/or relaying certain tasks. We also address tasks that must be performed by a coalition of robots in a coordinated manner. Given these features, we introduce a new class of heterogeneous MRTA problems which we analyze theoretically and optimally formulate as a Mixed-Integer Linear Program. We then contribute a heuristic algorithm to compute approximate solutions and integrate it into a mission planning and execution architecture capable of reacting to unexpected events by repairing or recomputing plans online. Our experimental results show the relevance of our newly formulated problem in a realistic use case for inspection with aerial robots. We assess the performance of our heuristic solver in comparison with other variants and with exact optimal solutions in small-scale scenarios. In addition, we evaluate the ability of our replanning framework to repair plans online.
comment: 20 pages, 10 figures
♻ ☆ Multi-Agent Monocular Dense SLAM With 3D Reconstruction Priors
Monocular Simultaneous Localization and Mapping (SLAM) aims to estimate a robot's pose while simultaneously reconstructing an unknown 3D scene using a single camera. While existing monocular SLAM systems generate detailed 3D geometry through dense scene representations, they are computationally expensive due to the need for iterative optimization. To address this challenge, MASt3R-SLAM utilizes learned 3D reconstruction priors, enabling more efficient and accurate estimation of both 3D structures and camera poses. However, MASt3R-SLAM is limited to single-agent operation. In this paper, we extend MASt3R-SLAM to introduce the first multi-agent monocular dense SLAM system. Each agent performs local SLAM using a 3D reconstruction prior, and their individual maps are fused into a globally consistent map through a loop-closure-based map fusion mechanism. Our approach improves computational efficiency compared to state-of-the-art methods, while maintaining similar mapping accuracy when evaluated on real-world datasets.
♻ ☆ Rapid and Safe Trajectory Planning over Diverse Scenes through Diffusion Composition
Safe trajectory planning in complex environments must balance stringent collision avoidance with real-time efficiency, which is a long-standing challenge in robotics. In this work, we present a diffusion-based trajectory planning framework that is both rapid and safe. First, we introduce a scene-agnostic, MPC-based data generation pipeline that efficiently produces large volumes of kinematically feasible trajectories. Building on this dataset, our integrated diffusion planner maps raw onboard sensor inputs directly to kinematically feasible trajectories, enabling efficient inference while maintaining strong collision avoidance. To generalize to diverse, previously unseen scenarios, we compose diffusion models at test time, enabling safe behavior without additional training. We further propose a lightweight, rule-based safety filter that, from the candidate set, selects the trajectory meeting safety and kinematic-feasibility requirements. Across seen and unseen settings, the proposed method delivers real-time-capable inference with high safety and stability. Experiments on an F1TENTH vehicle demonstrate practicality on real hardware. Project page: https://rstp-comp-diffuser.github.io/.
♻ ☆ Steering Flexible Linear Objects in Planar Environments by Two Robot Hands Using Euler's Elastica Solutions
The manipulation of flexible objects such as cables, wires and fresh food items by robot hands forms a special challenge in robot grasp mechanics. This paper considers the steering of flexible linear objects in planar environments by two robot hands. The flexible linear object, modeled as an elastic non-stretchable rod, is manipulated by varying the gripping endpoint positions while keeping equal endpoint tangents. The flexible linear object shape has a closed form solution in terms of the grasp endpoint positions and tangents, called Euler's elastica. This paper obtains the elastica solutions under the optimal control framework, then uses the elastica solutions to obtain closed-form criteria for non self-intersection, stability and obstacle avoidance of the flexible linear object. The new tools are incorporated into a planning scheme for steering flexible linear objects in planar environments populated by sparsely spaced obstacles. The scheme is fully implemented and demonstrated with detailed examples.
♻ ☆ A New Framework for Nonlinear Kalman Filters
The Kalman filter (KF) is a state estimation algorithm that optimally combines system knowledge and measurements to minimize the mean squared error of the estimated states. While KF was initially designed for linear systems, numerous extensions of it, such as extended Kalman filter (EKF), unscented Kalman filter (UKF), cubature Kalman filter (CKF), etc., have been proposed for nonlinear systems over the last sixty years. Although different types of nonlinear KFs have different pros and cons, they all use the same framework of linear KF. Yet, according to our theoretical and empirical analysis, the framework tends to give overconfident and less accurate state estimations when the measurement functions are nonlinear. Therefore, in this study, we designed a new framework that can be combined with any existing type of nonlinear KFs and showed theoretically and empirically that the new framework estimates the states and covariance more accurately than the old one. The new framework was tested on four different nonlinear KFs and five different tasks, showcasing its ability to reduce estimation errors by several orders of magnitude in low-measurement-noise conditions. The codes are available at https://github.com/Shida-Jiang/A-new-framework-for-nonlinear-Kalman-filters
comment: Massive revisions throughout the entire paper. More simulation results are added, and the mathematical expressions have become more rigorous and precise
♻ ☆ Connectivity-Preserving Multi-Agent Area Coverage via Optimal-Transport-Based Density-Driven Optimal Control (D2OC)
Multi-agent systems play a central role in area coverage tasks across search-and-rescue, environmental monitoring, and precision agriculture. Achieving non-uniform coverage, where spatial priorities vary across the domain, requires coordinating agents while respecting dynamic and communication constraints. Density-driven approaches can distribute agents according to a prescribed reference density, but existing methods do not ensure connectivity. This limitation often leads to communication loss, reduced coordination, and degraded coverage performance. This letter introduces a connectivity-preserving extension of the Density-Driven Optimal Control (D2OC) framework. The coverage objective, defined using the Wasserstein distance between the agent distribution and the reference density, admits a convex quadratic program formulation. Communication constraints are incorporated through a smooth connectivity penalty, which maintains strict convexity, supports distributed implementation, and preserves inter-agent communication without imposing rigid formations. Simulation studies show that the proposed method consistently maintains connectivity, improves convergence speed, and enhances non-uniform coverage quality compared with density-driven schemes that do not incorporate explicit connectivity considerations.
comment: Under review in IEEE Control Systems Letters (LCSS). 6 pages
♻ ☆ BRIC: Bridging Kinematic Plans and Physical Control at Test Time AAAI'26
We propose BRIC, a novel test-time adaptation (TTA) framework that enables long-term human motion generation by resolving execution discrepancies between diffusion-based kinematic motion planners and reinforcement learning-based physics controllers. While diffusion models can generate diverse and expressive motions conditioned on text and scene context, they often produce physically implausible outputs, leading to execution drift during simulation. To address this, BRIC dynamically adapts the physics controller to noisy motion plans at test time, while preserving pre-trained skills via a loss function that mitigates catastrophic forgetting. In addition, BRIC introduces a lightweight test-time guidance mechanism that steers the diffusion model in the signal space without updating its parameters. By combining both adaptation strategies, BRIC ensures consistent and physically plausible long-term executions across diverse environments in an effective and efficient manner. We validate the effectiveness of BRIC on a variety of long-term tasks, including motion composition, obstacle avoidance, and human-scene interaction, achieving state-of-the-art performance across all tasks.
comment: Accepted to AAAI'26
♻ ☆ Development of a Testbed for Autonomous Vehicles: Integrating MPC Control with Monocular Camera Lane Detection
Autonomous vehicles are becoming popular day by day not only for autonomous road traversal but also for industrial automation, farming and military. Most of the standard vehicles follow the Ackermann style steering mechanism. This has become to de facto standard for large and long faring vehicles. The local planner of an autonomous vehicle controls the low-level vehicle movement upon which the vehicle will perform its motor actuation. In our work, we focus on autonomous vehicles in road and perform experiments to analyze the effect of low-level controllers in the simulation and a real environment. To increase the precision and stability of trajectory tracking in autonomous cars, a novel method that combines lane identification with Model Predictive Control (MPC) is presented. The research focuses on camera-equipped autonomous vehicles and uses methods like edge recognition, sliding window-based straight-line identification for lane line extraction, and dynamic region of interest (ROI) extraction. Next, to follow the identified lane line, an MPC built on a bicycle vehicle dynamics model is created. A single-lane road simulation model is built using ROS Gazebo and tested in order to verify the controller's performance. The root mean square error between the optimal tracking trajectory and the target trajectory was reduced by 27.65% in the simulation results, demonstrating the high robustness and flexibility of the developed controller.
comment: 49 pages, 23 figures
♻ ☆ Analytical Solvers for Common Algebraic Equations Arising in Kinematics Problems
This paper presents analytical solvers for four common types of algebraic equations encountered in robot kinematics: single trigonometric equations, single-angle trigonometric systems, two-angle trigonometric systems, and bilinear two-angle systems. These equations arise frequently in the kinematics problems, particularly in robot kinematics. We provide detailed solution methods, including closed-form expressions, numerical algorithms, and robustness considerations. The solvers are designed to handle general coefficients, manage singularities, and enumerate all real solutions efficiently. These solvers are implemented in Python packages and can be reproduced by prompting Language Lanuage Models. Sampe prompts are also provided in the public code space Github repo. These prompts can generate a working solver code with one single prompt in coding agent such as OpenAI's Codex 5.1. This work serves as a foundation for developing complete inverse kinematics solvers for various robot architectures. Extensive validation and benchmarking demonstrate the effectiveness and reliability of the proposed methods.
♻ ☆ MAPF-HD: Multi-Agent Path Finding in High-Density Environments
Multi-agent path finding (MAPF) involves planning efficient paths for multiple agents to move simultaneously while avoiding collisions. In typical warehouse environments, agents are often sparsely distributed along aisles; however, increasing the agent density can improve space efficiency. When the agent density is high, it becomes necessary to optimize the paths not only for goal-assigned agents but also for those obstructing them. This study proposes a novel MAPF framework for high-density environments (MAPF-HD). Several studies have explored MAPF in similar settings using integer linear programming (ILP). However, ILP-based methods require substantial computation time to optimize all agent paths simultaneously. Even in small grid-based environments with fewer than $100$ cells, these computations can take tens to hundreds of seconds. Such high computational costs render these methods impractical for large-scale applications such as automated warehouses and valet parking. To address these limitations, we introduce the phased null-agent swapping (PHANS) method. PHANS employs a heuristic approach to incrementally swap positions between agents and empty vertices. This method solves the MAPF-HD problem within a few seconds, even in large environments containing more than $700$ cells. The proposed method has the potential to improve efficiency in various real-world applications such as warehouse logistics, traffic management, and crowd control. The implementation is available at https://github.com/ToyotaCRDL/MAPF-in-High-Density-Envs.
comment: 9 pages, 12 figures
♻ ☆ Simultaneous Calibration of Noise Covariance and Kinematics for State Estimation of Legged Robots via Bi-level Optimization
Accurate state estimation is critical for legged and aerial robots operating in dynamic, uncertain environments. A key challenge lies in specifying process and measurement noise covariances, which are typically unknown or manually tuned. In this work, we introduce a bi-level optimization framework that jointly calibrates covariance matrices and kinematic parameters in an estimator-in-the-loop manner. The upper level treats noise covariances and model parameters as optimization variables, while the lower level executes a full-information estimator. Differentiating through the estimator allows direct optimization of trajectory-level objectives, resulting in accurate and consistent state estimates. We validate our approach on quadrupedal and humanoid robots, demonstrating significantly improved estimation accuracy and uncertainty calibration compared to hand-tuned baselines. Our method unifies state estimation, sensor, and kinematics calibration into a principled, data-driven framework applicable across diverse robotic platforms.
♻ ☆ Efficient Learning of Object Placement with Intra-Category Transfer
Efficient learning from demonstration for long-horizon tasks remains an open challenge in robotics. While significant effort has been directed toward learning trajectories, a recent resurgence of object-centric approaches has demonstrated improved sample efficiency, enabling transferable robotic skills. Such approaches model tasks as a sequence of object poses over time. In this work, we propose a scheme for transferring observed object arrangements to novel object instances by learning these arrangements on canonical class frames. We then employ this scheme to enable a simple yet effective approach for training models from as few as five demonstrations to predict arrangements of a wide range of objects including tableware, cutlery, furniture, and desk spaces. We propose a method for optimizing the learned models to enable efficient learning of tasks such as setting a table or tidying up an office with intra-category transfer, even in the presence of distractors. We present extensive experimental results in simulation and on a real robotic system for table setting which, based on human evaluations, scored 73.3% compared to a human baseline. We make the code and trained models publicly available at https://oplict.cs.uni-freiburg.de.
comment: 12 pages, 8 figures, 3 tables, accepted at RA-L November 2025
♻ ☆ PointMapPolicy: Structured Point Cloud Processing for Multi-Modal Imitation Learning
Robotic manipulation systems benefit from complementary sensing modalities, where each provides unique environmental information. Point clouds capture detailed geometric structure, while RGB images provide rich semantic context. Current point cloud methods struggle to capture fine-grained detail, especially for complex tasks, which RGB methods lack geometric awareness, which hinders their precision and generalization. We introduce PointMapPolicy, a novel approach that conditions diffusion policies on structured grids of points without downsampling. The resulting data type makes it easier to extract shape and spatial relationships from observations, and can be transformed between reference frames. Yet due to their structure in a regular grid, we enable the use of established computer vision techniques directly to 3D data. Using xLSTM as a backbone, our model efficiently fuses the point maps with RGB data for enhanced multi-modal perception. Through extensive experiments on the RoboCasa and CALVIN benchmarks and real robot evaluations, we demonstrate that our method achieves state-of-the-art performance across diverse manipulation tasks. The overview and demos are available on our project page: https://point-map.github.io/Point-Map/
♻ ☆ The Role of Consequential and Functional Sound in Human-Robot Interaction: Toward Audio Augmented Reality Interfaces
As robots become increasingly integrated into everyday environments, understanding how they communicate with humans is critical. Sound offers a powerful channel for interaction, encompassing both operational noises and intentionally designed auditory cues. In this study, we examined the effects of consequential and functional sounds on human perception and behavior, including a novel exploration of spatial sound through localization and handover tasks. Results show that consequential sounds of the Kinova Gen3 manipulator did not negatively affect perceptions, spatial localization is highly accurate for lateral cues but declines for frontal cues, and spatial sounds can simultaneously convey task-relevant information while promoting warmth and reducing discomfort. These findings highlight the potential of functional and transformative auditory design to enhance human-robot collaboration and inform future sound-based interaction strategies.
comment: 9 pages, 6 figures
♻ ☆ Differentiable Skill Optimisation for Powder Manipulation in Laboratory Automation IROS 2025
Robotic automation is accelerating scientific discovery by reducing manual effort in laboratory workflows. However, precise manipulation of powders remains challenging, particularly in tasks such as transport that demand accuracy and stability. We propose a trajectory optimisation framework for powder transport in laboratory settings, which integrates differentiable physics simulation for accurate modelling of granular dynamics, low-dimensional skill-space parameterisation to reduce optimisation complexity, and a curriculum-based strategy that progressively refines task competence over long horizons. This formulation enables end-to-end optimisation of contact-rich robot trajectories while maintaining stability and convergence efficiency. Experimental results demonstrate that the proposed method achieves superior task success rates and stability compared to the reinforcement learning baseline.
comment: Accepted by IROS 2025 Workshop on Embodied AI and Robotics for Future Scientific Discovery
♻ ☆ Trust-Preserved Human-Robot Shared Autonomy enabled by Bayesian Relational Event Modeling
Shared autonomy functions as a flexible framework that empowers robots to operate across a spectrum of autonomy levels, allowing for efficient task execution with minimal human oversight. However, humans might be intimidated by the autonomous decision-making capabilities of robots due to perceived risks and a lack of trust. This paper proposed a trust-preserved shared autonomy strategy that allows robots to seamlessly adjust their autonomy level, striving to optimize team performance and enhance their acceptance among human collaborators. By enhancing the relational event modeling framework with Bayesian learning techniques, this paper enables dynamic inference of human trust based solely on time-stamped relational events communicated within human-robot teams. Adopting a longitudinal perspective on trust development and calibration in human-robot teams, the proposed trust-preserved shared autonomy strategy warrants robots to actively establish, maintain, and repair human trust, rather than merely passively adapting to it. We validate the effectiveness of the proposed approach through a user study on a human-robot collaborative search and rescue scenario. The objective and subjective evaluations demonstrate its merits on both task execution and user acceptability over the baseline approach that does not consider the preservation of trust.
Robotics 54
☆ Reinforcing Action Policies by Prophesying
Vision-Language-Action (VLA) policies excel in aligning language, perception, and robot control. However, most VLAs are trained purely by imitation, which overfits to demonstrations, and is brittle under distribution shift. Reinforcement learning (RL) directly optimizes task reward and thus addresses this misalignment, but real-robot interaction is expensive and conventional simulators are hard to engineer and transfer. We address both data efficiency and optimization stability in VLA post-training via a learned world model and an RL procedure tailored to flow-based action heads. Specifically, we introduce Prophet, a unified action-to-video robot actuation pretrained across large-scale, heterogeneous robot data to learn reusable action-outcome dynamics. It is able to few-shot adapt to new robots, objects, and environments, yielding a rollout-ready simulator. Upon Prophet, we reinforce action policies with Flow-action-GRPO (FA-GRPO), which adapts Flow-GRPO to operate on VLA actions, and with FlowScale, a stepwise reweighting that rescales per-step gradients in the flow head. Together, Prophet, FA-GRPO, and FlowScale constitute ProphRL, a practical, data- and compute-efficient path to VLA post-training. Experiments show 5-17% success gains on public benchmarks and 24-30% gains on real robots across different VLA variants.
comment: https://LogosRoboticsGroup.github.io/ProphRL
☆ Wanderland: Geometrically Grounded Simulation for Open-World Embodied AI
Reproducible closed-loop evaluation remains a major bottleneck in Embodied AI such as visual navigation. A promising path forward is high-fidelity simulation that combines photorealistic sensor rendering with geometrically grounded interaction in complex, open-world urban environments. Although recent video-3DGS methods ease open-world scene capturing, they are still unsuitable for benchmarking due to large visual and geometric sim-to-real gaps. To address these challenges, we introduce Wanderland, a real-to-sim framework that features multi-sensor capture, reliable reconstruction, accurate geometry, and robust view synthesis. Using this pipeline, we curate a diverse dataset of indoor-outdoor urban scenes and systematically demonstrate how image-only pipelines scale poorly, how geometry quality impacts novel view synthesis, and how all of these adversely affect navigation policy learning and evaluation reliability. Beyond serving as a trusted testbed for embodied navigation, Wanderland's rich raw sensor data further allows benchmarking of 3D reconstruction and novel view synthesis models. Our work establishes a new foundation for reproducible research in open-world embodied AI. Project website is at https://ai4ce.github.io/wanderland/.
☆ Safe and Stable Neural Network Dynamical Systems for Robot Motion Planning
Learning safe and stable robot motions from demonstrations remains a challenge, especially in complex, nonlinear tasks involving dynamic, obstacle-rich environments. In this paper, we propose Safe and Stable Neural Network Dynamical Systems S$^2$-NNDS, a learning-from-demonstration framework that simultaneously learns expressive neural dynamical systems alongside neural Lyapunov stability and barrier safety certificates. Unlike traditional approaches with restrictive polynomial parameterizations, S$^2$-NNDS leverages neural networks to capture complex robot motions providing probabilistic guarantees through split conformal prediction in learned certificates. Experimental results on various 2D and 3D datasets -- including LASA handwriting and demonstrations recorded kinesthetically from the Franka Emika Panda robot -- validate S$^2$-NNDS effectiveness in learning robust, safe, and stable motions from potentially unsafe demonstrations.
☆ Gated Uncertainty-Aware Runtime Dual Invariants for Neural Signal-Controlled Robotics NeurIPS 2025
Safety-critical assistive systems that directly decode user intent from neural signals require rigorous guarantees of reliability and trust. We present GUARDIAN (Gated Uncertainty-Aware Runtime Dual Invariants), a framework for real-time neuro-symbolic verification for neural signal-controlled robotics. GUARDIAN enforces both logical safety and physiological trust by coupling confidence-calibrated brain signal decoding with symbolic goal grounding and dual-layer runtime monitoring. On the BNCI2014 motor imagery electroencephalogram (EEG) dataset with 9 subjects and 5,184 trials, the system performs at a high safety rate of 94-97% even with lightweight decoder architectures with low test accuracies (27-46%) and high ECE confidence miscalibration (0.22-0.41). We demonstrate 1.7x correct interventions in simulated noise testing versus at baseline. The monitor operates at 100Hz and sub-millisecond decision latency, making it practically viable for closed-loop neural signal-based systems. Across 21 ablation results, GUARDIAN exhibits a graduated response to signal degradation, and produces auditable traces from intent, plan to action, helping to link neural evidence to verifiable robot action.
comment: Embodied and Safe-Assured Robotic Systems workshop at NeurIPS 2025
☆ MIMIC-MJX: Neuromechanical Emulation of Animal Behavior
The primary output of the nervous system is movement and behavior. While recent advances have democratized pose tracking during complex behavior, kinematic trajectories alone provide only indirect access to the underlying control processes. Here we present MIMIC-MJX, a framework for learning biologically-plausible neural control policies from kinematics. MIMIC-MJX models the generative process of motor control by training neural controllers that learn to actuate biomechanically-realistic body models in physics simulation to reproduce real kinematic trajectories. We demonstrate that our implementation is accurate, fast, data-efficient, and generalizable to diverse animal body models. Policies trained with MIMIC-MJX can be utilized to both analyze neural control strategies and simulate behavioral experiments, illustrating its potential as an integrative modeling framework for neuroscience.
☆ Metric, inertially aligned monocular state estimation via kinetodynamic priors
Accurate state estimation for flexible robotic systems poses significant challenges, particular for platforms with dynamically deforming structures that invalidate rigid-body assumptions. This paper tackles this problem and allows to extend existing rigid-body pose estimation methods to non-rigid systems. Our approach hinges on two core assumptions: first, the elastic properties are captured by an injective deformation-force model, efficiently learned via a Multi-Layer Perceptron; second, we solve the platform's inherently smooth motion using continuous-time B-spline kinematic models. By continuously applying Newton's Second Law, our method establishes a physical link between visually-derived trajectory acceleration and predicted deformation-induced acceleration. We demonstrate that our approach not only enables robust and accurate pose estimation on non-rigid platforms, but that the properly modeled platform physics instigate inertial sensing properties. We demonstrate this feasibility on a simple spring-camera system, and show how it robustly resolves the typically ill-posed problem of metric scale and gravity recovery in monocular visual odometry.
☆ Kleinkram: Open Robotic Data Management
We introduce Kleinkram, a free and open-source system designed to solve the challenge of managing massive, unstructured robotic datasets. Designed as a modular, on-premises cloud solution, Kleinkram enables scalable storage, indexing, and sharing of datasets, ranging from individual experiments to large-scale research collections. Kleinkram natively integrates with standard formats such as ROS bags and MCAP and utilises S3-compatible storage for flexibility. Beyond storage, Kleinkram features an integrated "Action Runner" that executes customizable Docker-based workflows for data validation, curation, and benchmarking. Kleinkram has successfully managed over 30 TB of data from diverse robotic systems, streamlining the research lifecycle through a modern web interface and a robust Command Line Interface (CLI).
comment: for associated source code, see https://github.com/leggedrobotics/kleinkram
☆ Power-Efficient Autonomous Mobile Robots
This paper presents pNav, a novel power-management system that significantly enhances the power/energy-efficiency of Autonomous Mobile Robots (AMRs) by jointly optimizing their physical/mechanical and cyber subsystems. By profiling AMRs' power consumption, we identify three challenges in achieving CPS (cyber-physical system) power-efficiency that involve both cyber (C) and physical (P) subsystems: (1) variabilities of system power consumption breakdown, (2) environment-aware navigation locality, and (3) coordination of C and P subsystems. pNav takes a multi-faceted approach to achieve power-efficiency of AMRs. First, it integrates millisecond-level power consumption prediction for both C and P subsystems. Second, it includes novel real-time modeling and monitoring of spatial and temporal navigation localities for AMRs. Third, it supports dynamic coordination of AMR software (navigation, detection) and hardware (motors, DVFS driver) configurations. pNav is prototyped using the Robot Operating System (ROS) Navigation Stack, 2D LiDAR, and camera. Our in-depth evaluation with a real robot and Gazebo environments demonstrates a >96% accuracy in predicting power consumption and a 38.1% reduction in power consumption without compromising navigation accuracy and safety.
comment: 13 pages, 16 figures
☆ BRIC: Bridging Kinematic Plans and Physical Control at Test Time
We propose BRIC, a novel test-time adaptation (TTA) framework that enables long-term human motion generation by resolving execution discrepancies between diffusion-based kinematic motion planners and reinforcement learning-based physics controllers. While diffusion models can generate diverse and expressive motions conditioned on text and scene context, they often produce physically implausible outputs, leading to execution drift during simulation. To address this, BRIC dynamically adapts the physics controller to noisy motion plans at test time, while preserving pre-trained skills via a loss function that mitigates catastrophic forgetting. In addition, BRIC introduces a lightweight test-time guidance mechanism that steers the diffusion model in the signal space without updating its parameters. By combining both adaptation strategies, BRIC ensures consistent and physically plausible long-term executions across diverse environments in an effective and efficient manner. We validate the effectiveness of BRIC on a variety of long-term tasks, including motion composition, obstacle avoidance, and human-scene interaction, achieving state-of-the-art performance across all tasks.
☆ VibraVerse: A Large-Scale Geometry-Acoustics Alignment Dataset for Physically-Consistent Multimodal Learning
Understanding the physical world requires perceptual models grounded in physical laws rather than mere statistical correlations. However, existing multimodal learning frameworks, focused on vision and language, lack physical consistency and overlook the intrinsic causal relationships among an object's geometry, material, vibration modes, and the sounds it produces. We introduce VibraVerse, a large-scale geometry-acoustics alignment dataset that explicitly bridges the causal chain from 3D geometry -> physical attributes -> modal parameters -> acoustic signals. Each 3D model has explicit physical properties (density, Young's modulus, Poisson's ratio) and volumetric geometry, from which modal eigenfrequencies and eigenvectors are computed for impact sound synthesis under controlled excitations. To establish this coherence, we introduce CLASP, a contrastive learning framework for cross-modal alignment that preserves the causal correspondence between an object's physical structure and its acoustic response. This framework enforces physically consistent alignment across modalities, ensuring that every sample is coherent, traceable to the governing equations, and embedded within a unified representation space spanning shape, image, and sound. Built upon VibraVerse, we define a suite of benchmark tasks for geometry-to-sound prediction, sound-guided shape reconstruction, and cross-modal representation learning. Extensive validations on these tasks demonstrate that models trained on VibraVerse exhibit superior accuracy, interpretability, and generalization across modalities. These results establish VibraVerse as a benchmark for physically consistent and causally interpretable multimodal learning, providing a foundation for sound-guided embodied perception and a deeper understanding of the physical world. The dataset will be open-sourced.
☆ Improved adaptive wind driven optimization algorithm for real-time path planning
Recently, path planning has achieved remarkable progress in enhancing global search capability and convergence accuracy through heuristic and learning-inspired optimization frameworks. However, real-time adaptability in dynamic environments remains a critical challenge for autonomous navigation, particularly when robots must generate collision-free, smooth, and efficient trajectories under complex constraints. By analyzing the difficulties in dynamic path planning, the Wind Driven Optimization (WDO) algorithm emerges as a promising framework owing to its physically interpretable search dynamics. Motivated by these observations, this work revisits the WDO principle and proposes a variant formulation, Multi-hierarchical adaptive wind driven optimization(MAWDO), that improves adaptability and robustness in time-varying environments. To mitigate instability and premature convergence, a hierarchical-guidance mechanism divides the population into multiple groups guided by individual, regional, and global leaders to balance exploration and exploitation. Extensive evaluations on sixteen benchmark functions show that MAWDO achieves superior optimization accuracy, convergence stability, and adaptability over state-of-the art metaheuristics. In dynamic path planning, MAWDO shortens the path length to 469.28 pixels, improving over Multi-strategy ensemble wind driven optimization(MEWDO), Adaptive wind driven optimization(AWDO) and WDO by 3.51\%, 11.63\% and 14.93\%, and achieves the smallest optimality gap (1.01) with smoothness 0.71 versus 13.50 and 15.67 for AWDO and WDO, leading to smoother, shorter, and collision-free trajectories that confirm its effectiveness for real-time path planning in complex environments.
comment: 23 pages, 4 figures
☆ Quality-guided UAV Surface Exploration for 3D Reconstruction
Reasons for mapping an unknown environment with autonomous robots are wide-ranging, but in practice, they are often overlooked when developing planning strategies. Rapid information gathering and comprehensive structural assessment of buildings have different requirements and therefore necessitate distinct methodologies. In this paper, we propose a novel modular Next-Best-View (NBV) planning framework for aerial robots that explicitly uses a reconstruction quality objective to guide the exploration planning. In particular, our approach introduces new and efficient methods for view generation and selection of viewpoint candidates that are adaptive to the user-defined quality requirements, fully exploiting the uncertainty encoded in a Truncated Signed Distance field (TSDF) representation of the environment. This results in informed and efficient exploration decisions tailored towards the predetermined objective. Finally, we validate our method via extensive simulations in realistic environments. We demonstrate that it successfully adjusts its behavior to the user goal while consistently outperforming conventional NBV strategies in terms of coverage, quality of the final 3D map and path efficiency.
☆ Material-informed Gaussian Splatting for 3D World Reconstruction in a Digital Twin
3D reconstruction for Digital Twins often relies on LiDAR-based methods, which provide accurate geometry but lack the semantics and textures naturally captured by cameras. Traditional LiDAR-camera fusion approaches require complex calibration and still struggle with certain materials like glass, which are visible in images but poorly represented in point clouds. We propose a camera-only pipeline that reconstructs scenes using 3D Gaussian Splatting from multi-view images, extracts semantic material masks via vision models, converts Gaussian representations to mesh surfaces with projected material labels, and assigns physics-based material properties for accurate sensor simulation in modern graphics engines and simulators. This approach combines photorealistic reconstruction with physics-based material assignment, providing sensor simulation fidelity comparable to LiDAR-camera fusion while eliminating hardware complexity and calibration requirements. We validate our camera-only method using an internal dataset from an instrumented test vehicle, leveraging LiDAR as ground truth for reflectivity validation alongside image similarity metrics.
comment: 8 pages, 5 figures. Submitted to IEEE Intelligent Vehicles Symposium (IV) 2026 for possible publication
☆ ArtiBench and ArtiBrain: Benchmarking Generalizable Vision-Language Articulated Object Manipulation
Interactive articulated manipulation requires long-horizon, multi-step interactions with appliances while maintaining physical consistency. Existing vision-language and diffusion-based policies struggle to generalize across parts, instances, and categories. We first introduce ArtiBench, a five-level benchmark covering kitchen, storage, office, and tool environments. ArtiBench enables structured evaluation from cross-part and cross-instance variation to long-horizon multi-object tasks, revealing the core generalization challenges of articulated object manipulation. Building on this benchmark, we propose ArtiBrain, a modular framework that unifies high-level reasoning with adaptive low-level control. ArtiBrain uses a VLM-based Task Reasoner (GPT-4.1) to decompose and validate subgoals, and employs a Hybrid Controller that combines geometry-aware keyframe execution with affordance-guided diffusion for precise and interpretable manipulation. An Affordance Memory Bank continually accumulates successful execution episodes and propagates part-level actionable affordances to unseen articulated parts and configurations. Extensive experiments on ArtiBench show that our ArtiBrain significantly outperforms state-of-the-art multimodal and diffusion-based methods in robustness and generalization. Code and dataset will be released upon acceptance.
☆ How Robot Kinematics Influence Human Performance in Virtual Robot-to-Human Handover Tasks
Recent advancements in robotics have increased the possibilities for integrating robotic systems into human-involved workplaces, highlighting the need to examine and optimize human-robot coordination in collaborative settings. This study explores human-robot interactions during handover tasks using Virtual Reality (VR) to investigate differences in human motor performance across various task dynamics and robot kinematics. A VR-based robot handover simulation afforded safe and controlled assessments of human-robot interactions. In separate experiments, four potential influences on human performance were examined (1) control over task initiation and robot movement synchrony (temporal and spatiotemporal); (2) partner appearance (human versus robotic); (3) robot velocity profiles (minimum jerk, constant velocity, constant acceleration, and biphasic); and (4) the timing of rotational object motion. Findings across experiments emphasize humans benefit from robots providing early and salient visual information about task-relevant object motion, and advantages of human-like smooth robot trajectories. To varying degrees, these manipulations improved predictive accuracy and synchronization during interaction. This suggests that human-robot interactions should be designed to allow humans to leverage their natural capabilities for detecting biological motion, which conversely may reduce the need for costly robotic computations or added cognitive adaptation on the human side.
☆ Dynamic-ICP: Doppler-Aware Iterative Closest Point Registration for Dynamic Scenes
Reliable odometry in highly dynamic environments remains challenging when it relies on ICP-based registration: ICP assumes near-static scenes and degrades in repetitive or low-texture geometry. We introduce Dynamic-ICP, a Doppler-aware registration framework. The method (i) estimates ego motion from per-point Doppler velocity via robust regression and builds a velocity filter, (ii) clusters dynamic objects and reconstructs object-wise translational velocities from ego-compensated radial measurements, (iii) predicts dynamic points with a constant-velocity model, and (iv) aligns scans using a compact objective that combines point-to-plane geometry residual with a translation-invariant, rotation-only Doppler residual. The approach requires no external sensors or sensor-vehicle calibration and operates directly on FMCW LiDAR range and Doppler velocities. We evaluate Dynamic-ICP on three datasets-HeRCULES, HeLiPR, AevaScenes-focusing on highly dynamic scenes. Dynamic-ICP consistently improves rotational stability and translation accuracy over the state-of-the-art methods. Our approach is also simple to integrate into existing pipelines, runs in real time, and provides a lightweight solution for robust registration in dynamic environments. To encourage further research, the code is available at: https://github.com/JMUWRobotics/Dynamic-ICP.
comment: 8 pages, 5 figures
☆ HAFO: Humanoid Force-Adaptive Control for Intense External Force Interaction Environments
Reinforcement learning controllers have made impressive progress in humanoid locomotion and light load manipulation. However, achieving robust and precise motion with strong force interaction remains a significant challenge. Based on the above limitations, this paper proposes HAFO, a dual-agent reinforcement learning control framework that simultaneously optimizes both a robust locomotion strategy and a precise upper-body manipulation strategy through coupled training under external force interaction environments. Simultaneously, we explicitly model the external pulling disturbances through a spring-damper system and achieve fine-grained force control by manipulating the virtual spring. During this process, the reinforcement-learning policy spontaneously generates disturbance-rejection response by exploiting environmental feedback. Moreover, HAFO employs an asymmetric Actor-Critic framework in which the Critic-network access to privileged spring-damping forces guides the actor-network to learn a generalizable, robust policy for resisting external disturbances. The experimental results demonstrate that HAFO achieves stable control of humanoid robot under various strong force interactions, showing remarkable performance in load tasks and ensuring stable robot operation under rope tension disturbances. Project website: hafo-robot.github.io.
☆ Toward generic control for soft robotic systems
Soft robotics has advanced rapidly, yet its control methods remain fragmented: different morphologies and actuation schemes still require task-specific controllers, hindering theoretical integration and large-scale deployment. A generic control framework is therefore essential, and a key obstacle lies in the persistent use of rigid-body control logic, which relies on precise models and strict low-level execution. Such a paradigm is effective for rigid robots but fails for soft robots, where the ability to tolerate and exploit approximate action representations, i.e., control compliance, is the basis of robustness and adaptability rather than a disturbance to be eliminated. Control should thus shift from suppressing compliance to explicitly exploiting it. Human motor control exemplifies this principle: instead of computing exact dynamics or issuing detailed muscle-level commands, it expresses intention through high-level movement tendencies, while reflexes and biomechanical mechanisms autonomously resolve local details. This architecture enables robustness, flexibility, and cross-task generalization. Motivated by this insight, we propose a generic soft-robot control framework grounded in control compliance and validate it across robots with diverse morphologies and actuation mechanisms. The results demonstrate stable, safe, and cross-platform transferable behavior, indicating that embracing control compliance, rather than resisting it, may provide a widely applicable foundation for unified soft-robot control.
☆ CostNav: A Navigation Benchmark for Cost-Aware Evaluation of Embodied Agents
Existing navigation benchmarks focus on task success metrics while overlooking economic viability -- critical for commercial deployment of autonomous delivery robots. We introduce \emph{CostNav}, a \textbf{Micro-Navigation Economic Testbed} that evaluates embodied agents through comprehensive cost-revenue analysis aligned with real-world business operations. CostNav models the complete economic lifecycle including hardware, training, energy, maintenance costs, and delivery revenue with service-level agreements, using industry-derived parameters. \textbf{To our knowledge, CostNav is the first work to quantitatively expose the gap between navigation research metrics and commercial viability}, revealing that optimizing for task success fundamentally differs from optimizing for economic deployment. Our cost model uses parameters derived from industry data sources (energy rates, delivery service pricing), and we project from a reduced-scale simulation to realistic deliveries. Under this projection, the baseline achieves 43.0\% SLA compliance but is \emph{not} commercially viable: yielding a loss of \$30.009 per run with no finite break-even point, because operating costs are dominated by collision-induced maintenance, which accounts for 99.7\% of per-run costs and highlights collision avoidance as a key optimization target. We demonstrate a learning-based on-device navigation baseline and establish a foundation for evaluating rule-based navigation, imitation learning, and cost-aware RL training. CostNav bridges the gap between navigation research and commercial deployment, enabling data-driven decisions about economic trade-offs across navigation paradigms.
☆ Hibikino-Musashi@Home 2025 Team Description Paper
This paper provides an overview of the techniques employed by Hibikino-Musashi@Home, which intends to participate in the domestic standard platform league. The team developed a dataset generator for training a robot vision system and an open-source development environment running on a Human Support Robot simulator. The large-language-model-powered task planner selects appropriate primitive skills to perform the task requested by the user. Moreover, the team has focused on research involving brain-inspired memory models for adaptation to individual home environments. This approach aims to provide intuitive and personalized assistance. Additionally, the team contributed to the reusability of the navigation system developed by Pumas in RoboCup2024. The team aimed to design a home service robot to assist humans in their homes and continuously attend competitions to evaluate and improve the developed system.
☆ Map-World: Masked Action planning and Path-Integral World Model for Autonomous Driving
Motion planning for autonomous driving must handle multiple plausible futures while remaining computationally efficient. Recent end-to-end systems and world-model-based planners predict rich multi-modal trajectories, but typically rely on handcrafted anchors or reinforcement learning to select a single best mode for training and control. This selection discards information about alternative futures and complicates optimization. We propose MAP-World, a prior-free multi-modal planning framework that couples masked action planning with a path-weighted world model. The Masked Action Planning (MAP) module treats future ego motion as masked sequence completion: past waypoints are encoded as visible tokens, future waypoints are represented as mask tokens, and a driving-intent path provides a coarse scaffold. A compact latent planning state is expanded into multiple trajectory queries with injected noise, yielding diverse, temporally consistent modes without anchor libraries or teacher policies. A lightweight world model then rolls out future BEV semantics conditioned on each candidate trajectory. During training, semantic losses are computed as an expectation over modes, using trajectory probabilities as discrete path weights, so the planner learns from the full distribution of plausible futures instead of a single selected path. On NAVSIM, our method matches anchor-based approaches and achieves state-of-the-art performance among world-model-based methods, while avoiding reinforcement learning and maintaining real-time inference latency.
☆ Active3D: Active High-Fidelity 3D Reconstruction via Hierarchical Uncertainty Quantification
In this paper, we present an active exploration framework for high-fidelity 3D reconstruction that incrementally builds a multi-level uncertainty space and selects next-best-views through an uncertainty-driven motion planner. We introduce a hybrid implicit-explicit representation that fuses neural fields with Gaussian primitives to jointly capture global structural priors and locally observed details. Based on this hybrid state, we derive a hierarchical uncertainty volume that quantifies both implicit global structure quality and explicit local surface confidence. To focus optimization on the most informative regions, we propose an uncertainty-driven keyframe selection strategy that anchors high-entropy viewpoints as sparse attention nodes, coupled with a viewpoint-space sliding window for uncertainty-aware local refinement. The planning module formulates next-best-view selection as an Expected Hybrid Information Gain problem and incorporates a risk-sensitive path planner to ensure efficient and safe exploration. Extensive experiments on challenging benchmarks demonstrate that our approach consistently achieves state-of-the-art accuracy, completeness, and rendering quality, highlighting its effectiveness for real-world active reconstruction and robotic perception tasks.
☆ ShapeForce: Low-Cost Soft Robotic Wrist for Contact-Rich Manipulation
Contact feedback is essential for contact-rich robotic manipulation, as it allows the robot to detect subtle interaction changes and adjust its actions accordingly. Six-axis force-torque sensors are commonly used to obtain contact feedback, but their high cost and fragility have discouraged many researchers from adopting them in contact-rich tasks. To offer a more cost-efficient and easy-accessible source of contact feedback, we present ShapeForce, a low-cost, plug-and-play soft wrist that provides force-like signals for contact-rich robotic manipulation. Inspired by how humans rely on relative force changes in contact rather than precise force magnitudes, ShapeForce converts external force and torque into measurable deformations of its compliant core, which are then estimated via marker-based pose tracking and converted into force-like signals. Our design eliminates the need for calibration or specialized electronics to obtain exact values, and instead focuses on capturing force and torque changes sufficient for enabling contact-rich manipulation. Extensive experiments across diverse contact-rich tasks and manipulation policies demonstrate that ShapeForce delivers performance comparable to six-axis force-torque sensors at an extremely low cost.
☆ Collaborate sim and real: Robot Bin Packing Learning in Real-world and Physical Engine
The 3D bin packing problem, with its diverse industrial applications, has garnered significant research attention in recent years. Existing approaches typically model it as a discrete and static process, while real-world applications involve continuous gravity-driven interactions. This idealized simplification leads to infeasible deployments (e.g., unstable packing) in practice. Simulations with physical engine offer an opportunity to emulate continuous gravity effects, enabling the training of reinforcement learning (RL) agents to address such limitations and improve packing stability. However, a simulation-to-reality gap persists due to dynamic variations in physical properties of real-world objects, such as various friction coefficients, elasticity, and non-uniform weight distributions. To bridge this gap, we propose a hybrid RL framework that collaborates with physical simulation with real-world data feedback. Firstly, domain randomization is applied during simulation to expose agents to a spectrum of physical parameters, enhancing their generalization capability. Secondly, the RL agent is fine-tuned with real-world deployment feedback, further reducing collapse rates. Extensive experiments demonstrate that our method achieves lower collapse rates in both simulated and real-world scenarios. Large-scale deployments in logistics systems validate the practical effectiveness, with a 35\% reduction in packing collapse compared to baseline methods.
☆ CoC-VLA: Delving into Adversarial Domain Transfer for Explainable Autonomous Driving via Chain-of-Causality Visual-Language-Action Model
Autonomous driving represents a prominent application of artificial intelligence. Recent approaches have shifted from focusing solely on common scenarios to addressing complex, long-tail situations such as subtle human behaviors, traffic accidents, and non-compliant driving patterns. Given the demonstrated capabilities of large language models (LLMs) in understanding visual and natural language inputs and following instructions, recent methods have integrated LLMs into autonomous driving systems to enhance reasoning, interpretability, and performance across diverse scenarios. However, existing methods typically rely either on real-world data, which is suitable for industrial deployment, or on simulation data tailored to rare or hard case scenarios. Few approaches effectively integrate the complementary advantages of both data sources. To address this limitation, we propose a novel VLM-guided, end-to-end adversarial transfer framework for autonomous driving that transfers long-tail handling capabilities from simulation to real-world deployment, named CoC-VLA. The framework comprises a teacher VLM model, a student VLM model, and a discriminator. Both the teacher and student VLM models utilize a shared base architecture, termed the Chain-of-Causality Visual-Language Model (CoC VLM), which integrates temporal information via an end-to-end text adapter. This architecture supports chain-of-thought reasoning to infer complex driving logic. The teacher and student VLM models are pre-trained separately on simulated and real-world datasets. The discriminator is trained adversarially to facilitate the transfer of long-tail handling capabilities from simulated to real-world environments by the student VLM model, using a novel backpropagation strategy.
☆ Reasoning-VLA: A Fast and General Vision-Language-Action Reasoning Model for Autonomous Driving
Vision-Language-Action (VLA) models have recently shown strong decision-making capabilities in autonomous driving. However, existing VLAs often struggle with achieving efficient inference and generalizing to novel autonomous vehicle configurations and driving scenarios. In this paper, we propose Reasoning-VLA, a general and fast action-generation VLA framework. The proposed model employs a set of learnable action queries, initialized via Gaussian sampling from ground-truth trajectories within the training corpus. These learnable queries interact with reasoning-enhanced vision-language features to generate continuous action trajectories in parallel. To promote robust generalization, we consolidate eight publicly available autonomous driving datasets into a standardized, Chain-of-Thought reasoning-based, and easy-to-use data format for model training. Leveraging both supervised learning and reinforcement learning fine-tuning, extensive empirical evaluations across multiple benchmarks demonstrate that Reasoning-VLA achieves state-of-the-art performance, superior generalization capability, and the excellent inference speed reported to date.
☆ Improved Linear-Time Construction of Minimal Dominating Set via Mobile Agents
Mobile agents have emerged as a powerful framework for solving fundamental graph problems in distributed settings in recent times. These agents, modelled as autonomous physical or software entities, possess local computation power, finite memory and have the ability to traverse a graph, offering efficient solutions to a range of classical problems. In this work, we focus on the problem of computing a \emph{minimal dominating set} (mDS) in anonymous graphs using mobile agents. Building on the recently proposed optimal dispersion algorithm on the synchronous mobile agent model, we design two new algorithms that achieve a \emph{linear-time} solution for this problem in the synchronous setting. Specifically, given a connected $n$-node graph with $n$ agents initially placed in either rooted or arbitrary configurations, we show that an mDS can be computed in $O(n)$ rounds using only $O(\log n)$ bits of memory per agent, without using any prior knowledge of any global parameters. This improves upon the best-known complexity results in the literature over the same model. In addition, as natural by-products of our methodology, our algorithms also construct a spanning tree and elect a unique leader in $O(n)$ rounds, which are also important results of independent interest in the mobile-agent framework.
☆ MAPS: Preserving Vision-Language Representations via Module-Wise Proximity Scheduling for Better Vision-Language-Action Generalization
Vision-Language-Action (VLA) models inherit strong priors from pretrained Vision-Language Models (VLMs), but naive fine-tuning often disrupts these representations and harms generalization. Existing fixes -- freezing modules or applying uniform regularization -- either overconstrain adaptation or ignore the differing roles of VLA components. We present MAPS (Module-Wise Proximity Scheduling), the first robust fine-tuning framework for VLAs. Through systematic analysis, we uncover an empirical order in which proximity constraints should be relaxed to balance stability and flexibility. MAPS linearly schedules this relaxation, enabling visual encoders to stay close to their pretrained priors while action-oriented language layers adapt more freely. MAPS introduces no additional parameters or data, and can be seamlessly integrated into existing VLAs. Across MiniVLA-VQ, MiniVLA-OFT, OpenVLA-OFT, and challenging benchmarks such as SimplerEnv, CALVIN, LIBERO, as well as real-world evaluations on the Franka Emika Panda platform, MAPS consistently boosts both in-distribution and out-of-distribution performance (up to +30%). Our findings highlight empirically guided proximity to pretrained VLMs as a simple yet powerful principle for preserving broad generalization in VLM-to-VLA transfer.
☆ Human-Centered Cooperative Control Coupling Autonomous and Haptic Shared Control via Control Barrier Function
Haptic shared control (HSC) is effective in teleoperation when full autonomy is limited by uncertainty or sensing constraints. However, autonomous control performance achieved by maximizing HSC strength is limited because the dynamics of the joystick and human arm affect the robot's behavior. We propose a cooperative framework coupling a joystick-independent autonomous controller with HSC. A control barrier function ignores joystick inputs within a safe region determined by the human operator in real-time, while HSC is engaged otherwise. A pilot experiment on simulated tasks with tele-operated underwater robot in virtual environment demonstrated improved accuracy and reduced required time over conventional HSC.
☆ GigaWorld-0: World Models as Data Engine to Empower Embodied AI
World models are emerging as a foundational paradigm for scalable, data-efficient embodied AI. In this work, we present GigaWorld-0, a unified world model framework designed explicitly as a data engine for Vision-Language-Action (VLA) learning. GigaWorld-0 integrates two synergistic components: GigaWorld-0-Video, which leverages large-scale video generation to produce diverse, texture-rich, and temporally coherent embodied sequences under fine-grained control of appearance, camera viewpoint, and action semantics; and GigaWorld-0-3D, which combines 3D generative modeling, 3D Gaussian Splatting reconstruction, physically differentiable system identification, and executable motion planning to ensure geometric consistency and physical realism. Their joint optimization enables the scalable synthesis of embodied interaction data that is visually compelling, spatially coherent, physically plausible, and instruction-aligned. Training at scale is made feasible through our efficient GigaTrain framework, which exploits FP8-precision and sparse attention to drastically reduce memory and compute requirements. We conduct comprehensive evaluations showing that GigaWorld-0 generates high-quality, diverse, and controllable data across multiple dimensions. Critically, VLA model (e.g., GigaBrain-0) trained on GigaWorld-0-generated data achieve strong real-world performance, significantly improving generalization and task success on physical robots without any real-world interaction during training.
comment: Project Page: https://gigaworld0.github.io/
☆ Unifying Perception and Action: A Hybrid-Modality Pipeline with Implicit Visual Chain-of-Thought for Robotic Action Generation
Vision-Language-Action (VLA) models built upon Chain-of-Thought (CoT) have achieved remarkable success in advancing general-purpose robotic agents, owing to its significant perceptual comprehension. Recently, since text-only CoT struggles to adequately capture scene details in complex spatial environments, a highly promising strategy involves leveraging visual priors to guide robotic action generation. Nevertheless, these strategies face two inherent challenges: (i) a modality gap between visual observations and low-level actions, and (ii) unstable training due to competing objectives between visual prediction and action generation. To address these challenges, we propose a Vision-Integrated Trajectory Alignment (VITA) framework that learns a shared discrete latent space for vision and action, enabling joint modeling of perception and motor control. VITA introduces a implicit visual CoT: autoregressively generated tokens is simultaneously decoded into future frames predictions and robot actions, thereby internalizing visual dynamics as an inductive bias for motion planning. Extensive experiments on simulated and real-world environments demonstrate state-of-the-art performance. VITA improves 14.5\%, 9.6\% and 12.1\% over existing baselines on CALVIN, LIBERO and SimplerEnv. Furthermore, VITA attains an average success rate of 80.5\% across six real-world tasks, demonstrating its potential as a generalist robotic manipulation model.
☆ Dynamic Test-Time Compute Scaling in Control Policy: Difficulty-Aware Stochastic Interpolant Policy
Diffusion- and flow-based policies deliver state-of-the-art performance on long-horizon robotic manipulation and imitation learning tasks. However, these controllers employ a fixed inference budget at every control step, regardless of task complexity, leading to computational inefficiency for simple subtasks while potentially underperforming on challenging ones. To address these issues, we introduce Difficulty-Aware Stochastic Interpolant Policy (DA-SIP), a framework that enables robotic controllers to adaptively adjust their integration horizon in real time based on task difficulty. Our approach employs a difficulty classifier that analyzes observations to dynamically select the step budget, the optimal solver variant, and ODE/SDE integration at each control cycle. DA-SIP builds upon the stochastic interpolant formulation to provide a unified framework that unlocks diverse training and inference configurations for diffusion- and flow-based policies. Through comprehensive benchmarks across diverse manipulation tasks, DA-SIP achieves 2.6-4.4x reduction in total computation time while maintaining task success rates comparable to fixed maximum-computation baselines. By implementing adaptive computation within this framework, DA-SIP transforms generative robot controllers into efficient, task-aware systems that intelligently allocate inference resources where they provide the greatest benefit.
☆ Efficient Greedy Algorithms for Feature Selection in Robot Visual Localization
Robot localization is a fundamental component of autonomous navigation in unknown environments. Among various sensing modalities, visual input from cameras plays a central role, enabling robots to estimate their position by tracking point features across image frames. However, image frames often contain a large number of features, many of which are redundant or uninformative for localization. Processing all features can introduce significant computational latency and inefficiency. This motivates the need for intelligent feature selection, identifying a subset of features that are most informative for localization over a prediction horizon. In this work, we propose two fast and memory-efficient feature selection algorithms that enable robots to actively evaluate the utility of visual features in real time. Unlike existing approaches with high computational and memory demands, the proposed methods are explicitly designed to reduce both time and memory complexity while achieving a favorable trade-off between computational efficiency and localization accuracy.
☆ ACE-F: A Cross Embodiment Foldable System with Force Feedback for Dexterous Teleoperation
Teleoperation systems are essential for efficiently collecting diverse and high-quality robot demonstration data, especially for complex, contact-rich tasks. However, current teleoperation platforms typically lack integrated force feedback, cross-embodiment generalization, and portable, user-friendly designs, limiting their practical deployment. To address these limitations, we introduce ACE-F, a cross embodiment foldable teleoperation system with integrated force feedback. Our approach leverages inverse kinematics (IK) combined with a carefully designed human-robot interface (HRI), enabling users to capture precise and high-quality demonstrations effortlessly. We further propose a generalized soft-controller pipeline integrating PD control and inverse dynamics to ensure robot safety and precise motion control across diverse robotic embodiments. Critically, to achieve cross-embodiment generalization of force feedback without additional sensors, we innovatively interpret end-effector positional deviations as virtual force signals, which enhance data collection and enable applications in imitation learning. Extensive teleoperation experiments confirm that ACE-F significantly simplifies the control of various robot embodiments, making dexterous manipulation tasks as intuitive as operating a computer mouse. The system is open-sourced at: https://acefoldable.github.io/
☆ NOIR 2.0: Neural Signal Operated Intelligent Robots for Everyday Activities
Neural Signal Operated Intelligent Robots (NOIR) system is a versatile brain-robot interface that allows humans to control robots for daily tasks using their brain signals. This interface utilizes electroencephalography (EEG) to translate human intentions regarding specific objects and desired actions directly into commands that robots can execute. We present NOIR 2.0, an enhanced version of NOIR. NOIR 2.0 includes faster and more accurate brain decoding algorithms, which reduce task completion time by 46%. NOIR 2.0 uses few-shot robot learning algorithms to adapt to individual users and predict their intentions. The new learning algorithms leverage foundation models for more sample-efficient learning and adaptation (15 demos vs. a single demo), significantly reducing overall human time by 65%.
comment: Conference on Robot Learning (CoRL 2024), CoRoboLearn
☆ OVAL-Grasp: Open-Vocabulary Affordance Localization for Task Oriented Grasping
To manipulate objects in novel, unstructured environments, robots need task-oriented grasps that target object parts based on the given task. Geometry-based methods often struggle with visually defined parts, occlusions, and unseen objects. We introduce OVAL-Grasp, a zero-shot open-vocabulary approach to task-oriented, affordance based grasping that uses large-language models and vision-language models to allow a robot to grasp objects at the correct part according to a given task. Given an RGB image and a task, OVAL-Grasp identifies parts to grasp or avoid with an LLM, segments them with a VLM, and generates a 2D heatmap of actionable regions on the object. During our evaluations, we found that our method outperformed two task oriented grasping baselines on experiments with 20 household objects with 3 unique tasks for each. OVAL-Grasp successfully identifies and segments the correct object part 95% of the time and grasps the correct actionable area 78.3% of the time in real-world experiments with the Fetch mobile manipulator. Additionally, OVAL-Grasp finds correct object parts under partial occlusions, demonstrating a part selection success rate of 80% in cluttered scenes. We also demonstrate OVAL-Grasp's efficacy in scenarios that rely on visual features for part selection, and show the benefit of a modular design through our ablation experiments. Our project webpage is available at https://ekjt.github.io/OVAL-Grasp/
comment: 10 pages, 7 figures, 3 tables. Presented at the 2025 International Symposium on Experimental Robotics (ISER)
☆ DeeAD: Dynamic Early Exit of Vision-Language Action for Efficient Autonomous Driving
Vision-Language Action (VLA) models unify perception, reasoning, and trajectory generation for autonomous driving, but suffer from significant inference latency due to deep transformer stacks. We present DeeAD, a training-free, action-guided early-exit framework that accelerates VLA planning by evaluating the physical feasibility of intermediate trajectories. Instead of relying on confidence scores, DeeAD terminates inference when predicted trajectories align with lightweight planning priors (e.g., Navigation or Low-precision Planning) within a tolerable deviation (<2m). To improve efficiency, we introduce a multi-hop controller that adaptively skips redundant layers based on the change rate of scores. DeeAD integrates into existing VLA models, such as ORION, without requiring retraining. Experiments on the Bench2Drive benchmark demonstrate up to 28% transformer-layer sparsity and 29% latency reduction, while preserving planning quality and safety.
♻ ☆ AutoFocus-IL: VLM-based Saliency Maps for Data-Efficient Visual Imitation Learning without Extra Human Annotations
AutoFocus-IL is a simple yet effective method to improve data efficiency and generalization in visual imitation learning by guiding policies to attend to task-relevant features rather than distractors and spurious correlations. Although saliency regularization has emerged as a promising way to achieve this, existing approaches typically require costly supervision such as human gaze data or manual saliency annotations. In contrast, AutoFocus-IL leverages vision-language models (VLMs) to automatically identify and track key objects in demonstrations, generating temporal saliency maps that highlight causal visual signals while suppressing distractors. These maps are then used to regularize behavior cloning policies, yielding stronger alignment between visual attention and task-relevant cues. Experiments in both the CARLA simulator and real-robot manipulation tasks demonstrate that AutoFocus-IL not only outperforms standard behavior cloning but also surpasses state-of-the-art baselines that assume privileged access to human supervision, such as gaze data. Code, datasets, and trained policy videos are available at https://AutoFocus-IL.github.io/.
comment: 8 pages, 6 figures. Code and datasets available at http://autofocus-il.github.io/
♻ ☆ OceanGym: A Benchmark Environment for Underwater Embodied Agents
We introduce OceanGym, the first comprehensive benchmark for ocean underwater embodied agents, designed to advance AI in one of the most demanding real-world environments. Unlike terrestrial or aerial domains, underwater settings present extreme perceptual and decision-making challenges, including low visibility, dynamic ocean currents, making effective agent deployment exceptionally difficult. OceanGym encompasses eight realistic task domains and a unified agent framework driven by Multi-modal Large Language Models (MLLMs), which integrates perception, memory, and sequential decision-making. Agents are required to comprehend optical and sonar data, autonomously explore complex environments, and accomplish long-horizon objectives under these harsh conditions. Extensive experiments reveal substantial gaps between state-of-the-art MLLM-driven agents and human experts, highlighting the persistent difficulty of perception, planning, and adaptability in ocean underwater environments. By providing a high-fidelity, rigorously designed platform, OceanGym establishes a testbed for developing robust embodied AI and transferring these capabilities to real-world autonomous ocean underwater vehicles, marking a decisive step toward intelligent agents capable of operating in one of Earth's last unexplored frontiers. The code and data are available at https://github.com/OceanGPT/OceanGym.
comment: Work in progress
♻ ☆ A Physics-informed Demonstration-guided Learning Framework for Granular Material Manipulation
Due to the complex physical properties of granular materials, research on robot learning for manipulating such materials predominantly either disregards the consideration of their physical characteristics or uses surrogate models to approximate their physical properties. Learning to manipulate granular materials based on physical information obtained through precise modelling remains an unsolved problem. In this paper, we propose to address this challenge by constructing a differentiable physics-based simulator for granular materials using the Taichi programming language and developing a learning framework accelerated by demonstrations generated through gradient-based optimisation on non-granular materials within our simulator, eliminating the costly data collection and model training of prior methods. Experimental results show that our method, with its flexible design, trains robust policies that are capable of executing the task of transporting granular materials in both simulated and real-world environments, beyond the capabilities of standard reinforcement learning, imitation learning, and prior task-specific granular manipulation methods.
comment: Accepted as a regular paper by IEEE Transactions on Neural Networks and Learning Systems (TNNLS)
♻ ☆ SafePR: Unified Approach for Safe Parallel Robots by Contact Detection and Reaction with Redundancy Resolution
Fast and safe motion is crucial for the successful deployment of physically interactive robots. Parallel robots (PRs) offer the potential for higher speeds while maintaining the same energy limits due to their low moving masses. However, they require methods for contact detection and reaction while avoiding singularities and self-collisions. We address this issue and present SafePR - a unified approach for the detection and localization, including the distinction between collision and clamping to perform a reaction that is safe for humans and feasible for PRs. Our approach uses information from the encoders and motor currents to estimate forces via a generalized-momentum observer. Neural networks and particle filters classify and localize the contacts. We introduce reactions with redundancy resolution to avoid self-collisions and type-II singularities. Our approach detected and terminated 72 real-world collision and clamping contacts with end-effector speeds of up to 1.5 m/s, each within 25-275 ms. The forces were below the thresholds from ISO/TS 15066. By using built-in sensors, SafePR enables safe interaction with already assembled PRs without the need for new hardware components.
♻ ☆ FSR-VLN: Fast and Slow Reasoning for Vision-Language Navigation with Hierarchical Multi-modal Scene Graph
Visual-Language Navigation (VLN) is a fundamental challenge in robotic systems, with broad applications for the deployment of embodied agents in real-world environments. Despite recent advances, existing approaches are limited in long-range spatial reasoning, often exhibiting low success rates and high inference latency, particularly in long-range navigation tasks. To address these limitations, we propose FSR-VLN, a vision-language navigation system that combines a Hierarchical Multi-modal Scene Graph (HMSG) with Fast-to-Slow Navigation Reasoning (FSR). The HMSG provides a multi-modal map representation supporting progressive retrieval, from coarse room-level localization to fine-grained goal view and object identification. Building on HMSG, FSR first performs fast matching to efficiently select candidate rooms, views, and objects, then applies VLM-driven refinement for final goal selection. We evaluated FSR-VLN across four comprehensive indoor datasets collected by humanoid robots, utilizing 87 instructions that encompass a diverse range of object categories. FSR-VLN achieves state-of-the-art (SOTA) performance in all datasets, measured by the retrieval success rate (RSR), while reducing the response time by 82% compared to VLM-based methods on tour videos by activating slow reasoning only when fast intuition fails. Furthermore, we integrate FSR-VLN with speech interaction, planning, and control modules on a Unitree-G1 humanoid robot, enabling natural language interaction and real-time navigation.
comment: Demo video are available at https://horizonrobotics.github.io/robot_lab/fsr-vln/
♻ ☆ An Efficient Closed-Form Solution to Full Visual-Inertial State Initialization
In this letter, we present a closed-form initialization method that recovers the full visual-inertial state without nonlinear optimization. Unlike previous approaches that rely on iterative solvers, our formulation yields analytical, easy-to-implement, and numerically stable solutions for reliable start-up. Our method builds on small-rotation and constant-velocity approximations, which keep the formulation compact while preserving the essential coupling between motion and inertial measurements. We further propose an observability-driven, two-stage initialization scheme that balances accuracy with initialization latency. Extensive experiments on the EuRoC dataset validate our assumptions: our method achieves 10-20% lower initialization error than optimization-based approaches, while using 4x shorter initialization windows and reducing computational cost by 5x.
comment: 8 pages, 2 figures, 10 tables. Submitted to RA-L
♻ ☆ RLZero: Direct Policy Inference from Language Without In-Domain Supervision NeurIPS 2025
The reward hypothesis states that all goals and purposes can be understood as the maximization of a received scalar reward signal. However, in practice, defining such a reward signal is notoriously difficult, as humans are often unable to predict the optimal behavior corresponding to a reward function. Natural language offers an intuitive alternative for instructing reinforcement learning (RL) agents, yet previous language-conditioned approaches either require costly supervision or test-time training given a language instruction. In this work, we present a new approach that uses a pretrained RL agent trained using only unlabeled, offline interactions--without task-specific supervision or labeled trajectories--to get zero-shot test-time policy inference from arbitrary natural language instructions. We introduce a framework comprising three steps: imagine, project, and imitate. First, the agent imagines a sequence of observations corresponding to the provided language description using video generative models. Next, these imagined observations are projected into the target environment domain. Finally, an agent pretrained in the target environment with unsupervised RL instantly imitates the projected observation sequence through a closed-form solution. To the best of our knowledge, our method, RLZero, is the first approach to show direct language-to-behavior generation abilities on a variety of tasks and environments without any in-domain supervision. We further show that components of RLZero can be used to generate policies zero-shot from cross-embodied videos, such as those available on YouTube, even for complex embodiments like humanoids.
comment: NeurIPS 2025, 26 pages
♻ ☆ LiHi-GS: LiDAR-Supervised Gaussian Splatting for Highway Driving Scene Reconstruction
Photorealistic 3D scene reconstruction plays an important role in autonomous driving, enabling the generation of novel data from existing datasets to simulate safety-critical scenarios and expand training data without additional acquisition costs. Gaussian Splatting (GS) facilitates real-time, photorealistic rendering with an explicit 3D Gaussian representation of the scene, providing faster processing and more intuitive scene editing than the implicit Neural Radiance Fields (NeRFs). While extensive GS research has yielded promising advancements in autonomous driving applications, they overlook two critical aspects: First, existing methods mainly focus on low-speed and feature-rich urban scenes and ignore the fact that highway scenarios play a significant role in autonomous driving. Second, while LiDARs are commonplace in autonomous driving platforms, existing methods learn primarily from images and use LiDAR only for initial estimates or without precise sensor modeling, thus missing out on leveraging the rich depth information LiDAR offers and limiting the ability to synthesize LiDAR data. In this paper, we propose a novel GS method for dynamic scene synthesis and editing with improved scene reconstruction through LiDAR supervision and support for LiDAR rendering. Unlike prior works that are tested mostly on urban datasets, to the best of our knowledge, we are the first to focus on the more challenging and highly relevant highway scenes for autonomous driving, with sparse sensor views and monotone backgrounds. Visit our project page at: https://umautobots.github.io/lihi_gs
comment: RA-L 2025
♻ ☆ Disentangled Control of Multi-Agent Systems
This paper develops a general framework for multi-agent control synthesis, which applies to a wide range of problems with convergence guarantees, regardless of the complexity of the underlying graph topology and the explicit time dependence of the objective function. The proposed framework systematically addresses a particularly challenging problem in multi-agent systems, i.e., decentralization of entangled dynamics among different agents, and it naturally supports multi-objective robotics and real-time implementations. To demonstrate its generality and effectiveness, the framework is implemented across three experiments, namely time-varying leader-follower formation control, decentralized coverage control for time-varying density functions without any approximations, which is a long-standing open problem, and safe formation navigation in dense environments.
♻ ☆ Hestia: Voxel-Face-Aware Hierarchical Next-Best-View Acquisition for Efficient 3D Reconstruction WACV
Advances in 3D reconstruction and novel view synthesis have enabled efficient and photorealistic rendering. However, images for reconstruction are still either largely manual or constrained by simple preplanned trajectories. To address this issue, recent works propose generalizable next-best-view planners that do not require online learning. Nevertheless, robustness and performance remain limited across various shapes. Hence, this study introduces Voxel-Face-Aware Hierarchical Next-Best-View Acquisition for Efficient 3D Reconstruction (Hestia), which addresses the shortcomings of the reinforcement learning-based generalizable approaches for five-degree-of-freedom viewpoint prediction. Hestia systematically improves the planners through four components: a more diverse dataset to promote robustness, a hierarchical structure to manage the high-dimensional continuous action search space, a close-greedy strategy to mitigate spurious correlations, and a face-aware design to avoid overlooking geometry. Experimental results show that Hestia achieves non-marginal improvements, with at least a 4% gain in coverage ratio, while reducing Chamfer Distance by 50% and maintaining real-time inference. In addition, Hestia outperforms prior methods by at least 12% in coverage ratio with a 5-image budget and remains robust to object placement variations. Finally, we demonstrate that Hestia, as a next-best-view planner, is feasible for the real-world application. Our project page is https://johnnylu305.github.io/hestia web.
comment: Accepted to the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2026
♻ ☆ GigaBrain-0: A World Model-Powered Vision-Language-Action Model
Training Vision-Language-Action (VLA) models for generalist robots typically requires large-scale real-world robot data, which is expensive and time-consuming to collect. The inefficiency of physical data collection severely limits the scalability, and generalization capacity of current VLA systems. To address this challenge, we introduce GigaBrain-0, a novel VLA foundation model empowered by world model-generated data (e.g., video generation, real2real transfer, human transfer, view transfer, sim2real transfer data). By leveraging world models to generate diverse data at scale, GigaBrain-0 significantly reduces reliance on real robot data while improving cross-task generalization. Our approach further improves policy robustness through RGBD input modeling and embodied Chain-of-Thought (CoT) supervision, enabling the model to reason about spatial geometry, object states, and long-horizon dependencies during task execution. This leads to substantial gains in real-world performance on dexterous, long-horizon, and mobile manipulation tasks. Extensive experiments demonstrate that GigaBrain-0 achieves superior generalization across variations in appearances (e.g., textures, colors), object placements, and camera viewpoints. Additionally, we present GigaBrain-0-Small, an optimized lightweight variant designed to run efficiently on devices such as the NVIDIA Jetson AGX Orin.
comment: https://gigabrain0.github.io/
♻ ☆ From Forecasting to Planning: Policy World Model for Collaborative State-Action Prediction
Despite remarkable progress in driving world models, their potential for autonomous systems remains largely untapped: the world models are mostly learned for world simulation and decoupled from trajectory planning. While recent efforts aim to unify world modeling and planning in a single framework, the synergistic facilitation mechanism of world modeling for planning still requires further exploration. In this work, we introduce a new driving paradigm named Policy World Model (PWM), which not only integrates world modeling and trajectory planning within a unified architecture, but is also able to benefit planning using the learned world knowledge through the proposed action-free future state forecasting scheme. Through collaborative state-action prediction, PWM can mimic the human-like anticipatory perception, yielding more reliable planning performance. To facilitate the efficiency of video forecasting, we further introduce a dynamically enhanced parallel token generation mechanism, equipped with a context-guided tokenizer and an adaptive dynamic focal loss. Despite utilizing only front camera input, our method matches or exceeds state-of-the-art approaches that rely on multi-view and multi-modal inputs. Code and model weights will be released at https://github.com/6550Zhao/Policy-World-Model.
comment: Accepted by NuerIPS 2025 (Poster)
♻ ☆ ARBoids: Adaptive Residual Reinforcement Learning With Boids Model for Cooperative Multi-USV Target Defense
The target defense problem (TDP) for unmanned surface vehicles (USVs) concerns intercepting an adversarial USV before it breaches a designated target region, using one or more defending USVs. A particularly challenging scenario arises when the attacker exhibits superior maneuverability compared to the defenders, significantly complicating effective interception. To tackle this challenge, this letter introduces ARBoids, a novel adaptive residual reinforcement learning framework that integrates deep reinforcement learning (DRL) with the biologically inspired, force-based Boids model. Within this framework, the Boids model serves as a computationally efficient baseline policy for multi-agent coordination, while DRL learns a residual policy to adaptively refine and optimize the defenders' actions. The proposed approach is validated in a high-fidelity Gazebo simulation environment, demonstrating superior performance over traditional interception strategies, including pure force-based approaches and vanilla DRL policies. Furthermore, the learned policy exhibits strong adaptability to attackers with diverse maneuverability profiles, highlighting its robustness and generalization capability. The code of ARBoids will be released upon acceptance of this letter.
♻ ☆ SVBRD-LLM: Self-Verifying Behavioral Rule Discovery for Autonomous Vehicle Identification
As more autonomous vehicles operate on public roads, understanding real-world behavior of autonomous vehicles is critical to analyzing traffic safety, making policies, and public acceptance. This paper proposes SVBRD-LLM, a framework that automatically discovers, verifies, and applies interpretable behavioral rules from real traffic videos through zero-shot prompt engineering. The framework extracts vehicle trajectories using YOLOv8 and ByteTrack, computes kinematic features, and employs GPT-5 zero-shot prompting to compare autonomous and human-driven vehicles, generating 35 structured behavioral rule hypotheses. These rules are tested on a validation set, iteratively refined based on failure cases to filter spurious correlations, and compiled into a high-confidence rule library. The framework is evaluated on an independent test set for speed change prediction, lane change prediction, and autonomous vehicle identification tasks. Experiments on over 1500 hours of real traffic videos show that the framework achieves 90.0% accuracy and 93.3% F1-score in autonomous vehicle identification. The discovered rules clearly reveal distinctive characteristics of autonomous vehicles in speed control smoothness, lane change conservativeness, and acceleration stability, with each rule accompanied by semantic description, applicable context, and validation confidence.
♻ ☆ GRAM: Generalization in Deep RL with a Robust Adaptation Module
The reliable deployment of deep reinforcement learning in real-world settings requires the ability to generalize across a variety of conditions, including both in-distribution scenarios seen during training as well as novel out-of-distribution scenarios. In this work, we present a framework for dynamics generalization in deep reinforcement learning that unifies these two distinct types of generalization within a single architecture. We introduce a robust adaptation module that provides a mechanism for identifying and reacting to both in-distribution and out-of-distribution environment dynamics, along with a joint training pipeline that combines the goals of in-distribution adaptation and out-of-distribution robustness. Our algorithm GRAM achieves strong generalization performance across in-distribution and out-of-distribution scenarios upon deployment, which we demonstrate through extensive simulation and hardware locomotion experiments on a quadruped robot.
comment: Accepted for publication in IEEE Robotics and Automation Letters (RA-L)
♻ ☆ Continually Evolving Skill Knowledge in Vision Language Action Model
Developing general robot intelligence in open environments requires continual skill learning. Recent Vision-Language-Action (VLA) models leverage massive pretraining data to support diverse manipulation tasks, but they still depend heavily on task-specific fine-tuning, revealing a lack of continual learning capability. Existing continual learning methods are also resource-intensive to scale to VLA models. We propose Stellar VLA, a knowledge-driven continual learning framework with two variants: T-Stellar, modeling task-centric knowledge space, and TS-Stellar, capturing hierarchical task-skill structure. Stellar VLA enables self-supervised knowledge evolution through joint learning of task latent representation and the knowledge space, reducing annotation needs. Knowledge-guided expert routing provide task specialization without extra network parameters, lowering training overhead. Experiments on the LIBERO benchmark and real-world tasks show over 50 percentage average improvement in final success rates relative to baselines. TS-Stellar further excels in complex action inference, and in-depth analyses verify effective knowledge retention and discovery. Our code will be released soon.
♻ ☆ Count Every Rotation and Every Rotation Counts: Exploring Drone Dynamics via Propeller Sensing
As drone-based applications proliferate, paramount contactless sensing of airborne drones from the ground becomes indispensable. This work demonstrates concentrating on propeller rotational speed will substantially improve drone sensing performance and proposes an event-camera-based solution, \sysname. \sysname features two components: \textit{Count Every Rotation} achieves accurate, real-time propeller speed estimation by mitigating ultra-high sensitivity of event cameras to environmental noise. \textit{Every Rotation Counts} leverages these speeds to infer both internal and external drone dynamics. Extensive evaluations in real-world drone delivery scenarios show that \sysname achieves a sensing latency of 3$ms$ and a rotational speed estimation error of merely 0.23\%. Additionally, \sysname infers drone flight commands with 96.5\% precision and improves drone tracking accuracy by over 22\% when combined with other sensing modalities. \textit{ Demo: {\color{blue}https://eventpro25.github.io/EventPro/.} }
Robotics 64
☆ Mixture of Horizons in Action Chunking
Vision-language-action (VLA) models have shown remarkable capabilities in robotic manipulation, but their performance is sensitive to the $\textbf{action chunk length}$ used during training, termed $\textbf{horizon}$. Our empirical study reveals an inherent trade-off: longer horizons provide stronger global foresight but degrade fine-grained accuracy, while shorter ones sharpen local control yet struggle on long-term tasks, implying fixed choice of single horizons being suboptimal. To mitigate the trade-off, we propose a $\textbf{mixture of horizons (MoH)}$ strategy. MoH rearranges the action chunk into several segments with different horizons, processes them in parallel with a shared action transformer, and fuses outputs with a light linear gate. It has three appealing benefits. 1) MoH exploits long-term foresight and short-term precision jointly within a single model, improving both performance and generalizability to complex tasks. 2) MoH is plug-and-play for full-attention action modules with minimal training or inference overhead. 3) MoH enables dynamic inference with adaptive horizons, which selects stable actions through cross-horizon consensus, achieving 2.5$\times$ higher throughput than baselines while preserving superior performance. Extensive experiments over flow-based policies $π_0$, $π_{0.5}$, and one-step regression policy $π_{\text{reg}}$ demonstrate that MoH yields consistent and significant gains on both simulations and real-world tasks. Notably, under mixed-task setting, $π_{0.5}$ with MoH reaches a new state-of-the-art with 99$\%$ average success rate on LIBERO after only $30k$ training iterations. Project page: https://github.com/Timsty1/MixtureOfHorizons
comment: 15 pages, 14 figures
☆ Deployment Dynamics and Optimization of Novel Space Antenna Deployable Mechanism
Given the increasing need for large aperture antennas in space missions, the difficulty of fitting such structures into small launch vehicles has prompted the design of deployable antenna systems. The thesis introduces a new Triple Scissors Deployable Truss Mechanism (TSDTM) for space antenna missions. The new mechanism is to be stowed during launch and efficiently deploy in orbit, offering maximum aperture size while taking up minimal launch volume. The thesis covers the entire design process from geometric modeling, kinematic analysis with screw theory and Newtonian approaches, dynamic analysis by eigenvalue and simulation methods, and verification with SolidWorks. In addition, optimization routines were coded based on Support Vector Machines for material choice in LEO environments and machine learning method for geometric setup. The TSDTM presented has enhanced structural dynamics with good comparison between simulation and analytical predictions. The structure optimized proved highly accurate, with a deviation of just 1.94% between machine learning-predicted and simulated natural frequencies, demonstrating the potential of incorporating AI-based methods in space structural design.
☆ Leveraging LLMs for reward function design in reinforcement learning control tasks
The challenge of designing effective reward functions in reinforcement learning (RL) represents a significant bottleneck, often requiring extensive human expertise and being time-consuming. Previous work and recent advancements in large language models (LLMs) have demonstrated their potential for automating the generation of reward functions. However, existing methodologies often require preliminary evaluation metrics, human-engineered feedback for the refinement process, or the use of environmental source code as context. To address these limitations, this paper introduces LEARN-Opt (LLM-based Evaluator and Analyzer for Reward functioN Optimization). This LLM-based, fully autonomous, and model-agnostic framework eliminates the need for preliminary metrics and environmental source code as context to generate, execute, and evaluate reward function candidates from textual descriptions of systems and task objectives. LEARN-Opt's main contribution lies in its ability to autonomously derive performance metrics directly from the system description and the task objective, enabling unsupervised evaluation and selection of reward functions. Our experiments indicate that LEARN-Opt achieves performance comparable to or better to that of state-of-the-art methods, such as EUREKA, while requiring less prior knowledge. We find that automated reward design is a high-variance problem, where the average-case candidate fails, requiring a multi-run approach to find the best candidates. Finally, we show that LEARN-Opt can unlock the potential of low-cost LLMs to find high-performing candidates that are comparable to, or even better than, those of larger models. This demonstrated performance affirms its potential to generate high-quality reward functions without requiring any preliminary human-defined metrics, thereby reducing engineering overhead and enhancing generalizability.
☆ Rethinking Intermediate Representation for VLM-based Robot Manipulation
Vision-Language Model (VLM) is an important component to enable robust robot manipulation. Yet, using it to translate human instructions into an action-resolvable intermediate representation often needs a tradeoff between VLM-comprehensibility and generalizability. Inspired by context-free grammar, we design the Semantic Assembly representation named SEAM, by decomposing the intermediate representation into vocabulary and grammar. Doing so leads us to a concise vocabulary of semantically-rich operations and a VLM-friendly grammar for handling diverse unseen tasks. In addition, we design a new open-vocabulary segmentation paradigm with a retrieval-augmented few-shot learning strategy to localize fine-grained object parts for manipulation, effectively with the shortest inference time over all state-of-the-art parallel works. Also, we formulate new metrics for action-generalizability and VLM-comprehensibility, demonstrating the compelling performance of SEAM over mainstream representations on both aspects. Extensive real-world experiments further manifest its SOTA performance under varying settings and tasks.
☆ SENTINEL: A Fully End-to-End Language-Action Model for Humanoid Whole Body Control
Existing humanoid control systems often rely on teleoperation or modular generation pipelines that separate language understanding from physical execution. However, the former is entirely human-driven, and the latter lacks tight alignment between language commands and physical behaviors. In this paper, we present SENTINEL, a fully end-to-end language-action model for humanoid whole-body control. We construct a large-scale dataset by tracking human motions in simulation using a pretrained whole body controller, combined with their text annotations. The model directly maps language commands and proprioceptive inputs to low-level actions without any intermediate representation. The model generates action chunks using flow matching, which can be subsequently refined by a residual action head for real-world deployment. Our method exhibits strong semantic understanding and stable execution on humanoid robots in both simulation and real-world deployment, and also supports multi-modal extensions by converting inputs into texts.
comment: 23 pages, 8 figures, 11 tables
☆ Soft pneumatic grippers: Topology optimization, 3D-printing and experimental validation
This paper presents a systematic topology optimization framework for designing a soft pneumatic gripper (SPG), explicitly considering the design-dependent nature of the actuating load. The load is modeled using Darcy's law with an added drainage term. A 2D soft arm unit is optimized by formulating it as a compliant mechanism design problem using the robust formulation. The problem is posed as a min-max optimization, where the output deformations of blueprint and eroded designs are considered. A volume constraint is imposed on the blueprint part, while a strain-energy constraint is enforced on the eroded part. The MMA is employed to solve the optimization problem and obtain the optimized soft unit. Finite element analysis with the Ogden material model confirms that the optimized 2D unit outperforms a conventional rectangular design under pneumatic loading. The optimized 2D unit is extruded to obtain a 3D module, and ten such units are assembled to create a soft arm. Deformation profiles of the optimized arm are analysed under different pressure loads. Four arms are 3D-printed and integrated with a supporting structure to realize the proposed SPG. The gripping performance of the SPG is demonstrated on objects with different weights, sizes, stiffness, and shapes.
comment: 9 Figures
☆ Reference-Free Sampling-Based Model Predictive Control
We present a sampling-based model predictive control (MPC) framework that enables emergent locomotion without relying on handcrafted gait patterns or predefined contact sequences. Our method discovers diverse motion patterns, ranging from trotting to galloping, robust standing policies, jumping, and handstand balancing, purely through the optimization of high-level objectives. Building on model predictive path integral (MPPI), we propose a dual-space spline parameterization that operates on position and velocity control points. Our approach enables contact-making and contact-breaking strategies that adapt automatically to task requirements, requiring only a limited number of sampled trajectories. This sample efficiency allows us to achieve real-time control on standard CPU hardware, eliminating the need for GPU acceleration typically required by other state-of-the-art MPPI methods. We validate our approach on the Go2 quadrupedal robot, demonstrating various emergent gaits and basic jumping capabilities. In simulation, we further showcase more complex behaviors, such as backflips, dynamic handstand balancing and locomotion on a Humanoid, all without requiring reference tracking or offline pre-training.
☆ Efficient Optimization of a Permanent Magnet Array for a Stable 2D Trap ICRA
Untethered magnetic manipulation of biomedical millirobots has a high potential for minimally invasive surgical applications. However, it is still challenging to exert high actuation forces on the small robots over a large distance. Permanent magnets offer stronger magnetic torques and forces than electromagnetic coils, however, feedback control is more difficult. As proven by Earnshaw's theorem, it is not possible to achieve a stable magnetic trap in 3D by static permanent magnets. Here, we report a stable 2D magnetic force trap by an array of permanent magnets to control a millirobot. The trap is located in an open space with a tunable distance to the magnet array in the range of 20 - 120mm, which is relevant to human anatomical scales. The design is achieved by a novel GPU-accelerated optimization algorithm that uses mean squared error (MSE) and Adam optimizer to efficiently compute the optimal angles for any number of magnets in the array. The algorithm is verified using numerical simulation and physical experiments with an array of two magnets. A millirobot is successfully trapped and controlled to follow a complex trajectory. The algorithm demonstrates high scalability by optimizing the angles for 100 magnets in under three seconds. Moreover, the optimization workflow can be adapted to optimize a permanent magnet array to achieve the desired force vector fields.
comment: 6 pages, 6 figures, IEEE International Conference on Robotics and Automation (ICRA)
☆ Three-Dimensional Anatomical Data Generation Based on Artificial Neural Networks IROS
Surgical planning and training based on machine learning requires a large amount of 3D anatomical models reconstructed from medical imaging, which is currently one of the major bottlenecks. Obtaining these data from real patients and during surgery is very demanding, if even possible, due to legal, ethical, and technical challenges. It is especially difficult for soft tissue organs with poor imaging contrast, such as the prostate. To overcome these challenges, we present a novel workflow for automated 3D anatomical data generation using data obtained from physical organ models. We additionally use a 3D Generative Adversarial Network (GAN) to obtain a manifold of 3D models useful for other downstream machine learning tasks that rely on 3D data. We demonstrate our workflow using an artificial prostate model made of biomimetic hydrogels with imaging contrast in multiple zones. This is used to physically simulate endoscopic surgery. For evaluation and 3D data generation, we place it into a customized ultrasound scanner that records the prostate before and after the procedure. A neural network is trained to segment the recorded ultrasound images, which outperforms conventional, non-learning-based computer vision techniques in terms of intersection over union (IoU). Based on the segmentations, a 3D mesh model is reconstructed, and performance feedback is provided.
comment: 6 pages, 4 figures, 1 table, IEEE International Conference on Intelligent Robots and Systems (IROS)
☆ First-order Sobolev Reinforcement Learning
We propose a refinement of temporal-difference learning that enforces first-order Bellman consistency: the learned value function is trained to match not only the Bellman targets in value but also their derivatives with respect to states and actions. By differentiating the Bellman backup through differentiable dynamics, we obtain analytically consistent gradient targets. Incorporating these into the critic objective using a Sobolev-type loss encourages the critic to align with both the value and local geometry of the target function. This first-order TD matching principle can be seamlessly integrated into existing algorithms, such as Q-learning or actor-critic methods (e.g., DDPG, SAC), potentially leading to faster critic convergence and more stable policy gradients without altering their overall structure.
comment: Workshop paper at Differentiable Systems and Scientific Machine Learning, EurIPS 2025
☆ Autonomous Docking of Multi-Rotor UAVs on Blimps under the Influence of Wind Gusts
Multi-rotor UAVs face limited flight time due to battery constraints. Autonomous docking on blimps with onboard battery recharging and data offloading offers a promising solution for extended UAV missions. However, the vulnerability of blimps to wind gusts causes trajectory deviations, requiring precise, obstacle-aware docking strategies. To this end, this work introduces two key novelties: (i) a temporal convolutional network that predicts blimp responses to wind gusts, enabling rapid gust detection and estimation of points where the wind gust effect has subsided; (ii) a model predictive controller (MPC) that leverages these predictions to compute collision-free trajectories for docking, enabled by a novel obstacle avoidance method for close-range manoeuvres near the blimp. Simulation results show our method outperforms a baseline constant-velocity model of the blimp significantly across different scenarios. We further validate the approach in real-world experiments, demonstrating the first autonomous multi-rotor docking control strategy on blimps shown outside simulation. Source code is available here https://github.com/robot-perception-group/multi_rotor_airship_docking.
comment: 13 pages, 8 figures, 8 tables
☆ Analysis of Deep-Learning Methods in an ISO/TS 15066-Compliant Human-Robot Safety Framework
Over the last years collaborative robots have gained great success in manufacturing applications where human and robot work together in close proximity. However, current ISO/TS-15066-compliant implementations often limit the efficiency of collaborative tasks due to conservative speed restrictions. For this reason, this paper introduces a deep-learning-based human-robot-safety framework (HRSF) that aims at a dynamical adaptation of robot velocities depending on the separation distance between human and robot while respecting maximum biomechanical force and pressure limits. The applicability of the framework was investigated for four different deep learning approaches that can be used for human body extraction: human body recognition, human body segmentation, human pose estimation, and human body part segmentation. Unlike conventional industrial safety systems, the proposed HRSF differentiates individual human body parts from other objects, enabling optimized robot process execution. Experiments demonstrated a quantitative reduction in cycle time of up to 15% compared to conventional safety technology.
comment: MDPI Sensors, published 22 November 2025
☆ Multi-Agent Monocular Dense SLAM With 3D Reconstruction Priors
Monocular Simultaneous Localization and Mapping (SLAM) aims to estimate a robot's pose while simultaneously reconstructing an unknown 3D scene using a single camera. While existing monocular SLAM systems generate detailed 3D geometry through dense scene representations, they are computationally expensive due to the need for iterative optimization. To address this challenge, MASt3R-SLAM utilizes learned 3D reconstruction priors, enabling more efficient and accurate estimation of both 3D structures and camera poses. However, MASt3R-SLAM is limited to single-agent operation. In this paper, we extend MASt3R-SLAM to introduce the first multi-agent monocular dense SLAM system. Each agent performs local SLAM using a 3D reconstruction prior, and their individual maps are fused into a globally consistent map through a loop-closure-based map fusion mechanism. Our approach improves computational efficiency compared to state-of-the-art methods, while maintaining similar mapping accuracy when evaluated on real-world datasets.
☆ End-to-end Autonomous Vehicle Following System using Monocular Fisheye Camera
The increase in vehicle ownership has led to increased traffic congestion, more accidents, and higher carbon emissions. Vehicle platooning is a promising solution to address these issues by improving road capacity and reducing fuel consumption. However, existing platooning systems face challenges such as reliance on lane markings and expensive high-precision sensors, which limits their general applicability. To address these issues, we propose a vehicle following framework that expands its capability from restricted scenarios to general scenario applications using only a camera. This is achieved through our newly proposed end-to-end method, which improves overall driving performance. The method incorporates a semantic mask to address causal confusion in multi-frame data fusion. Additionally, we introduce a dynamic sampling mechanism to precisely track the trajectories of preceding vehicles. Extensive closed-loop validation in real-world vehicle experiments demonstrates the system's ability to follow vehicles in various scenarios, outperforming traditional multi-stage algorithms. This makes it a promising solution for cost-effective autonomous vehicle platooning. A complete real-world vehicle experiment is available at https://youtu.be/zL1bcVb9kqQ.
☆ AVA-VLA: Improving Vision-Language-Action models with Active Visual Attention
Vision-Language-Action (VLA) models have demonstrated remarkable capabilities in embodied AI tasks. However, existing VLA models, often built upon Vision-Language Models (VLMs), typically process dense visual inputs independently at each timestep. This approach implicitly models the task as a Markov Decision Process (MDP). However, this history-agnostic design is suboptimal for effective visual token processing in dynamic sequential decision-making, as it fails to leverage the context of history. To address this limitation, we reformulate the problem from a Partially Observable Markov Decision Process (POMDP) perspective and propose a novel framework named AVA-VLA. Inspired by the POMDP that the action generation should be conditioned on the belief state. AVA-VLA introduces Active Visual Attention (AVA) to dynamically modulate visual processing. It achieves this by leveraging the recurrent state, which is a neural approximation of the agent's belief state derived from the previous decision step. Specifically, the AVA module uses the recurrent state to compute the soft weights to actively process task-relevant visual tokens based on its historical context. Comprehensive evaluations demonstrate that AVA-VLA achieves state-of-the-art performance across popular robotic benchmarks, including LIBERO and CALVIN. Furthermore, real-world deployments on a dual-arm robot platform validate the framework's practical applicability and robust sim-to-real transferability.
comment: 18 pages, 10 figures
☆ Compressor-VLA: Instruction-Guided Visual Token Compression for Efficient Robotic Manipulation
Vision-Language-Action (VLA) models have emerged as a powerful paradigm in Embodied AI. However, the significant computational overhead of processing redundant visual tokens remains a critical bottleneck for real-time robotic deployment. While standard token pruning techniques can alleviate this, these task-agnostic methods struggle to preserve task-critical visual information. To address this challenge, simultaneously preserving both the holistic context and fine-grained details for precise action, we propose Compressor-VLA, a novel hybrid instruction-conditioned token compression framework designed for efficient, task-oriented compression of visual information in VLA models. The proposed Compressor-VLA framework consists of two token compression modules: a Semantic Task Compressor (STC) that distills holistic, task-relevant context, and a Spatial Refinement Compressor (SRC) that preserves fine-grained spatial details. This compression is dynamically modulated by the natural language instruction, allowing for the adaptive condensation of task-relevant visual information. Experimentally, extensive evaluations demonstrate that Compressor-VLA achieves a competitive success rate on the LIBERO benchmark while reducing FLOPs by 59% and the visual token count by over 3x compared to its baseline. The real-robot deployments on a dual-arm robot platform validate the model's sim-to-real transferability and practical applicability. Moreover, qualitative analyses reveal that our instruction guidance dynamically steers the model's perceptual focus toward task-relevant objects, thereby validating the effectiveness of our approach.
comment: 11 pages, 5 figures
☆ An Efficient Closed-Form Solution to Full Visual-Inertial State Initialization
In this letter, we present a closed-form initialization method that recovers the full visual-inertial state without nonlinear optimization. Unlike previous approaches that rely on iterative solvers, our formulation yields analytical, easy-to-implement, and numerically stable solutions for reliable start-up. Our method builds on small-rotation and constant-velocity approximations, which keep the formulation compact while preserving the essential coupling between motion and inertial measurements. We further propose an observability-driven, two-stage initialization scheme that balances accuracy with initialization latency. Extensive experiments on the EuRoC dataset validate our assumptions: our method achieves 10-20% lower initialization error than optimization-based approaches, while using 4x shorter initialization windows and reducing computational cost by 5x.
comment: 8 pages, 2 figures, 10 tables. Submitted to RA-L
☆ Accelerating Reinforcement Learning via Error-Related Human Brain Signals
In this work, we investigate how implicit neural feed back can accelerate reinforcement learning in complex robotic manipulation settings. While prior electroencephalogram (EEG) guided reinforcement learning studies have primarily focused on navigation or low-dimensional locomotion tasks, we aim to understand whether such neural evaluative signals can improve policy learning in high-dimensional manipulation tasks involving obstacles and precise end-effector control. We integrate error related potentials decoded from offline-trained EEG classifiers into reward shaping and systematically evaluate the impact of human-feedback weighting. Experiments on a 7-DoF manipulator in an obstacle-rich reaching environment show that neural feedback accelerates reinforcement learning and, depending on the human-feedback weighting, can yield task success rates that at times exceed those of sparse-reward baselines. Moreover, when applying the best-performing feedback weighting across all sub jects, we observe consistent acceleration of reinforcement learning relative to the sparse-reward setting. Furthermore, leave-one subject-out evaluations confirm that the proposed framework remains robust despite the intrinsic inter-individual variability in EEG decodability. Our findings demonstrate that EEG-based reinforcement learning can scale beyond locomotion tasks and provide a viable pathway for human-aligned manipulation skill acquisition.
☆ GContextFormer: A global context-aware hybrid multi-head attention approach with scaled additive aggregation for multimodal trajectory prediction
Multimodal trajectory prediction generates multiple plausible future trajectories to address vehicle motion uncertainty from intention ambiguity and execution variability. However, HD map-dependent models suffer from costly data acquisition, delayed updates, and vulnerability to corrupted inputs, causing prediction failures. Map-free approaches lack global context, with pairwise attention over-amplifying straight patterns while suppressing transitional patterns, resulting in motion-intention misalignment. This paper proposes GContextFormer, a plug-and-play encoder-decoder architecture with global context-aware hybrid attention and scaled additive aggregation achieving intention-aligned multimodal prediction without map reliance. The Motion-Aware Encoder builds scene-level intention prior via bounded scaled additive aggregation over mode-embedded trajectory tokens and refines per-mode representations under shared global context, mitigating inter-mode suppression and promoting intention alignment. The Hierarchical Interaction Decoder decomposes social reasoning into dual-pathway cross-attention: a standard pathway ensures uniform geometric coverage over agent-mode pairs while a neighbor-context-enhanced pathway emphasizes salient interactions, with gating module mediating their contributions to maintain coverage-focus balance. Experiments on eight highway-ramp scenarios from TOD-VT dataset show GContextFormer outperforms state-of-the-art baselines. Compared to existing transformer models, GContextFormer achieves greater robustness and concentrated improvements in high-curvature and transition zones via spatial distributions. Interpretability is achieved through motion mode distinctions and neighbor context modulation exposing reasoning attribution. The modular architecture supports extensibility toward cross-domain multimodal reasoning tasks. Source: https://fenghy-chen.github.io/sources/.
☆ AutoOdom: Learning Auto-regressive Proprioceptive Odometry for Legged Locomotion
Accurate proprioceptive odometry is fundamental for legged robot navigation in GPS-denied and visually degraded environments where conventional visual odometry systems fail. Current approaches face critical limitations: analytical filtering methods suffer from modeling uncertainties and cumulative drift, hybrid learning-filtering approaches remain constrained by their analytical components, while pure learning-based methods struggle with simulation-to-reality transfer and demand extensive real-world data collection. This paper introduces AutoOdom, a novel autoregressive proprioceptive odometry system that overcomes these challenges through an innovative two-stage training paradigm. Stage 1 employs large-scale simulation data to learn complex nonlinear dynamics and rapidly changing contact states inherent in legged locomotion, while Stage 2 introduces an autoregressive enhancement mechanism using limited real-world data to effectively bridge the sim-to-real gap. The key innovation lies in our autoregressive training approach, where the model learns from its own predictions to develop resilience against sensor noise and improve robustness in highly dynamic environments. Comprehensive experimental validation on the Booster T1 humanoid robot demonstrates that AutoOdom significantly outperforms state-of-the-art methods across all evaluation metrics, achieving 57.2% improvement in absolute trajectory error, 59.2% improvement in Umeyama-aligned error, and 36.2% improvement in relative pose error compared to the Legolas baseline. Extensive ablation studies provide critical insights into sensor modality selection and temporal modeling, revealing counterintuitive findings about IMU acceleration data and validating our systematic design choices for robust proprioceptive odometry in challenging locomotion scenarios.
☆ MergeVLA: Cross-Skill Model Merging Toward a Generalist Vision-Language-Action Agent
Recent Vision-Language-Action (VLA) models reformulate vision-language models by tuning them with millions of robotic demonstrations. While they perform well when fine-tuned for a single embodiment or task family, extending them to multi-skill settings remains challenging: directly merging VLA experts trained on different tasks results in near-zero success rates. This raises a fundamental question: what prevents VLAs from mastering multiple skills within one model? With an empirical decomposition of learnable parameters during VLA fine-tuning, we identify two key sources of non-mergeability: (1) Finetuning drives LoRA adapters in the VLM backbone toward divergent, task-specific directions beyond the capacity of existing merging methods to unify. (2) Action experts develop inter-block dependencies through self-attention feedback, causing task information to spread across layers and preventing modular recombination. To address these challenges, we present MergeVLA, a merging-oriented VLA architecture that preserves mergeability by design. MergeVLA introduces sparsely activated LoRA adapters via task masks to retain consistent parameters and reduce irreconcilable conflicts in the VLM. Its action expert replaces self-attention with cross-attention-only blocks to keep specialization localized and composable. When the task is unknown, it uses a test-time task router to adaptively select the appropriate task mask and expert head from the initial observation, enabling unsupervised task inference. Across LIBERO, LIBERO-Plus, RoboTwin, and multi-task experiments on the real SO101 robotic arm, MergeVLA achieves performance comparable to or even exceeding individually finetuned experts, demonstrating robust generalization across tasks, embodiments, and environments.
☆ SP-VINS: A Hybrid Stereo Visual Inertial Navigation System based on Implicit Environmental Map
Filter-based visual inertial navigation system (VINS) has attracted mobile-robot researchers for the good balance between accuracy and efficiency, but its limited mapping quality hampers long-term high-accuracy state estimation. To this end, we first propose a novel filter-based stereo VINS, differing from traditional simultaneous localization and mapping (SLAM) systems based on 3D map, which performs efficient loop closure constraints with implicit environmental map composed of keyframes and 2D keypoints. Secondly, we proposed a hybrid residual filter framework that combines landmark reprojection and ray constraints to construct a unified Jacobian matrix for measurement updates. Finally, considering the degraded environment, we incorporated the camera-IMU extrinsic parameters into visual description to achieve online calibration. Benchmark experiments demonstrate that the proposed SP-VINS achieves high computational efficiency while maintaining long-term high-accuracy localization performance, and is superior to existing state-of-the-art (SOTA) methods.
☆ AIRHILT: A Human-in-the-Loop Testbed for Multimodal Conflict Detection in Aviation
We introduce AIRHILT (Aviation Integrated Reasoning, Human-in-the-Loop Testbed), a modular and lightweight simulation environment designed to evaluate multimodal pilot and air traffic control (ATC) assistance systems for aviation conflict detection. Built on the open-source Godot engine, AIRHILT synchronizes pilot and ATC radio communications, visual scene understanding from camera streams, and ADS-B surveillance data within a unified, scalable platform. The environment supports pilot- and controller-in-the-loop interactions, providing a comprehensive scenario suite covering both terminal area and en route operational conflicts, including communication errors and procedural mistakes. AIRHILT offers standardized JSON-based interfaces that enable researchers to easily integrate, swap, and evaluate automatic speech recognition (ASR), visual detection, decision-making, and text-to-speech (TTS) models. We demonstrate AIRHILT through a reference pipeline incorporating fine-tuned Whisper ASR, YOLO-based visual detection, ADS-B-based conflict logic, and GPT-OSS-20B structured reasoning, and present preliminary results from representative runway-overlap scenarios, where the assistant achieves an average time-to-first-warning of approximately 7.7 s, with average ASR and vision latencies of approximately 5.9 s and 0.4 s, respectively. The AIRHILT environment and scenario suite are openly available, supporting reproducible research on multimodal situational awareness and conflict detection in aviation; code and scenarios are available at https://github.com/ogarib3/airhilt.
comment: 9 pages, 4 figures, 1 table, 1 algorithm
☆ Head Stabilization for Wheeled Bipedal Robots via Force-Estimation-Based Admittance Control
Wheeled bipedal robots are emerging as flexible platforms for field exploration. However, head instability induced by uneven terrain can degrade the accuracy of onboard sensors or damage fragile payloads. Existing research primarily focuses on stabilizing the mobile platform but overlooks active stabilization of the head in the world frame, resulting in vertical oscillations that undermine overall stability. To address this challenge, we developed a model-based ground force estimation method for our 6-degree-of-freedom wheeled bipedal robot. Leveraging these force estimates, we implemented an admittance control algorithm to enhance terrain adaptability. Simulation experiments validated the real-time performance of the force estimator and the robot's robustness when traversing uneven terrain.
☆ Autonomous Surface Selection For Manipulator-Based UV Disinfection In Hospitals Using Foundation Models IROS
Ultraviolet (UV) germicidal radiation is an established non-contact method for surface disinfection in medical environments. Traditional approaches require substantial human intervention to define disinfection areas, complicating automation, while deep learning-based methods often need extensive fine-tuning and large datasets, which can be impractical for large-scale deployment. Additionally, these methods often do not address scene understanding for partial surface disinfection, which is crucial for avoiding unintended UV exposure. We propose a solution that leverages foundation models to simplify surface selection for manipulator-based UV disinfection, reducing human involvement and removing the need for model training. Additionally, we propose a VLM-assisted segmentation refinement to detect and exclude thin and small non-target objects, showing that this reduces mis-segmentation errors. Our approach achieves over 92\% success rate in correctly segmenting target and non-target surfaces, and real-world experiments with a manipulator and simulated UV light demonstrate its practical potential for real-world applications.
comment: 7 pages, 7 figures; This paper has been accepted by IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
☆ GVD-TG: Topological Graph based on Fast Hierarchical GVD Sampling for Robot Exploration
Topological maps are more suitable than metric maps for robotic exploration tasks. However, real-time updating of accurate and detail-rich environmental topological maps remains a challenge. This paper presents a topological map updating method based on the Generalized Voronoi Diagram (GVD). First, the newly observed areas are denoised to avoid low-efficiency GVD nodes misleading the topological structure. Subsequently, a multi-granularity hierarchical GVD generation method is designed to control the sampling granularity at both global and local levels. This not only ensures the accuracy of the topological structure but also enhances the ability to capture detail features, reduces the probability of path backtracking, and ensures no overlap between GVDs through the maintenance of a coverage map, thereby improving GVD utilization efficiency. Second, a node clustering method with connectivity constraints and a connectivity method based on a switching mechanism are designed to avoid the generation of unreachable nodes and erroneous nodes caused by obstacle attraction. A special cache structure is used to store all connectivity information, thereby improving exploration efficiency. Finally, to address the issue of frontiers misjudgment caused by obstacles within the scope of GVD units, a frontiers extraction method based on morphological dilation is designed to effectively ensure the reachability of frontiers. On this basis, a lightweight cost function is used to assess and switch to the next viewpoint in real time. This allows the robot to quickly adjust its strategy when signs of path backtracking appear, thereby escaping the predicament and increasing exploration flexibility. And the performance of system for exploration task is verified through comparative tests with SOTA methods.
comment: 12 pages, 10 figures
☆ Asynchronous Distributed Multi-Robot Motion Planning Under Imperfect Communication
This paper addresses the challenge of coordinating multi-robot systems under realistic communication delays using distributed optimization. We focus on consensus ADMM as a scalable framework for generating collision-free, dynamically feasible motion plans in both trajectory optimization and receding-horizon control settings. In practice, however, these algorithms are sensitive to penalty tuning or adaptation schemes (e.g. residual balancing and adaptive parameter heuristics) that do not explicitly consider delays. To address this, we introduce a Delay-Aware ADMM (DA-ADMM) variant that adapts penalty parameters based on real-time delay statistics, allowing agents to down-weight stale information and prioritize recent updates during consensus and dual updates. Through extensive simulations in 2D and 3D environments with double-integrator, Dubins-car, and drone dynamics, we show that DA-ADMM significantly improves robustness, success rate, and solution quality compared to fixed-parameter, residual-balancing, and fixed-constraint baselines. Our results highlight that performance degradation is not solely determined by delay length or frequency, but by the optimizer's ability to contextually reason over delayed information. The proposed DA-ADMM achieves consistently better coordination performance across a wide range of delay conditions, offering a principled and efficient mechanism for resilient multi-robot motion planning under imperfect communication.
comment: 9 pages, 5 figures
☆ CNN-Based Camera Pose Estimation and Localisation of Scan Images for Aircraft Visual Inspection
General Visual Inspection is a manual inspection process regularly used to detect and localise obvious damage on the exterior of commercial aircraft. There has been increasing demand to perform this process at the boarding gate to minimise the downtime of the aircraft and automating this process is desired to reduce the reliance on human labour. Automating this typically requires estimating a camera's pose with respect to the aircraft for initialisation but most existing localisation methods require infrastructure, which is very challenging in uncontrolled outdoor environments and within the limited turnover time (approximately 2 hours) on an airport tarmac. Additionally, many airlines and airports do not allow contact with the aircraft's surface or using UAVs for inspection between flights, and restrict access to commercial aircraft. Hence, this paper proposes an on-site method that is infrastructure-free and easy to deploy for estimating a pan-tilt-zoom camera's pose and localising scan images. This method initialises using the same pan-tilt-zoom camera used for the inspection task by utilising a Deep Convolutional Neural Network fine-tuned on only synthetic images to predict its own pose. We apply domain randomisation to generate the dataset for fine-tuning the network and modify its loss function by leveraging aircraft geometry to improve accuracy. We also propose a workflow for initialisation, scan path planning, and precise localisation of images captured from a pan-tilt-zoom camera. We evaluate and demonstrate our approach through experiments with real aircraft, achieving root-mean-square camera pose estimation errors of less than 0.24 m and 2 degrees for all real scenes.
comment: 12 pages, 12 figures
☆ Stable Multi-Drone GNSS Tracking System for Marine Robots
Accurate localization is essential for marine robotics, yet Global Navigation Satellite System (GNSS) signals are unreliable or unavailable even at a very short distance below the water surface. Traditional alternatives, such as inertial navigation, Doppler Velocity Loggers (DVL), SLAM, and acoustic methods, suffer from error accumulation, high computational demands, or infrastructure dependence. In this work, we present a scalable multi-drone GNSS-based tracking system for surface and near-surface marine robots. Our approach combines efficient visual detection, lightweight multi-object tracking, GNSS-based triangulation, and a confidence-weighted Extended Kalman Filter (EKF) to provide stable GNSS estimation in real time. We further introduce a cross-drone tracking ID alignment algorithm that enforces global consistency across views, enabling robust multi-robot tracking with redundant aerial coverage. We validate our system in diversified complex settings to show the scalability and robustness of the proposed algorithm.
☆ Beyond Description: Cognitively Benchmarking Fine-Grained Action for Embodied Agents
Multimodal Large Language Models (MLLMs) show promising results as decision-making engines for embodied agents operating in complex, physical environments. However, existing benchmarks often prioritize high-level planning or spatial reasoning, leaving the fine-grained action intelligence required for embodied physical interaction underexplored. To address this gap, we introduce CFG-Bench, a new benchmark designed to systematically evaluate this crucial capability. CFG-Bench consists of 1,368 curated videos paired with 19,562 three-modalities question-answer pairs targeting four cognitive abilities: 1) Physical Interaction, 2) Temporal-Causal Relation, 3) Intentional Understanding, and 4) Evaluative Judgment. Together, these dimensions provide a systematic framework for assessing a model's ability to translate visual observations into actionable knowledge, moving beyond mere surface-level recognition. Our comprehensive evaluation on CFG-Bench reveals that leading MLLMs struggle to produce detailed instructions for physical interactions and exhibit profound limitations in the higher-order reasoning of intention and evaluation. Moreover, supervised fine-tuning (SFT) on our data demonstrates that teaching an MLLMs to articulate fine-grained actions directly translates to significant performance gains on established embodied benchmarks. Our analysis highlights these limitations and offers insights for developing more capable and grounded embodied agents.
☆ Online Learning-Enhanced Lie Algebraic MPC for Robust Trajectory Tracking of Autonomous Surface Vehicles
Autonomous surface vehicles (ASVs) are easily influenced by environmental disturbances such as wind and waves, making accurate trajectory tracking a persistent challenge in dynamic marine conditions. In this paper, we propose an efficient controller for trajectory tracking of marine vehicles under unknown disturbances by combining a convex error-state MPC on the Lie group with an online learning module to compensate for these disturbances in real time. This design enables adaptive and robust control while maintaining computational efficiency. Extensive evaluations in numerical simulations, the Virtual RobotX (VRX) simulator, and real-world field experiments demonstrate that our method achieves superior tracking accuracy under various disturbance scenarios compared with existing approaches.
☆ Prune-Then-Plan: Step-Level Calibration for Stable Frontier Exploration in Embodied Question Answering
Large vision-language models (VLMs) have improved embodied question answering (EQA) agents by providing strong semantic priors for open-vocabulary reasoning. However, when used directly for step-level exploration, VLMs often exhibit frontier oscillations, unstable back-and-forth movements caused by overconfidence and miscalibration, leading to inefficient navigation and degraded answer quality. We propose Prune-Then-Plan, a simple and effective framework that stabilizes exploration through step-level calibration. Instead of trusting raw VLM scores, our method prunes implausible frontier choices using a Holm-Bonferroni inspired pruning procedure and then delegates final decisions to a coverage-based planner. This separation converts overconfident predictions into conservative, interpretable actions by relying on human-level judgments to calibrate the step-level behavior of VLMs. Integrated into the 3D-Mem EQA framework, our approach achieves relative improvements of up to 49% and 33% in visually grounded SPL and LLM-Match metrics respectively over baselines. Overall, our method achieves better scene coverage under equal exploration budgets on both OpenEQA and EXPRESS-Bench datasets.
comment: webpage: https://noahfrahm.github.io/Prune-Then-Plan-project-page/
☆ Maritime Small Object Detection from UAVs using Deep Learning with Altitude-Aware Dynamic Tiling
Unmanned Aerial Vehicles (UAVs) are crucial in Search and Rescue (SAR) missions due to their ability to monitor vast maritime areas. However, small objects often remain difficult to detect from high altitudes due to low object-to-background pixel ratios. We propose an altitude-aware dynamic tiling method that scales and adaptively subdivides the image into tiles for enhanced small object detection. By integrating altitude-dependent scaling with an adaptive tiling factor, we reduce unnecessary computation while maintaining detection performance. Tested on the SeaDronesSee dataset [1] with YOLOv5 [2] and Slicing Aided Hyper Inference (SAHI) framework [3], our approach improves Mean Average Precision (mAP) for small objects by 38% compared to a baseline and achieves more than double the inference speed compared to static tiling. This approach enables more efficient and accurate UAV-based SAR operations under diverse conditions.
comment: This is the author's accepted version of an article that has been published by IEEE. The final published version is available at IEEE Xplore
☆ Whole-Body Inverse Dynamics MPC for Legged Loco-Manipulation
Loco-manipulation demands coordinated whole-body motion to manipulate objects effectively while maintaining locomotion stability, presenting significant challenges for both planning and control. In this work, we propose a whole-body model predictive control (MPC) framework that directly optimizes joint torques through full-order inverse dynamics, enabling unified motion and force planning and execution within a single predictive layer. This approach allows emergent, physically consistent whole-body behaviors that account for the system's dynamics and physical constraints. We implement our MPC formulation using open software frameworks (Pinocchio and CasADi), along with the state-of-the-art interior-point solver Fatrop. In real-world experiments on a Unitree B2 quadruped equipped with a Unitree Z1 manipulator arm, our MPC formulation achieves real-time performance at 80 Hz. We demonstrate loco-manipulation tasks that demand fine control over the end-effector's position and force to perform real-world interactions like pulling heavy loads, pushing boxes, and wiping whiteboards.
comment: 9 pages, 6 figures, to be published in IEEE Robotics and Automation Letters (Special Issue: Advancements in MPC and Learning Algorithms for Legged Robots)
☆ Multi-Agent gatekeeper: Safe Flight Planning and Formation Control for Urban Air Mobility
We present Multi-Agent gatekeeper, a framework that provides provable safety guarantees for leader-follower formation control in cluttered 3D environments. Existing methods face a trad-off: online planners and controllers lack formal safety guarantees, while offline planners lack adaptability to changes in the number of agents or desired formation. To address this gap, we propose a hybrid architecture where a single leader tracks a pre-computed, safe trajectory, which serves as a shared trajectory backup set for all follower agents. Followers execute a nominal formation-keeping tracking controller, and are guaranteed to remain safe by always possessing a known-safe backup maneuver along the leader's path. We formally prove this method ensures collision avoidance with both static obstacles and other agents. The primary contributions are: (1) the multi-agent gatekeeper algorithm, which extends our single-agent gatekeeper framework to multi-agent systems; (2) the trajectory backup set for provably safe inter-agent coordination for leader-follower formation control; and (3) the first application of the gatekeeper framework in a 3D environment. We demonstrate our approach in a simulated 3D urban environment, where it achieved a 100% collision-avoidance success rate across 100 randomized trials, significantly outperforming baseline CBF and NMPC methods. Finally, we demonstrate the physical feasibility of the resulting trajectories on a team of quadcopters.
comment: 13 pages, 4 figures, to appear AIAA SciTech 2026
☆ IndEgo: A Dataset of Industrial Scenarios and Collaborative Work for Egocentric Assistants NeurIPS 2025
We introduce IndEgo, a multimodal egocentric and exocentric dataset addressing common industrial tasks, including assembly/disassembly, logistics and organisation, inspection and repair, woodworking, and others. The dataset contains 3,460 egocentric recordings (approximately 197 hours), along with 1,092 exocentric recordings (approximately 97 hours). A key focus of the dataset is collaborative work, where two workers jointly perform cognitively and physically intensive tasks. The egocentric recordings include rich multimodal data and added context via eye gaze, narration, sound, motion, and others. We provide detailed annotations (actions, summaries, mistake annotations, narrations), metadata, processed outputs (eye gaze, hand pose, semi-dense point cloud), and benchmarks on procedural and non-procedural task understanding, Mistake Detection, and reasoning-based Question Answering. Baseline evaluations for Mistake Detection, Question Answering and collaborative task understanding show that the dataset presents a challenge for the state-of-the-art multimodal models. Our dataset is available at: https://huggingface.co/datasets/FraunhoferIPK/IndEgo
comment: Accepted to NeurIPS 2025 D&B Track. Project Page: https://indego-dataset.github.io/
☆ Anytime-Feasible First-Order Optimization via Safe Sequential QCQP
This paper presents the Safe Sequential Quadratically Constrained Quadratic Programming (SS-QCQP) algorithm, a first-order method for smooth inequality-constrained nonconvex optimization that guarantees feasibility at every iteration. The method is derived from a continuous-time dynamical system whose vector field is obtained by solving a convex QCQP that enforces monotonic descent of the objective and forward invariance of the feasible set. The resulting continuous-time dynamics achieve an $O(1/t)$ convergence rate to first-order stationary points under standard constraint qualification conditions. We then propose a safeguarded Euler discretization with adaptive step-size selection that preserves this convergence rate while maintaining both descent and feasibility in discrete time. To enhance scalability, we develop an active-set variant (SS-QCQP-AS) that selectively enforces constraints near the boundary, substantially reducing computational cost without compromising theoretical guarantees. Numerical experiments on a multi-agent nonlinear optimal control problem demonstrate that SS-QCQP and SS-QCQP-AS maintain feasibility, exhibit the predicted convergence behavior, and deliver solution quality comparable to second-order solvers such as SQP and IPOPT.
☆ Development of a Testbed for Autonomous Vehicles: Integrating MPC Control with Monocular Camera Lane Detection
Autonomous vehicles are becoming popular day by day not only for autonomous road traversal but also for industrial automation, farming and military. Most of the standard vehicles follow the Ackermann style steering mechanism. This has become to de facto standard for large and long faring vehicles. The local planner of an autonomous vehicle controls the low-level vehicle movement upon which the vehicle will perform its motor actuation. In our work, we focus on autonomous vehicles in road and perform experiments to analyze the effect of low-level controllers in the simulation and a real environment. To increase the precision and stability of trajectory tracking in autonomous cars, a novel method that combines lane identification with Model Predictive Control (MPC) is presented. The research focuses on camera-equipped autonomous vehicles and uses methods like edge recognition, sliding window-based straight-line identification for lane line extraction, and dynamic region of interest (ROI) extraction. Next, to follow the identified lane line, an MPC built on a bicycle vehicle dynamics model is created. A single-lane road simulation model is built using ROS Gazebo and tested in order to verify the controller's performance. The root mean square error between the optimal tracking trajectory and the target trajectory was reduced by 27.65% in the simulation results, demonstrating the high robustness and flexibility of the developed controller.
comment: 49 pages, 23 figures
☆ Flow-Based Path Planning for Multiple Homogenous UAVs for Outdoor Formation-Flying
Collision-free path planning is the most crucial component in multi-UAV formation-flying (MFF). We use unlabeled homogenous quadcopters (UAVs) to demonstrate the use of a flow network to create complete (inter-UAV) collision-free paths. This procedure has three main parts: 1) Creating a flow network graph from physical GPS coordinates, 2) Finding a path of minimum cost (least distance) using any graph-based path-finding algorithm, and 3) Implementing the Ford-Fulkerson Method to find the paths with the maximum flow (no collision). Simulations of up to 64 UAVs were conducted for various formations, followed by a practical experiment with 3 quadcopters for testing physical plausibility and feasibility. The results of these tests show the efficacy of this method's ability to produce safe, collision-free paths.
comment: 9 pages, 15 figures, conference
☆ Online Learning-Enhanced High Order Adaptive Safety Control
Control barrier functions (CBFs) are an effective model-based tool to formally certify the safety of a system. With the growing complexity of modern control problems, CBFs have received increasing attention in both optimization-based and learning-based control communities as a safety filter, owing to their provable guarantees. However, success in transferring these guarantees to real-world systems is critically tied to model accuracy. For example, payloads or wind disturbances can significantly influence the dynamics of an aerial vehicle and invalidate the safety guarantee. In this work, we propose an efficient yet flexible online learning-enhanced high-order adaptive control barrier function using Neural ODEs. Our approach improves the safety of a CBF-certified system on the fly, even under complex time-varying model perturbations. In particular, we deploy our hybrid adaptive CBF controller on a 38g nano quadrotor, keeping a safe distance from the obstacle, against 18km/h wind.
comment: 8 pages, 7 figures, submitted to RA-L
☆ Robot-Powered Data Flywheels: Deploying Robots in the Wild for Continual Data Collection and Foundation Model Adaptation
Foundation models (FM) have unlocked powerful zero-shot capabilities in vision and language, yet their reliance on internet pretraining data leaves them brittle in unstructured, real-world settings. The messy, real-world data encountered during deployment (e.g. occluded or multilingual text) remains massively underrepresented in existing corpora. Robots, as embodied agents, are uniquely positioned to close this gap: they can act in physical environments to collect large-scale, real-world data that enriches FM training with precisely the examples current models lack. We introduce the Robot-Powered Data Flywheel, a framework that transforms robots from FM consumers into data generators. By deploying robots equipped with FMs in the wild, we enable a virtuous cycle: robots perform useful tasks while collecting real-world data that improves both domain-specific adaptation and domain-adjacent generalization. We instantiate this framework with Scanford, a mobile manipulator deployed in the East Asia Library for 2 weeks. Scanford autonomously scans shelves, identifies books using a vision-language model (VLM), and leverages the library catalog to label images without human annotation. This deployment both aids librarians and produces a dataset to finetune the underlying VLM, improving performance on the domain-specific in-the-wild library setting and on domain-adjacent multilingual OCR benchmarks. Using data collected from 2103 shelves, Scanford improves VLM performance on book identification from 32.0% to 71.8% and boosts domain-adjacent multilingual OCR from 24.8% to 46.6% (English) and 30.8% to 38.0% (Chinese), while saving an ~18.7 hrs of human time. These results highlight how robot-powered data flywheels can both reduce human effort in real deployments and unlock new pathways for continually adapting FMs to the messiness of reality. More details are at: https://scanford-robot.github.io
☆ Learning Massively Multitask World Models for Continuous Control
General-purpose control demands agents that act across many tasks and embodiments, yet research on reinforcement learning (RL) for continuous control remains dominated by single-task or offline regimes, reinforcing a view that online RL does not scale. Inspired by the foundation model recipe (large-scale pretraining followed by light RL) we ask whether a single agent can be trained on hundreds of tasks with online interaction. To accelerate research in this direction, we introduce a new benchmark with 200 diverse tasks spanning many domains and embodiments, each with language instructions, demonstrations, and optionally image observations. We then present \emph{Newt}, a language-conditioned multitask world model that is first pretrained on demonstrations to acquire task-aware representations and action priors, and then jointly optimized with online interaction across all tasks. Experiments show that Newt yields better multitask performance and data-efficiency than a set of strong baselines, exhibits strong open-loop control, and enables rapid adaptation to unseen tasks. We release our environments, demonstrations, code for training and evaluation, as well as 200+ checkpoints.
comment: Webpage: https://www.nicklashansen.com/NewtWM
☆ A Virtual Mechanical Interaction Layer Enables Resilient Human-to-Robot Object Handovers
Object handover is a common form of interaction that is widely present in collaborative tasks. However, achieving it efficiently remains a challenge. We address the problem of ensuring resilient robotic actions that can adapt to complex changes in object pose during human-to-robot object handovers. We propose the use of Virtual Model Control to create an interaction layer that controls the robot and adapts to the dynamic changes in the handover process. Additionally, we propose the use of augmented reality to facilitate bidirectional communication between humans and robots during handovers. We assess the performance of our controller in a set of experiments that demonstrate its resilience to various sources of uncertainties, including complex changes to the object's pose during the handover. Finally, we performed a user study with 16 participants to understand human preferences for different robot control profiles and augmented reality visuals in object handovers. Our results showed a general preference for the proposed approach and revealed insights that can guide further development in adapting the interaction with the user.
☆ Discover, Learn, and Reinforce: Scaling Vision-Language-Action Pretraining with Diverse RL-Generated Trajectories
Scaling vision-language-action (VLA) model pre-training requires large volumes of diverse, high-quality manipulation trajectories. Most current data is obtained via human teleoperation, which is expensive and difficult to scale. Reinforcement learning (RL) methods learn useful skills through autonomous exploration, making them a viable approach for generating data. However, standard RL training collapses to a narrow execution pattern, limiting its utility for large-scale pre-training. We propose Discover, Lea rn and Reinforce (DLR), an information-theoretic pattern discovery framework that generates multiple distinct, high-success behavioral patterns for VLA pretraining. Empirically, DLR generates a markedly more diverse trajectory corpus on LIBERO. Specifically, it learns multiple distinct, high-success strategies for the same task where standard RL discovers only one, and hence it covers substantially broader regions of the state-action space. When adapted to unseen downstream task suites, VLA models pretrained on our diverse RL data surpass counterparts trained on equal-sized standard RL datasets. Moreover, DLR exhibits positive data-scaling behavior that single-pattern RL lacks. These results position multi-pattern RL as a practical, scalable data engine for embodied foundation models.
♻ ☆ Robotic World Model: A Neural Network Simulator for Robust Policy Optimization in Robotics
Learning robust and generalizable world models is crucial for enabling efficient and scalable robotic control in real-world environments. In this work, we introduce a novel framework for learning world models that accurately capture complex, partially observable, and stochastic dynamics. The proposed method employs a dual-autoregressive mechanism and self-supervised training to achieve reliable long-horizon predictions without relying on domain-specific inductive biases, ensuring adaptability across diverse robotic tasks. We further propose a policy optimization framework that leverages world models for efficient training in imagined environments and seamless deployment in real-world systems. This work advances model-based reinforcement learning by addressing the challenges of long-horizon prediction, error accumulation, and sim-to-real transfer. By providing a scalable and robust framework, the introduced methods pave the way for adaptive and efficient robotic systems in real-world applications.
♻ ☆ Extremum Seeking Controlled Wiggling for Tactile Insertion
When humans perform complex insertion tasks such as pushing a cup into a cupboard, routing a cable, or putting a key in a lock, they wiggle the object and adapt the process through tactile feedback. A similar robotic approach has not been developed. We study an extremum seeking control law that wiggles end effector pose to maximize insertion depth while minimizing strain measured by a GelSight Mini sensor. Evaluation is conducted on four keys featuring complex geometry and five assembly tasks featuring basic geometry. On keys, the algorithm achieves 71% success rate over 120 trials with 6-DOF perturbations, 84% over 240 trials with 1-DOF perturbations, and 75% over 40 trials initialized with vision. It significantly outperforms a baseline optimizer, CMA-ES, that replaces wiggling with random sampling. When tested on a state-of-the-art assembly benchmark featuring basic geometry, it achieves 98% over 50 vision-initialized trials. The benchmark's most similar baseline, which was trained on the objects, achieved 86%. These results, realized without contact modeling or learning, show that closed loop wiggling based on tactile feedback is a robust paradigm for robotic insertion.
comment: 8 pages, 6 figures, 4 tables
♻ ☆ A Target-based Multi-LiDAR Multi-Camera Extrinsic Calibration System
Extrinsic Calibration represents the cornerstone of autonomous driving. Its accuracy plays a crucial role in the perception pipeline, as any errors can have implications for the safety of the vehicle. Modern sensor systems collect different types of data from the environment, making it harder to align the data. To this end, we propose a target-based extrinsic calibration system tailored for a multi-LiDAR and multi-camera sensor suite. This system enables cross-calibration between LiDARs and cameras with limited prior knowledge using a custom ChArUco board and a tailored nonlinear optimization method. We test the system with real-world data gathered in a warehouse. Results demonstrated the effectiveness of the proposed method, highlighting the feasibility of a unique pipeline tailored for various types of sensors.
comment: RiTA 2025 Accepted, 13 Pages, 6 Figures and 2 Tables
♻ ☆ Evo-0: Vision-Language-Action Model with Implicit Spatial Understanding
Vision-Language-Action (VLA) models have emerged as a promising framework for enabling generalist robots capable of perceiving, reasoning, and acting in the real world. These models usually build upon pretrained Vision-Language Models (VLMs), which excel at semantic understanding due to large-scale image and text pretraining. However, existing VLMs typically lack precise spatial understanding capabilities, as they are primarily tuned on 2D image-text pairs without 3D supervision. To address this limitation, recent approaches have incorporated explicit 3D inputs such as point clouds or depth maps, but this necessitates additional depth sensors or pre-trained depth estimation models, which may yield defective results. In contrast, our work introduces a plug-and-play module that implicitly incorporates 3D geometry features into VLA models by leveraging an off-the-shelf visual geometry foundation model. This integration provides the model with depth-aware visual representations, improving its ability to understand the geometric structure of the scene and the spatial relationships among objects from RGB images alone. We evaluate our method on a set of spatially challenging tasks in both simulation and the real world. Extensive evaluations show that our method significantly improves the performance of state-of-the-art VLA models across diverse scenarios.
♻ ☆ M2R2: MultiModal Robotic Representation for Temporal Action Segmentation
Temporal action segmentation (TAS) has long been a key area of research in both robotics and computer vision. In robotics, algorithms have primarily focused on leveraging proprioceptive information to determine skill boundaries, with recent approaches in surgical robotics incorporating vision. In contrast, computer vision typically relies on exteroceptive sensors, such as cameras. Existing multimodal TAS models in robotics integrate feature fusion within the model, making it difficult to reuse learned features across different models. Meanwhile, pretrained vision-only feature extractors commonly used in computer vision struggle in scenarios with limited object visibility. In this work, we address these challenges by proposing M2R2, a multimodal feature extractor tailored for TAS, which combines information from both proprioceptive and exteroceptive sensors. We introduce a novel pretraining strategy that enables the reuse of learned features across multiple TAS models. Our method achieves state-of-the-art performance on the REASSEMBLE dataset, a challenging multimodal robotic assembly dataset, outperforming existing robotic action segmentation models by 46.6%. Additionally, we conduct an extensive ablation study to evaluate the contribution of different modalities in robotic TAS tasks.
comment: 8 pages, 6 figures, 2 tables
♻ ☆ Autonomous Vehicle Path Planning by Searching With Differentiable Simulation
Planning allows an agent to safely refine its actions before executing them in the real world. In autonomous driving, this is crucial to avoid collisions and navigate in complex, dense traffic scenarios. One way to plan is to search for the best action sequence. However, this is challenging when all necessary components - policy, next-state predictor, and critic - have to be learned. Here we propose Differentiable Simulation for Search (DSS), a framework that leverages the differentiable simulator Waymax as both a next state predictor and a critic. It relies on the simulator's hardcoded dynamics, making state predictions highly accurate, while utilizing the simulator's differentiability to effectively search across action sequences. Our DSS agent optimizes its actions using gradient descent over imagined future trajectories. We show experimentally that DSS - the combination of planning gradients and stochastic search - significantly improves tracking and path planning accuracy compared to sequence prediction, imitation learning, model-free RL, and other planning methods.
♻ ☆ Agility Meets Stability: Versatile Humanoid Control with Heterogeneous Data
Humanoid robots are envisioned to perform a wide range of tasks in human-centered environments, requiring controllers that combine agility with robust balance. Recent advances in locomotion and whole-body tracking have enabled impressive progress in either agile dynamic skills or stability-critical behaviors, but existing methods remain specialized, focusing on one capability while compromising the other. In this work, we introduce AMS (Agility Meets Stability), the first framework that unifies both dynamic motion tracking and extreme balance maintenance in a single policy. Our key insight is to leverage heterogeneous data sources: human motion capture datasets that provide rich, agile behaviors, and physically constrained synthetic balance motions that capture stability configurations. To reconcile the divergent optimization goals of agility and stability, we design a hybrid reward scheme that applies general tracking objectives across all data while injecting balance-specific priors only into synthetic motions. Further, an adaptive learning strategy with performance-driven sampling and motion-specific reward shaping enables efficient training across diverse motion distributions. We validate AMS extensively in simulation and on a real Unitree G1 humanoid. Experiments demonstrate that a single policy can execute agile skills such as dancing and running, while also performing zero-shot extreme balance motions like Ip Man's Squat, highlighting AMS as a versatile control paradigm for future humanoid applications.
♻ ☆ Monocular Person Localization under Camera Ego-motion IROS2025
Localizing a person from a moving monocular camera is critical for Human-Robot Interaction (HRI). To estimate the 3D human position from a 2D image, existing methods either depend on the geometric assumption of a fixed camera or use a position regression model trained on datasets containing little camera ego-motion. These methods are vulnerable to severe camera ego-motion, resulting in inaccurate person localization. We consider person localization as a part of a pose estimation problem. By representing a human with a four-point model, our method jointly estimates the 2D camera attitude and the person's 3D location through optimization. Evaluations on both public datasets and real robot experiments demonstrate our method outperforms baselines in person localization accuracy. Our method is further implemented into a person-following system and deployed on an agile quadruped robot.
comment: Accepted by IROS2025. Project page: https://medlartea.github.io/rpf-quadruped/
♻ ☆ Towards Sensor Data Abstraction of Autonomous Vehicle Perception Systems SC2
Full-stack autonomous driving perception modules usually consist of data-driven models based on multiple sensor modalities. However, these models might be biased to the sensor setup used for data acquisition. This bias can seriously impair the perception models' transferability to new sensor setups, which continuously occur due to the market's competitive nature. We envision sensor data abstraction as an interface between sensor data and machine learning applications for highly automated vehicles (HAD). For this purpose, we review the primary sensor modalities, camera, lidar, and radar, published in autonomous-driving related datasets, examine single sensor abstraction and abstraction of sensor setups, and identify critical paths towards an abstraction of sensor data from multiple perception configurations.
comment: Hannes Reichert, Lukas Lang, Kevin Rösch and Daniel Bogdoll contributed equally. Accepted for publication at ISC2 2021
♻ ☆ Description of Corner Cases in Automated Driving: Goals and Challenges ICCV 2021
Scaling the distribution of automated vehicles requires handling various unexpected and possibly dangerous situations, termed corner cases (CC). Since many modules of automated driving systems are based on machine learning (ML), CC are an essential part of the data for their development. However, there is only a limited amount of CC data in large-scale data collections, which makes them challenging in the context of ML. With a better understanding of CC, offline applications, e.g., dataset analysis, and online methods, e.g., improved performance of automated driving systems, can be improved. While there are knowledge-based descriptions and taxonomies for CC, there is little research on machine-interpretable descriptions. In this extended abstract, we will give a brief overview of the challenges and goals of such a description.
comment: Daniel Bogdoll, Jasmin Breitenstein and Florian Heidecker contributed equally. Accepted for publication at ICCV 2021 ERCVAD Workshop
♻ ☆ Anomaly Detection in Autonomous Driving: A Survey CVPR 2022
Nowadays, there are outstanding strides towards a future with autonomous vehicles on our roads. While the perception of autonomous vehicles performs well under closed-set conditions, they still struggle to handle the unexpected. This survey provides an extensive overview of anomaly detection techniques based on camera, lidar, radar, multimodal and abstract object level data. We provide a systematization including detection approach, corner case level, ability for an online application, and further attributes. We outline the state-of-the-art and point out current research gaps.
comment: Daniel Bogdoll and Maximilian Nitsche contributed equally. Accepted for publication at CVPR 2022 WAD workshop
♻ ☆ Greedy Heuristics for Sampling-Based Motion Planning in High-Dimensional State Spaces
Informed sampling techniques accelerate the convergence of sampling-based motion planners by biasing sampling toward regions of the state space that are most likely to yield better solutions. However, when the current solution path contains redundant or tortuous segments, the resulting informed subset may remain unnecessarily large, slowing convergence. Our prior work addressed this issue by introducing the greedy informed set, which reduces the sampling region based on the maximum heuristic cost along the current solution path. In this article, we formally characterize the behavior of the greedy informed set within Rapidly-exploring Random Tree (RRT*)-like planners and analyze how greedy sampling affects exploration and asymptotic optimality. We then present Greedy RRT* (G-RRT*), a bi-directional anytime variant of RRT* that leverages the greedy informed set to focus sampling in the most promising regions of the search space. Experiments on abstract planning benchmarks, manipulation tasks from the MotionBenchMaker dataset, and a dual-arm Barrett WAM problem demonstrate that G-RRT* rapidly finds initial solutions and converges asymptotically to optimal paths, outperforming state-of-the-art sampling-based planners.
comment: Submitted to the Autonomous Robots journal
♻ ☆ Yummy Operations Robot Initiative: Autonomous Cooking System Utilizing a Modular Robotic Kitchen and a Dual-Arm Proprioceptive Manipulator
This paper presents Yummy Operations Robot Initiative (YORI), a proprioceptive dual-arm robotic system that demonstrates autonomous multi-dish cooking for scalable food service applications. YORI integrates a dual-arm manipulator equipped with proprioceptive actuators, custom-designed tools, appliances, and a structured kitchen environment to address the complexities of cooking tasks. The proprioceptive actuators enable fast, precise, force-controlled movements while mitigating the risks associated with cooking-related impacts. The system's modular kitchen design and flexible tool-changing mechanism support simultaneous multi-dish preparation through torque control and optimization-based motion planning and scheduling. A comprehensive scheduling framework with dynamic rescheduling ensures reliable adaptation to new orders and delays. The system was publicly validated through live demonstrations, reliably preparing steak-frites across multiple convention sessions. This paper details YORI's design and explores future directions in kitchen optimization, task planning, and food quality control, demonstrating its potential as a scalable robotic cooking solution. A system introduction and cooking videos are available online.
♻ ☆ RynnVLA-002: A Unified Vision-Language-Action and World Model
We introduce RynnVLA-002, a unified Vision-Language-Action (VLA) and world model. The world model leverages action and visual inputs to predict future image states, learning the underlying physics of the environment to refine action generation. Conversely, the VLA model produces subsequent actions from image observations, enhancing visual understanding and supporting the world model's image generation. The unified framework of RynnVLA-002 enables joint learning of environmental dynamics and action planning. Our experiments show that RynnVLA-002 surpasses individual VLA and world models, demonstrating their mutual enhancement. We evaluate RynnVLA-002 in both simulation and real-world robot tasks. RynnVLA-002 achieves 97.4% success rate on the LIBERO simulation benchmark without pretraining, while in real-world LeRobot experiments, its integrated world model boosts the overall success rate by 50%.
♻ ☆ Doppler Correspondence: Non-Iterative Scan Matching With Doppler Velocity-Based Correspondence
Achieving successful scan matching is essential for LiDAR odometry. However, in challenging environments with adverse weather conditions or repetitive geometric patterns, LiDAR odometry performance is degraded due to incorrect scan matching. Recently, the emergence of frequency-modulated continuous wave 4D LiDAR and 4D radar technologies has provided the potential to address these unfavorable conditions. The term 4D refers to point cloud data characterized by range, azimuth, and elevation along with Doppler velocity. Although 4D data is available, most scan matching methods for 4D LiDAR and 4D radar still establish correspondence by repeatedly identifying the closest points between consecutive scans, overlooking the Doppler information. This paper introduces, for the first time, a simple Doppler velocity-based correspondence -- Doppler Correspondence -- that is invariant to translation and small rotation of the sensor, with its geometric and kinematic foundations. Extensive experiments demonstrate that the proposed method enables the direct matching of consecutive point clouds without an iterative process, making it computationally efficient. Additionally, it provides a more robust correspondence estimation in environments with repetitive geometric patterns.The implementation of our proposed method is publicly available at https://github.com/Tars0523/Doppler Correspondence.
♻ ☆ Learning Primitive Embodied World Models: Towards Scalable Robotic Learning
While video-generation-based embodied world models have gained increasing attention, their reliance on large-scale embodied interaction data remains a key bottleneck. The scarcity, difficulty of collection, and high dimensionality of embodied data fundamentally limit the alignment granularity between language and actions and exacerbate the challenge of long-horizon video generation--hindering generative models from achieving a "GPT moment" in the embodied domain. There is a naive observation: the diversity of embodied data far exceeds the relatively small space of possible primitive motions. Based on this insight, we propose a novel paradigm for world modeling--Primitive Embodied World Models (PEWM). By restricting video generation to fixed short horizons, our approach 1) enables fine-grained alignment between linguistic concepts and visual representations of robotic actions, 2) reduces learning complexity, 3) improves data efficiency in embodied data collection, and 4) decreases inference latency. By equipping with a modular Vision-Language Model (VLM) planner and a Start-Goal heatmap Guidance mechanism (SGG), PEWM further enables flexible closed-loop control and supports compositional generalization of primitive-level policies over extended, complex tasks. Our framework leverages the spatiotemporal vision priors in video models and the semantic awareness of VLMs to bridge the gap between fine-grained physical interaction and high-level reasoning, paving the way toward scalable, interpretable, and general-purpose embodied intelligence.
♻ ☆ DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning AAAI 2026
Autonomous vehicles must navigate safely in complex driving environments. Imitating a single expert trajectory, as in regression-based approaches, usually does not explicitly assess the safety of the predicted trajectory. Selection-based methods address this by generating and scoring multiple trajectory candidates and predicting the safety score for each. However, they face optimization challenges in precisely selecting the best option from thousands of candidates and distinguishing subtle but safety-critical differences, especially in rare and challenging scenarios. We propose DriveSuprim to overcome these challenges and advance the selection-based paradigm through a coarse-to-fine paradigm for progressive candidate filtering, a rotation-based augmentation method to improve robustness in out-of-distribution scenarios, and a self-distillation framework to stabilize training. DriveSuprim achieves state-of-the-art performance, reaching 93.5% PDMS in NAVSIM v1 and 87.1% EPDMS in NAVSIM v2 without extra data, with 83.02 Driving Score and 60.00 Success Rate on the Bench2Drive benchmark, demonstrating superior planning capabilities in various driving scenarios.
comment: Accepted to AAAI 2026
♻ ☆ Learning to See and Act: Task-Aware Virtual View Exploration for Robotic Manipulation
Recent vision-language-action (VLA) models for multi-task robotic manipulation commonly rely on static viewpoints and shared visual encoders, which limit 3D perception and cause task interference, hindering robustness and generalization. In this work, we propose Task-aware Virtual View Exploration (TVVE), a framework designed to overcome these challenges by integrating virtual view exploration with task-specific representation learning. TVVE employs an efficient exploration policy, accelerated by a novel pseudo-environment, to acquire informative views. Furthermore, we introduce a Task-aware Mixture-of-Experts (TaskMoE) visual encoder to disentangle features across different tasks, boosting both representation fidelity and task generalization. By learning to see the world in a task-aware way, TVVE generates more complete and discriminative visual representations, demonstrating significantly enhanced action prediction across a wide array of manipulation challenges. To further validate the robustness and generalization capability of TVVE under out-of-distribution (OOD) settings, we construct a challenging benchmark, RLBench-OG, covering various visual perturbations and camera pose variations. Extensive experiments on RLBench and RLBench-OG show that our TVVE achieves superior performance over state-of-the-art approaches. In real-robot experiments, TVVE demonstrates exceptional performance and generalizes robustly in multiple OOD settings, including visual disturbances and unseen instructions. Visual results and code are provided at: https://hcplab-sysu.github.io/TAVP.
comment: 24 pages, 15 figures, project page: https://hcplab-sysu.github.io/TAVP
♻ ☆ Nauplius Optimisation for Autonomous Hydrodynamics
Autonomous Underwater vehicles must operate in strong currents, limited acoustic bandwidth, and persistent sensing requirements where conventional swarm optimisation methods are unreliable. This paper formulates an irreversible hydrodynamic deployment problem for Autonomous Underwater Vehicle (AUV) swarms and presents Nauplius Optimisation for Autonomous Hydrodynamics (NOAH), a novel nature-inspired swarm optimisation algorithm that combines current-aware drift, irreversible settlement in persistent sensing nodes, and colony-based communication. Drawing inspiration from the behaviour of barnacle nauplii, NOAH addresses the critical limitations of existing swarm algorithms by providing hydrodynamic awareness, irreversible anchoring mechanisms, and colony-based communication capabilities essential for underwater exploration missions. The algorithm establishes a comprehensive foundation for scalable and energy-efficient underwater swarm robotics with validated performance analysis. Validation studies demonstrate an 86% success rate for permanent anchoring scenarios, providing a unified formulation for hydrodynamic constraints and irreversible settlement behaviours with an empirical study under flow.
♻ ☆ Vision Language Models Can Parse Floor Plan Maps
Vision language models (VLMs) can simultaneously reason about images and texts to tackle many tasks, from visual question answering to image captioning. This paper focuses on map parsing, a novel task that is unexplored within the VLM context and particularly useful to mobile robots. Map parsing requires understanding not only the labels but also the geometric configurations of a map, i.e., what areas are like and how they are connected. To evaluate the performance of VLMs on map parsing, we prompt VLMs with floor plan maps to generate task plans for complex indoor navigation. Our results demonstrate the remarkable capability of VLMs in map parsing, with a success rate of 0.96 in tasks requiring a sequence of nine navigation actions, e.g., approaching and going through doors. Other than intuitive observations, e.g., VLMs do better in smaller maps and simpler navigation tasks, there was a very interesting observation that its performance drops in large open areas. We provide practical suggestions to address such challenges as validated by our experimental results. Webpage: https://sites.google.com/view/vlm-floorplan/
Robotics 27
☆ AutoFocus-IL: VLM-based Saliency Maps for Data-Efficient Visual Imitation Learning without Extra Human Annotations
AutoFocus-IL is a simple yet effective method to improve data efficiency and generalization in visual imitation learning by guiding policies to attend to task-relevant features rather than distractors and spurious correlations. Although saliency regularization has emerged as a promising way to achieve this, existing approaches typically require costly supervision such as human gaze data or manual saliency annotations. In contrast, AutoFocus-IL leverages vision-language models (VLMs) to automatically identify and track key objects in demonstrations, generating temporal saliency maps that highlight causal visual signals while suppressing distractors. These maps are then used to regularize behavior cloning policies, yielding stronger alignment between visual attention and task-relevant cues. Experiments in both the CARLA simulator and real-robot manipulation tasks demonstrate that AutoFocus-IL not only outperforms standard behavior cloning but also surpasses state-of-the-art baselines that assume privileged access to human supervision, such as gaze data. Code, datasets, and trained policy videos are available at https://AutoFocus-IL.github.io/.
comment: 8 pages, 6 figures. Code and datasets available at http://autofocus-il.github.io/
☆ How to Train Your Latent Control Barrier Function: Smooth Safety Filtering Under Hard-to-Model Constraints
Latent safety filters extend Hamilton-Jacobi (HJ) reachability to operate on latent state representations and dynamics learned directly from high-dimensional observations, enabling safe visuomotor control under hard-to-model constraints. However, existing methods implement "least-restrictive" filtering that discretely switch between nominal and safety policies, potentially undermining the task performance that makes modern visuomotor policies valuable. While reachability value functions can, in principle, be adapted to be control barrier functions (CBFs) for smooth optimization-based filtering, we theoretically and empirically show that current latent-space learning methods produce fundamentally incompatible value functions. We identify two sources of incompatibility: First, in HJ reachability, failures are encoded via a "margin function" in latent space, whose sign indicates whether or not a latent is in the constraint set. However, representing the margin function as a classifier yields saturated value functions that exhibit discontinuous jumps. We prove that the value function's Lipschitz constant scales linearly with the margin function's Lipschitz constant, revealing that smooth CBFs require smooth margins. Second, reinforcement learning (RL) approximations trained solely on safety policy data yield inaccurate value estimates for nominal policy actions, precisely where CBF filtering needs them. We propose the LatentCBF, which addresses both challenges through gradient penalties that lead to smooth margin functions without additional labeling, and a value-training procedure that mixes data from both nominal and safety policy distributions. Experiments on simulated benchmarks and hardware with a vision-based manipulation policy demonstrate that LatentCBF enables smooth safety filtering while doubling the task-completion rate over prior switching methods.
comment: 3 figures, 10 tables, 22 pages
☆ An Analysis of Constraint-Based Multi-Agent Pathfinding Algorithms
This study informs the design of future multi-agent pathfinding (MAPF) and multi-robot motion planning (MRMP) algorithms by guiding choices based on constraint classification for constraint-based search algorithms. We categorize constraints as conservative or aggressive and provide insights into their search behavior, focusing specifically on vanilla Conflict-Based Search (CBS) and Conflict-Based Search with Priorities (CBSw/P). Under a hybrid grid-roadmap representation with varying resolution, we observe that aggressive (priority constraint) formulations tend to solve more instances as agent count or resolution increases, whereas conservative (motion constraint) formulations yield stronger solution quality when both succeed. Findings are synthesized in a decision flowchart, aiding users in selecting suitable constraints. Recommendations extend to Multi-Robot Motion Planning (MRMP), emphasizing the importance of considering topological features alongside problem, solution, and representation features. A comprehensive exploration of the study, including raw data and map performance, is available in our public GitHub Repository: https://GitHub.com/hannahjmlee/constraint-mapf-analysis
☆ Connectivity-Preserving Multi-Agent Area Coverage via Optimal-Transport-Based Density-Driven Optimal Control (D2OC)
Multi-agent systems play a central role in area coverage tasks across search-and-rescue, environmental monitoring, and precision agriculture. Achieving non-uniform coverage, where spatial priorities vary across the domain, requires coordinating agents while respecting dynamic and communication constraints. Density-driven approaches can distribute agents according to a prescribed reference density, but existing methods do not ensure connectivity. This limitation often leads to communication loss, reduced coordination, and degraded coverage performance. This letter introduces a connectivity-preserving extension of the Density-Driven Optimal Control (D2OC) framework. The coverage objective, defined using the Wasserstein distance between the agent distribution and the reference density, admits a convex quadratic program formulation. Communication constraints are incorporated through a smooth connectivity penalty, which maintains strict convexity, supports distributed implementation, and preserves inter-agent communication without imposing rigid formations. Simulation studies show that the proposed method consistently maintains connectivity, improves convergence speed, and enhances non-uniform coverage quality compared with density-driven schemes that do not incorporate explicit connectivity considerations.
comment: Under review in IEEE Control Systems Letters (LCSS). 6 pages
☆ PhysGS: Bayesian-Inferred Gaussian Splatting for Physical Property Estimation CVPR 2026
Understanding physical properties such as friction, stiffness, hardness, and material composition is essential for enabling robots to interact safely and effectively with their surroundings. However, existing 3D reconstruction methods focus on geometry and appearance and cannot infer these underlying physical properties. We present PhysGS, a Bayesian-inferred extension of 3D Gaussian Splatting that estimates dense, per-point physical properties from visual cues and vision--language priors. We formulate property estimation as Bayesian inference over Gaussian splats, where material and property beliefs are iteratively refined as new observations arrive. PhysGS also models aleatoric and epistemic uncertainties, enabling uncertainty-aware object and scene interpretation. Across object-scale (ABO-500), indoor, and outdoor real-world datasets, PhysGS improves accuracy of the mass estimation by up to 22.8%, reduces Shore hardness error by up to 61.2%, and lowers kinetic friction error by up to 18.1% compared to deterministic baselines. Our results demonstrate that PhysGS unifies 3D reconstruction, uncertainty modeling, and physical reasoning in a single, spatially continuous framework for dense physical property estimation. Additional results are available at https://samchopra2003.github.io/physgs.
comment: Submitted to CVPR 2026
☆ Object-centric Task Representation and Transfer using Diffused Orientation Fields
Curved objects pose a fundamental challenge for skill transfer in robotics: unlike planar surfaces, they do not admit a global reference frame. As a result, task-relevant directions such as "toward" or "along" the surface vary with position and geometry, making object-centric tasks difficult to transfer across shapes. To address this, we introduce an approach using Diffused Orientation Fields (DOF), a smooth representation of local reference frames, for transfer learning of tasks across curved objects. By expressing manipulation tasks in these smoothly varying local frames, we reduce the problem of transferring tasks across curved objects to establishing sparse keypoint correspondences. DOF is computed online from raw point cloud data using diffusion processes governed by partial differential equations, conditioned on keypoints. We evaluate DOF under geometric, topological, and localization perturbations, and demonstrate successful transfer of tasks requiring continuous physical interaction such as inspection, slicing, and peeling across varied objects. We provide our open-source codes at our website https://github.com/idiap/diffused_fields_robotics
☆ Splatblox: Traversability-Aware Gaussian Splatting for Outdoor Robot Navigation ICRA 2026
We present Splatblox, a real-time system for autonomous navigation in outdoor environments with dense vegetation, irregular obstacles, and complex terrain. Our method fuses segmented RGB images and LiDAR point clouds using Gaussian Splatting to construct a traversability-aware Euclidean Signed Distance Field (ESDF) that jointly encodes geometry and semantics. Updated online, this field enables semantic reasoning to distinguish traversable vegetation (e.g., tall grass) from rigid obstacles (e.g., trees), while LiDAR ensures 360-degree geometric coverage for extended planning horizons. We validate Splatblox on a quadruped robot and demonstrate transfer to a wheeled platform. In field trials across vegetation-rich scenarios, it outperforms state-of-the-art methods with over 50% higher success rate, 40% fewer freezing incidents, 5% shorter paths, and up to 13% faster time to goal, while supporting long-range missions up to 100 meters. Experiment videos and more details can be found on our project page: https://splatblox.github.io
comment: Submitted to ICRA 2026
☆ SafeFall: Learning Protective Control for Humanoid Robots
Bipedal locomotion makes humanoid robots inherently prone to falls, causing catastrophic damage to the expensive sensors, actuators, and structural components of full-scale robots. To address this critical barrier to real-world deployment, we present \method, a framework that learns to predict imminent, unavoidable falls and execute protective maneuvers to minimize hardware damage. SafeFall is designed to operate seamlessly alongside existing nominal controller, ensuring no interference during normal operation. It combines two synergistic components: a lightweight, GRU-based fall predictor that continuously monitors the robot's state, and a reinforcement learning policy for damage mitigation. The protective policy remains dormant until the predictor identifies a fall as unavoidable, at which point it activates to take control and execute a damage-minimizing response. This policy is trained with a novel, damage-aware reward function that incorporates the robot's specific structural vulnerabilities, learning to shield critical components like the head and hands while absorbing energy with more robust parts of its body. Validated on a full-scale Unitree G1 humanoid, SafeFall demonstrated significant performance improvements over unprotected falls. It reduced peak contact forces by 68.3\%, peak joint torques by 78.4\%, and eliminated 99.3\% of collisions with vulnerable components. By enabling humanoids to fail safely, SafeFall provides a crucial safety net that allows for more aggressive experiments and accelerates the deployment of these robots in complex, real-world environments.
☆ Expanding the Workspace of Electromagnetic Navigation Systems Using Dynamic Feedback for Single- and Multi-agent Control
Electromagnetic navigation systems (eMNS) enable a number of magnetically guided surgical procedures. A challenge in magnetically manipulating surgical tools is that the effective workspace of an eMNS is often severely constrained by power and thermal limits. We show that system-level control design significantly expands this workspace by reducing the currents needed to achieve a desired motion. We identified five key system approaches that enable this expansion: (i) motion-centric torque/force objectives, (ii) energy-optimal current allocation, (iii) real-time pose estimation, (iv) dynamic feedback, and (v) high-bandwidth eMNS components. As a result, we stabilize a 3D inverted pendulum on an eight-coil OctoMag eMNS with significantly lower currents (0.1-0.2 A vs. 8-14 A), by replacing a field-centric field-alignment strategy with a motion-centric torque/force-based approach. We generalize to multi-agent control by simultaneously stabilizing two inverted pendulums within a shared workspace, exploiting magnetic-field nonlinearity and coil redundancy for independent actuation. A structured analysis compares the electromagnetic workspaces of both paradigms and examines current-allocation strategies that map motion objectives to coil currents. Cross-platform evaluation of the clinically oriented Navion eMNS further demonstrates substantial workspace expansion by maintaining stable balancing at distances up to 50 cm from the coils. The results demonstrate that feedback is a practical path to scalable, efficient, and clinically relevant magnetic manipulation.
☆ Categorical Equivariant Deep Learning: Category-Equivariant Neural Networks and Universal Approximation Theorems
We develop a theory of category-equivariant neural networks (CENNs) that unifies group/groupoid-equivariant networks, poset/lattice-equivariant networks, graph and sheaf neural networks. Equivariance is formulated as naturality in a topological category with Radon measures, formulating linear and nonlinear layers in the categorical setup. We prove the equivariant universal approximation theorem in the general setting: the class of finite-depth CENNs is dense in the space of continuous equivariant transformations. We instantiate the framework for groups/groupoids, posets/lattices, graphs and cellular sheaves, deriving universal approximation theorems for them in a systematic manner. Categorical equivariant deep learning thus allows us to expand the horizons of equivariant deep learning beyond group actions, encompassing not only geometric symmetries but also contextual and compositional symmetries.
☆ Explicit Bounds on the Hausdorff Distance for Truncated mRPI Sets via Norm-Dependent Contraction Rates
This paper establishes the first explicit and closed-form upper bound on the Hausdorff distance between the truncated minimal robust positively invariant (mRPI) set and its infinite-horizon limit. While existing mRPI approximations guarantee asymptotic convergence through geometric or norm-based arguments, none provides a computable expression that quantifies the truncation error for a given horizon. We show that the error satisfies \( d_H(\mathcal{E}_N,\mathcal{E}_\infty) \le r_W\,γ^{N+1}/(1-γ), \) where $γ<1$ is the induced-norm contraction factor and $r_W$ depends only on the disturbance set. The bound is fully analytic, requires no iterative set computations, and directly characterizes the decay rate of the truncated Minkowski series. We further demonstrate that the choice of vector norm serves as a design parameter that accelerates convergence, enabling substantially tighter horizon selection for robust invariant-set computations and tube-based MPC. Numerical experiments validate the sharpness, scalability, and practical relevance of the proposed bound.
☆ Enhancing UAV Search under Occlusion using Next Best View Planning
Search and rescue missions are often critical following sudden natural disasters or in high-risk environmental situations. The most challenging search and rescue missions involve difficult-to-access terrains, such as dense forests with high occlusion. Deploying unmanned aerial vehicles for exploration can significantly enhance search effectiveness, facilitate access to challenging environments, and reduce search time. However, in dense forests, the effectiveness of unmanned aerial vehicles depends on their ability to capture clear views of the ground, necessitating a robust search strategy to optimize camera positioning and perspective. This work presents an optimized planning strategy and an efficient algorithm for the next best view problem in occluded environments. Two novel optimization heuristics, a geometry heuristic, and a visibility heuristic, are proposed to enhance search performance by selecting optimal camera viewpoints. Comparative evaluations in both simulated and real-world settings reveal that the visibility heuristic achieves greater performance, identifying over 90% of hidden objects in simulated forests and offering 10% better detection rates than the geometry heuristic. Additionally, real-world experiments demonstrate that the visibility heuristic provides better coverage under the canopy, highlighting its potential for improving search and rescue missions in occluded environments.
comment: Submitted to IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
☆ Learning Visually Interpretable Oscillator Networks for Soft Continuum Robots from Video
Data-driven learning of soft continuum robot (SCR) dynamics from high-dimensional observations offers flexibility but often lacks physical interpretability, while model-based approaches require prior knowledge and can be computationally expensive. We bridge this gap by introducing (1) the Attention Broadcast Decoder (ABCD), a plug-and-play module for autoencoder-based latent dynamics learning that generates pixel-accurate attention maps localizing each latent dimension's contribution while filtering static backgrounds. (2) By coupling these attention maps to 2D oscillator networks, we enable direct on-image visualization of learned dynamics (masses, stiffness, and forces) without prior knowledge. We validate our approach on single- and double-segment SCRs, demonstrating that ABCD-based models significantly improve multi-step prediction accuracy: 5.7x error reduction for Koopman operators and 3.5x for oscillator networks on the two-segment robot. The learned oscillator network autonomously discovers a chain structure of oscillators. Unlike standard methods, ABCD models enable smooth latent space extrapolation beyond training data. This fully data-driven approach yields compact, physically interpretable models suitable for control applications.
☆ MicCheck: Repurposing Off-the-Shelf Pin Microphones for Easy and Low-Cost Contact Sensing
Robotic manipulation tasks are contact-rich, yet most imitation learning (IL) approaches rely primarily on vision, which struggles to capture stiffness, roughness, slip, and other fine interaction cues. Tactile signals can address this gap, but existing sensors often require expensive, delicate, or integration-heavy hardware. In this work, we introduce MicCheck, a plug-and-play acoustic sensing approach that repurposes an off-the-shelf Bluetooth pin microphone as a low-cost contact sensor. The microphone clips into a 3D-printed gripper insert and streams audio via a standard USB receiver, requiring no custom electronics or drivers. Despite its simplicity, the microphone provides signals informative enough for both perception and control. In material classification, it achieves 92.9% accuracy on a 10-class benchmark across four interaction types (tap, knock, slow press, drag). For manipulation, integrating pin microphone into an IL pipeline with open source hardware improves the success rate on picking and pouring task from 0.40 to 0.80 and enables reliable execution of contact-rich skills such as unplugging and sound-based sorting. Compared with high-resolution tactile sensors, pin microphones trade spatial detail for cost and ease of integration, offering a practical pathway for deploying acoustic contact sensing in low-cost robot setups.
☆ AIA-UltraNeRF:Acoustic-Impedance-Aware Neural Radiance Field with Hash Encodings for Robotic Ultrasound Reconstruction and Localization
Neural radiance field (NeRF) is a promising approach for reconstruction and new view synthesis. However, previous NeRF-based reconstruction methods overlook the critical role of acoustic impedance in ultrasound imaging. Localization methods face challenges related to local minima due to the selection of initial poses. In this study, we design a robotic ultrasound system (RUSS) with an acoustic-impedance-aware ultrasound NeRF (AIA-UltraNeRF) to decouple the scanning and diagnostic processes. Specifically, AIA-UltraNeRF models a continuous function of hash-encoded spatial coordinates for the 3D ultrasound map, allowing for the storage of acoustic impedance without dense sampling. This approach accelerates both reconstruction and inference speeds. We then propose a dual-supervised network that leverages teacher and student models to hash-encode the rendered ultrasound images from the reconstructed map. AIA-UltraNeRF retrieves the most similar hash values without the need to render images again, providing an offline initial image position for localization. Moreover, we develop a RUSS with a spherical remote center of motion mechanism to hold the probe, implementing operator-independent scanning modes that separate image acquisition from diagnostic workflows. Experimental results on a phantom and human subjects demonstrate the effectiveness of acoustic impedance in implicitly characterizing the color of ultrasound images. AIAUltraNeRF achieves both reconstruction and localization with inference speeds that are 9.9 faster than those of vanilla NeRF.
☆ Skypilot: Fine-Tuning LLM with Physical Grounding for AAV Coverage Search
Autonomous aerial vehicles (AAVs) have played a pivotal role in coverage operations and search missions. Recent advances in large language models (LLMs) offer promising opportunities to augment AAV intelligence. These advances help address complex challenges like area coverage optimization, dynamic path planning, and adaptive decision-making. However, the absence of physical grounding in LLMs leads to hallucination and reproducibility problems in spatial reasoning and decision-making. To tackle these issues, we present Skypilot, an LLM-enhanced two-stage framework that grounds language models in physical reality by integrating monte carlo tree search (MCTS). In the first stage, we introduce a diversified action space that encompasses generate, regenerate, fine-tune, and evaluate operations, coupled with physics-informed reward functions to ensure trajectory feasibility. In the second stage, we fine-tune Qwen3-4B on 23,000 MCTS-generated samples, achieving substantial inference acceleration while maintaining solution quality. Extensive numerical simulations and real-world flight experiments validate the efficiency and superiority of our proposed approach. Detailed information and experimental results are accessible at https://sky-pilot.top.
☆ Dreaming Falcon: Physics-Informed Model-Based Reinforcement Learning for Quadcopters
Current control algorithms for aerial robots struggle with robustness in dynamic environments and adverse conditions. Model-based reinforcement learning (RL) has shown strong potential in handling these challenges while remaining sample-efficient. Additionally, Dreamer has demonstrated that online model-based RL can be achieved using a recurrent world model trained on replay buffer data. However, applying Dreamer to aerial systems has been quite challenging due to its sample inefficiency and poor generalization of dynamics models. Our work explores a physics-informed approach to world model learning and improves policy performance. The world model treats the quadcopter as a free-body system and predicts the net forces and moments acting on it, which are then passed through a 6-DOF Runge-Kutta integrator (RK4) to predict future state rollouts. In this paper, we compare this physics-informed method to a standard RNN-based world model. Although both models perform well on the training data, we observed that they fail to generalize to new trajectories, leading to rapid divergence in state rollouts, preventing policy convergence.
☆ APULSE: A Scalable Hybrid Algorithm for the RCSPP on Large-Scale Dense Graphs
The resource-constrained shortest path problem (RCSPP) is a fundamental NP-hard optimization challenge with broad applications, from network routing to autonomous navigation. This problem involves finding a path that minimizes a primary cost subject to a budget on a secondary resource. While various RCSPP solvers exist, they often face critical scalability limitations when applied to the large, dense graphs characteristic of complex, real-world scenarios, making them impractical for time-critical planning. This challenge is particularly acute in domains like mission planning for unmanned ground vehicles (UGVs), which demand solutions on large-scale terrain graphs. This paper introduces APULSE, a hybrid label-setting algorithm designed to efficiently solve the RCSPP on such challenging graphs. APULSE integrates a best-first search guided by an A* heuristic with aggressive, Pulse-style pruning mechanisms and a time-bucketing strategy for effective state-space reduction. A computational study, using a large-scale UGV planning scenario, benchmarks APULSE against state-of-the-art algorithms. The results demonstrate that APULSE consistently finds near-optimal solutions while being orders of magnitude faster and more robust, particularly on large problem instances where competing methods fail. This superior scalability establishes APULSE as an effective solution for RCSPP in complex, large-scale environments, enabling capabilities such as interactive decision support and dynamic replanning.
comment: 9 pages
♻ ☆ ReactEMG: Stable, Low-Latency Intent Detection from sEMG via Masked Modeling
Surface electromyography (sEMG) signals show promise for effective human-machine interfaces, particularly in rehabilitation and prosthetics. However, challenges remain in developing systems that respond quickly to user intent, produce stable flicker-free output suitable for device control, and work across different subjects without time-consuming calibration. In this work, we propose a framework for EMG-based intent detection that addresses these challenges. We cast intent detection as per-timestep segmentation of continuous sEMG streams, assigning labels as gestures unfold in real time. We introduce a masked modeling training strategy that aligns muscle activations with their corresponding user intents, enabling rapid onset detection and stable tracking of ongoing gestures. In evaluations against baseline methods, using metrics that capture accuracy, latency and stability for device control, our approach achieves state-of-the-art performance in zero-shot conditions. These results demonstrate its potential for wearable robotics and next-generation prosthetic systems. Our project website, video, code, and dataset are available at: https://reactemg.github.io/
♻ ☆ Simultaneous Localization and 3D-Semi Dense Mapping for Micro Drones Using Monocular Camera and Inertial Sensors
Monocular simultaneous localization and mapping (SLAM) algorithms estimate drone poses and build a 3D map using a single camera. Current algorithms include sparse methods that lack detailed geometry, while learning-driven approaches produce dense maps but are computationally intensive. Monocular SLAM also faces scale ambiguities, which affect its accuracy. To address these challenges, we propose an edge-aware lightweight monocular SLAM system combining sparse keypoint-based pose estimation with dense edge reconstruction. Our method employs deep learning-based depth prediction and edge detection, followed by optimization to refine keypoints and edges for geometric consistency, without relying on global loop closure or heavy neural computations. We fuse inertial data with vision by using an extended Kalman filter to resolve scale ambiguity and improve accuracy. The system operates in real time on low-power platforms, as demonstrated on a DJI Tello drone with a monocular camera and inertial sensors. In addition, we demonstrate robust autonomous navigation and obstacle avoidance in indoor corridors and on the TUM RGBD dataset. Our approach offers an effective, practical solution to real-time mapping and navigation in resource-constrained environments.
♻ ☆ MonoMPC: Monocular Vision Based Navigation with Learned Collision Model and Risk-Aware Model Predictive Control
Navigating unknown environments with a single RGB camera is challenging, as the lack of depth information prevents reliable collision-checking. While some methods use estimated depth to build collision maps, we found that depth estimates from vision foundation models are too noisy for zero-shot navigation in cluttered environments. We propose an alternative approach: instead of using noisy estimated depth for direct collision-checking, we use it as a rich context input to a learned collision model. This model predicts the distribution of minimum obstacle clearance that the robot can expect for a given control sequence. At inference, these predictions inform a risk-aware MPC planner that minimizes estimated collision risk. We proposed a joint learning pipeline that co-trains the collision model and risk metric using both safe and unsafe trajectories. Crucially, our joint-training ensures well calibrated uncertainty in our collision model that improves navigation in highly cluttered environments. Consequently, real-world experiments show reductions in collision-rate and improvements in goal reaching and speed over several strong baselines.
♻ ☆ Unreal Robotics Lab: A High-Fidelity Robotics Simulator with Advanced Physics and Rendering
High-fidelity simulation is essential for robotics research, enabling safe and efficient testing of perception, control, and navigation algorithms. However, achieving both photorealistic rendering and accurate physics modeling remains a challenge. This paper presents a novel simulation framework, the Unreal Robotics Lab (URL), that integrates the advanced rendering capabilities of the Unreal Engine with MuJoCo's high-precision physics simulation. Our approach enables realistic robotic perception while maintaining accurate physical interactions, facilitating benchmarking and dataset generation for vision-based robotics applications. The system supports complex environmental effects, such as smoke, fire, and water dynamics, which are critical to evaluating robotic performance under adverse conditions. We benchmark visual navigation and SLAM methods within our framework, demonstrating its utility for testing real-world robustness in controlled yet diverse scenarios. By bridging the gap between physics accuracy and photorealistic rendering, our framework provides a powerful tool for advancing robotics research and sim-to-real transfer. Our open-source framework is available at https://unrealroboticslab.github.io/.
♻ ☆ Advancing Autonomous Driving: DepthSense with Radar and Spatial Attention
Depth perception is crucial for spatial understanding and has traditionally been achieved through stereoscopic imaging. However, the precision of depth estimation using stereoscopic methods depends on the accurate calibration of binocular vision sensors. Monocular cameras, while more accessible, often suffer from reduced accuracy, especially under challenging imaging conditions. Optical sensors, too, face limitations in adverse environments, leading researchers to explore radar technology as a reliable alternative. Although radar provides coarse but accurate signals, its integration with fine-grained monocular camera data remains underexplored. In this research, we propose DepthSense, a novel radar-assisted monocular depth enhancement approach. DepthSense employs an encoder-decoder architecture, a Radar Residual Network, feature fusion with a spatial attention mechanism, and an ordinal regression layer to deliver precise depth estimations. We conducted extensive experiments on the nuScenes dataset to validate the effectiveness of DepthSense. Our methodology not only surpasses existing approaches in quantitative performance but also reduces parameter complexity and inference times. Our findings demonstrate that DepthSense represents a significant advancement over traditional stereo methods, offering a robust and efficient solution for depth estimation in autonomous driving. By leveraging the complementary strengths of radar and monocular camera data, DepthSense sets a new benchmark in the field, paving the way for more reliable and accurate spatial perception systems.
♻ ☆ Continuous Gaussian Process Pre-Optimization for Asynchronous Event-Inertial Odometry
Event cameras, as bio-inspired sensors, are asynchronously triggered with high-temporal resolution compared to intensity cameras. Recent work has focused on fusing the event measurements with inertial measurements to enable ego-motion estimation in high-speed and HDR environments. However, existing methods predominantly rely on IMU preintegration designed mainly for synchronous sensors and discrete-time frameworks. In this paper, we propose a continuous-time preintegration method based on the Temporal Gaussian Process (TGP) called GPO. Concretely, we model the preintegration as a time-indexed motion trajectory and leverage an efficient two-step optimization to initialize the precision preintegration pseudo-measurements. Our method realizes a linear and constant time cost for initialization and query, respectively. To further validate the proposal, we leverage the GPO to design an asynchronous event-inertial odometry and compare with other asynchronous fusion schemes within the same odometry system. Experiments conducted on both public and own-collected datasets demonstrate that the proposed GPO offers significant advantages in terms of precision and efficiency, outperforming existing approaches in handling asynchronous sensor fusion.
comment: 8pages
♻ ☆ AsynEIO: Asynchronous Monocular Event-Inertial Odometry Using Gaussian Process Regression
Event cameras, when combined with inertial sensors, show significant potential for motion estimation in challenging scenarios, such as high-speed maneuvers and low-light environments. There are many methods for producing such estimations, but most boil down to a synchronous discrete-time fusion problem. However, the asynchronous nature of event cameras and their unique fusion mechanism with inertial sensors remain underexplored. In this paper, we introduce a monocular event-inertial odometry method called AsynEIO, designed to fuse asynchronous event and inertial data within a unified Gaussian Process (GP) regression framework. Our approach incorporates an event-driven frontend that tracks feature trajectories directly from raw event streams at a high temporal resolution. These tracked feature trajectories, along with various inertial factors, are integrated into the same GP regression framework to enable asynchronous fusion. With deriving analytical residual Jacobians and noise models, our method constructs a factor graph that is iteratively optimized and pruned using a sliding-window optimizer. Comparative assessments highlight the performance of different inertial fusion strategies, suggesting optimal choices for varying conditions. Experimental results on both public datasets and our own event-inertial sequences indicate that AsynEIO outperforms existing methods, especially in high-speed and low-illumination scenarios.
comment: 20 pages, 20 figures
♻ ☆ Leverage Cross-Attention for End-to-End Open-Vocabulary Panoptic Reconstruction
Open-vocabulary panoptic reconstruction offers comprehensive scene understanding, enabling advances in embodied robotics and photorealistic simulation. In this paper, we propose PanopticRecon++, an end-to-end method that formulates panoptic reconstruction through a novel cross-attention perspective. This perspective models the relationship between 3D instances (as queries) and the scene's 3D embedding field (as keys) through their attention map. Unlike existing methods that separate the optimization of queries and keys or overlook spatial proximity, PanopticRecon++ introduces learnable 3D Gaussians as instance queries. This formulation injects 3D spatial priors to preserve proximity while maintaining end-to-end optimizability. Moreover, this query formulation facilitates the alignment of 2D open-vocabulary instance IDs across frames by leveraging optimal linear assignment with instance masks rendered from the queries. Additionally, we ensure semantic-instance segmentation consistency by fusing query-based instance segmentation probabilities with semantic probabilities in a novel panoptic head supervised by a panoptic loss. During training, the number of instance query tokens dynamically adapts to match the number of objects. PanopticRecon++ shows competitive performance in terms of 3D and 2D segmentation and reconstruction performance on both simulation and real-world datasets, and demonstrates a user case as a robot simulator. Our project website is at: https://yuxuan1206.github.io/panopticrecon_pp/
comment: 18 pages, 10 figures
♻ ☆ Situationally-Aware Dynamics Learning
Autonomous robots operating in complex, unstructured environments face significant challenges due to latent, unobserved factors that obscure their understanding of both their internal state and the external world. Addressing this challenge would enable robots to develop a more profound grasp of their operational context. To tackle this, we propose a novel framework for online learning of hidden state representations, with which the robots can adapt in real-time to uncertain and dynamic conditions that would otherwise be ambiguous and result in suboptimal or erroneous behaviors. Our approach is formalized as a Generalized Hidden Parameter Markov Decision Process, which explicitly models the influence of unobserved parameters on both transition dynamics and reward structures. Our core innovation lies in learning online the joint distribution of state transitions, which serves as an expressive representation of latent ego- and environmental-factors. This probabilistic approach supports the identification and adaptation to different operational situations, improving robustness and safety. Through a multivariate extension of Bayesian Online Changepoint Detection, our method segments changes in the underlying data generating process governing the robot's dynamics. The robot's transition model is then informed with a symbolic representation of the current situation derived from the joint distribution of latest state transitions, enabling adaptive and context-aware decision-making. To showcase the real-world effectiveness, we validate our approach in the challenging task of unstructured terrain navigation, where unmodeled and unmeasured terrain characteristics can significantly impact the robot's motion. Extensive experiments in both simulation and real world reveal significant improvements in data efficiency, policy performance, and the emergence of safer, adaptive navigation strategies.
Robotics 28
☆ AFT: Appearance-Based Feature Tracking for Markerless and Training-Free Shape Reconstruction of Soft Robots
Accurate shape reconstruction is essential for precise control and reliable operation of soft robots. Compared to sensor-based approaches, vision-based methods offer advantages in cost, simplicity, and ease of deployment. However, existing vision-based methods often rely on complex camera setups, specific backgrounds, or large-scale training datasets, limiting their practicality in real-world scenarios. In this work, we propose a vision-based, markerless, and training-free framework for soft robot shape reconstruction that directly leverages the robot's natural surface appearance. These surface features act as implicit visual markers, enabling a hierarchical matching strategy that decouples local partition alignment from global kinematic optimization. Requiring only an initial 3D reconstruction and kinematic alignment, our method achieves real-time shape tracking across diverse environments while maintaining robustness to occlusions and variations in camera viewpoints. Experimental validation on a continuum soft robot demonstrates an average tip error of 2.6% during real-time operation, as well as stable performance in practical closed-loop control tasks. These results highlight the potential of the proposed approach for reliable, low-cost deployment in dynamic real-world settings.
☆ SkillWrapper: Generative Predicate Invention for Skill Abstraction
Generalizing from individual skill executions to solving long-horizon tasks remains a core challenge in building autonomous agents. A promising direction is learning high-level, symbolic abstractions of the low-level skills of the agents, enabling reasoning and planning independent of the low-level state space. Among possible high-level representations, object-centric skill abstraction with symbolic predicates has been proven to be efficient because of its compatibility with domain-independent planners. Recent advances in foundation models have made it possible to generate symbolic predicates that operate on raw sensory inputs, a process we call generative predicate invention, to facilitate downstream abstraction learning. However, it remains unclear which formal properties the learned representations must satisfy, and how they can be learned to guarantee these properties. In this paper, we address both questions by presenting a formal theory of generative predicate invention for skill abstraction, resulting in symbolic operators that can be used for provably sound and complete planning. Within this framework, we propose SkillWrapper, a method that leverages foundation models to actively collect robot data and learn human-interpretable, plannable representations of black-box skills, using only RGB image observations. Our extensive empirical evaluation in simulation and on real robots shows that SkillWrapper learns abstract representations that enable solving unseen, long-horizon tasks in the real world with black-box skills.
☆ Off-Road Navigation via Implicit Neural Representation of Terrain Traversability
Autonomous off-road navigation requires robots to estimate terrain traversability from onboard sensors and plan accordingly. Conventional approaches typically rely on sampling-based planners such as MPPI to generate short-term control actions that aim to minimize traversal time and risk measures derived from the traversability estimates. These planners can react quickly but optimize only over a short look-ahead window, limiting their ability to reason about the full path geometry, which is important for navigating in challenging off-road environments. Moreover, they lack the ability to adjust speed based on the terrain bumpiness, which is important for smooth navigation on challenging terrains. In this paper, we introduce TRAIL (Traversability with an Implicit Learned Representation), an off-road navigation framework that leverages an implicit neural representation to continuously parameterize terrain properties. This representation yields spatial gradients that enable integration with a novel gradient-based trajectory optimization method that adapts the path geometry and speed profile based on terrain traversability.
comment: 9 pages
☆ Time-aware Motion Planning in Dynamic Environments with Conformal Prediction
Safe navigation in dynamic environments remains challenging due to uncertain obstacle behaviors and the lack of formal prediction guarantees. We propose two motion planning frameworks that leverage conformal prediction (CP): a global planner that integrates Safe Interval Path Planning (SIPP) for uncertainty-aware trajectory generation, and a local planner that performs online reactive planning. The global planner offers distribution-free safety guarantees for long-horizon navigation, while the local planner mitigates inaccuracies in obstacle trajectory predictions through adaptive CP, enabling robust and responsive motion in dynamic environments. To further enhance trajectory feasibility, we introduce an adaptive quantile mechanism in the CP-based uncertainty quantification. Instead of using a fixed confidence level, the quantile is automatically tuned to the optimal value that preserves trajectory feasibility, allowing the planner to adaptively tighten safety margins in regions with higher uncertainty. We validate the proposed framework through numerical experiments conducted in dynamic and cluttered environments. The project page is available at https://time-aware-planning.github.io
☆ scipy.spatial.transform: Differentiable Framework-Agnostic 3D Transformations in Python
Three-dimensional rigid-body transforms, i.e. rotations and translations, are central to modern differentiable machine learning pipelines in robotics, vision, and simulation. However, numerically robust and mathematically correct implementations, particularly on SO(3), are error-prone due to issues such as axis conventions, normalizations, composition consistency and subtle errors that only appear in edge cases. SciPy's spatial.transform module is a rigorously tested Python implementation. However, it historically only supported NumPy, limiting adoption in GPU-accelerated and autodiff-based workflows. We present a complete overhaul of SciPy's spatial.transform functionality that makes it compatible with any array library implementing the Python array API, including JAX, PyTorch, and CuPy. The revised implementation preserves the established SciPy interface while enabling GPU/TPU execution, JIT compilation, vectorized batching, and differentiation via native autodiff of the chosen backend. We demonstrate how this foundation supports differentiable scientific computing through two case studies: (i) scalability of 3D transforms and rotations and (ii) a JAX drone simulation that leverages SciPy's Rotation for accurate integration of rotational dynamics. Our contributions have been merged into SciPy main and will ship in the next release, providing a framework-agnostic, production-grade basis for 3D spatial math in differentiable systems and ML.
comment: Accepted as oral at the 1st Workshop on Differentiable Systems and Scientific Machine Learning @ EurIPS 2025
☆ A Coordinated Dual-Arm Framework for Delicate Snap-Fit Assemblies
Delicate snap-fit assemblies, such as inserting a lens into an eye-wear frame or during electronics assembly, demand timely engagement detection and rapid force attenuation to prevent overshoot-induced component damage or assembly failure. We address these challenges with two key contributions. First, we introduce SnapNet, a lightweight neural network that detects snap-fit engagement from joint-velocity transients in real-time, showing that reliable detection can be achieved using proprioceptive signals without external sensors. Second, we present a dynamical-systems-based dual-arm coordination framework that integrates SnapNet driven detection with an event-triggered impedance modulation, enabling accurate alignment and compliant insertion during delicate snap-fit assemblies. Experiments across diverse geometries on a heterogeneous bimanual platform demonstrate high detection accuracy (over 96% recall) and up to a 30% reduction in peak impact forces compared to standard impedance control.
comment: 10 pages, 9 figures
☆ Observer Actor: Active Vision Imitation Learning with Sparse View Gaussian Splatting
We propose Observer Actor (ObAct), a novel framework for active vision imitation learning in which the observer moves to optimal visual observations for the actor. We study ObAct on a dual-arm robotic system equipped with wrist-mounted cameras. At test time, ObAct dynamically assigns observer and actor roles: the observer arm constructs a 3D Gaussian Splatting (3DGS) representation from three images, virtually explores this to find an optimal camera pose, then moves to this pose; the actor arm then executes a policy using the observer's observations. This formulation enhances the clarity and visibility of both the object and the gripper in the policy's observations. As a result, we enable the training of ambidextrous policies on observations that remain closer to the occlusion-free training distribution, leading to more robust policies. We study this formulation with two existing imitation learning methods -- trajectory transfer and behavior cloning -- and experiments show that ObAct significantly outperforms static-camera setups: trajectory transfer improves by 145% without occlusion and 233% with occlusion, while behavior cloning improves by 75% and 143%, respectively. Videos are available at https://obact.github.io.
comment: Videos are available on our project webpage at https://obact.github.io
☆ EchoVLA: Robotic Vision-Language-Action Model with Synergistic Declarative Memory for Mobile Manipulation
Recent progress in Vision-Language-Action (VLA) models has enabled embodied agents to interpret multimodal instructions and perform complex tasks. However, existing VLAs are mostly confined to short-horizon, table-top manipulation, lacking the memory and reasoning capability required for long-horizon mobile manipulation, where agents must coordinate navigation and manipulation under changing spatial contexts. In this work, we present EchoVLA, a memory-aware VLA model for long-horizon mobile manipulation. EchoVLA incorporates a synergistic declarative memory inspired by the human brain, consisting of a scene memory that maintains a collection of spatial-semantic maps and an episodic memory that stores task-level experiences with multimodal contextual features. During both training and inference, the two memories are individually stored, updated, and retrieved based on current observations, task history, and instructions, and their retrieved representations are fused via coarse- and fine-grained attention to guide mobile-arm diffusion policies. To support large-scale training and evaluation, we further introduce MoMani, an automated benchmark that generates expert-level long-horizon trajectories through multimodal large language model (MLLM)-guided planning and feedback-driven refinement, supplemented with real-robot demonstrations. Experiments in simulated and real-world settings show that EchoVLA improves long-horizon performance, reaching 0.52 SR on manipulation/navigation and 0.31 on mobile manipulation, exceeding $π_{0.5}$ by +0.08 and +0.11.
☆ A Unified Multi-Dynamics Framework for Perception-Oriented Modeling in Tendon-Driven Continuum Robots
Tendon-driven continuum robots offer intrinsically safe and contact-rich interactions owing to their kinematic redundancy and structural compliance. However, their perception often depends on external sensors, which increase hardware complexity and limit scalability. This work introduces a unified multi-dynamics modeling framework for tendon-driven continuum robotic systems, exemplified by a spiral-inspired robot named Spirob. The framework integrates motor electrical dynamics, motor-winch dynamics, and continuum robot dynamics into a coherent system model. Within this framework, motor signals such as current and angular displacement are modeled to expose the electromechanical signatures of external interactions, enabling perception grounded in intrinsic dynamics. The model captures and validates key physical behaviors of the real system, including actuation hysteresis and self-contact at motion limits. Building on this foundation, the framework is applied to environmental interaction: first for passive contact detection, verified experimentally against simulation data; then for active contact sensing, where control and perception strategies from simulation are successfully applied to the real robot; and finally for object size estimation, where a policy learned in simulation is directly deployed on hardware. The results demonstrate that the proposed framework provides a physically grounded way to interpret interaction signatures from intrinsic motor signals in tendon-driven continuum robots.
☆ Anti-Jamming based on Null-Steering Antennas and Intelligent UAV Swarm Behavior
Unmanned Aerial Vehicle (UAV) swarms represent a key advancement in autonomous systems, enabling coordinated missions through inter-UAV communication. However, their reliance on wireless links makes them vulnerable to jamming, which can disrupt coordination and mission success. This work investigates whether a UAV swarm can effectively overcome jamming while maintaining communication and mission efficiency. To address this, a unified optimization framework combining Genetic Algorithms (GA), Supervised Learning (SL), and Reinforcement Learning (RL) is proposed. The mission model, structured into epochs and timeslots, allows dynamic path planning, antenna orientation, and swarm formation while progressively enforcing collision rules. Null-steering antennas enhance resilience by directing antenna nulls toward interference sources. Results show that the GA achieved stable, collision-free trajectories but with high computational cost. SL models replicated GA-based configurations but struggled to generalize under dynamic or constrained settings. RL, trained via Proximal Policy Optimization (PPO), demonstrated adaptability and real-time decision-making with consistent communication and lower computational demand. Additionally, the Adaptive Movement Model generalized UAV motion to arbitrary directions through a rotation-based mechanism, validating the scalability of the proposed system. Overall, UAV swarms equipped with null-steering antennas and guided by intelligent optimization algorithms effectively mitigate jamming while maintaining communication stability, formation cohesion, and collision safety. The proposed framework establishes a unified, flexible, and reproducible basis for future research on resilient swarm communication systems.
comment: 10 pages
☆ Continually Evolving Skill Knowledge in Vision Language Action Model
Developing general robot intelligence in open environments requires continual skill learning. Recent Vision-Language-Action (VLA) models leverage massive pretraining data to support diverse manipulation tasks, but they still depend heavily on task-specific fine-tuning, revealing a lack of continual learning capability. Existing continual learning methods are also resource-intensive to scale to VLA models. We propose Stellar VLA, a knowledge-driven continual learning framework with two variants: T-Stellar, modeling task-centric knowledge space, and TS-Stellar, capturing hierarchical task-skill structure. Stellar VLA enables self-supervised knowledge evolution through joint learning of task latent representation and the knowledge space, reducing annotation needs. Knowledge-guided expert routing provide task specialization without extra network parameters, lowering training overhead.Experiments on the LIBERO benchmark and real-world tasks show over 50 percentage average improvement in final success rates relative to baselines. TS-Stellar further excels in complex action inference, and in-depth analyses verify effective knowledge retention and discovery. Our code will be released soon.
☆ ActDistill: General Action-Guided Self-Derived Distillation for Efficient Vision-Language-Action Models
Recent Vision-Language-Action (VLA) models have shown impressive flexibility and generalization, yet their deployment in robotic manipulation remains limited by heavy computational overhead and inference latency. In this work, we present ActDistill, a general action-guided self-derived distillation framework that transfers the action prediction capability of any existing VLA model to a lightweight counterpart. Unlike previous efficiency strategies that primarily emphasize vision-language correlations, ActDistill leverages action priors to guide knowledge transfer and model compression, achieving action-oriented efficiency for VLA models. Specifically, we employ a well-trained VLA model as the teacher and introduce a graph-structured encapsulation strategy to explicitly model the hierarchical evolution of action prediction. The student model, derived from the graph-encapsulated teacher, is further equipped with a dynamic router that adaptively selects computation paths based on action prediction demands, guided by hierarchical graph-informed supervision to ensure smooth and efficient evolution. During inference, graph-related auxiliary components are removed, allowing the student to execute only dynamically routed layers and predict high-precision actions with minimal computation and latency. Experiments on embodied benchmarks demonstrate that ActDistill achieves comparable or superior performance to full-scale VLA models while reducing computation by over 50% with up to 1.67 times speedup, thereby establishing a general paradigm toward efficient embodied intelligence.
☆ Unobservable Subspace Evolution and Alignment for Consistent Visual-Inertial Navigation
The inconsistency issue in the Visual-Inertial Navigation System (VINS) is a long-standing and fundamental challenge. While existing studies primarily attribute the inconsistency to observability mismatch, these analyses are often based on simplified theoretical formulations that consider only prediction and SLAM correction. Such formulations fail to cover the non-standard estimation steps, such as MSCKF correction and delayed initialization, which are critical for practical VINS estimators. Furthermore, the lack of a comprehensive understanding of how inconsistency dynamically emerges across estimation steps has hindered the development of precise and efficient solutions. As a result, current approaches often face a trade-off between estimator accuracy, consistency, and implementation complexity. To address these limitations, this paper proposes a novel analysis framework termed Unobservable Subspace Evolution (USE), which systematically characterizes how the unobservable subspace evolves throughout the entire estimation pipeline by explicitly tracking changes in its evaluation points. This perspective sheds new light on how individual estimation steps contribute to inconsistency. Our analysis reveals that observability misalignment induced by certain steps is the antecedent of observability mismatch. Guided by this insight, we propose a simple yet effective solution paradigm, Unobservable Subspace Alignment (USA), which eliminates inconsistency by selectively intervening only in those estimation steps that induce misalignment. We design two USA methods: transformation-based and re-evaluation-based, both offering accurate and computationally lightweight solutions. Extensive simulations and real-world experiments validate the effectiveness of the proposed methods.
comment: 20 pages, 16 figures
☆ RoboArmGS: High-Quality Robotic Arm Splatting via Bézier Curve Refinement
Building high-quality digital assets of robotic arms is crucial yet challenging for the Real2Sim2Real pipeline. Current approaches naively bind static 3D Gaussians according to URDF links, forcing them to follow an URDF-rigged motion passively. However, real-world arm motion is noisy, and the idealized URDF-rigged motion cannot accurately model it, leading to severe rendering artifacts in 3D Gaussians. To address these challenges, we propose RoboArmGS, a novel hybrid representation that refines the URDF-rigged motion with learnable Bézier curves, enabling more accurate real-world motion modeling. To be more specific, we present a learnable Bézier Curve motion refiner that corrects per-joint residuals to address mismatches between real-world motion and URDF-rigged motion. RoboArmGS enables the learning of more accurate real-world motion while achieving a coherent binding of 3D Gaussians across arm parts. To support future research, we contribute a carefully collected dataset named RoboArm4D, which comprises several widely used robotic arms for evaluating the quality of building high-quality digital assets. We evaluate our approach on RoboArm4D, and RoboArmGS achieves state-of-the-art performance in real-world motion modeling and rendering quality. The code and dataset will be released.
☆ Switch-JustDance: Benchmarking Whole Body Motion Tracking Policies Using a Commercial Console Game
Recent advances in whole-body robot control have enabled humanoid and legged robots to perform increasingly agile and coordinated motions. However, standardized benchmarks for evaluating these capabilities in real-world settings, and in direct comparison to humans, remain scarce. Existing evaluations often rely on pre-collected human motion datasets or simulation-based experiments, which limit reproducibility, overlook hardware factors, and hinder fair human-robot comparisons. We present Switch-JustDance, a low-cost and reproducible benchmarking pipeline that leverages motion-sensing console games, Just Dance on the Nintendo Switch, to evaluate robot whole-body control. Using Just Dance on the Nintendo Switch as a representative platform, Switch-JustDance converts in-game choreography into robot-executable motions through streaming, motion reconstruction, and motion retargeting modules and enables users to evaluate controller performance through the game's built-in scoring system. We first validate the evaluation properties of Just Dance, analyzing its reliability, validity, sensitivity, and potential sources of bias. Our results show that the platform provides consistent and interpretable performance measures, making it a suitable tool for benchmarking embodied AI. Building on this foundation, we benchmark three state-of-the-art humanoid whole-body controllers on hardware and provide insights into their relative strengths and limitations.
☆ CUS-GS: A Compact Unified Structured Gaussian Splatting Framework for Multimodal Scene Representation
Recent advances in Gaussian Splatting based 3D scene representation have shown two major trends: semantics-oriented approaches that focus on high-level understanding but lack explicit 3D geometry modeling, and structure-oriented approaches that capture spatial structures yet provide limited semantic abstraction. To bridge this gap, we present CUS-GS, a compact unified structured Gaussian Splatting representation, which connects multimodal semantic features with structured 3D geometry. Specifically, we design a voxelized anchor structure that constructs a spatial scaffold, while extracting multimodal semantic features from a set of foundation models (e.g., CLIP, DINOv2, SEEM). Moreover, we introduce a multimodal latent feature allocation mechanism to unify appearance, geometry, and semantics across heterogeneous feature spaces, ensuring a consistent representation across multiple foundation models. Finally, we propose a feature-aware significance evaluation strategy to dynamically guide anchor growing and pruning, effectively removing redundant or invalid anchors while maintaining semantic integrity. Extensive experiments show that CUS-GS achieves competitive performance compared to state-of-the-art methods using as few as 6M parameters - an order of magnitude smaller than the closest rival at 35M - highlighting the excellent trade off between performance and model efficiency of the proposed framework.
comment: 15 pages, 8 figures, 4 tables
☆ L1 Sample Flow for Efficient Visuomotor Learning
Denoising-based models, such as diffusion and flow matching, have been a critical component of robotic manipulation for their strong distribution-fitting and scaling capacity. Concurrently, several works have demonstrated that simple learning objectives, such as L1 regression, can achieve performance comparable to denoising-based methods on certain tasks, while offering faster convergence and inference. In this paper, we focus on how to combine the advantages of these two paradigms: retaining the ability of denoising models to capture multi-modal distributions and avoid mode collapse while achieving the efficiency of the L1 regression objective. To achieve this vision, we reformulate the original v-prediction flow matching and transform it into sample-prediction with the L1 training objective. We empirically show that the multi-modality can be expressed via a single ODE step. Thus, we propose \textbf{L1 Flow}, a two-step sampling schedule that generates a suboptimal action sequence via a single integration step and then reconstructs the precise action sequence through a single prediction. The proposed method largely retains the advantages of flow matching while reducing the iterative neural function evaluations to merely two and mitigating the potential performance degradation associated with direct sample regression. We evaluate our method with varying baselines and benchmarks, including 8 tasks in MimicGen, 5 tasks in RoboMimic \& PushT Bench, and one task in the real-world scenario. The results show the advantages of the proposed method with regard to training efficiency, inference speed, and overall performance. \href{https://song-wx.github.io/l1flow.github.io/}{Project Website.}
☆ MobileVLA-R1: Reinforcing Vision-Language-Action for Mobile Robots
Grounding natural-language instructions into continuous control for quadruped robots remains a fundamental challenge in vision language action. Existing methods struggle to bridge high-level semantic reasoning and low-level actuation, leading to unstable grounding and weak generalization in the real world. To address these issues, we present MobileVLA-R1, a unified vision-language-action framework that enables explicit reasoning and continuous control for quadruped robots. We construct MobileVLA-CoT, a large-scale dataset of multi-granularity chain-of-thought (CoT) for embodied trajectories, providing structured reasoning supervision for alignment. Built upon this foundation, we introduce a two-stage training paradigm that combines supervised CoT alignment with GRPO reinforcement learning to enhance reasoning consistency, control stability, and long-horizon execution. Extensive evaluations on VLN and VLA tasks demonstrate superior performance over strong baselines, with approximately a 5% improvement. Real-world deployment on a quadruped robot validates robust performance in complex environments. Code: https://github.com/AIGeeksGroup/MobileVLA-R1. Website: https://aigeeksgroup.github.io/MobileVLA-R1.
☆ ArticFlow: Generative Simulation of Articulated Mechanisms
Recent advances in generative models have produced strong results for static 3D shapes, whereas articulated 3D generation remains challenging due to action-dependent deformations and limited datasets. We introduce ArticFlow, a two-stage flow matching framework that learns a controllable velocity field from noise to target point sets under explicit action control. ArticFlow couples (i) a latent flow that transports noise to a shape-prior code and (ii) a point flow that transports points conditioned on the action and the shape prior, enabling a single model to represent diverse articulated categories and generalize across actions. On MuJoCo Menagerie, ArticFlow functions both as a generative model and as a neural simulator: it predicts action-conditioned kinematics from a compact prior and synthesizes novel morphologies via latent interpolation. Compared with object-specific simulators and an action-conditioned variant of static point-cloud generators, ArticFlow achieves higher kinematic accuracy and better shape quality. Results show that action-conditioned flow matching is a practical route to controllable and high-quality articulated mechanism generation.
comment: 8 pages, 8 figures
☆ QuickLAP: Quick Language-Action Preference Learning for Autonomous Driving Agents
Robots must learn from both what people do and what they say, but either modality alone is often incomplete: physical corrections are grounded but ambiguous in intent, while language expresses high-level goals but lacks physical grounding. We introduce QuickLAP: Quick Language-Action Preference learning, a Bayesian framework that fuses physical and language feedback to infer reward functions in real time. Our key insight is to treat language as a probabilistic observation over the user's latent preferences, clarifying which reward features matter and how physical corrections should be interpreted. QuickLAP uses Large Language Models (LLMs) to extract reward feature attention masks and preference shifts from free-form utterances, which it integrates with physical feedback in a closed-form update rule. This enables fast, real-time, and robust reward learning that handles ambiguous feedback. In a semi-autonomous driving simulator, QuickLAP reduces reward learning error by over 70% compared to physical-only and heuristic multimodal baselines. A 15-participant user study further validates our approach: participants found QuickLAP significantly more understandable and collaborative, and preferred its learned behavior over baselines. Code is available at https://github.com/MIT-CLEAR-Lab/QuickLAP.
♻ ☆ Text to Robotic Assembly of Multi Component Objects using 3D Generative AI and Vision Language Models NeurIPS 2025
Advances in 3D generative AI have enabled the creation of physical objects from text prompts, but challenges remain in creating objects involving multiple component types. We present a pipeline that integrates 3D generative AI with vision-language models (VLMs) to enable the robotic assembly of multi-component objects from natural language. Our method leverages VLMs for zero-shot, multi-modal reasoning about geometry and functionality to decompose AI-generated meshes into multi-component 3D models using predefined structural and panel components. We demonstrate that a VLM is capable of determining which mesh regions need panel components in addition to structural components, based on the object's geometry and functionality. Evaluation across test objects shows that users preferred the VLM-generated assignments 90.6% of the time, compared to 59.4% for rule-based and 2.5% for random assignment. Lastly, the system allows users to refine component assignments through conversational feedback, enabling greater human control and agency in making physical objects with generative AI and robotics.
comment: Accepted to NeurIPS 2025, Conference on Neural Information Processing Systems, Creative AI Track
♻ ☆ Vision-Only Gaussian Splatting for Collaborative Semantic Occupancy Prediction AAAI 2026
Collaborative perception enables connected vehicles to share information, overcoming occlusions and extending the limited sensing range inherent in single-agent (non-collaborative) systems. Existing vision-only methods for 3D semantic occupancy prediction commonly rely on dense 3D voxels, which incur high communication costs, or 2D planar features, which require accurate depth estimation or additional supervision, limiting their applicability to collaborative scenarios. To address these challenges, we propose the first approach leveraging sparse 3D semantic Gaussian splatting for collaborative 3D semantic occupancy prediction. By sharing and fusing intermediate Gaussian primitives, our method provides three benefits: a neighborhood-based cross-agent fusion that removes duplicates and suppresses noisy or inconsistent Gaussians; a joint encoding of geometry and semantics in each primitive, which reduces reliance on depth supervision and allows simple rigid alignment; and sparse, object-centric messages that preserve structural information while reducing communication volume. Extensive experiments demonstrate that our approach outperforms single-agent perception and baseline collaborative methods by +8.42 and +3.28 points in mIoU, and +5.11 and +22.41 points in IoU, respectively. When further reducing the number of transmitted Gaussians, our method still achieves a +1.9 improvement in mIoU, using only 34.6% communication volume, highlighting robust performance under limited communication budgets.
comment: Accepted by AAAI 2026 (Oral)
♻ ☆ Unified Generation-Refinement Planning: Bridging Guided Flow Matching and Sampling-Based MPC for Social Navigation
Planning safe and effective robot behavior in dynamic, human-centric environments remains a core challenge due to the need to handle multimodal uncertainty, adapt in real-time, and ensure safety. Optimization-based planners offer explicit constraint handling but performance relies on initialization quality. Learning-based planners better capture multimodal possible solutions but struggle to enforce constraints such as safety. In this paper, we introduce a unified generation-refinement framework bridging learning and optimization with a novel reward-guided conditional flow matching (CFM) model and model predictive path integral (MPPI) control. Our key innovation is in the incorporation of a bidirectional information exchange: samples from a reward-guided CFM model provide informed priors for MPPI refinement, while the optimal trajectory from MPPI warm-starts the next CFM generation. Using autonomous social navigation as a motivating application, we demonstrate that our approach can flexibly adapt to dynamic environments to satisfy safety requirements in real-time.
♻ ☆ RGBSQGrasp: Inferring Local Superquadric Primitives from Single RGB Image for Graspability-Aware Bin Picking IROS2025
Bin picking is a challenging robotic task due to occlusions and physical constraints that limit visual information for object recognition and grasping. Existing approaches often rely on known CAD models or prior object geometries, restricting generalization to novel or unknown objects. Other methods directly regress grasp poses from RGB-D data without object priors, but the inherent noise in depth sensing and the lack of object understanding make grasp synthesis and evaluation more difficult. Superquadrics (SQ) offer a compact, interpretable shape representation that captures the physical and graspability understanding of objects. However, recovering them from limited viewpoints is challenging, as existing methods rely on multiple perspectives for near-complete point cloud reconstruction, limiting their effectiveness in bin-picking. To address these challenges, we propose \textbf{RGBSQGrasp}, a grasping framework that leverages superquadric shape primitives and foundation metric depth estimation models to infer grasp poses from a monocular RGB camera -- eliminating the need for depth sensors. Our framework integrates a universal, cross-platform dataset generation pipeline, a foundation model-based object point cloud estimation module, a global-local superquadric fitting network, and an SQ-guided grasp pose sampling module. By integrating these components, RGBSQGrasp reliably infers grasp poses through geometric reasoning, enhancing grasp stability and adaptability to unseen objects. Real-world robotic experiments demonstrate a 92% grasp success rate, highlighting the effectiveness of RGBSQGrasp in packed bin-picking environments.
comment: 8 pages, 6 figures, IROS2025 RGMCW Best Workshop Paper
♻ ☆ ResAD: Normalized Residual Trajectory Modeling for End-to-End Autonomous Driving
End-to-end autonomous driving (E2EAD) systems, which learn to predict future trajectories directly from sensor data, are fundamentally challenged by the inherent spatio-temporal imbalance of trajectory data. This imbalance creates a significant optimization burden, causing models to learn spurious correlations instead of robust driving logic, while also prioritizing uncertain, distant predictions, thereby compromising immediate safety. To address these issues, we propose ResAD, a novel Normalized Residual Trajectory Modeling framework. Instead of predicting the future trajectory directly, our approach reframes and simplifies the learning task by predicting the residual deviation from a deterministic inertial reference. This inertial reference serves as a strong physical prior, compelling the model to move beyond simple pattern-matching and instead focus its capacity on learning the necessary, context-driven deviations (e.g., traffic rules, obstacles) from this default, inertially-guided path. To mitigate the optimization imbalance caused by uncertain, long-term horizons, ResAD further incorporates Point-wise Normalization of the predicted residual. This technique re-weights the optimization objective, preventing large-magnitude errors associated with distant, uncertain waypoints from dominating the learning signal. On the NAVSIM v1 and v2 benchmarks, ResAD achieves state-of-the-art results of 88.8 PDMS and 85.5 EPDMS with only two denoising steps, demonstrating that ResAD significantly simplifies the learning task and improves planning performance. The code will be released to facilitate further research.
♻ ☆ Multi-Timescale Hierarchical Reinforcement Learning for Unified Behavior and Control of Autonomous Driving
Reinforcement Learning (RL) is increasingly used in autonomous driving (AD) and shows clear advantages. However, most RL-based AD methods overlook policy structure design. An RL policy that only outputs short-timescale vehicle control commands results in fluctuating driving behavior due to fluctuations in network outputs, while one that only outputs long-timescale driving goals cannot achieve unified optimality of driving behavior and control. Therefore, we propose a multi-timescale hierarchical reinforcement learning approach. Our approach adopts a hierarchical policy structure, where high- and low-level RL policies are unified-trained to produce long-timescale motion guidance and short-timescale control commands, respectively. Therein, motion guidance is explicitly represented by hybrid actions to capture multimodal driving behaviors on structured road and support incremental low-level extend-state updates. Additionally, a hierarchical safety mechanism is designed to ensure multi-timescale safety. Evaluation in simulator-based and HighD dataset-based highway multi-lane scenarios demonstrates that our approach significantly improves AD performance, effectively increasing driving efficiency, action consistency and safety.
comment: 8 pages, accepted for publication in IEEE Robotics and Automation Letters (RAL)
♻ ☆ Occlusion-Aware Contingency Safety-Critical Planning for Autonomous Driving
Ensuring safe driving while maintaining travel efficiency for autonomous vehicles in dynamic and occluded environments is a critical challenge. This paper proposes an occlusion-aware contingency safety-critical planning approach for real-time autonomous driving. Leveraging reachability analysis for risk assessment, forward reachable sets of phantom vehicles are used to derive risk-aware dynamic velocity boundaries. These velocity boundaries are incorporated into a biconvex nonlinear programming (NLP) formulation that formally enforces safety using spatiotemporal barrier constraints, while simultaneously optimizing exploration and fallback trajectories within a receding horizon planning framework. To enable real-time computation and coordination between trajectories, we employ the consensus alternating direction method of multipliers (ADMM) to decompose the biconvex NLP problem into low-dimensional convex subproblems. The effectiveness of the proposed approach is validated through simulations and real-world experiments in occluded intersections. Experimental results demonstrate enhanced safety and improved travel efficiency, enabling real-time safe trajectory generation in dynamic occluded intersections under varying obstacle conditions. The project page is available at https://zack4417.github.io/oacp-website/.
comment: 14 pages, 9 figures
♻ ☆ DragonFly: Single mmWave Radar 3D Localization of Highly Dynamic Tags in GPS-Denied Environments
The accurate localization and tracking of dynamic targets, such as equipment, people, vehicles, drones, robots, and the assets that they interact with in GPS-denied indoor environments is critical to enabling safe and efficient operations in the next generation of spatially aware industrial facilities. This paper presents DragonFly , a 3D localization system of highly dynamic backscatter tags using a single MIMO mmWave radar. The system delivers the first demonstration of a mmWave backscatter system capable of exploiting the capabilities of MIMO radars for the 3D localization of mmID tags moving at high speeds and accelerations at long ranges by introducing a critical Doppler disambiguation algorithm and a fully integrated cross-polarized dielectric lens-based mmID tag consuming a mere 68 uW. DragonFly was extensively evaluated in static and dynamic configurations, including on a flying quadcopter, and benchmarked against multiple baselines, demonstrating its ability to track the positions of multiple tags with a median 3D accuracy of 12 cm at speeds and acceleration on the order of 10 m/s and 4 m/s^2 and at ranges of up to 50m.
comment: Accepted in ACM Mobicom'25
Robotics 50
☆ RynnVLA-002: A Unified Vision-Language-Action and World Model
We introduce RynnVLA-002, a unified Vision-Language-Action (VLA) and world model. The world model leverages action and visual inputs to predict future image states, learning the underlying physics of the environment to refine action generation. Conversely, the VLA model produces subsequent actions from image observations, enhancing visual understanding and supporting the world model's image generation. The unified framework of RynnVLA-002 enables joint learning of environmental dynamics and action planning. Our experiments show that RynnVLA-002 surpasses individual VLA and world models, demonstrating their mutual enhancement. We evaluate RynnVLA-002 in both simulation and real-world robot tasks. RynnVLA-002 achieves 97.4% success rate on the LIBERO simulation benchmark without pretraining, while in real-world LeRobot experiments, its integrated world model boosts the overall success rate by 50%.
☆ HALO: High-Altitude Language-Conditioned Monocular Aerial Exploration and Navigation
We demonstrate real-time high-altitude aerial metric-semantic mapping and exploration using a monocular camera paired with a global positioning system (GPS) and an inertial measurement unit (IMU). Our system, named HALO, addresses two key challenges: (i) real-time dense 3D reconstruction using vision at large distances, and (ii) mapping and exploration of large-scale outdoor environments with accurate scene geometry and semantics. We demonstrate that HALO can plan informative paths that exploit this information to complete missions with multiple tasks specified in natural language. In simulation-based evaluation across large-scale environments of size up to 78,000 sq. m., HALO consistently completes tasks with less exploration time and achieves up to 68% higher competitive ratio in terms of the distance traveled compared to the state-of-the-art semantic exploration baseline. We use real-world experiments on a custom quadrotor platform to demonstrate that (i) all modules can run onboard the robot, and that (ii) in diverse environments HALO can support effective autonomous execution of missions covering up to 24,600 sq. m. area at an altitude of 40 m. Experiment videos and more details can be found on our project page: https://tyuezhan.github.io/halo/.
☆ MDG: Masked Denoising Generation for Multi-Agent Behavior Modeling in Traffic Environments
Modeling realistic and interactive multi-agent behavior is critical to autonomous driving and traffic simulation. However, existing diffusion and autoregressive approaches are limited by iterative sampling, sequential decoding, or task-specific designs, which hinder efficiency and reuse. We propose Masked Denoising Generation (MDG), a unified generative framework that reformulates multi-agent behavior modeling as the reconstruction of independently noised spatiotemporal tensors. Instead of relying on diffusion time steps or discrete tokenization, MDG applies continuous, per-agent and per-timestep noise masks that enable localized denoising and controllable trajectory generation in a single or few forward passes. This mask-driven formulation generalizes across open-loop prediction, closed-loop simulation, motion planning, and conditional generation within one model. Trained on large-scale real-world driving datasets, MDG achieves competitive closed-loop performance on the Waymo Sim Agents and nuPlan Planning benchmarks, while providing efficient, consistent, and controllable open-loop multi-agent trajectory generation. These results position MDG as a simple yet versatile paradigm for multi-agent behavior modeling.
☆ RoboCOIN: An Open-Sourced Bimanual Robotic Data COllection for INtegrated Manipulation
Bimanual manipulation is essential for achieving human-like dexterity in robots, but the large-scale and diverse bimanual robot datasets remain scarce due to hardware heterogeneity across robotic platforms. To address the challenge, we present RoboCOIN, a comprehensive multi-embodiment bimanual manipulation dataset with over 180,000 demonstrations collected from 15 distinct robotic platforms. The dataset covers 16 scenarios, including residential, commercial, and working environments, with 421 tasks systematically organized by bimanual coordination patterns and object properties. Our key innovation is a hierarchical capability pyramid that provides multi-level annotations, spanning trajectory-level concepts, segment-level subtasks, and frame-level kinematics. We further develop CoRobot, a comprehensive processing framework featuring Robot Trajectory Markup Language (RTML) for quality assessment, automated annotation generation, and unified multi-embodiment management. Extensive experiments demonstrate the reliability and effectiveness of RoboCOIN in multi-embodiment bimanual learning, with significant performance improvements across various model architectures and robotic platforms. The complete dataset and framework are open-sourced and publicly available for further research purposes. Project website: https://FlagOpen.github.io/RoboCOIN/.
☆ SPEAR-1: Scaling Beyond Robot Demonstrations via 3D Understanding
Robotic Foundation Models (RFMs) hold great promise as generalist, end-to-end systems for robot control. Yet their ability to generalize across new environments, tasks, and embodiments remains limited. We argue that a major bottleneck lies in their foundations: most RFMs are built by fine-tuning internet-pretrained Vision-Language Models (VLMs). However, these VLMs are trained on 2D image-language tasks and lack the 3D spatial reasoning inherently required for embodied control in the 3D world. Bridging this gap directly with large-scale robotic data is costly and difficult to scale. Instead, we propose to enrich easy-to-collect non-robotic image data with 3D annotations and enhance a pretrained VLM with 3D understanding capabilities. Following this strategy, we train SPEAR-VLM, a 3D-aware VLM that infers object coordinates in 3D space from a single 2D image. Building on SPEAR-VLM, we introduce our main contribution, $~\textbf{SPEAR-1}$: a robotic foundation model that integrates grounded 3D perception with language-instructed embodied control. Trained on $\sim$45M frames from 24 Open X-Embodiment datasets, SPEAR-1 outperforms or matches state-of-the-art models such as $π_0$-FAST and $π_{0.5}$, while it uses 20$\times$ fewer robot demonstrations. This carefully-engineered training strategy unlocks new VLM capabilities and as a consequence boosts the reliability of embodied control beyond what is achievable with only robotic data. We make our model weights and 3D-annotated datasets publicly available.
☆ Feasibility of Embodied Dynamics Based Bayesian Learning for Continuous Pursuit Motion Control of Assistive Mobile Robots in the Built Environment
Non-invasive electroencephalography (EEG)-based brain-computer interfaces (BCIs) offer an intuitive means for individuals with severe motor impairments to independently operate assistive robotic wheelchairs and navigate built environments. Despite considerable progress in BCI research, most current motion control systems are limited to discrete commands, rather than supporting continuous pursuit, where users can freely adjust speed and direction in real time. Such natural mobility control is, however, essential for wheelchair users to navigate complex public spaces, such as transit stations, airports, hospitals, and indoor corridors, to interact socially with the dynamic populations with agility, and to move flexibly and comfortably as autonomous driving is refined to allow movement at will. In this study, we address the gap of continuous pursuit motion control in BCIs by proposing and validating a brain-inspired Bayesian inference framework, where embodied dynamics in acceleration-based motor representations are decoded. This approach contrasts with conventional kinematics-level decoding and deep learning-based methods. Using a public dataset with sixteen hours of EEG from four subjects performing motor imagery-based target-following, we demonstrate that our method, utilizing Automatic Relevance Determination for feature selection and continual online learning, reduces the normalized mean squared error between predicted and true velocities by 72% compared to autoregressive and EEGNet-based methods in a session-accumulative transfer learning setting. Theoretically, these findings empirically support embodied cognition theory and reveal the brain's intrinsic motor control dynamics in an embodied and predictive nature. Practically, grounding EEG decoding in the same dynamical principles that govern biological motion offers a promising path toward more stable and intuitive BCI control.
comment: 37 pages, 9 figures, and 7 tables
☆ Human Imitated Bipedal Locomotion with Frequency Based Gait Generator Network
Learning human-like, robust bipedal walking remains difficult due to hybrid dynamics and terrain variability. We propose a lightweight framework that combines a gait generator network learned from human motion with Proximal Policy Optimization (PPO) controller for torque control. Despite being trained only on flat or mildly sloped ground, the learned policies generalize to steeper ramps and rough surfaces. Results suggest that pairing spectral motion priors with Deep Reinforcement Learning (DRL) offers a practical path toward natural and robust bipedal locomotion with modest training cost.
☆ IndustryNav: Exploring Spatial Reasoning of Embodied Agents in Dynamic Industrial Navigation
While Visual Large Language Models (VLLMs) show great promise as embodied agents, they continue to face substantial challenges in spatial reasoning. Existing embodied benchmarks largely focus on passive, static household environments and evaluate only isolated capabilities, failing to capture holistic performance in dynamic, real-world complexity. To fill this gap, we present IndustryNav, the first dynamic industrial navigation benchmark for active spatial reasoning. IndustryNav leverages 12 manually created, high-fidelity Unity warehouse scenarios featuring dynamic objects and human movement. Our evaluation employs a PointGoal navigation pipeline that effectively combines egocentric vision with global odometry to assess holistic local-global planning. Crucially, we introduce the "collision rate" and "warning rate" metrics to measure safety-oriented behaviors and distance estimation. A comprehensive study of nine state-of-the-art VLLMs (including models such as GPT-5-mini, Claude-4.5, and Gemini-2.5) reveals that closed-source models maintain a consistent advantage; however, all agents exhibit notable deficiencies in robust path planning, collision avoidance and active exploration. This highlights a critical need for embodied research to move beyond passive perception and toward tasks that demand stable planning, active exploration, and safe behavior in dynamic, real-world environment.
☆ Vector Cost Behavioral Planning for Autonomous Robotic Systems with Contemporary Validation Strategies
The vector cost bimatrix game is a method for multi-objective decision making that enables autonomous robotic systems to optimize for multiple goals at once while avoiding worst-case scenarios in neglected objectives. We expand this approach to arbitrary numbers of objectives and compare its performance to scalar weighted sum methods during competitive motion planning. Explainable Artificial Intelligence (XAI) software is used to aid in the analysis of high dimensional decision-making data. State-space Exploration of Multidimensional Boundaries using Adherence Strategies (SEMBAS) is applied to explore performance modes in the parameter space as a sensitivity study for the baseline and proposed frameworks. While some works have explored aspects of game theoretic planning and intelligent systems validation separately, we combine each of these into a novel and comprehensive simulation pipeline. This integration demonstrates a dramatic improvement of the vector cost method over scalarization and offers an interpretable and generalizable framework for robotic behavioral planning. Code available at https://github.com/toazbenj/race_simulation. The video companion to this work is available at https://tinyurl.com/vectorcostvideo.
comment: Technical report associated with submission to Journal of Intelligent & Robotic Systems, currently under review
☆ Agility Meets Stability: Versatile Humanoid Control with Heterogeneous Data
Humanoid robots are envisioned to perform a wide range of tasks in human-centered environments, requiring controllers that combine agility with robust balance. Recent advances in locomotion and whole-body tracking have enabled impressive progress in either agile dynamic skills or stability-critical behaviors, but existing methods remain specialized, focusing on one capability while compromising the other. In this work, we introduce AMS (Agility Meets Stability), the first framework that unifies both dynamic motion tracking and extreme balance maintenance in a single policy. Our key insight is to leverage heterogeneous data sources: human motion capture datasets that provide rich, agile behaviors, and physically constrained synthetic balance motions that capture stability configurations. To reconcile the divergent optimization goals of agility and stability, we design a hybrid reward scheme that applies general tracking objectives across all data while injecting balance-specific priors only into synthetic motions. Further, an adaptive learning strategy with performance-driven sampling and motion-specific reward shaping enables efficient training across diverse motion distributions. We validate AMS extensively in simulation and on a real Unitree G1 humanoid. Experiments demonstrate that a single policy can execute agile skills such as dancing and running, while also performing zero-shot extreme balance motions like Ip Man's Squat, highlighting AMS as a versatile control paradigm for future humanoid applications.
☆ METIS: Multi-Source Egocentric Training for Integrated Dexterous Vision-Language-Action Model
Building a generalist robot that can perceive, reason, and act across diverse tasks remains an open challenge, especially for dexterous manipulation. A major bottleneck lies in the scarcity of large-scale, action-annotated data for dexterous skills, as teleoperation is difficult and costly. Human data, with its vast scale and diverse manipulation behaviors, provides rich priors for learning robotic actions. While prior works have explored leveraging human demonstrations, they are often constrained by limited scenarios and a large visual gap between human and robots. To eliminate these limitations, we propose METIS, a vision-language-action (VLA) model for dexterous manipulation pretrained on multi-source egocentric datasets. We first construct EgoAtlas, which integrates large-scale human and robotic data from multiple sources, all unified under a consistent action space. We further extract motion-aware dynamics, a compact and discretized motion representation, which provides efficient and expressive supervision for VLA training. Built upon them, METIS integrates reasoning and acting into a unified framework, enabling effective deployment to downstream dexterous manipulation tasks. Our method demonstrates exceptional dexterous manipulation capabilities, achieving highest average success rate in six real-world tasks. Experimental results also highlight the superior generalization and robustness to out-of-distribution scenarios. These findings emphasize METIS as a promising step toward a generalist model for dexterous manipulation.
☆ Robot Confirmation Generation and Action Planning Using Long-context Q-Former Integrated with Multimodal LLM
Human-robot collaboration towards a shared goal requires robots to understand human action and interaction with the surrounding environment. This paper focuses on human-robot interaction (HRI) based on human-robot dialogue that relies on the robot action confirmation and action step generation using multimodal scene understanding. The state-of-the-art approach uses multimodal transformers to generate robot action steps aligned with robot action confirmation from a single clip showing a task composed of multiple micro steps. Although actions towards a long-horizon task depend on each other throughout an entire video, the current approaches mainly focus on clip-level processing and do not leverage long-context information. This paper proposes a long-context Q-former incorporating left and right context dependency in full videos. Furthermore, this paper proposes a text-conditioning approach to feed text embeddings directly into the LLM decoder to mitigate the high abstraction of the information in text by Q-former. Experiments with the YouCook2 corpus show that the accuracy of confirmation generation is a major factor in the performance of action planning. Furthermore, we demonstrate that the long-context Q-former improves the confirmation and action planning by integrating VideoLLaMA3.
comment: Accepted to ASRU 2025
☆ FORWARD: Dataset of a forwarder operating in rough terrain
We present FORWARD, a high-resolution multimodal dataset of a cut-to-length forwarder operating in rough terrain on two harvest sites in the middle part of Sweden. The forwarder is a large Komatsu model equipped with a variety of sensors, including RTK-GNSS, 360-camera, operator vibration sensors, internal CAN-bus signal recording, and multiple IMUs. The data includes event time logs recorded in 5 Hz with e.g., driving speed, fuel consumption, vehicle position with centimeter accuracy, and crane use while the vehicle operates in forest areas laser-scanned with very high-resolution, $\sim$1500 points per square meter. Production log files (StanForD standard) with time-stamped machine events, extensive video material, and terrain data in various formats are included as well. About 18 hours of regular wood extraction work during three days is annotated from 360-video material into individual work elements and included in the dataset. We also include scenario specifications of conducted experiments on forest roads and in terrain. Scenarios include repeatedly driving the same routes with and without steel tracks, different load weight, and different target driving speeds. The dataset is intended for developing models and algorithms for trafficability, perception, and autonomous control of forest machines using artificial intelligence, simulation, and experiments on physical testbeds. In part, we focus on forwarders traversing terrain, avoiding obstacles, and loading or unloading logs, with consideration for efficiency, fuel consumption, safety, and environmental impact. Other benefits of the open dataset include the ability to explore auto-generation and calibration of forestry machine simulators and automation scenario descriptions using the data recorded in the field.
comment: 25 pages, 22 figures
☆ MonoSpheres: Large-Scale Monocular SLAM-Based UAV Exploration through Perception-Coupled Mapping and Planning
Autonomous exploration of unknown environments is a key capability for mobile robots, but it is largely unsolved for robots equipped with only a single monocular camera and no dense range sensors. In this paper, we present a novel approach to monocular vision-based exploration that can safely cover large-scale unstructured indoor and outdoor 3D environments by explicitly accounting for the properties of a sparse monocular SLAM frontend in both mapping and planning. The mapping module solves the problems of sparse depth data, free-space gaps, and large depth uncertainty by oversampling free space in texture-sparse areas and keeping track of obstacle position uncertainty. The planning module handles the added free-space uncertainty through rapid replanning and perception-aware heading control. We further show that frontier-based exploration is possible with sparse monocular depth data when parallax requirements and the possibility of textureless surfaces are taken into account. We evaluate our approach extensively in diverse real-world and simulated environments, including ablation studies. To the best of the authors' knowledge, the proposed method is the first to achieve 3D monocular exploration in real-world unstructured outdoor environments. We open-source our implementation to support future research.
comment: 8 pages, 9 figures, submitted to RA-L
☆ Leveraging CVAE for Joint Configuration Estimation of Multifingered Grippers from Point Cloud Data
This paper presents an efficient approach for determining the joint configuration of a multifingered gripper solely from the point cloud data of its poly-articulated chain, as generated by visual sensors, simulations or even generative neural networks. Well-known inverse kinematics (IK) techniques can provide mathematically exact solutions (when they exist) for joint configuration determination based solely on the fingertip pose, but often require post-hoc decision-making by considering the positions of all intermediate phalanges in the gripper's fingers, or rely on algorithms to numerically approximate solutions for more complex kinematics. In contrast, our method leverages machine learning to implicitly overcome these challenges. This is achieved through a Conditional Variational Auto-Encoder (CVAE), which takes point cloud data of key structural elements as input and reconstructs the corresponding joint configurations. We validate our approach on the MultiDex grasping dataset using the Allegro Hand, operating within 0.05 milliseconds and achieving accuracy comparable to state-of-the-art methods. This highlights the effectiveness of our pipeline for joint configuration estimation within the broader context of AI-driven techniques for grasp planning.
☆ Simulation of Active Soft Nets for Capture of Space Debris
In this work, we propose a simulator, based on the open-source physics engine MuJoCo, for the design and control of soft robotic nets for the autonomous removal of space debris. The proposed simulator includes net dynamics, contact between the net and the debris, self-contact of the net, orbital mechanics, and a controller that can actuate thrusters on the four satellites at the corners of the net. It showcases the case of capturing Envisat, a large ESA satellite that remains in orbit as space debris following the end of its mission. This work investigates different mechanical models, which can be used to simulate the net dynamics, simulating various degrees of compliance, and different control strategies to achieve the capture of the debris, depending on the relative position of the net and the target. Unlike previous works on this topic, we do not assume that the net has been previously ballistically thrown toward the target, and we start from a relatively static configuration. The results show that a more compliant net achieves higher performance when attempting the capture of Envisat. Moreover, when paired with a sliding mode controller, soft nets are able to achieve successful capture in 100% of the tested cases, whilst also showcasing a higher effective area at contact and a higher number of contact points between net and Envisat.
☆ A ROS2 Interface for Universal Robots Collaborative Manipulators Based on ur_rtde
In this paper a novel ROS2 driver for UR robot manipulators is presented, based on the ur_rtde C++ library. The proposed driver aims to be a flexible solution, adaptable to a wide range of applications. The driver exposes the high-level commands of Universal Robots URScripts, and custom commands can be added using a plugin system. Several commands have been implemented, including motion execution along a waypoint-based path. The driver is published as open source.
☆ TP-MDDN: Task-Preferenced Multi-Demand-Driven Navigation with Autonomous Decision-Making NeurIPS 2025
In daily life, people often move through spaces to find objects that meet their needs, posing a key challenge in embodied AI. Traditional Demand-Driven Navigation (DDN) handles one need at a time but does not reflect the complexity of real-world tasks involving multiple needs and personal choices. To bridge this gap, we introduce Task-Preferenced Multi-Demand-Driven Navigation (TP-MDDN), a new benchmark for long-horizon navigation involving multiple sub-demands with explicit task preferences. To solve TP-MDDN, we propose AWMSystem, an autonomous decision-making system composed of three key modules: BreakLLM (instruction decomposition), LocateLLM (goal selection), and StatusMLLM (task monitoring). For spatial memory, we design MASMap, which combines 3D point cloud accumulation with 2D semantic mapping for accurate and efficient environmental understanding. Our Dual-Tempo action generation framework integrates zero-shot planning with policy-based fine control, and is further supported by an Adaptive Error Corrector that handles failure cases in real time. Experiments demonstrate that our approach outperforms state-of-the-art baselines in both perception accuracy and navigation robustness.
comment: Accepted at NeurIPS 2025
☆ QueryOcc: Query-based Self-Supervision for 3D Semantic Occupancy
Learning 3D scene geometry and semantics from images is a core challenge in computer vision and a key capability for autonomous driving. Since large-scale 3D annotation is prohibitively expensive, recent work explores self-supervised learning directly from sensor data without manual labels. Existing approaches either rely on 2D rendering consistency, where 3D structure emerges only implicitly, or on discretized voxel grids from accumulated lidar point clouds, limiting spatial precision and scalability. We introduce QueryOcc, a query-based self-supervised framework that learns continuous 3D semantic occupancy directly through independent 4D spatio-temporal queries sampled across adjacent frames. The framework supports supervision from either pseudo-point clouds derived from vision foundation models or raw lidar data. To enable long-range supervision and reasoning under constant memory, we introduce a contractive scene representation that preserves near-field detail while smoothly compressing distant regions. QueryOcc surpasses previous camera-based methods by 26% in semantic RayIoU on the self-supervised Occ3D-nuScenes benchmark while running at 11.6 FPS, demonstrating that direct 4D query supervision enables strong self-supervised occupancy learning. https://research.zenseact.com/publications/queryocc/
☆ SING3R-SLAM: Submap-based Indoor Monocular Gaussian SLAM with 3D Reconstruction Priors
Recent advances in dense 3D reconstruction enable the accurate capture of local geometry; however, integrating them into SLAM is challenging due to drift and redundant point maps, which limit efficiency and downstream tasks, such as novel view synthesis. To address these issues, we propose SING3R-SLAM, a globally consistent and compact Gaussian-based dense RGB SLAM framework. The key idea is to combine locally consistent 3D reconstructions with a unified global Gaussian representation that jointly refines scene geometry and camera poses, enabling efficient and versatile 3D mapping for multiple downstream applications. SING3R-SLAM first builds locally consistent submaps through our lightweight tracking and reconstruction module, and then progressively aligns and fuses them into a global Gaussian map that enforces cross-view geometric consistency. This global map, in turn, provides feedback to correct local drift and enhance the robustness of tracking. Extensive experiments demonstrate that SING3R-SLAM achieves state-of-the-art tracking, 3D reconstruction, and novel view rendering, resulting in over 12% improvement in tracking and producing finer, more detailed geometry, all while maintaining a compact and memory-efficient global representation on real-world datasets.
☆ Efficient Robot Design with Multi-Objective Black-Box Optimization and Large Language Models
Various methods for robot design optimization have been developed so far. These methods are diverse, ranging from numerical optimization to black-box optimization. While numerical optimization is fast, it is not suitable for cases involving complex structures or discrete values, leading to frequent use of black-box optimization instead. However, black-box optimization suffers from low sampling efficiency and takes considerable sampling iterations to obtain good solutions. In this study, we propose a method to enhance the efficiency of robot body design based on black-box optimization by utilizing large language models (LLMs). In parallel with the sampling process based on black-box optimization, sampling is performed using LLMs, which are provided with problem settings and extensive feedback. We demonstrate that this method enables more efficient exploration of design solutions and discuss its characteristics and limitations.
☆ Reflection-Based Relative Localization for Cooperative UAV Teams Using Active Markers
Reflections of active markers in the environment are a common source of ambiguity in onboard visual relative localization. This work presents a novel approach for onboard relative localization in multi-robot teams that exploits these typically unwanted reflections of active markers in the environment. It operates without prior knowledge of robot size or predefined marker configurations and remains independent of surface properties, an essential feature for heterogeneous micro-aerial swarms cooperating in unknown environments. It explicitly accounts for uncertainties caused by non-flat surfaces, with a particular focus on dynamic water surfaces, which are especially relevant for marine deployments. We validated the approach in both indoor and outdoor experiments, demonstrating that the proposed reflection-based localization system operates reliably without prior knowledge of team member size and achieves greater effective range (above 30 m) and accuracy than state-of-the-art methods. The video and source code of this work will be made publicly available after publication.
☆ Progress-Think: Semantic Progress Reasoning for Vision-Language Navigation
Vision-Language Navigation requires agents to act coherently over long horizons by understanding not only local visual context but also how far they have advanced within a multi-step instruction. However, recent Vision-Language-Action models focus on direct action prediction and earlier progress methods predict numeric achievements; both overlook the monotonic co-progression property of the observation and instruction sequences. Building on this insight, Progress-Think introduces semantic progress reasoning, predicting instruction-style progress from visual observations to enable more accurate navigation. To achieve this without expensive annotations, we propose a three-stage framework. In the initial stage, Self-Aligned Progress Pretraining bootstraps a reasoning module via a novel differentiable alignment between visual history and instruction prefixes. Then, Progress-Guided Policy Pretraining injects learned progress states into the navigation context, guiding the policy toward consistent actions. Finally, Progress-Policy Co-Finetuning jointly optimizes both modules with tailored progress-aware reinforcement objectives. Experiments on R2R-CE and RxR-CE show state-of-the-art success and efficiency, demonstrating that semantic progress yields a more consistent representation of navigation advancement.
☆ H-GAR: A Hierarchical Interaction Framework via Goal-Driven Observation-Action Refinement for Robotic Manipulation AAAI 2026
Unified video and action prediction models hold great potential for robotic manipulation, as future observations offer contextual cues for planning, while actions reveal how interactions shape the environment. However, most existing approaches treat observation and action generation in a monolithic and goal-agnostic manner, often leading to semantically misaligned predictions and incoherent behaviors. To this end, we propose H-GAR, a Hierarchical interaction framework via Goal-driven observation-Action Refinement.To anchor prediction to the task objective, H-GAR first produces a goal observation and a coarse action sketch that outline a high-level route toward the goal. To enable explicit interaction between observation and action under the guidance of the goal observation for more coherent decision-making, we devise two synergistic modules. (1) Goal-Conditioned Observation Synthesizer (GOS) synthesizes intermediate observations based on the coarse-grained actions and the predicted goal observation. (2) Interaction-Aware Action Refiner (IAAR) refines coarse actions into fine-grained, goal-consistent actions by leveraging feedback from the intermediate observations and a Historical Action Memory Bank that encodes prior actions to ensure temporal consistency. By integrating goal grounding with explicit action-observation interaction in a coarse-to-fine manner, H-GAR enables more accurate manipulation. Extensive experiments on both simulation and real-world robotic manipulation tasks demonstrate that H-GAR achieves state-of-the-art performance.
comment: Accepted to AAAI 2026 (Oral), Project Page: https://github.com/JiuTian-VL/H-GAR
☆ A segment anchoring-based balancing algorithm for agricultural multi-robot task allocation with energy constraints
Multi-robot systems have emerged as a key technology for addressing the efficiency and cost challenges in labor-intensive industries. In the representative scenario of smart farming, planning efficient harvesting schedules for a fleet of electric robots presents a highly challenging frontier problem. The complexity arises not only from the need to find Pareto-optimal solutions for the conflicting objectives of makespan and transportation cost, but also from the necessity to simultaneously manage payload constraints and finite battery capacity. When robot loads are dynamically updated during planned multi-trip operations, a mandatory recharge triggered by energy constraints introduces an unscheduled load reset. This interaction creates a complex cascading effect that disrupts the entire schedule and renders traditional optimization methods ineffective. To address this challenge, this paper proposes the segment anchoring-based balancing algorithm (SABA). The core of SABA lies in the organic combination of two synergistic mechanisms: the sequential anchoring and balancing mechanism, which leverages charging decisions as `anchors' to systematically reconstruct disrupted routes, while the proportional splitting-based rebalancing mechanism is responsible for the fine-grained balancing and tuning of the final solutions' makespans. Extensive comparative experiments, conducted on a real-world case study and a suite of benchmark instances, demonstrate that SABA comprehensively outperforms 6 state-of-the-art algorithms in terms of both solution convergence and diversity. This research provides a novel theoretical perspective and an effective solution for the multi-robot task allocation problem under energy constraints.
☆ MfNeuPAN: Proactive End-to-End Navigation in Dynamic Environments via Direct Multi-Frame Point Constraints
Obstacle avoidance in complex and dynamic environments is a critical challenge for real-time robot navigation. Model-based and learning-based methods often fail in highly dynamic scenarios because traditional methods assume a static environment and cannot adapt to real-time changes, while learning-based methods rely on single-frame observations for motion constraint estimation, limiting their adaptability. To overcome these limitations, this paper proposes a novel framework that leverages multi-frame point constraints, including current and future frames predicted by a dedicated module, to enable proactive end-to-end navigation. By incorporating a prediction module that forecasts the future path of moving obstacles based on multi-frame observations, our method allows the robot to proactively anticipate and avoid potential dangers. This proactive planning capability significantly enhances navigation robustness and efficiency in unknown dynamic environments. Simulations and real-world experiments validate the effectiveness of our approach.
comment: 6 pages, 9 figures, accepted at IEEE ROBIO 2025
☆ Stable Offline Hand-Eye Calibration for any Robot with Just One Mark
Imitation learning has achieved remarkable success in a variety of robotic tasks by learning a mapping function from camera-space observations to robot-space actions. Recent work indicates that the use of robot-to-camera transformation information ({\ie}, camera extrinsics) benefits the learning process and produces better results. However, camera extrinsics are oftentimes unavailable and estimation methods usually suffer from local minima and poor generalizations. In this paper, we present CalibAll, a simple yet effective method that \textbf{requires only a single mark} and performs training-free, stable, and accurate camera extrinsic estimation across diverse robots and datasets through a coarse-to-fine calibration pipeline. In particular, we annotate a single mark on an end-effector (EEF), and leverage the correspondence ability emerged from vision foundation models (VFM) to automatically localize the corresponding mark across robots in diverse datasets. Using this mark, together with point tracking and the 3D EEF trajectory, we obtain a coarse camera extrinsic via temporal Perspective-n-Point (PnP). This estimate is further refined through a rendering-based optimization that aligns rendered and ground-true masks, yielding accurate and stable camera extrinsic. Experimental results demonstrate that our method outperforms state-of-the-art approaches, showing strong robustness and general effectiveness across three robot platforms. It also produces useful auxiliary annotations such as depth maps, link-wise masks, and end-effector 2D trajectories, which can further support downstream tasks.
☆ MobileOcc: A Human-Aware Semantic Occupancy Dataset for Mobile Robots
Dense 3D semantic occupancy perception is critical for mobile robots operating in pedestrian-rich environments, yet it remains underexplored compared to its application in autonomous driving. To address this gap, we present MobileOcc, a semantic occupancy dataset for mobile robots operating in crowded human environments. Our dataset is built using an annotation pipeline that incorporates static object occupancy annotations and a novel mesh optimization framework explicitly designed for human occupancy modeling. It reconstructs deformable human geometry from 2D images and subsequently refines and optimizes it using associated LiDAR point data. Using MobileOcc, we establish benchmarks for two tasks, i) Occupancy prediction and ii) Pedestrian velocity prediction, using different methods including monocular, stereo, and panoptic occupancy, with metrics and baseline implementations for reproducible comparison. Beyond occupancy prediction, we further assess our annotation method on 3D human pose estimation datasets. Results demonstrate that our method exhibits robust performance across different datasets.
☆ Multi-UAV Swarm Obstacle Avoidance Based on Potential Field Optimization
In multi UAV scenarios,the traditional Artificial Potential Field (APF) method often leads to redundant flight paths and frequent abrupt heading changes due to unreasonable obstacle avoidance path planning,and is highly prone to inter UAV collisions during the obstacle avoidance process.To address these issues,this study proposes a novel hybrid algorithm that combines the improved Multi-Robot Formation Obstacle Avoidance (MRF IAPF) algorithm with an enhanced APF optimized for single UAV path planning.Its core ideas are as follows:first,integrating three types of interaction forces from MRF IAPF obstacle repulsion force,inter UAV interaction force,and target attraction force;second,incorporating a refined single UAV path optimization mechanism,including collision risk assessment and an auxiliary sub goal strategy.When a UAV faces a high collision threat,temporary waypoints are generated to guide obstacle avoidance,ensuring eventual precise arrival at the actual target.Simulation results demonstrate that compared with traditional APF based formation algorithms,the proposed algorithm achieves significant improvements in path length optimization and heading stability,can effectively avoid obstacles and quickly restore the formation configuration,thus verifying its applicability and effectiveness in static environments with unknown obstacles.
comment: 12 pages, 13 figures, and 2 tables
☆ Single-Pixel Tactile Skin via Compressive Sampling
Development of large-area, high-speed electronic skins is a grand challenge for robotics, prosthetics, and human-machine interfaces, but is fundamentally limited by wiring complexity and data bottlenecks. Here, we introduce Single-Pixel Tactile Skin (SPTS), a paradigm that uses compressive sampling to reconstruct rich tactile information from an entire sensor array via a single output channel. This is achieved through a direct circuit-level implementation where each sensing element, equipped with a miniature microcontroller, contributes a dynamically weighted analog signal to a global sum, performing distributed compressed sensing in hardware. Our flexible, daisy-chainable design simplifies wiring to a few input lines and one output, and significantly reduces measurement requirements compared to raster scanning methods. We demonstrate the system's performance by achieving object classification at an effective 3500 FPS and by capturing transient dynamics, resolving an 8 ms projectile impact into 23 frames. A key feature is the support for adaptive reconstruction, where sensing fidelity scales with measurement time. This allows for rapid contact localization using as little as 7% of total data, followed by progressive refinement to a high-fidelity image - a capability critical for responsive robotic systems. This work offers an efficient pathway towards large-scale tactile intelligence for robotics and human-machine interfaces.
comment: 24 pages, 6 main figures, 6 supplemental figures
☆ QAL: A Loss for Recall Precision Balance in 3D Reconstruction WACV 2026
Volumetric learning underpins many 3D vision tasks such as completion, reconstruction, and mesh generation, yet training objectives still rely on Chamfer Distance (CD) or Earth Mover's Distance (EMD), which fail to balance recall and precision. We propose Quality-Aware Loss (QAL), a drop-in replacement for CD/EMD that combines a coverage-weighted nearest-neighbor term with an uncovered-ground-truth attraction term, explicitly decoupling recall and precision into tunable components. Across diverse pipelines, QAL achieves consistent coverage gains, improving by an average of +4.3 pts over CD and +2.8 pts over the best alternatives. Though modest in percentage, these improvements reliably recover thin structures and under-represented regions that CD/EMD overlook. Extensive ablations confirm stable performance across hyperparameters and across output resolutions, while full retraining on PCN and ShapeNet demonstrates generalization across datasets and backbones. Moreover, QAL-trained completions yield higher grasp scores under GraspNet evaluation, showing that improved coverage translates directly into more reliable robotic manipulation. QAL thus offers a principled, interpretable, and practical objective for robust 3D vision and safety-critical robotics pipelines
comment: Accepted to WACV 2026. Camera-ready version to appear
☆ SM2ITH: Safe Mobile Manipulation with Interactive Human Prediction via Task-Hierarchical Bilevel Model Predictive Control
Mobile manipulators are designed to perform complex sequences of navigation and manipulation tasks in human-centered environments. While recent optimization-based methods such as Hierarchical Task Model Predictive Control (HTMPC) enable efficient multitask execution with strict task priorities, they have so far been applied mainly to static or structured scenarios. Extending these approaches to dynamic human-centered environments requires predictive models that capture how humans react to the actions of the robot. This work introduces Safe Mobile Manipulation with Interactive Human Prediction via Task-Hierarchical Bilevel Model Predictive Control (SM$^2$ITH), a unified framework that combines HTMPC with interactive human motion prediction through bilevel optimization that jointly accounts for robot and human dynamics. The framework is validated on two different mobile manipulators, the Stretch 3 and the Ridgeback-UR10, across three experimental settings: (i) delivery tasks with different navigation and manipulation priorities, (ii) sequential pick-and-place tasks with different human motion prediction models, and (iii) interactions involving adversarial human behavior. Our results highlight how interactive prediction enables safe and efficient coordination, outperforming baselines that rely on weighted objectives or open-loop human models.
☆ Target-Bench: Can World Models Achieve Mapless Path Planning with Semantic Targets?
While recent world models generate highly realistic videos, their ability to perform robot path planning remains unclear and unquantified. We introduce Target-Bench, the first benchmark specifically designed to evaluate world models on mapless path planning toward semantic targets in real-world environments. Target-Bench provides 450 robot-collected video sequences spanning 45 semantic categories with SLAM-based ground truth trajectories. Our evaluation pipeline recovers camera motion from generated videos and measures planning performance using five complementary metrics that quantify target-reaching capability, trajectory accuracy, and directional consistency. We evaluate state-of-the-art models including Sora 2, Veo 3.1, and the Wan series. The best off-the-shelf model (Wan2.2-Flash) achieves only 0.299 overall score, revealing significant limitations in current world models for robotic planning tasks. We show that fine-tuning an open-source 5B-parameter model on only 325 scenarios from our dataset achieves 0.345 overall score -- an improvement of more than 400% over its base version (0.066) and 15% higher than the best off-the-shelf model. We will open-source the code and dataset.
comment: 10 pages
☆ SAFE-SMART: Safety Analysis and Formal Evaluation using STL Metrics for Autonomous RoboTs
We present a novel, regulator-driven approach for post hoc safety evaluation of learning-based, black-box autonomous mobile robots, ensuring ongoing compliance with evolving, human-defined safety rules. In our iterative workflow, human safety requirements are translated by regulators into Signal Temporal Logic (STL) specifications. Rollout traces from the black-box model are externally verified for compliance, yielding quantitative safety metrics, Total Robustness Value (TRV) and Largest Robustness Value (LRV), which measure average and worst-case specification adherence. These metrics inform targeted retraining and iterative improvement by model designers. We apply our method across two different applications: a virtual driving scenario and an autonomous mobile robot navigating a complex environment, and observe statistically significant improvements across both scenarios. In the virtual driving scenario, we see a 177% increase in traces adhering to the simulation speed limit, a 1138% increase in traces minimizing off-road driving, and a 16% increase in traces successfully reaching the goal within the time limit. In the autonomous navigation scenario, there is a 300% increase in traces avoiding sharp turns, a 200% increase in traces reaching the goal within the time limit, and a 49% increase in traces minimizing time spent near obstacles. Finally, we validate our approach on a TurtleBot3 robot in the real world, and demonstrate improved obstacle navigation with safety buffers.
☆ See, Plan, Cut: MPC-Based Autonomous Volumetric Robotic Laser Surgery with OCT Guidance
Robotic laser systems offer the potential for sub-millimeter, non-contact, high-precision tissue resection, yet existing platforms lack volumetric planning and intraoperative feedback. We present RATS (Robot-Assisted Tissue Surgery), an intelligent opto-mechanical, optical coherence tomography (OCT)-guided robotic platform designed for autonomous volumetric soft tissue resection in surgical applications. RATS integrates macro-scale RGB-D imaging, micro-scale OCT, and a fiber-coupled surgical laser, calibrated through a novel multistage alignment pipeline that achieves OCT-to-laser calibration accuracy of 0.161+-0.031mm on tissue phantoms and ex vivo porcine tissue. A super-Gaussian laser-tissue interaction (LTI) model characterizes ablation crater morphology with an average RMSE of 0.231+-0.121mm, outperforming Gaussian baselines. A sampling-based model predictive control (MPC) framework operates directly on OCT voxel data to generate constraint-aware resection trajectories with closed-loop feedback, achieving 0.842mm RMSE and improving intersection-over-union agreement by 64.8% compared to feedforward execution. With OCT, RATS detects subsurface structures and modifies the planner's objective to preserve them, demonstrating clinical feasibility.
comment: 9 pages, 8 figures
☆ Learning Diffusion Policies for Robotic Manipulation of Timber Joinery under Fabrication Uncertainty
Construction uncertainties such as fabrication inaccuracies and material imperfections pose a significant challenge to contact-rich robotic manipulation by hindering precise and robust assembly. In this paper, we explore the performance and robustness of diffusion policy learning as a promising solution for contact-sensitive robotic assembly at construction scale, using timber mortise and tenon joints as a case study. A two-phase study is conducted: first, to evaluate policy performance and applicability; second, to assess robustness in handling fabrication uncertainties simulated as randomized perturbations to the mortise position. The best-performing policy achieved a total average success rate of 75% with perturbations up to 10 mm, including 100% success in unperturbed cases. The results demonstrate the potential of sensory-motor diffusion policies to generalize to a wide range of complex, contact-rich assembly tasks across construction and manufacturing, advancing robotic construction under uncertainty and contributing to safer, more efficient building practices.
☆ LEARN: Learning End-to-End Aerial Resource-Constrained Multi-Robot Navigation
Nano-UAV teams offer great agility yet face severe navigation challenges due to constrained onboard sensing, communication, and computation. Existing approaches rely on high-resolution vision or compute-intensive planners, rendering them infeasible for these platforms. We introduce LEARN, a lightweight, two-stage safety-guided reinforcement learning (RL) framework for multi-UAV navigation in cluttered spaces. Our system combines low-resolution Time-of-Flight (ToF) sensors and a simple motion planner with a compact, attention-based RL policy. In simulation, LEARN outperforms two state-of-the-art planners by $10\%$ while using substantially fewer resources. We demonstrate LEARN's viability on six Crazyflie quadrotors, achieving fully onboard flight in diverse indoor and outdoor environments at speeds up to $2.0 m/s$ and traversing $0.2 m$ gaps.
comment: 20 pages, 15 figures
☆ Vision-Guided Optic Flow Navigation for Small Lunar Missions
Private lunar missions are faced with the challenge of robust autonomous navigation while operating under stringent constraints on mass, power, and computational resources. This work proposes a motion-field inversion framework that uses optical flow and rangefinder-based depth estimation as a lightweight CPU-based solution for egomotion estimation during lunar descent. We extend classical optical flow formulations by integrating them with depth modeling strategies tailored to the geometry for lunar/planetary approach, descent, and landing, specifically, planar and spherical terrain approximations parameterized by a laser rangefinder. Motion field inversion is performed through a least-squares framework, using sparse optical flow features extracted via the pyramidal Lucas-Kanade algorithm. We verify our approach using synthetically generated lunar images over the challenging terrain of the lunar south pole, using CPU budgets compatible with small lunar landers. The results demonstrate accurate velocity estimation from approach to landing, with sub-10% error for complex terrain and on the order of 1% for more typical terrain, as well as performances suitable for real-time applications. This framework shows promise for enabling robust, lightweight on-board navigation for small lunar missions.
♻ ☆ Perception, Control and Hardware for In-Hand Slip-Aware Object Manipulation with Parallel Grippers
Dexterous in-hand manipulation offers significant potential to enhance robotic manipulator capabilities. This paper presents a sensori-motor architecture for in-hand slip-aware control, being embodied in a sensorized gripper. The gripper in our architecture features rapid closed-loop, low-level force control, and is equipped with sensors capable of independently measuring contact forces and sliding velocities. Our system can quickly estimate essential object properties during pick-up using only in-hand sensing, without relying on prior object information. We introduce four distinct slippage controllers: gravity-assisted trajectory following for both rotational and linear slippage, a hinge controller that maintains the object's orientation while the gripper rotates, and a slip-avoidance controller. The gripper is mounted on a robot arm and validated through extensive experiments involving a diverse range of objects, demonstrating the architecture's novel capabilities for manipulating objects with flat surfaces.
♻ ☆ CleverDistiller: Simple and Spatially Consistent Cross-modal Distillation BMVC 2025
Vision foundation models (VFMs) such as DINO have led to a paradigm shift in 2D camera-based perception towards extracting generalized features to support many downstream tasks. Recent works introduce self-supervised cross-modal knowledge distillation (KD) as a way to transfer these powerful generalization capabilities into 3D LiDAR-based models. However, they either rely on highly complex distillation losses, pseudo-semantic maps, or limit KD to features useful for semantic segmentation only. In this work, we propose CleverDistiller, a self-supervised, cross-modal 2D-to-3D KD framework introducing a set of simple yet effective design choices: Unlike contrastive approaches relying on complex loss design choices, our method employs a direct feature similarity loss in combination with a multi layer perceptron (MLP) projection head to allow the 3D network to learn complex semantic dependencies throughout the projection. Crucially, our approach does not depend on pseudo-semantic maps, allowing for direct knowledge transfer from a VFM without explicit semantic supervision. Additionally, we introduce the auxiliary self-supervised spatial task of occupancy prediction to enhance the semantic knowledge, obtained from a VFM through KD, with 3D spatial reasoning capabilities. Experiments on standard autonomous driving benchmarks for 2D-to-3D KD demonstrate that CleverDistiller achieves state-of-the-art performance in both semantic segmentation and 3D object detection (3DOD) by up to 10% mIoU, especially when fine tuning on really low data amounts, showing the effectiveness of our simple yet powerful KD strategy
comment: Accepted to BMVC 2025
♻ ☆ Bootstrap Off-policy with World Model NeurIPS 2025
Online planning has proven effective in reinforcement learning (RL) for improving sample efficiency and final performance. However, using planning for environment interaction inevitably introduces a divergence between the collected data and the policy's actual behaviors, degrading both model learning and policy improvement. To address this, we propose BOOM (Bootstrap Off-policy with WOrld Model), a framework that tightly integrates planning and off-policy learning through a bootstrap loop: the policy initializes the planner, and the planner refines actions to bootstrap the policy through behavior alignment. This loop is supported by a jointly learned world model, which enables the planner to simulate future trajectories and provides value targets to facilitate policy improvement. The core of BOOM is a likelihood-free alignment loss that bootstraps the policy using the planner's non-parametric action distribution, combined with a soft value-weighted mechanism that prioritizes high-return behaviors and mitigates variability in the planner's action quality within the replay buffer. Experiments on the high-dimensional DeepMind Control Suite and Humanoid-Bench show that BOOM achieves state-of-the-art results in both training stability and final performance. The code is accessible at https://github.com/molumitu/BOOM_MBRL.
comment: NeurIPS 2025
♻ ☆ PhyBlock: A Progressive Benchmark for Physical Understanding and Planning via 3D Block Assembly
While vision-language models (VLMs) have demonstrated promising capabilities in reasoning and planning for embodied agents, their ability to comprehend physical phenomena, particularly within structured 3D environments, remains severely limited. To close this gap, we introduce PhyBlock, a progressive benchmark designed to assess VLMs on physical understanding and planning through robotic 3D block assembly tasks. PhyBlock integrates a novel four-level cognitive hierarchy assembly task alongside targeted Visual Question Answering (VQA) samples, collectively aimed at evaluating progressive spatial reasoning and fundamental physical comprehension, including object properties, spatial relationships, and holistic scene understanding. PhyBlock includes 2600 block tasks (400 assembly tasks, 2200 VQA tasks) and evaluates models across three key dimensions: partial completion, failure diagnosis, and planning robustness. We benchmark 21 state-of-the-art VLMs, highlighting their strengths and limitations in physically grounded, multi-step planning. Our empirical findings indicate that the performance of VLMs exhibits pronounced limitations in high-level planning and reasoning capabilities, leading to a notable decline in performance for the growing complexity of the tasks. Error analysis reveals persistent difficulties in spatial orientation and dependency reasoning. Surprisingly, chain-of-thought prompting offers minimal improvements, suggesting spatial tasks heavily rely on intuitive model comprehension. We position PhyBlock as a unified testbed to advance embodied reasoning, bridging vision-language understanding and real-world physical problem-solving.
♻ ☆ Mask2IV: Interaction-Centric Video Generation via Mask Trajectories AAAI 2026
Generating interaction-centric videos, such as those depicting humans or robots interacting with objects, is crucial for embodied intelligence, as they provide rich and diverse visual priors for robot learning, manipulation policy training, and affordance reasoning. However, existing methods often struggle to model such complex and dynamic interactions. While recent studies show that masks can serve as effective control signals and enhance generation quality, obtaining dense and precise mask annotations remains a major challenge for real-world use. To overcome this limitation, we introduce Mask2IV, a novel framework specifically designed for interaction-centric video generation. It adopts a decoupled two-stage pipeline that first predicts plausible motion trajectories for both actor and object, then generates a video conditioned on these trajectories. This design eliminates the need for dense mask inputs from users while preserving the flexibility to manipulate the interaction process. Furthermore, Mask2IV supports versatile and intuitive control, allowing users to specify the target object of interaction and guide the motion trajectory through action descriptions or spatial position cues. To support systematic training and evaluation, we curate two benchmarks covering diverse action and object categories across both human-object interaction and robotic manipulation scenarios. Extensive experiments demonstrate that our method achieves superior visual realism and controllability compared to existing baselines.
comment: AAAI 2026. Project page: https://reagan1311.github.io/mask2iv
♻ ☆ VLM-SFD: VLM-Assisted Siamese Flow Diffusion Framework for Dual-Arm Cooperative Manipulation
Dual-arm cooperative manipulation holds great promise for tackling complex real-world tasks that demand seamless coordination and adaptive dynamics. Despite substantial progress in learning-based motion planning, most approaches struggle to generalize across diverse manipulation tasks and adapt to dynamic, unstructured environments, particularly in scenarios involving interactions between two objects such as assembly, tool use, and bimanual grasping. To address these challenges, we introduce a novel VLM-Assisted Siamese Flow Diffusion (VLM-SFD) framework for efficient imitation learning in dual-arm cooperative manipulation. The proposed VLM-SFD framework exhibits outstanding adaptability, significantly enhancing the ability to rapidly adapt and generalize to diverse real-world tasks from only a minimal number of human demonstrations. Specifically, we propose a Siamese Flow Diffusion Network (SFDNet) employs a dual-encoder-decoder Siamese architecture to embed two target objects into a shared latent space, while a diffusion-based conditioning process - conditioned by task instructions - generates two-stream object-centric motion flows that guide dual-arm coordination. We further design a dynamic task assignment strategy that seamlessly maps the predicted 2D motion flows into 3D space and incorporates a pre-trained vision-language model (VLM) to adaptively assign the optimal motion to each robotic arm over time. Experiments validate the effectiveness of the proposed method, demonstrating its ability to generalize to diverse manipulation tasks while maintaining high efficiency and adaptability. The code and demo videos are publicly available on our project website https://sites.google.com/view/vlm-sfd/.
comment: Accepted by IEEE RA-L
♻ ☆ AgriChrono: A Multi-modal Dataset Capturing Crop Growth and Lighting Variability with a Field Robot
Advances in AI and Robotics have accelerated significant initiatives in agriculture, particularly in the areas of robot navigation and 3D digital twin creation. A significant bottleneck impeding this progress is the critical lack of "in-the-wild" datasets that capture the full complexities of real farmland, including non-rigid motion from wind, drastic illumination variance, and morphological changes resulting from growth. This data gap fundamentally limits research on robust AI models for autonomous field navigation and scene-level dynamic 3D reconstruction. In this paper, we present AgriChrono, a modular robotic data collection platform and multi-modal dataset designed to capture these dynamic farmland conditions. Our platform integrates multiple sensors, enabling remote, time-synchronized acquisition of RGB, Depth, LiDAR, IMU, and Pose data for efficient and repeatable long-term data collection in real-world agricultural environments. We successfully collected 18TB of data over one month, documenting the entire growth cycle of Canola under diverse illumination conditions. We benchmark state-of-the-art 3D reconstruction methods on AgriChrono, revealing the profound challenge of reconstructing high-fidelity, dynamic non-rigid scenes in such farmland settings. This benchmark validates AgriChrono as a critical asset for advancing model generalization, and its public release is expected to significantly accelerate research and development in precision agriculture. The code and dataset are publicly available at: https://github.com/StructuresComp/agri-chrono
♻ ☆ OpenVLN: Open-world Aerial Vision-Language Navigation
Vision-language models (VLMs) have been widely-applied in ground-based vision-language navigation (VLN). However, the vast complexity of outdoor aerial environments compounds data acquisition challenges and imposes long-horizon trajectory planning requirements on Unmanned Aerial Vehicles (UAVs), introducing novel complexities for aerial VLN. To address these challenges, we propose a data-efficient Open-world aerial Vision-Language Navigation (i.e., OpenVLN) framework, which could execute language-guided flight with limited data constraints and enhance long-horizon trajectory planning capabilities in complex aerial environments. Specifically, we reconfigure a reinforcement learning framework to optimize the VLM for UAV navigation tasks, which can efficiently fine-tune VLM by using rule-based policies under limited training data. Concurrently, we introduce a long-horizon planner for trajectory synthesis that dynamically generates precise UAV actions via value-based rewards. To the end, we conduct sufficient navigation experiments on the TravelUAV benchmark with dataset scaling across diverse reward settings. Our method demonstrates consistent performance gains of up to 4.34% in Success Rate, 6.19% in Oracle Success Rate, and 4.07% in Success weighted by Path Length over baseline methods, validating its deployment efficacy for long-horizon UAV navigation in complex aerial environments.
comment: Content: 8 pages 4 figures, conference paper under review
♻ ☆ Learning to Drive Anywhere with Model-Based Reannotation
Developing broadly generalizable visual navigation policies for robots is a significant challenge, primarily constrained by the availability of large-scale, diverse training data. While curated datasets collected by researchers offer high quality, their limited size restricts policy generalization. To overcome this, we explore leveraging abundant, passively collected data sources, including large volumes of crowd-sourced teleoperation data and unlabeled YouTube videos, despite their potential for lower quality or missing action labels. We propose Model-Based ReAnnotation (MBRA), a framework that utilizes a learned short-horizon, model-based expert model to relabel or generate high-quality actions for these passive datasets. This relabeled data is then distilled into LogoNav, a long-horizon navigation policy conditioned on visual goals or GPS waypoints. We demonstrate that LogoNav, trained using MBRA-processed data, achieves state-of-the-art performance, enabling robust navigation over distances exceeding 300 meters in previously unseen indoor and outdoor environments. Our extensive real-world evaluations, conducted across a fleet of robots (including quadrupeds) in six cities on three continents, validate the policy's ability to generalize and navigate effectively even amidst pedestrians in crowded settings.
comment: 9 pages, 8 figures, 6 tables
♻ ☆ FalconWing: An Ultra-Light Indoor Fixed-Wing UAV Platform for Vision-Based Autonomy
We introduce FalconWing, an ultra-light (150 g) indoor fixed-wing UAV platform for vision-based autonomy. Controlled indoor environment enables year-round repeatable UAV experiment but imposes strict weight and maneuverability limits on the UAV, motivating our ultra-light FalconWing design. FalconWing couples a lightweight hardware stack (137g airframe with a 9g camera) and offboard computation with a software stack featuring a photorealistic 3D Gaussian Splat (GSplat) simulator for developing and evaluating vision-based controllers. We validate FalconWing on two challenging vision-based aerial case studies. In the leader-follower case study, our best vision-based controller, trained via imitation learning on GSplat-rendered data augmented with domain randomization, achieves 100% tracking success across 3 types of leader maneuvers over 30 trials and shows robustness to leader's appearance shifts in simulation. In the autonomous landing case study, our vision-based controller trained purely in simulation transfers zero-shot to real hardware, achieving an 80% success rate over ten landing trials. We will release hardware designs, GSplat scenes, and dynamics models upon publication to make FalconWing an open-source flight kit for engineering students and research labs.
♻ ☆ DeepFleet: Multi-Agent Foundation Models for Mobile Robots
We introduce DeepFleet, a suite of foundation models designed to support coordination and planning for large-scale mobile robot fleets. These models are trained on fleet movement data, including robot positions, goals, and interactions, from hundreds of thousands of robots in Amazon warehouses worldwide. DeepFleet consists of four architectures that each embody a distinct inductive bias and collectively explore key points in the design space for multi-agent foundation models: the robot-centric (RC) model is an autoregressive decision transformer operating on neighborhoods of individual robots; the robot-floor (RF) model uses a transformer with cross-attention between robots and the warehouse floor; the image-floor (IF) model applies convolutional encoding to a multi-channel image representation of the full fleet; and the graph-floor (GF) model combines temporal attention with graph neural networks for spatial relationships. In this paper, we describe these models and present our evaluation of the impact of these design choices on prediction task performance. We find that the robot-centric and graph-floor models, which both use asynchronous robot state updates and incorporate the localized structure of robot interactions, show the most promise. We also present experiments that show that these two models can make effective use of larger warehouses operation datasets as the models are scaled up.
comment: 27 pages, 10 figures, 2 tables
♻ ☆ Ionospheric and Plasmaspheric Delay Characterization for Lunar Terrestrial GNSS Receivers with Global Core Plasma Model
Recent advancements in lunar positioning, navigation, and timing (PNT) have demonstrated that terrestrial GNSS signals, including weak sidelobe transmissions, can be exploited for lunar spacecraft positioning and timing. While GNSS-based navigation at the Moon has been validated recently, unmodeled ionospheric and plasmaspheric delays remain a significant error source, particularly given the unique signal geometry and extended propagation paths. This paper characterizes these delays using the Global Core Plasma Model (GCPM) and a custom low-cost ray-tracing algorithm that iteratively solves for bent signal paths. We simulate first-, second-, and third-order group delays, as well as excess path length from ray bending, for GNSS signals received at both lunar orbit and the lunar south pole under varying solar and geomagnetic conditions. Results show that mean group delays are typically on the order of 1 m, but can exceed 100 m for low-altitude ray paths during high solar activity, while bending delays are generally smaller but non-negligible for low-altitude ray paths. We also quantify the influence of signal frequency, geomagnetic $K_p$ index, and solar R12 index. These findings inform the design of robust positioning and timing algorithms that utilize terrestrial GNSS signals.
comment: Submitted NAVIGATION: Journal of the Institute of Navigation
Robotics 39
☆ Dexterity from Smart Lenses: Multi-Fingered Robot Manipulation with In-the-Wild Human Demonstrations
Learning multi-fingered robot policies from humans performing daily tasks in natural environments has long been a grand goal in the robotics community. Achieving this would mark significant progress toward generalizable robot manipulation in human environments, as it would reduce the reliance on labor-intensive robot data collection. Despite substantial efforts, progress toward this goal has been bottle-necked by the embodiment gap between humans and robots, as well as by difficulties in extracting relevant contextual and motion cues that enable learning of autonomous policies from in-the-wild human videos. We claim that with simple yet sufficiently powerful hardware for obtaining human data and our proposed framework AINA, we are now one significant step closer to achieving this dream. AINA enables learning multi-fingered policies from data collected by anyone, anywhere, and in any environment using Aria Gen 2 glasses. These glasses are lightweight and portable, feature a high-resolution RGB camera, provide accurate on-board 3D head and hand poses, and offer a wide stereo view that can be leveraged for depth estimation of the scene. This setup enables the learning of 3D point-based policies for multi-fingered hands that are robust to background changes and can be deployed directly without requiring any robot data (including online corrections, reinforcement learning, or simulation). We compare our framework against prior human-to-robot policy learning approaches, ablate our design choices, and demonstrate results across nine everyday manipulation tasks. Robot rollouts are best viewed on our website: https://aina-robot.github.io.
☆ InternData-A1: Pioneering High-Fidelity Synthetic Data for Pre-training Generalist Policy
Recent works explore how real and synthetic data contribute to Vision-Language-Action (VLA) models' generalization. While current VLA models have shown the strong effectiveness of large-scale real-robot pre-training, synthetic data has not previously demonstrated comparable capability at scale. This paper provides the first evidence that synthetic data alone can match the performance of the strongest $π$-dataset in pre-training a VLA model, revealing the substantial value of large-scale simulation. The resulting model also exhibits surprisingly zero-shot sim-to-real transfer on several challenging tasks. Our synthetic dataset, InternData-A1, contains over 630k trajectories and 7,433 hours across 4 embodiments, 18 skills, 70 tasks, and 227 scenes, covering rigid, articulated, deformable, and fluid-object manipulation. It is generated through a highly autonomous, fully decoupled, and compositional simulation pipeline that enables long-horizon skill composition, flexible task assembly, and heterogeneous embodiments with minimal manual tuning. Using the same architecture as $π_0$, we pre-train a model entirely on InternData-A1 and find that it matches the official $π_0$ across 49 simulation tasks, 5 real-world tasks, and 4 long-horizon dexterous tasks. We release the dataset and will open-source the generation pipeline to broaden access to large-scale robotic data and to lower the barrier to scalable data creation for embodied AI research.
☆ Green Resilience of Cyber-Physical Systems: Doctoral Dissertation
Cyber-physical systems (CPS) combine computational and physical components. Online Collaborative AI System (OL-CAIS) is a type of CPS that learn online in collaboration with humans to achieve a common goal, which makes it vulnerable to disruptive events that degrade performance. Decision-makers must therefore restore performance while limiting energy impact, creating a trade-off between resilience and greenness. This research addresses how to balance these two properties in OL-CAIS. It aims to model resilience for automatic state detection, develop agent-based policies that optimize the greenness-resilience trade-off, and understand catastrophic forgetting to maintain performance consistency. We model OL-CAIS behavior through three operational states: steady, disruptive, and final. To support recovery during disruptions, we introduce the GResilience framework, which provides recovery strategies through multi-objective optimization (one-agent), game-theoretic decision-making (two-agent), and reinforcement learning (RL-agent). We also design a measurement framework to quantify resilience and greenness. Empirical evaluation uses real and simulated experiments with a collaborative robot learning object classification from human demonstrations. Results show that the resilience model captures performance transitions during disruptions, and that GResilience policies improve green recovery by shortening recovery time, stabilizing performance, and reducing human dependency. RL-agent policies achieve the strongest results, although with a marginal increase in CO2 emissions. We also observe catastrophic forgetting after repeated disruptions, while our policies help maintain steadiness. A comparison with containerized execution shows that containerization cuts CO2 emissions by half. Overall, this research provides models, metrics, and policies that ensure the green recovery of OL-CAIS.
☆ MiMo-Embodied: X-Embodied Foundation Model Technical Report
We open-source MiMo-Embodied, the first cross-embodied foundation model to successfully integrate and achieve state-of-the-art performance in both Autonomous Driving and Embodied AI. MiMo-Embodied sets new records across 17 embodied AI benchmarks in Task Planning, Affordance Prediction and Spatial Understanding, while also excelling in 12 autonomous driving benchmarks across Environmental Perception, Status Prediction, and Driving Planning. Across these tasks, MiMo-Embodied significantly outperforms existing open-source, closed-source, and specialized baselines. Our results indicate that through multi-stage learning, curated data construction, and CoT/RL fine-tuning, these two domains exhibit strong positive transfer and mutually reinforce one another. We provide a detailed analysis of our model design and training methodologies to facilitate further research. Code and models are available at https://github.com/XiaomiMiMo/MiMo-Embodied.
comment: Code: https://github.com/XiaomiMiMo/MiMo-Embodied Model: https://huggingface.co/XiaomiMiMo/MiMo-Embodied-7B
☆ From Prompts to Printable Models: Support-Effective 3D Generation via Offset Direct Preference Optimization
The transition from digital 3D models to physical objects via 3D printing often requires support structures to prevent overhanging features from collapsing during the fabrication process. While current slicing technologies offer advanced support strategies, they focus on post-processing optimizations rather than addressing the underlying need for support-efficient design during the model generation phase. This paper introduces SEG (\textit{\underline{S}upport-\underline{E}ffective \underline{G}eneration}), a novel framework that integrates Direct Preference Optimization with an Offset (ODPO) into the 3D generation pipeline to directly optimize models for minimal support material usage. By incorporating support structure simulation into the training process, SEG encourages the generation of geometries that inherently require fewer supports, thus reducing material waste and production time. We demonstrate SEG's effectiveness through extensive experiments on two benchmark datasets, Thingi10k-Val and GPT-3DP-Val, showing that SEG significantly outperforms baseline models such as TRELLIS, DPO, and DRO in terms of support volume reduction and printability. Qualitative results further reveal that SEG maintains high fidelity to input prompts while minimizing the need for support structures. Our findings highlight the potential of SEG to transform 3D printing by directly optimizing models during the generative process, paving the way for more sustainable and efficient digital fabrication practices.
comment: Technical report (7 pages)
☆ LAOF: Robust Latent Action Learning with Optical Flow Constraints
Learning latent actions from large-scale videos is crucial for the pre-training of scalable embodied foundation models, yet existing methods often struggle with action-irrelevant distractors. Although incorporating action supervision can alleviate these distractions, its effectiveness is restricted by the scarcity of available action labels. Optical flow represents pixel-level motion between consecutive frames, naturally suppressing background elements and emphasizing moving objects. Motivated by this, we propose robust Latent Action learning with Optical Flow constraints, called LAOF, a pseudo-supervised framework that leverages the agent's optical flow as an action-driven signal to learn latent action representations robust to distractors. Experimental results show that the latent representations learned by LAOF outperform existing methods on downstream imitation learning and reinforcement learning tasks. This superior performance arises from optical flow constraints, which substantially stabilize training and improve the quality of latent representations under extremely label-scarce conditions, while remaining effective as the proportion of action labels increases to 10 percent. Importantly, even without action supervision, LAOF matches or surpasses action-supervised methods trained with 1 percent of action labels.
comment: Code can be found at https://github.com/XizoB/LAOF
☆ Homogeneous Proportional-Integral-Derivative Controller in Mobile Robotic Manipulators
Mobile robotic manipulators (MRMs), which integrate mobility and manipulation capabilities, present significant control challenges due to their nonlinear dynamics, underactuation, and coupling between the base and manipulator subsystems. This paper proposes a novel homogeneous Proportional-Integral-Derivative (hPID) control strategy tailored for MRMs to achieve robust and coordinated motion control. Unlike classical PID controllers, the hPID controller leverages the mathematical framework of homogeneous control theory to systematically enhance the stability and convergence properties of the closed-loop system, even in the presence of dynamic uncertainties and external disturbances involved into a system in a homogeneous way. A homogeneous PID structure is designed, ensuring improved convergence of tracking errors through a graded homogeneity approach that generalizes traditional PID gains to nonlinear, state-dependent functions. Stability analysis is conducted using Lyapunov-based methods, demonstrating that the hPID controller guarantees global asymptotic stability and finite-time convergence under mild assumptions. Experimental results on a representative MRM model validate the effectiveness of the hPID controller in achieving high-precision trajectory tracking for both the mobile base and manipulator arm, outperforming conventional linear PID controllers in terms of response time, steady-state error, and robustness to model uncertainties. This research contributes a scalable and analytically grounded control framework for enhancing the autonomy and reliability of next-generation mobile manipulation systems in structured and unstructured environments.
☆ Robot Metacognition: Decision Making with Confidence for Tool Invention
Robots today often miss a key ingredient of truly intelligent behavior: the ability to reflect on their own cognitive processes and decisions. In humans, this self-monitoring or metacognition is crucial for learning, decision making and problem solving. For instance, they can evaluate how confident they are in performing a task, thus regulating their own behavior and allocating proper resources. Taking inspiration from neuroscience, we propose a robot metacognition architecture centered on confidence (a second-order judgment on decisions) and we demonstrate it on the use case of autonomous tool invention. We propose the use of confidence as a metacognitive measure within the robot decision making scheme. Confidence-informed robots can evaluate the reliability of their decisions, improving their robustness during real-world physical deployment. This form of robotic metacognition emphasizes embodied action monitoring as a means to achieve better informed decisions. We also highlight potential applications and research directions for robot metacognition.
comment: under review
☆ Flow-Aided Flight Through Dynamic Clutters From Point To Motion
Challenges in traversing dynamic clutters lie mainly in the efficient perception of the environmental dynamics and the generation of evasive behaviors considering obstacle movement. Previous solutions have made progress in explicitly modeling the dynamic obstacle motion for avoidance, but this key dependency of decision-making is time-consuming and unreliable in highly dynamic scenarios with occlusions. On the contrary, without introducing object detection, tracking, and prediction, we empower the reinforcement learning (RL) with single LiDAR sensing to realize an autonomous flight system directly from point to motion. For exteroception, a depth sensing distance map achieving fixed-shape, low-resolution, and detail-safe is encoded from raw point clouds, and an environment change sensing point flow is adopted as motion features extracted from multi-frame observations. These two are integrated into a lightweight and easy-to-learn representation of complex dynamic environments. For action generation, the behavior of avoiding dynamic threats in advance is implicitly driven by the proposed change-aware sensing representation, where the policy optimization is indicated by the relative motion modulated distance field. With the deployment-friendly sensing simulation and dynamics model-free acceleration control, the proposed system shows a superior success rate and adaptability to alternatives, and the policy derived from the simulator can drive a real-world quadrotor with safe maneuvers.
comment: Accepted to IEEE Robotics and Automation Letters (RA-L), November, 2025
☆ The Shawshank Redemption of Embodied AI: Understanding and Benchmarking Indirect Environmental Jailbreaks
The adoption of Vision-Language Models (VLMs) in embodied AI agents, while being effective, brings safety concerns such as jailbreaking. Prior work have explored the possibility of directly jailbreaking the embodied agents through elaborated multi-modal prompts. However, no prior work has studied or even reported indirect jailbreaks in embodied AI, where a black-box attacker induces a jailbreak without issuing direct prompts to the embodied agent. In this paper, we propose, for the first time, indirect environmental jailbreak (IEJ), a novel attack to jailbreak embodied AI via indirect prompt injected into the environment, such as malicious instructions written on a wall. Our key insight is that embodied AI does not ''think twice'' about the instructions provided by the environment -- a blind trust that attackers can exploit to jailbreak the embodied agent. We further design and implement open-source prototypes of two fully-automated frameworks: SHAWSHANK, the first automatic attack generation framework for the proposed attack IEJ; and SHAWSHANK-FORGE, the first automatic benchmark generation framework for IEJ. Then, using SHAWSHANK-FORGE, we automatically construct SHAWSHANK-BENCH, the first benchmark for indirectly jailbreaking embodied agents. Together, our two frameworks and one benchmark answer the questions of what content can be used for malicious IEJ instructions, where they should be placed, and how IEJ can be systematically evaluated. Evaluation results show that SHAWSHANK outperforms eleven existing methods across 3,957 task-scene combinations and compromises all six tested VLMs. Furthermore, current defenses only partially mitigate our attack, and we have responsibly disclosed our findings to all affected VLM vendors.
☆ Safe and Optimal Variable Impedance Control via Certified Reinforcement Learning
Reinforcement learning (RL) offers a powerful approach for robots to learn complex, collaborative skills by combining Dynamic Movement Primitives (DMPs) for motion and Variable Impedance Control (VIC) for compliant interaction. However, this model-free paradigm often risks instability and unsafe exploration due to the time-varying nature of impedance gains. This work introduces Certified Gaussian Manifold Sampling (C-GMS), a novel trajectory-centric RL framework that learns combined DMP and VIC policies while guaranteeing Lyapunov stability and actuator feasibility by construction. Our approach reframes policy exploration as sampling from a mathematically defined manifold of stable gain schedules. This ensures every policy rollout is guaranteed to be stable and physically realizable, thereby eliminating the need for reward penalties or post-hoc validation. Furthermore, we provide a theoretical guarantee that our approach ensures bounded tracking error even in the presence of bounded model errors and deployment-time uncertainties. We demonstrate the effectiveness of C-GMS in simulation and verify its efficacy on a real robot, paving the way for reliable autonomous interaction in complex environments.
☆ InEKFormer: A Hybrid State Estimator for Humanoid Robots
Humanoid robots have great potential for a wide range of applications, including industrial and domestic use, healthcare, and search and rescue missions. However, bipedal locomotion in different environments is still a challenge when it comes to performing stable and dynamic movements. This is where state estimation plays a crucial role, providing fast and accurate feedback of the robot's floating base state to the motion controller. Although classical state estimation methods such as Kalman filters are widely used in robotics, they require expert knowledge to fine-tune the noise parameters. Due to recent advances in the field of machine learning, deep learning methods are increasingly used for state estimation tasks. In this work, we propose the InEKFormer, a novel hybrid state estimation method that incorporates an invariant extended Kalman filter (InEKF) and a Transformer network. We compare our method with the InEKF and the KalmanNet approaches on datasets obtained from the humanoid robot RH5. The results indicate the potential of Transformers in humanoid state estimation, but also highlight the need for robust autoregressive training in these high-dimensional problems.
comment: Accepted at The 22nd International Conference on Advanced Robotics (ICAR 2025)
☆ Funabot-Upper: McKibben Actuated Haptic Suit Inducing Kinesthetic Perceptions in Trunk, Shoulder, Elbow, and Wrist
This paper presents Funabot-Upper, a wearable haptic suit that enables users to perceive 14 upper-body motions, including those of the trunk, shoulder, elbow, and wrist. Inducing kinesthetic perception through wearable haptic devices has attracted attention, and various devices have been developed in the past. However, these have been limited to verifications on single body parts, and few have applied the same method to multiple body parts as well. In our previous study, we developed a technology that uses the contraction of artificial muscles to deform clothing in three dimensions. Using this technology, we developed a haptic suit that induces kinesthetic perception of 7 motions in multiple upper body. However, perceptual mixing caused by stimulating multiple human muscles has occurred between the shoulder and the elbow. In this paper, we established a new, simplified design policy and developed a novel haptic suit that induces kinesthetic perceptions in the trunk, shoulder, elbow, and wrist by stimulating joints and muscles independently. We experimentally demonstrated the induced kinesthetic perception and examined the relationship between stimulation and perceived kinesthetic perception under the new design policy. Experiments confirmed that Funabot-Upper successfully induces kinesthetic perception across multiple joints while reducing perceptual mixing observed in previous designs. The new suit improved recognition accuracy from 68.8% to 94.6% compared to the previous Funabot-Suit, demonstrating its superiority and potential for future haptic applications.
comment: 8 pages, 8 figures. This work has been submitted to the IEEE for possible publication
☆ How Robot Dogs See the Unseeable
Peering, a side-to-side motion used by animals to estimate distance through motion parallax, offers a powerful bio-inspired strategy to overcome a fundamental limitation in robotic vision: partial occlusion. Conventional robot cameras, with their small apertures and large depth of field, render both foreground obstacles and background objects in sharp focus, causing occluders to obscure critical scene information. This work establishes a formal connection between animal peering and synthetic aperture (SA) sensing from optical imaging. By having a robot execute a peering motion, its camera describes a wide synthetic aperture. Computational integration of the captured images synthesizes an image with an extremely shallow depth of field, effectively blurring out occluding elements while bringing the background into sharp focus. This efficient, wavelength-independent technique enables real-time, high-resolution perception across various spectral bands. We demonstrate that this approach not only restores basic scene understanding but also empowers advanced visual reasoning in large multimodal models, which fail with conventionally occluded imagery. Unlike feature-dependent multi-view 3D vision methods or active sensors like LiDAR, SA sensing via peering is robust to occlusion, computationally efficient, and immediately deployable on any mobile robot. This research bridges animal behavior and robotics, suggesting that peering motions for synthetic aperture sensing are a key to advanced scene understanding in complex, cluttered environments.
☆ FT-NCFM: An Influence-Aware Data Distillation Framework for Efficient VLA Models AAAI
The powerful generalization of Vision-Language-Action (VLA) models is bottlenecked by their heavy reliance on massive, redundant, and unevenly valued datasets, hindering their widespread application. Existing model-centric optimization paths, such as model compression (which often leads to performance degradation) or policy distillation (whose products are model-dependent and lack generality), fail to fundamentally address this data-level challenge. To this end, this paper introduces FT-NCFM, a fundamentally different, data-centric generative data distillation framework. Our framework employs a self-contained Fact-Tracing (FT) engine that combines causal attribution with programmatic contrastive verification to assess the intrinsic value of samples. Guided by these assessments, an adversarial NCFM process synthesizes a model-agnostic, information-dense, and reusable data asset. Experimental results on several mainstream VLA benchmarks show that models trained on just 5% of our distilled coreset achieve a success rate of 85-90% compared with training on the full dataset, while reducing training time by over 80%. Our work demonstrates that intelligent data distillation is a highly promising new path for building efficient, high-performance VLA models.
comment: Accepted at the AAAI Conference on Artificial Intelligence (AAAI-26)
☆ DynaMimicGen: A Data Generation Framework for Robot Learning of Dynamic Tasks
Learning robust manipulation policies typically requires large and diverse datasets, the collection of which is time-consuming, labor-intensive, and often impractical for dynamic environments. In this work, we introduce DynaMimicGen (D-MG), a scalable dataset generation framework that enables policy training from minimal human supervision while uniquely supporting dynamic task settings. Given only a few human demonstrations, D-MG first segments the demonstrations into meaningful sub-tasks, then leverages Dynamic Movement Primitives (DMPs) to adapt and generalize the demonstrated behaviors to novel and dynamically changing environments. Improving prior methods that rely on static assumptions or simplistic trajectory interpolation, D-MG produces smooth, realistic, and task-consistent Cartesian trajectories that adapt in real time to changes in object poses, robot states, or scene geometry during task execution. Our method supports different scenarios - including scene layouts, object instances, and robot configurations - making it suitable for both static and highly dynamic manipulation tasks. We show that robot agents trained via imitation learning on D-MG-generated data achieve strong performance across long-horizon and contact-rich benchmarks, including tasks like cube stacking and placing mugs in drawers, even under unpredictable environment changes. By eliminating the need for extensive human demonstrations and enabling generalization in dynamic settings, D-MG offers a powerful and efficient alternative to manual data collection, paving the way toward scalable, autonomous robot learning.
☆ PIPHEN: Physical Interaction Prediction with Hamiltonian Energy Networks AAAI
Multi-robot systems in complex physical collaborations face a "shared brain dilemma": transmitting high-dimensional multimedia data (e.g., video streams at ~30MB/s) creates severe bandwidth bottlenecks and decision-making latency. To address this, we propose PIPHEN, an innovative distributed physical cognition-control framework. Its core idea is to replace "raw data communication" with "semantic communication" by performing "semantic distillation" at the robot edge, reconstructing high-dimensional perceptual data into compact, structured physical representations. This idea is primarily realized through two key components: (1) a novel Physical Interaction Prediction Network (PIPN), derived from large model knowledge distillation, to generate this representation; and (2) a Hamiltonian Energy Network (HEN) controller, based on energy conservation, to precisely translate this representation into coordinated actions. Experiments show that, compared to baseline methods, PIPHEN can compress the information representation to less than 5% of the original data volume and reduce collaborative decision-making latency from 315ms to 76ms, while significantly improving task success rates. This work provides a fundamentally efficient paradigm for resolving the "shared brain dilemma" in resource-constrained multi-robot systems.
comment: Accepted at the AAAI Conference on Artificial Intelligence (AAAI-26)
☆ MagBotSim: Physics-Based Simulation and Reinforcement Learning Environments for Magnetic Robotics
Magnetic levitation is about to revolutionize in-machine material flow in industrial automation. Such systems are flexibly configurable and can include a large number of independently actuated shuttles (movers) that dynamically rebalance production capacity. Beyond their capabilities for dynamic transportation, these systems possess the inherent yet unexploited potential to perform manipulation. By merging the fields of transportation and manipulation into a coordinated swarm of magnetic robots (MagBots), we enable manufacturing systems to achieve significantly higher efficiency, adaptability, and compactness. To support the development of intelligent algorithms for magnetic levitation systems, we introduce MagBotSim (Magnetic Robotics Simulation): a physics-based simulation for magnetic levitation systems. By framing magnetic levitation systems as robot swarms and providing a dedicated simulation, this work lays the foundation for next generation manufacturing systems powered by Magnetic Robotics. MagBotSim's documentation, videos, experiments, and code are available at: https://ubi-coro.github.io/MagBotSim/
☆ LEGO-SLAM: Language-Embedded Gaussian Optimization SLAM
Recent advances in 3D Gaussian Splatting (3DGS) have enabled Simultaneous Localization and Mapping (SLAM) systems to build photorealistic maps. However, these maps lack the open-vocabulary semantic understanding required for advanced robotic interaction. Integrating language features into SLAM remains a significant challenge, as storing high-dimensional features demands excessive memory and rendering overhead, while existing methods with static models lack adaptability for novel environments. To address these limitations, we propose LEGO-SLAM (Language-Embedded Gaussian Optimization SLAM), the first framework to achieve real-time, open-vocabulary mapping within a 3DGS-based SLAM system. At the core of our method is a scene-adaptive encoder-decoder that distills high-dimensional language embeddings into a compact 16-dimensional feature space. This design reduces the memory per Gaussian and accelerates rendering, enabling real-time performance. Unlike static approaches, our encoder adapts online to unseen scenes. These compact features also enable a language-guided pruning strategy that identifies semantic redundancy, reducing the map's Gaussian count by over 60\% while maintaining rendering quality. Furthermore, we introduce a language-based loop detection approach that reuses these mapping features, eliminating the need for a separate detection model. Extensive experiments demonstrate that LEGO-SLAM achieves competitive mapping quality and tracking accuracy, all while providing open-vocabulary capabilities at 15 FPS.
comment: 18 pages
☆ Heterogeneous Stroke: Using Unique Vibration Cues to Improve the Wrist-Worn Spatiotemporal Tactile Display
Beyond a simple notification of incoming calls or messages, more complex information such as alphabets and digits can be delivered through spatiotemporal tactile patterns (STPs) on a wrist-worn tactile display (WTD) with multiple tactors. However, owing to the limited skin area and spatial acuity of the wrist, frequent confusions occur between closely located tactors, resulting in a low recognition accuracy. Furthermore, the accuracies reported in previous studies have mostly been measured for a specific posture and could further decrease with free arm postures in real life. Herein, we present Heterogeneous Stroke, a design concept for improving the recognition accuracy of STPs on a WTD. By assigning unique vibrotactile stimuli to each tactor, the confusion between tactors can be reduced. Through our implementation of Heterogeneous Stroke, the alphanumeric characters could be delivered with high accuracy (93.8% for 26 alphabets and 92.4% for 10 digits) across different arm postures.
comment: ACM CHI 2021
☆ Bi-AQUA: Bilateral Control-Based Imitation Learning for Underwater Robot Arms via Lighting-Aware Action Chunking with Transformers
Underwater robotic manipulation is fundamentally challenged by extreme lighting variations, color distortion, and reduced visibility. We introduce Bi-AQUA, the first underwater bilateral control-based imitation learning framework that integrates lighting-aware visual processing for underwater robot arms. Bi-AQUA employs a hierarchical three-level lighting adaptation mechanism: a Lighting Encoder that extracts lighting representations from RGB images without manual annotation and is implicitly supervised by the imitation objective, FiLM modulation of visual backbone features for adaptive, lighting-aware feature extraction, and an explicit lighting token added to the transformer encoder input for task-aware conditioning. Experiments on a real-world underwater pick-and-place task under diverse static and dynamic lighting conditions show that Bi-AQUA achieves robust performance and substantially outperforms a bilateral baseline without lighting modeling. Ablation studies further confirm that all three lighting-aware components are critical. This work bridges terrestrial bilateral control-based imitation learning and underwater manipulation, enabling force-sensitive autonomous operation in challenging marine environments. For additional material, please check: https://mertcookimg.github.io/bi-aqua
☆ Semantic Glitch: Agency and Artistry in an Autonomous Pixel Cloud NeurIPS 2025
While mainstream robotics pursues metric precision and flawless performance, this paper explores the creative potential of a deliberately "lo-fi" approach. We present the "Semantic Glitch," a soft flying robotic art installation whose physical form, a 3D pixel style cloud, is a "physical glitch" derived from digital archaeology. We detail a novel autonomous pipeline that rejects conventional sensors like LiDAR and SLAM, relying solely on the qualitative, semantic understanding of a Multimodal Large Language Model to navigate. By authoring a bio-inspired personality for the robot through a natural language prompt, we create a "narrative mind" that complements the "weak," historically, loaded body. Our analysis begins with a 13-minute autonomous flight log, and a follow-up study statistically validates the framework's robustness for authoring quantifiably distinct personas. The combined analysis reveals emergent behaviors, from landmark-based navigation to a compelling "plan to execution" gap, and a character whose unpredictable, plausible behavior stems from a lack of precise proprioception. This demonstrates a lo-fi framework for creating imperfect companions whose success is measured in character over efficiency.
comment: NeurIPS 2025 Creative AI Track, The Thirty-Ninth Annual Conference on Neural Information Processing Systems
☆ Towards a Safer and Sustainable Manufacturing Process: Material classification in Laser Cutting Using Deep Learning
Laser cutting is a widely adopted technology in material processing across various industries, but it generates a significant amount of dust, smoke, and aerosols during operation, posing a risk to both the environment and workers' health. Speckle sensing has emerged as a promising method to monitor the cutting process and identify material types in real-time. This paper proposes a material classification technique using a speckle pattern of the material's surface based on deep learning to monitor and control the laser cutting process. The proposed method involves training a convolutional neural network (CNN) on a dataset of laser speckle patterns to recognize distinct material types for safe and efficient cutting. Previous methods for material classification using speckle sensing may face issues when the color of the laser used to produce the speckle pattern is changed. Experiments conducted in this study demonstrate that the proposed method achieves high accuracy in material classification, even when the laser color is changed. The model achieved an accuracy of 98.30 % on the training set and 96.88% on the validation set. Furthermore, the model was evaluated on a set of 3000 new images for 30 different materials, achieving an F1-score of 0.9643. The proposed method provides a robust and accurate solution for material-aware laser cutting using speckle sensing.
☆ PushingBots: Collaborative Pushing via Neural Accelerated Combinatorial Hybrid Optimization
Many robots are not equipped with a manipulator and many objects are not suitable for prehensile manipulation (such as large boxes and cylinders). In these cases, pushing is a simple yet effective non-prehensile skill for robots to interact with and further change the environment. Existing work often assumes a set of predefined pushing modes and fixed-shape objects. This work tackles the general problem of controlling a robotic fleet to push collaboratively numerous arbitrary objects to respective destinations, within complex environments of cluttered and movable obstacles. It incorporates several characteristic challenges for multi-robot systems such as online task coordination under large uncertainties of cost and duration, and for contact-rich tasks such as hybrid switching among different contact modes, and under-actuation due to constrained contact forces. The proposed method is based on combinatorial hybrid optimization over dynamic task assignments and hybrid execution via sequences of pushing modes and associated forces. It consists of three main components: (I) the decomposition, ordering and rolling assignment of pushing subtasks to robot subgroups; (II) the keyframe guided hybrid search to optimize the sequence of parameterized pushing modes for each subtask; (III) the hybrid control to execute these modes and transit among them. Last but not least, a diffusion-based accelerator is adopted to predict the keyframes and pushing modes that should be prioritized during hybrid search; and further improve planning efficiency. The framework is complete under mild assumptions. Its efficiency and effectiveness under different numbers of robots and general-shaped objects are validated extensively in simulations and hardware experiments, as well as generalizations to heterogeneous robots, planar assembly and 6D pushing.
comment: 20 pages, 24 figures. Accepted to IEEE Transactions on Robotics (T-RO), 2025
☆ The Role of Consequential and Functional Sound in Human-Robot Interaction: Toward Audio Augmented Reality Interfaces
As robots become increasingly integrated into everyday environments, understanding how they communicate with humans is critical. Sound offers a powerful channel for interaction, encompassing both operational noises and intentionally designed auditory cues. In this study, we examined the effects of consequential and functional sounds on human perception and behavior, including a novel exploration of spatial sound through localization and handover tasks. Results show that consequential sounds of the Kinova Gen3 manipulator did not negatively affect perceptions, spatial localization is highly accurate for lateral cues but declines for frontal cues, and spatial sounds can simultaneously convey task-relevant information while promoting warmth and reducing discomfort. These findings highlight the potential of functional and transformative auditory design to enhance human-robot collaboration and inform future sound-based interaction strategies.
comment: 9 pages, 6 figures
☆ BOP-ASK: Object-Interaction Reasoning for Vision-Language Models
Vision Language Models (VLMs) have achieved impressive performance on spatial reasoning benchmarks, yet these evaluations mask critical weaknesses in understanding object interactions. Current benchmarks test high level relationships ('left of,' 'behind', etc.) but ignore fine-grained spatial understanding needed for real world applications: precise 3D localization, physical compatibility between objects, object affordances and multi step spatial planning. In this work, we present BOP-ASK, a novel large scale dataset for object interaction reasoning for both training and benchmarking. Our data generation pipeline leverages 6D object poses from the Benchmark for Object Pose Estimation (BOP) datasets from which we derive fine grained annotations such as grasp poses, referred object poses, path planning trajectories, relative spatial and depth relationships, and object-to-object relationships. BOP-ASK comprises over 150k images and 33M question answer pairs spanning six tasks (four novel), providing a rich resource for training and evaluating VLMs. We evaluate proprietary and open sourced VLMs, and conduct human evaluations on BOP-ASK-core, a contributed test benchmark. We also release BOP-ASK-lab, an out-of-distribution benchmark with images not sourced from BOP, enabling testing of generalization. Our experiments demonstrate that models trained on BOP-ASK outperform baselines and exhibit emergent capabilities such as precise object and grasp pose estimation, trajectory planning, and fine-grained object-centric spatial reasoning in cluttered environments. We will publicly release our datasets and dataset generation pipeline.
☆ A*-based Temporal Logic Path Planning with User Preferences on Relaxed Task Satisfaction
In this work, we consider the problem of planning for temporal logic tasks in large robot environments. When full task compliance is unattainable, we aim to achieve the best possible task satisfaction by integrating user preferences for relaxation into the planning process. Utilizing the automata-based representations for temporal logic goals and user preferences, we propose an A*-based planning framework. This approach effectively tackles large-scale problems while generating near-optimal high-level trajectories. To facilitate this, we propose a simple, efficient heuristic that allows for planning over large robot environments in a fraction of time and search memory as compared to uninformed search algorithms. We present extensive case studies to demonstrate the scalability, runtime analysis as well as empirical bounds on the suboptimality of the proposed heuristic.
♻ ☆ CleverDistiller: Simple and Spatially Consistent Cross-modal Distillation BMVC 2025
Vision foundation models (VFMs) such as DINO have led to a paradigm shift in 2D camera-based perception towards extracting generalized features to support many downstream tasks. Recent works introduce self-supervised cross-modal knowledge distillation (KD) as a way to transfer these powerful generalization capabilities into 3D LiDAR-based models. However, they either rely on highly complex distillation losses, pseudo-semantic maps, or limit KD to features useful for semantic segmentation only. In this work, we propose CleverDistiller, a self-supervised, cross-modal 2D-to-3D KD framework introducing a set of simple yet effective design choices: Unlike contrastive approaches relying on complex loss design choices, our method employs a direct feature similarity loss in combination with a multi layer perceptron (MLP) projection head to allow the 3D network to learn complex semantic dependencies throughout the projection. Crucially, our approach does not depend on pseudo-semantic maps, allowing for direct knowledge transfer from a VFM without explicit semantic supervision. Additionally, we introduce the auxiliary self-supervised spatial task of occupancy prediction to enhance the semantic knowledge, obtained from a VFM through KD, with 3D spatial reasoning capabilities. Experiments on standard autonomous driving benchmarks for 2D-to-3D KD demonstrate that CleverDistiller achieves state-of-the-art performance in both semantic segmentation and 3D object detection (3DOD) by up to 10% mIoU, especially when fine tuning on really low data amounts, showing the effectiveness of our simple yet powerful KD strategy
comment: Accepted to BMVC 2025
♻ ☆ Non-Gaited Legged Locomotion with Monte-Carlo Tree Search and Supervised Learning
Legged robots are able to navigate complex terrains by continuously interacting with the environment through careful selection of contact sequences and timings. However, the combinatorial nature behind contact planning hinders the applicability of such optimization problems on hardware. In this work, we present a novel approach that optimizes gait sequences and respective timings for legged robots in the context of optimization-based controllers through the use of sampling-based methods and supervised learning techniques. We propose to bootstrap the search by learning an optimal value function in order to speed-up the gait planning procedure making it applicable in real-time. To validate our proposed method, we showcase its performance both in simulation and on hardware using a 22 kg electric quadruped robot. The method is assessed on different terrains, under external perturbations, and in comparison to a standard control approach where the gait sequence is fixed a priori.
♻ ☆ A Continuous sEMG-Based Prosthetic Hand Control System Without Motion or Force Sensors
Regressively-based surface electromyography (sEMG) prosthetics are widely used for their ability to continuously convert muscle activity into finger force and motion. However, they typically require additional kinematic or dynamic sensors, which increases complexity and limits practical application. To address this, this paper proposes a method based on the simplified near-linear relationship between sEMG and finger force, using the near-linear model ResDD proposed in this work. By applying the principle that a line can be determined by two points, we eliminate the need for complex sensor calibration. Specifically, by recording the sEMG during maximum finger flexion and extension, and assigning corresponding forces of 1 and -1, the ResDD model can fit the simplified relationship between sEMG signals and force, enabling continuous prediction and control of finger force and gestures. Offline experiments were conducted to evaluate the model's classification accuracy and its ability to learn sufficient information. It uses interpolation analysis to open up the internal structure of the trained model and checks whether the fitted curve of the model conforms to the nearly linear relationship between sEMG and force. Finally, online control and sine wave tracking experiments were carried out to further verify the practicality of the proposed method. The results show that the method effectively extracts meaningful information from sEMG and accurately decodes them. The near-linear model sufficiently reflects the expected relationship between sEMG and finger force. Fitting this simplified near-linear relationship is adequate to achieve continuous and smooth control of finger force and gestures, confirming the feasibility and effectiveness of the proposed approach.
comment: 12 pages
♻ ☆ Risk Map As Middleware: Towards Interpretable Cooperative End-to-end Autonomous Driving for Risk-Aware Planning
End-to-end paradigm has emerged as a promising approach to autonomous driving. However, existing single-agent end-to-end pipelines are often constrained by occlusion and limited perception range, resulting in hazardous driving. Furthermore, their black-box nature prevents the interpretability of the driving behavior, leading to an untrustworthiness system. To address these limitations, we introduce Risk Map as Middleware (RiskMM) and propose an interpretable cooperative end-to-end driving framework. The risk map learns directly from the driving data and provides an interpretable spatiotemporal representation of the scenario from the upstream perception and the interactions between the ego vehicle and the surrounding environment for downstream planning. RiskMM first constructs a multi-agent spatiotemporal representation with unified Transformer-based architecture, then derives risk-aware representations by modeling interactions among surrounding environments with attention. These representations are subsequently fed into a learning-based Model Predictive Control (MPC) module. The MPC planner inherently accommodates physical constraints and different vehicle types and can provide interpretation by aligning learned parameters with explicit MPC elements. Evaluations conducted on the real-world V2XPnP-Seq dataset confirm that RiskMM achieves superior and robust performance in risk-aware trajectory planning, significantly enhancing the interpretability of the cooperative end-to-end driving framework. The codebase will be released to facilitate future research in this field.
comment: IEEE RA-L
♻ ☆ UltraDP: Generalizable Carotid Ultrasound Scanning with Force-Aware Diffusion Policy
Ultrasound scanning is a critical imaging technique for real-time, non-invasive diagnostics. However, variations in patient anatomy and complex human-in-the-loop interactions pose significant challenges for autonomous robotic scanning. Existing ultrasound scanning robots are commonly limited to relatively low generalization and inefficient data utilization. To overcome these limitations, we present UltraDP, a Diffusion-Policy-based method that receives multi-sensory inputs (ultrasound images, wrist camera images, contact wrench, and probe pose) and generates actions that are fit for multi-modal action distributions in autonomous ultrasound scanning of carotid artery. We propose a specialized guidance module to enable the policy to output actions that center the artery in ultrasound images. To ensure stable contact and safe interaction between the robot and the human subject, a hybrid force-impedance controller is utilized to drive the robot to track such trajectories. Also, we have built a large-scale training dataset for carotid scanning comprising 210 scans with 460k sample pairs from 21 volunteers of both genders. By exploring our guidance module and DP's strong generalization ability, UltraDP achieves a 95% success rate in transverse scanning on previously unseen subjects, demonstrating its effectiveness.
♻ ☆ Relative Pose Estimation for Nonholonomic Robot Formation with UWB-IO Measurements (Extended version)
This article studies the problem of distributed formation control for multiple robots by using onboard ultra wide band (UWB) distance and inertial odometer (IO) measurements. Although this problem has been widely studied, a fundamental limitation of most works is that they require each robot's pose and sensor measurements are expressed in a common reference frame. However, it is inapplicable for nonholonomic robot formations due to the practical difficulty of aligning IO measurements of individual robot in a common frame. To address this problem, firstly, a concurrent-learning based estimator is firstly proposed to achieve relative localization between neighboring robots in a local frame. Different from most relative localization methods in a global frame, both relative position and orientation in a local frame are estimated with only UWB ranging and IO measurements. Secondly, to deal with information loss caused by directed communication topology, a cooperative localization algorithm is introduced to estimate the relative pose to the leader robot. Thirdly, based on the theoretical results on relative pose estimation, a distributed formation tracking controller is proposed for nonholonomic robots. Both 3D and 2D real-world experiments conducted on aerial robots and grounded robots are provided to demonstrate the effectiveness of the proposed method.
comment: 17 pages, 26 figures
♻ ☆ Barrier-Riccati Synthesis for Nonlinear Safe Control with Expanded Region of Attraction
We present a Riccati-based framework for safety-critical nonlinear control that integrates the barrier states (BaS) methodology with the State-Dependent Riccati Equation (SDRE) approach. The BaS formulation embeds safety constraints into the system dynamics via auxiliary states, enabling safety to be treated as a control objective. To overcome the limited region of attraction in linear BaS controllers, we extend the framework to nonlinear systems using SDRE synthesis applied to the barrier-augmented dynamics and derive a matrix inequality condition that certifies forward invariance of a large region of attraction and guarantees asymptotic safe stabilization. The resulting controller is computed online via pointwise Riccati solutions. We validate the method on an unstable constrained system and cluttered quadrotor navigation tasks, demonstrating improved constraint handling, scalability, and robustness near safety boundaries. This framework offers a principled and computationally tractable solution for synthesizing nonlinear safe feedback in safety-critical environments.
♻ ☆ Statistically Assuring Safety of Control Systems using Ensembles of Safety Filters and Conformal Prediction
Safety assurance is a fundamental requirement for deploying learning-enabled autonomous systems. Hamilton-Jacobi (HJ) reachability analysis is a fundamental method for formally verifying safety and generating safe controllers. However, computing the HJ value function that characterizes the backward reachable set (BRS) of a set of user-defined failure states is computationally expensive, especially for high-dimensional systems, motivating the use of reinforcement learning approaches to approximate the value function. Unfortunately, a learned value function and its corresponding safe policy are not guaranteed to be correct. The learned value function evaluated at a given state may not be equal to the actual safety return achieved by following the learned safe policy. To address this challenge, we introduce a conformal prediction-based (CP) framework that bounds such uncertainty. We leverage CP to provide probabilistic safety guarantees when using learned HJ value functions and policies to prevent control systems from reaching failure states. Specifically, we use CP to calibrate the switching between the unsafe nominal controller and the learned HJ-based safe policy and to derive safety guarantees under this switched policy. We also investigate using an ensemble of independently trained HJ value functions as a safety filter and compare this ensemble approach to using individual value functions alone.
♻ ☆ Grounding LLMs For Robot Task Planning Using Closed-loop State Feedback
Planning algorithms decompose complex problems into intermediate steps that can be sequentially executed by robots to complete tasks. Recent works have employed Large Language Models (LLMs) for task planning, using natural language to generate robot policies in both simulation and real-world environments. LLMs like GPT-4 have shown promising results in generalizing to unseen tasks, but their applicability is limited due to hallucinations caused by insufficient grounding in the robot environment. The robustness of LLMs in task planning can be enhanced with environmental state information and feedback. In this paper, we introduce a novel approach to task planning that utilizes two separate LLMs for high-level planning and low-level control, improving task-related success rates and goal condition recall. Our algorithm, \textit{BrainBody-LLM}, draws inspiration from the human neural system, emulating its brain-body architecture by dividing planning across two LLMs in a structured, hierarchical manner. BrainBody-LLM implements a closed-loop feedback mechanism, enabling learning from simulator errors to resolve execution errors in complex settings. We demonstrate the successful application of BrainBody-LLM in the VirtualHome simulation environment, achieving a 29\% improvement in task-oriented success rates over competitive baselines with the GPT-4 backend. Additionally, we evaluate our algorithm on seven complex tasks using a realistic physics simulator and the Franka Research 3 robotic arm, comparing it with various state-of-the-art LLMs. Our results show advancements in the reasoning capabilities of recent LLMs, which enable them to learn from raw simulator/controller errors to correct plans, making them highly effective in robotic task planning.
comment: Preprint version. Accepted full paper available here: https://advanced.onlinelibrary.wiley.com/doi/10.1002/adrr.202500072
♻ ☆ Searching in Space and Time: Unified Memory-Action Loops for Open-World Object Retrieval
Service robots must retrieve objects in dynamic, open-world settings where requests may reference attributes ("the red mug"), spatial context ("the mug on the table"), or past states ("the mug that was here yesterday"). Existing approaches capture only parts of this problem: scene graphs capture spatial relations but ignore temporal grounding, temporal reasoning methods model dynamics but do not support embodied interaction, and dynamic scene graphs handle both but remain closed-world with fixed vocabularies. We present STAR (SpatioTemporal Active Retrieval), a framework that unifies memory queries and embodied actions within a single decision loop. STAR leverages non-parametric long-term memory and a working memory to support efficient recall, and uses a vision-language model to select either temporal or spatial actions at each step. We introduce STARBench, a benchmark of spatiotemporal object search tasks across simulated and real environments. Experiments in STARBench and on a Tiago robot show that STAR consistently outperforms scene-graph and memory-only baselines, demonstrating the benefits of treating search in time and search in space as a unified problem. For more information: https://amrl.cs.utexas.edu/STAR.
comment: https://amrl.cs.utexas.edu/STAR/
♻ ☆ DynoSAM: Open-Source Smoothing and Mapping Framework for Dynamic SLAM
Traditional Visual Simultaneous Localization and Mapping (vSLAM) systems focus solely on static scene structures, overlooking dynamic elements in the environment. Although effective for accurate visual odometry in complex scenarios, these methods discard crucial information about moving objects. By incorporating this information into a Dynamic SLAM framework, the motion of dynamic entities can be estimated, enhancing navigation whilst ensuring accurate localization. However, the fundamental formulation of Dynamic SLAM remains an open challenge, with no consensus on the optimal approach for accurate motion estimation within a SLAM pipeline. Therefore, we developed DynoSAM, an open-source framework for Dynamic SLAM that enables the efficient implementation, testing, and comparison of various Dynamic SLAM optimization formulations. DynoSAM integrates static and dynamic measurements into a unified optimization problem solved using factor graphs, simultaneously estimating camera poses, static scene, object motion or poses, and object structures. We evaluate DynoSAM across diverse simulated and real-world datasets, achieving state-of-the-art motion estimation in indoor and outdoor environments, with substantial improvements over existing systems. Additionally, we demonstrate DynoSAM utility in downstream applications, including 3D reconstruction of dynamic scenes and trajectory prediction, thereby showcasing potential for advancing dynamic object-aware SLAM systems. DynoSAM is open-sourced at https://github.com/ACFR-RPG/DynOSAM.
comment: 20 pages, 10 figures. Submitted to T-RO Visual SLAM SI 2025
♻ ☆ AeroVerse: UAV-Agent Benchmark Suite for Simulating, Pre-training, Finetuning, and Evaluating Aerospace Embodied World Models
Aerospace embodied intelligence aims to empower unmanned aerial vehicles (UAVs) and other aerospace platforms to achieve autonomous perception, cognition, and action, as well as egocentric active interaction with humans and the environment. The aerospace embodied world model serves as an effective means to realize the autonomous intelligence of UAVs and represents a necessary pathway toward aerospace embodied intelligence. However, existing embodied world models primarily focus on ground-level intelligent agents in indoor scenarios, while research on UAV intelligent agents remains unexplored. To address this gap, we construct the first large-scale real-world image-text pre-training dataset, AerialAgent-Ego10k, featuring urban drones from a first-person perspective. We also create a virtual image-text-pose alignment dataset, CyberAgent Ego500k, to facilitate the pre-training of the aerospace embodied world model. For the first time, we clearly define 5 downstream tasks, i.e., aerospace embodied scene awareness, spatial reasoning, navigational exploration, task planning, and motion decision, and construct corresponding instruction datasets, i.e., SkyAgent-Scene3k, SkyAgent-Reason3k, SkyAgent-Nav3k and SkyAgent-Plan3k, and SkyAgent-Act3k, for fine-tuning the aerospace embodiment world model. Simultaneously, we develop SkyAgentEval, the downstream task evaluation metrics based on GPT-4, to comprehensively, flexibly, and objectively assess the results, revealing the potential and limitations of 2D/3D visual language models in UAV-agent tasks. Furthermore, we integrate over 10 2D/3D visual-language models, 2 pre-training datasets, 5 finetuning datasets, more than 10 evaluation metrics, and a simulator into the benchmark suite, i.e., AeroVerse, which will be released to the community to promote exploration and development of aerospace embodied intelligence.
Robotics 51
☆ In-N-On: Scaling Egocentric Manipulation with in-the-wild and on-task Data
Egocentric videos are a valuable and scalable data source to learn manipulation policies. However, due to significant data heterogeneity, most existing approaches utilize human data for simple pre-training, which does not unlock its full potential. This paper first provides a scalable recipe for collecting and using egocentric data by categorizing human data into two categories: in-the-wild and on-task alongside with systematic analysis on how to use the data. We first curate a dataset, PHSD, which contains over 1,000 hours of diverse in-the-wild egocentric data and over 20 hours of on-task data directly aligned to the target manipulation tasks. This enables learning a large egocentric language-conditioned flow matching policy, Human0. With domain adaptation techniques, Human0 minimizes the gap between humans and humanoids. Empirically, we show Human0 achieves several novel properties from scaling human data, including language following of instructions from only human data, few-shot learning, and improved robustness using on-task data. Project website: https://xiongyicai.github.io/In-N-On/
comment: Project webpage: https://xiongyicai.github.io/In-N-On/
☆ MambaIO: Global-Coordinate Inertial Odometry for Pedestrians via Multi-Scale Frequency-Decoupled Modeling
Inertial Odometry (IO) enables real-time localization using only acceleration and angular velocity measurements from an Inertial Measurement Unit (IMU), making it a promising solution for localization in consumer-grade applications. Traditionally, IMU measurements in IO have been processed under two coordinate system paradigms: the body coordinate frame and the global coordinate frame, with the latter being widely adopted. However, recent studies in drone scenarios have demonstrated that the body frame can significantly improve localization accuracy, prompting a re-evaluation of the suitability of the global frame for pedestrian IO. To address this issue, this paper systematically evaluates the effectiveness of the global coordinate frame in pedestrian IO through theoretical analysis, qualitative inspection, and quantitative experiments. Building upon these findings, we further propose MambaIO, which decomposes IMU measurements into high-frequency and low-frequency components using a Laplacian pyramid. The low-frequency component is processed by a Mamba architecture to extract implicit contextual motion cues, while the high-frequency component is handled by a convolutional structure to capture fine-grained local motion details. Experiments on multiple public datasets show that MambaIO substantially reduces localization error and achieves state-of-the-art (SOTA) performance. To the best of our knowledge, this is the first application of the Mamba architecture to the inertial odometry task.
☆ Optimus-Q: Utilizing Federated Learning in Adaptive Robots for Intelligent Nuclear Power Plant Operations through Quantum Cryptography
The integration of advanced robotics in nuclear power plants (NPPs) presents a transformative opportunity to enhance safety, efficiency, and environmental monitoring in high-stakes environments. Our paper introduces the Optimus-Q robot, a sophisticated system designed to autonomously monitor air quality and detect contamination while leveraging adaptive learning techniques and secure quantum communication. Equipped with advanced infrared sensors, the Optimus-Q robot continuously streams real-time environmental data to predict hazardous gas emissions, including carbon dioxide (CO$_2$), carbon monoxide (CO), and methane (CH$_4$). Utilizing a federated learning approach, the robot collaborates with other systems across various NPPs to improve its predictive capabilities without compromising data privacy. Additionally, the implementation of Quantum Key Distribution (QKD) ensures secure data transmission, safeguarding sensitive operational information. Our methodology combines systematic navigation patterns with machine learning algorithms to facilitate efficient coverage of designated areas, thereby optimizing contamination monitoring processes. Through simulations and real-world experiments, we demonstrate the effectiveness of the Optimus-Q robot in enhancing operational safety and responsiveness in nuclear facilities. This research underscores the potential of integrating robotics, machine learning, and quantum technologies to revolutionize monitoring systems in hazardous environments.
☆ SRPO: Self-Referential Policy Optimization for Vision-Language-Action Models
Vision-Language-Action (VLA) models excel in robotic manipulation but are constrained by their heavy reliance on expert demonstrations, leading to demonstration bias and limiting performance. Reinforcement learning (RL) is a vital post-training strategy to overcome these limits, yet current VLA-RL methods, including group-based optimization approaches, are crippled by severe reward sparsity. Relying on binary success indicators wastes valuable information in failed trajectories, resulting in low training efficiency. To solve this, we propose Self-Referential Policy Optimization (SRPO), a novel VLA-RL framework. SRPO eliminates the need for external demonstrations or manual reward engineering by leveraging the model's own successful trajectories, generated within the current training batch, as a self-reference. This allows us to assign a progress-wise reward to failed attempts. A core innovation is the use of latent world representations to measure behavioral progress robustly. Instead of relying on raw pixels or requiring domain-specific fine-tuning, we utilize the compressed, transferable encodings from a world model's latent space. These representations naturally capture progress patterns across environments, enabling accurate, generalized trajectory comparison. Empirical evaluations on the LIBERO benchmark demonstrate SRPO's efficiency and effectiveness. Starting from a supervised baseline with 48.9% success, SRPO achieves a new state-of-the-art success rate of 99.2% in just 200 RL steps, representing a 103% relative improvement without any extra supervision. Furthermore, SRPO shows substantial robustness, achieving a 167% performance improvement on the LIBERO-Plus benchmark.
☆ UltraDP: Generalizable Carotid Ultrasound Scanning with Force-Aware Diffusion Policy
Ultrasound scanning is a critical imaging technique for real-time, non-invasive diagnostics. However, variations in patient anatomy and complex human-in-the-loop interactions pose significant challenges for autonomous robotic scanning. Existing ultrasound scanning robots are commonly limited to relatively low generalization and inefficient data utilization. To overcome these limitations, we present UltraDP, a Diffusion-Policy-based method that receives multi-sensory inputs (ultrasound images, wrist camera images, contact wrench, and probe pose) and generates actions that are fit for multi-modal action distributions in autonomous ultrasound scanning of carotid artery. We propose a specialized guidance module to enable the policy to output actions that center the artery in ultrasound images. To ensure stable contact and safe interaction between the robot and the human subject, a hybrid force-impedance controller is utilized to drive the robot to track such trajectories. Also, we have built a large-scale training dataset for carotid scanning comprising 210 scans with 460k sample pairs from 21 volunteers of both genders. By exploring our guidance module and DP's strong generalization ability, UltraDP achieves a 95% success rate in transverse scanning on previously unseen subjects, demonstrating its effectiveness.
☆ NMPC-based Motion Planning with Adaptive Weighting for Dynamic Object Interception
Catching fast-moving objects serves as a benchmark for robotic agility, posing significant coordination challenges for cooperative manipulator systems holding a catcher, particularly due to inherent closed-chain constraints. This paper presents a nonlinear model predictive control (MPC)-based motion planner that bridges high-level interception planning with real-time joint space control, enabling dynamic object interception for systems comprising two cooperating arms. We introduce an Adaptive- Terminal (AT) MPC formulation featuring cost shaping, which contrasts with a simpler Primitive-Terminal (PT) approach relying heavily on terminal penalties for rapid convergence. The proposed AT formulation is shown to effectively mitigate issues related to actuator power limit violations frequently encountered with the PT strategy, yielding trajectories and significantly reduced control effort. Experimental results on a robotic platform with two cooperative arms, demonstrating excellent real time performance, with an average planner cycle computation time of approximately 19 ms-less than half the 40 ms system sampling time. These results indicate that the AT formulation achieves significantly improved motion quality and robustness with minimal computational overhead compared to the PT baseline, making it well-suited for dynamic, cooperative interception tasks.
comment: This work has been submitted to the IFAC World Congress for possible publication. Under review
☆ Decentralized Gaussian Process Classification and an Application in Subsea Robotics IROS 2025
Teams of cooperating autonomous underwater vehicles (AUVs) rely on acoustic communication for coordination, yet this communication medium is constrained by limited range, multi-path effects, and low bandwidth. One way to address the uncertainty associated with acoustic communication is to learn the communication environment in real-time. We address the challenge of a team of robots building a map of the probability of communication success from one location to another in real-time. This is a decentralized classification problem -- communication events are either successful or unsuccessful -- where AUVs share a subset of their communication measurements to build the map. The main contribution of this work is a rigorously derived data sharing policy that selects measurements to be shared among AUVs. We experimentally validate our proposed sharing policy using real acoustic communication data collected from teams of Virginia Tech 690 AUVs, demonstrating its effectiveness in underwater environments.
comment: 8 pages, 8 figures, IROS 2025 conference
☆ PCARNN-DCBF: Minimal-Intervention Geofence Enforcement for Ground Vehicles
Runtime geofencing for ground vehicles is rapidly emerging as a critical technology for enforcing Operational Design Domains (ODDs). However, existing solutions struggle to reconcile high-fidelity learning with the structural requirements of verifiable control. We address this by introducing PCARNN-DCBF, a novel pipeline integrating a Physics-encoded Control-Affine Residual Neural Network with a preview-based Discrete Control Barrier Function. Unlike generic learned models, PCARNN explicitly preserves the control-affine structure of vehicle dynamics, ensuring the linearity required for reliable optimization. This enables the DCBF to enforce polygonal keep-in constraints via a real-time Quadratic Program (QP) that handles high relative degree and mitigates actuator saturation. Experiments in CARLA across electric and combustion platforms demonstrate that this structure-preserving approach significantly outperforms analytical and unstructured neural baselines.
☆ Theoretical Closed-loop Stability Bounds for Dynamical System Coupled with Diffusion Policies
Diffusion Policy has shown great performance in robotic manipulation tasks under stochastic perturbations, due to its ability to model multimodal action distributions. Nonetheless, its reliance on a computationally expensive reverse-time diffusion (denoising) process, for action inference, makes it challenging to use for real-time applications where quick decision-making is mandatory. This work studies the possibility of conducting the denoising process only partially before executing an action, allowing the plant to evolve according to its dynamics in parallel to the reverse-time diffusion dynamics ongoing on the computer. In a classical diffusion policy setting, the plant dynamics are usually slow and the two dynamical processes are uncoupled. Here, we investigate theoretical bounds on the stability of closed-loop systems using diffusion policies when the plant dynamics and the denoising dynamics are coupled. The contribution of this work gives a framework for faster imitation learning and a metric that yields if a controller will be stable based on the variance of the demonstrations.
comment: 5 pages, 3 figures
☆ Discovering Optimal Natural Gaits of Dissipative Systems via Virtual Energy Injection
Legged robots offer several advantages when navigating unstructured environments, but they often fall short of the efficiency achieved by wheeled robots. One promising strategy to improve their energy economy is to leverage their natural (unactuated) dynamics using elastic elements. This work explores that concept by designing energy-optimal control inputs through a unified, multi-stage framework. It starts with a novel energy injection technique to identify passive motion patterns by harnessing the system's natural dynamics. This enables the discovery of passive solutions even in systems with energy dissipation caused by factors such as friction or plastic collisions. Building on these passive solutions, we then employ a continuation approach to derive energy-optimal control inputs for the fully actuated, dissipative robotic system. The method is tested on simulated models to demonstrate its applicability in both single- and multi-legged robotic systems. This analysis provides valuable insights into the design and operation of elastic legged robots, offering pathways to improve their efficiency and adaptability by exploiting the natural system dynamics.
comment: Preprint Version, IEEE Robotics and Automation Letters (RA-L), accepted November 2025
☆ RRT*former: Environment-Aware Sampling-Based Motion Planning using Transformer IROS 2025
We investigate the sampling-based optimal path planning problem for robotics in complex and dynamic environments. Most existing sampling-based algorithms neglect environmental information or the information from previous samples. Yet, these pieces of information are highly informative, as leveraging them can provide better heuristics when sampling the next state. In this paper, we propose a novel sampling-based planning algorithm, called \emph{RRT*former}, which integrates the standard RRT* algorithm with a Transformer network in a novel way. Specifically, the Transformer is used to extract features from the environment and leverage information from previous samples to better guide the sampling process. Our extensive experiments demonstrate that, compared to existing sampling-based approaches such as RRT*, Neural RRT*, and their variants, our algorithm achieves considerable improvements in both the optimality of the path and sampling efficiency. The code for our implementation is available on https://github.com/fengmingyang666/RRTformer.
comment: Accepted to IROS 2025
☆ Platform-Agnostic Reinforcement Learning Framework for Safe Exploration of Cluttered Environments with Graph Attention
Autonomous exploration of obstacle-rich spaces requires strategies that ensure efficiency while guaranteeing safety against collisions with obstacles. This paper investigates a novel platform-agnostic reinforcement learning framework that integrates a graph neural network-based policy for next-waypoint selection, with a safety filter ensuring safe mobility. Specifically, the neural network is trained using reinforcement learning through the Proximal Policy Optimization (PPO) algorithm to maximize exploration efficiency while minimizing safety filter interventions. Henceforth, when the policy proposes an infeasible action, the safety filter overrides it with the closest feasible alternative, ensuring consistent system behavior. In addition, this paper introduces a reward function shaped by a potential field that accounts for both the agent's proximity to unexplored regions and the expected information gain from reaching them. The proposed framework combines the adaptability of reinforcement learning-based exploration policies with the reliability provided by explicit safety mechanisms. This feature plays a key role in enabling the deployment of learning-based policies on robotic platforms operating in real-world environments. Extensive evaluations in both simulations and experiments performed in a lab environment demonstrate that the approach achieves efficient and safe exploration in cluttered spaces.
comment: 8 pages, 6 figures, submitted to the 2026 IEEE International Conference on Robotics & Automation
☆ Fast Post-Hoc Confidence Fusion for 3-Class Open-Set Aerial Object Detection
Developing reliable UAV navigation systems requires robust air-to-air object detectors capable of distinguishing between objects seen during training and previously unseen objects. While many methods address closed-set detection and achieve high-confidence recognition of in-domain (ID) targets, they generally do not tackle open-set detection, which requires simultaneous handling of both ID and out-of-distribution (OOD) objects. Existing open-set approaches typically rely on a single uncertainty score with thresholding, limiting flexibility and often conflating OOD objects with background clutter. In contrast, we propose a lightweight, model-agnostic post-processing framework that explicitly separates background from unknown objects while preserving the base detector's performance. Our approach extends open-set detection beyond binary ID/OOD classification to real-time three-way classification among ID targets, OOD objects, and background. To this end, we employ a fusion scheme that aggregates multiple confidence estimates and per-detection features using a compact multilayer perceptron (MLP). Incorporating different logit variants into the MLP consistently enhances performance across both binary and three-class classification without compromising throughput. Extensive ablation and comparative experiments confirm that our method surpasses threshold-based baselines in two-class classification by an average of 2.7% AUROC, while retaining or improving open-set mAP. Furthermore, our study uniquely enables robust three-class classification, a critical capability for safe UAV navigation, where OOD objects must be actively avoided and background regions safely ignored. Comparative analysis highlights that our method surpasses competitive techniques in AUROC across datasets, while improving closed-set mAP by up to 9 points, an 18% relative gain.
☆ C2F-Space: Coarse-to-Fine Space Grounding for Spatial Instructions using Vision-Language Models
Space grounding refers to localizing a set of spatial references described in natural language instructions. Traditional methods often fail to account for complex reasoning -- such as distance, geometry, and inter-object relationships -- while vision-language models (VLMs), despite strong reasoning abilities, struggle to produce a fine-grained region of outputs. To overcome these limitations, we propose C2F-Space, a novel coarse-to-fine space-grounding framework that (i) estimates an approximated yet spatially consistent region using a VLM, then (ii) refines the region to align with the local environment through superpixelization. For the coarse estimation, we design a grid-based visual-grounding prompt with a propose-validate strategy, maximizing VLM's spatial understanding and yielding physically and semantically valid canonical region (i.e., ellipses). For the refinement, we locally adapt the region to surrounding environment without over-relaxed to free space. We construct a new space-grounding benchmark and compare C2F-Space with five state-of-the-art baselines using success rate and intersection-over-union. Our C2F-Space significantly outperforms all baselines. Our ablation study confirms the effectiveness of each module in the two-step process and their synergistic effect of the combined framework. We finally demonstrate the applicability of C2F-Space to simulated robotic pick-and-place tasks.
comment: 16 pages, 12 figures
☆ MSA - Technique for Stiffness Modeling of Manipulators with Complex and Hybrid Structures
The paper presents a systematic approach for stiffness modeling of manipulators with complex and hybrid structures using matrix structural analysis. In contrast to previous results, it is suitable for mixed architectures containing closed-loops, flexible links, rigid connections, passive and elastic joints with external loadings and preloadings. The proposed approach produces the Cartesian stiffness matrices in a semi-analytical manner. It presents the manipulator stiffness model as a set of conventional equations describing the link elasticities that are supplemented by a set of constraints describing connections between links. Its allows user straightforward aggregation of stiffness model equations avoiding traditional column/row merging procedures in the extended stiffness matrix. Advantages of this approach are illustrated by stiffness analysis of NaVaRo manipulator.
☆ Optimizing Robot Positioning Against Placement Inaccuracies: A Study on the Fanuc CRX10iA/L
This study presents a methodology for determining the optimal base placement of a Fanuc CRX10iA/L collaborative robot for a desired trajectory corresponding to an industrial task. The proposed method uses a particle swarm optimization algorithm that explores the search space to find positions for performing the trajectory. An $α$-shape algorithm is then used to draw the borders of the feasibility areas, and the largest circle inscribed is calculated from the Voronoi diagrams. The aim of this approach is to provide a robustness criterion in the context of robot placement inaccuracies that may be encountered, for example, if the robot is placed on a mobile base when the system is deployed by an operator. The approach developed uses an inverse kinematics model to evaluate all initial configurations, then moves the robot end-effector along the reference trajectory using the Jacobian matrix and assigns a score to the attempt. For the Fanuc CRX10iA/L robot, there can be up to 16 solutions to the inverse kinematics model. The calculation of these solutions is not trivial and requires a specific study that planning tools such as MoveIt cannot fully take into account. Additionally, the optimization process must consider constraints such as joint limits, singularities, and workspace limitations to ensure feasible and efficient trajectory execution.
☆ Path Planning through Multi-Agent Reinforcement Learning in Dynamic Environments
Path planning in dynamic environments is a fundamental challenge in intelligent transportation and robotics, where obstacles and conditions change over time, introducing uncertainty and requiring continuous adaptation. While existing approaches often assume complete environmental unpredictability or rely on global planners, these assumptions limit scalability and practical deployment in real-world settings. In this paper, we propose a scalable, region-aware reinforcement learning (RL) framework for path planning in dynamic environments. Our method builds on the observation that environmental changes, although dynamic, are often localized within bounded regions. To exploit this, we introduce a hierarchical decomposition of the environment and deploy distributed RL agents that adapt to changes locally. We further propose a retraining mechanism based on sub-environment success rates to determine when policy updates are necessary. Two training paradigms are explored: single-agent Q-learning and multi-agent federated Q-learning, where local Q-tables are aggregated periodically to accelerate the learning process. Unlike prior work, we evaluate our methods in more realistic settings, where multiple simultaneous obstacle changes and increasing difficulty levels are present. Results show that the federated variants consistently outperform their single-agent counterparts and closely approach the performance of A* Oracle while maintaining shorter adaptation times and robust scalability. Although initial training remains time-consuming in large environments, our decentralized framework eliminates the need for a global planner and lays the groundwork for future improvements using deep RL and flexible environment decomposition.
☆ Look, Zoom, Understand: The Robotic Eyeball for Embodied Perception
In embodied AI perception systems, visual perception should be active: the goal is not to passively process static images, but to actively acquire more informative data within pixel and spatial budget constraints. Existing vision models and fixed RGB-D camera systems fundamentally fail to reconcile wide-area coverage with fine-grained detail acquisition, severely limiting their efficacy in open-world robotic applications. To address this issue, we propose EyeVLA, a robotic eyeball for active visual perception that can take proactive actions based on instructions, enabling clear observation of fine-grained target objects and detailed information across a wide spatial extent. EyeVLA discretizes action behaviors into action tokens and integrates them with vision-language models (VLMs) that possess strong open-world understanding capabilities, enabling joint modeling of vision, language, and actions within a single autoregressive sequence. By using the 2D bounding box coordinates to guide the reasoning chain and applying reinforcement learning to refine the viewpoint selection policy, we transfer the open-world scene understanding capability of the VLM to a vision language action (VLA) policy using only minimal real-world data. Experiments show that our system efficiently performs instructed scenes in real-world environments and actively acquires more accurate visual information through instruction-driven actions of rotation and zoom, thereby achieving strong environmental perception capabilities. EyeVLA introduces a novel robotic vision system that leverages detailed and spatially rich, large-scale embodied data, and actively acquires highly informative visual observations for downstream embodied tasks.
☆ Behavior Trees vs Executable Ontologies: a Comparative Analysis of Robot Control Paradigms
This paper compares two distinct approaches to modeling robotic behavior: imperative Behavior Trees (BTs) and declarative Executable Ontologies (EO), implemented through the boldsea framework. BTs structure behavior hierarchically using control-flow, whereas EO represents the domain as a temporal, event-based semantic graph driven by dataflow rules. We demonstrate that EO achieves comparable reactivity and modularity to BTs through a fundamentally different architecture: replacing polling-based tick execution with event-driven state propagation. We propose that EO offers an alternative framework, moving from procedural programming to semantic domain modeling, to address the semantic-process gap in traditional robotic control. EO supports runtime model modification, full temporal traceability, and a unified representation of data, logic, and interface - features that are difficult or sometimes impossible to achieve with BTs, although BTs excel in established, predictable scenarios. The comparison is grounded in a practical mobile manipulation task. This comparison highlights the respective operational strengths of each approach in dynamic, evolving robotic systems.
comment: 22 pages, 8 figures
☆ Symmetry-Breaking in Multi-Agent Navigation: Winding Number-Aware MPC with a Learned Topological Strategy
We address the fundamental challenge of resolving symmetry-induced deadlocks in distributed multi-agent navigation by proposing a new hierarchical navigation method. When multiple agents interact, it is inherently difficult for them to autonomously break the symmetry of deciding how to pass each other. To tackle this problem, we introduce an approach that quantifies cooperative symmetry-breaking strategies using a topological invariant called the winding number, and learns the strategies themselves through reinforcement learning. Our method features a hierarchical policy consisting of a learning-based Planner, which plans topological cooperative strategies, and a model-based Controller, which executes them. Through reinforcement learning, the Planner learns to produce two types of parameters for the Controller: one is the topological cooperative strategy represented by winding numbers, and the other is a set of dynamic weights that determine which agent interaction to prioritize in dense scenarios where multiple agents cross simultaneously. The Controller then generates collision-free and efficient motions based on the strategy and weights provided by the Planner. This hierarchical structure combines the flexible decision-making ability of learning-based methods with the reliability of model-based approaches. Simulation and real-world robot experiments demonstrate that our method outperforms existing baselines, particularly in dense environments, by efficiently avoiding collisions and deadlocks while achieving superior navigation performance. The code for the experiments is available at https://github.com/omron-sinicx/WNumMPC.
comment: 11 pages, 5 figures
☆ Modelling and Model-Checking a ROS2 Multi-Robot System using Timed Rebeca
Model-based development enables quicker prototyping, earlier experimentation and validation of design intents. For a multi-agent system with complex asynchronous interactions and concurrency, formal verification, model-checking in particular, offers an automated mechanism for verifying desired properties. Timed Rebeca is an actor-based modelling language supporting reactive, concurrent and time semantics, accompanied with a model-checking compiler. These capabilities allow using Timed Rebeca to correctly model ROS2 node topographies, recurring physical signals, motion primitives and other timed and time-convertible behaviors. The biggest challenges in modelling and verifying a multi-robot system lie in abstracting complex information, bridging the gap between a discrete model and a continuous system and compacting the state space, while maintaining the model's accuracy. We develop different discretization strategies for different kinds of information, identifying the 'enough' thresholds of abstraction, and applying efficient optimization techniques to boost computations. With this work we demonstrate how to use models to design and verify a multi-robot system, how to discretely model a continuous system to do model-checking efficiently, and the round-trip engineering flow between the model and the implementation. The released Rebeca and ROS2 codes can serve as a foundation for modelling multiple autonomous robots systems.
☆ A Class of Dual-Frame Passively-Tilting Fully-Actuated Hexacopter
This paper proposed a novel fully-actuated hexacopter. It features a dual-frame passive tilting structure and achieves independent control of translational motion and attitude with minimal actuators. Compared to previous fully-actuated UAVs, it liminates internal force cancellation, resulting in higher flight efficiency and endurance under equivalent payload conditions. Based on the dynamic model of fully-actuated hexacopter, a full-actuation controller is designed to achieve efficient and stable control. Finally, simulation is conducted, validating the superior fully-actuated motion capability of fully-actuated hexacopter and the effectiveness of the proposed control strategy.
☆ VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation
A key barrier to the real-world deployment of humanoid robots is the lack of autonomous loco-manipulation skills. We introduce VIRAL, a visual sim-to-real framework that learns humanoid loco-manipulation entirely in simulation and deploys it zero-shot to real hardware. VIRAL follows a teacher-student design: a privileged RL teacher, operating on full state, learns long-horizon loco-manipulation using a delta action space and reference state initialization. A vision-based student policy is then distilled from the teacher via large-scale simulation with tiled rendering, trained with a mixture of online DAgger and behavior cloning. We find that compute scale is critical: scaling simulation to tens of GPUs (up to 64) makes both teacher and student training reliable, while low-compute regimes often fail. To bridge the sim-to-real gap, VIRAL combines large-scale visual domain randomization over lighting, materials, camera parameters, image quality, and sensor delays--with real-to-sim alignment of the dexterous hands and cameras. Deployed on a Unitree G1 humanoid, the resulting RGB-based policy performs continuous loco-manipulation for up to 54 cycles, generalizing to diverse spatial and appearance variations without any real-world fine-tuning, and approaching expert-level teleoperation performance. Extensive ablations dissect the key design choices required to make RGB-based humanoid loco-manipulation work in practice.
comment: Project website: https://viral-humanoid.github.io/
☆ Eq.Bot: Enhance Robotic Manipulation Learning via Group Equivariant Canonicalization
Robotic manipulation systems are increasingly deployed across diverse domains. Yet existing multi-modal learning frameworks lack inherent guarantees of geometric consistency, struggling to handle spatial transformations such as rotations and translations. While recent works attempt to introduce equivariance through bespoke architectural modifications, these methods suffer from high implementation complexity, computational cost, and poor portability. Inspired by human cognitive processes in spatial reasoning, we propose Eq.Bot, a universal canonicalization framework grounded in SE(2) group equivariant theory for robotic manipulation learning. Our framework transforms observations into a canonical space, applies an existing policy, and maps the resulting actions back to the original space. As a model-agnostic solution, Eq.Bot aims to endow models with spatial equivariance without requiring architectural modifications. Extensive experiments demonstrate the superiority of Eq.Bot under both CNN-based (e.g., CLIPort) and Transformer-based (e.g., OpenVLA-OFT) architectures over existing methods on various robotic manipulation tasks, where the most significant improvement can reach 50.0%.
comment: 12 pages, 4 figures and 3 tables
☆ Nonholonomic Robot Parking by Feedback -- Part I: Modular Strict CLF Designs
It has been known in the robotics literature since about 1995 that, in polar coordinates, the nonholonomic unicycle is asymptotically stabilizable by smooth feedback, even globally. We introduce a modular design framework that selects the forward velocity to decouple the radial coordinate, allowing the steering subsystem to be stabilized independently. Within this structure, we develop families of feedback laws using passivity, backstepping, and integrator forwarding. Each law is accompanied by a strict control Lyapunov function, including barrier variants that enforce angular constraints. These strict CLFs provide constructive class KL convergence estimates and enable eigenvalue assignment at the target equilibrium. The framework generalizes and extends prior modular and nonmodular approaches, while preparing the ground for inverse optimal and adaptive redesigns in the sequel paper.
☆ Painted Heart Beats ICRA 2025
In this work we present AURA, a framework for synergistic human-artist painting. We developed a robot arm that collaboratively paints with a human artist. The robot has an awareness of the artist's heartbeat through the EmotiBit sensor, which provides the arousal levels of the painter. Given the heartbeat detected, the robot decides to increase proximity to the artist's workspace or retract. If a higher heartbeat is detected, which is associated with increased arousal in human artists, the robot will move away from that area of the canvas. If the artist's heart rate is detected as neutral, indicating the human artist's baseline state, the robot will continue its painting actions across the entire canvas. We also demonstrate and propose alternative robot-artist interactions using natural language and physical touch. This work combines the biometrics of a human artist to inform fluent artistic interactions.
comment: 4 pages, 2 figures, ICRA 2025
☆ Learning Human-Like RL Agents Through Trajectory Optimization With Action Quantization NeurIPS 2025
Human-like agents have long been one of the goals in pursuing artificial intelligence. Although reinforcement learning (RL) has achieved superhuman performance in many domains, relatively little attention has been focused on designing human-like RL agents. As a result, many reward-driven RL agents often exhibit unnatural behaviors compared to humans, raising concerns for both interpretability and trustworthiness. To achieve human-like behavior in RL, this paper first formulates human-likeness as trajectory optimization, where the objective is to find an action sequence that closely aligns with human behavior while also maximizing rewards, and adapts the classic receding-horizon control to human-like learning as a tractable and efficient implementation. To achieve this, we introduce Macro Action Quantization (MAQ), a human-like RL framework that distills human demonstrations into macro actions via Vector-Quantized VAE. Experiments on D4RL Adroit benchmarks show that MAQ significantly improves human-likeness, increasing trajectory similarity scores, and achieving the highest human-likeness rankings among all RL agents in the human evaluation study. Our results also demonstrate that MAQ can be easily integrated into various off-the-shelf RL algorithms, opening a promising direction for learning human-like RL agents. Our code is available at https://rlg.iis.sinica.edu.tw/papers/MAQ.
comment: Accepted by the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025)
☆ Lie Group Control Architectures for UAVs: a Comparison of SE2(3)-Based Approaches in Simulation and Hardware
This paper presents the integration and experimental validation of advanced control strategies for quadcopters based on Lie groups. We build upon recent theoretical developments on SE2(3)-based controllers and introduce a novel SE2(3) model predictive controller (MPC) that combines the predictive capabilities and constraint-handling of optimal control with the geometric properties of Lie group formulations. We evaluated this MPC against a state-of-the-art SE2(3)-based LQR approach and obtained comparable performance in simulation. Both controllers where also deployed on the Quanser QDrone platform and compared to each other and an industry standard control architecture. Results show that the SE_2(3) MPC achieves superior trajectory tracking performance and robustness across a range of scenarios. This work demonstrates the practical effectiveness of Lie group-based controllers and offers comparative insights into their impact on system behaviour and real-time performance
☆ Communication-Aware Asynchronous Distributed Trajectory Optimization for UAV Swarm
Distributed optimization offers a promising paradigm for trajectory planning in Unmanned Aerial Vehicle (UAV) swarms, yet its deployment in communication-constrained environments remains challenging due to unreliable links and limited data exchange. This paper addresses this issue via a two-tier architecture explicitly designed for operation under communication constraints. We develop a Communication-Aware Asynchronous Distributed Trajectory Optimization (CA-ADTO) framework that integrates Parameterized Differential Dynamic Programming (PDDP) for local trajectory optimization of individual UAVs with an asynchronous Alternating Direction Method of Multipliers (async-ADMM) for swarm-level coordination. The proposed architecture enables fully distributed optimization while substantially reducing communication overhead, making it suitable for real-world scenarios in which reliable connectivity cannot be guaranteed. The method is particularly effective in handling nonlinear dynamics and spatio-temporal coupling under communication constraints.
☆ An Alignment-Based Approach to Learning Motions from Demonstrations
Learning from Demonstration (LfD) has shown to provide robots with fundamental motion skills for a variety of domains. Various branches of LfD research (e.g., learned dynamical systems and movement primitives) can generally be classified into ''time-dependent'' or ''time-independent'' systems. Each provides fundamental benefits and drawbacks -- time-independent methods cannot learn overlapping trajectories, while time-dependence can result in undesirable behavior under perturbation. This paper introduces Cluster Alignment for Learned Motions (CALM), an LfD framework dependent upon an alignment with a representative ''mean" trajectory of demonstrated motions rather than pure time- or state-dependence. We discuss the convergence properties of CALM, introduce an alignment technique able to handle the shifts in alignment possible under perturbation, and utilize demonstration clustering to generate multi-modal behavior. We show how CALM mitigates the drawbacks of time-dependent and time-independent techniques on 2D datasets and implement our system on a 7-DoF robot learning tasks in three domains.
comment: 8 pages, 8 figures, originally published in the IEEE Robotics and Automation Letters
☆ I've Changed My Mind: Robots Adapting to Changing Human Goals during Collaboration
For effective human-robot collaboration, a robot must align its actions with human goals, even as they change mid-task. Prior approaches often assume fixed goals, reducing goal prediction to a one-time inference. However, in real-world scenarios, humans frequently shift goals, making it challenging for robots to adapt without explicit communication. We propose a method for detecting goal changes by tracking multiple candidate action sequences and verifying their plausibility against a policy bank. Upon detecting a change, the robot refines its belief in relevant past actions and constructs Receding Horizon Planning (RHP) trees to actively select actions that assist the human while encouraging Differentiating Actions to reveal their updated goal. We evaluate our approach in a collaborative cooking environment with up to 30 unique recipes and compare it to three comparable human goal prediction algorithms. Our method outperforms all baselines, quickly converging to the correct goal after a switch, reducing task completion time, and improving collaboration efficiency.
comment: Accepted to RA-L
☆ Gimballed Rotor Mechanism for Omnidirectional Quadrotors
This paper presents the design of a gimballed rotor mechanism as a modular and efficient solution for constructing omnidirectional quadrotors. Unlike conventional quadrotors, which are underactuated, this class of quadrotors achieves full actuation, enabling independent motion in all six degrees of freedom. While existing omnidirectional quadrotor designs often require significant structural modifications, the proposed gimballed rotor system maintains a lightweight and easy-to-integrate design by incorporating servo motors within the rotor platforms, allowing independent tilting of each rotor without major alterations to the central structure of a quadrotor. To accommodate this unconventional design, we develop a new control allocation scheme in PX4 Autopilot and present successful flight tests, validating the effectiveness of the proposed approach.
comment: 6 pages, 7 figures, CASE 2025
♻ ☆ Inverse k-visibility for RSSI-based Indoor Geometric Mapping
In recent years, the increased availability of WiFi in indoor environments has gained interest in the robotics community to utilize WiFi signals for indoor simultaneous localization and mapping algorithms. This paper discusses the challenges of achieving high-accuracy geometric map building using WiFi signals. The paper introduces the concept of inverse k-visibility, developed from the k-visibility algorithm, to identify free space in an unknown environment, used for planning, navigation, and obstacle avoidance. Comprehensive experiments, including those utilizing single and multiple RSSI signals, were conducted in both simulated and real-world environments to demonstrate the robustness of the proposed algorithm. Additionally, a detailed analysis comparing the resulting maps with ground-truth LiDAR-based maps is provided to highlight the algorithm's accuracy and reliability.
♻ ☆ Optimizing the flight path for a scouting Uncrewed Aerial Vehicle
Post-disaster situations pose unique navigation challenges. One of those challenges is the unstructured nature of the environment, which makes it hard to layout paths for rescue vehicles. We propose the use of Uncrewed Aerial Vehicle (UAV) in such scenario to perform reconnaissance across the environment. To accomplish this, we propose an optimization-based approach to plan a path for the UAV at optimal height where the sensors of the UAV can cover the most area and collect data with minimum uncertainty.
comment: This paper was prepared as an end of semester project for ME8710: Engineering Optimization, Clemson University. Consists of 7 pages and 8 figures
♻ ☆ Class-Aware PillarMix: Can Mixed Sample Data Augmentation Enhance 3D Object Detection with Radar Point Clouds? IROS 2025
Due to the significant effort required for data collection and annotation in 3D perception tasks, mixed sample data augmentation (MSDA) has been widely studied to generate diverse training samples by mixing existing data. Recently, many MSDA techniques have been developed for point clouds, but they mainly target LiDAR data, leaving their application to radar point clouds largely unexplored. In this paper, we examine the feasibility of applying existing MSDA methods to radar point clouds and identify several challenges in adapting these techniques. These obstacles stem from the radar's irregular angular distribution, deviations from a single-sensor polar layout in multi-radar setups, and point sparsity. To address these issues, we propose Class-Aware PillarMix (CAPMix), a novel MSDA approach that applies MixUp at the pillar level in 3D point clouds, guided by class labels. Unlike methods that rely a single mix ratio to the entire sample, CAPMix assigns an independent ratio to each pillar, boosting sample diversity. To account for the density of different classes, we use class-specific distributions: for dense objects (e.g., large vehicles), we skew ratios to favor points from another sample, while for sparse objects (e.g., pedestrians), we sample more points from the original. This class-aware mixing retains critical details and enriches each sample with new information, ultimately generating more diverse training data. Experimental results demonstrate that our method not only significantly boosts performance but also outperforms existing MSDA approaches across two datasets (Bosch Street and K-Radar). We believe that this straightforward yet effective approach will spark further investigation into MSDA techniques for radar data.
comment: 8 pages, 6 figures, 4 tables, accepted to 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2025). Code: https://github.com/boschresearch/CAPMIX
♻ ☆ Robust Adaptive Safe Robotic Grasping with Tactile Sensing
Robotic grasping requires safe force interaction to prevent a grasped object from being damaged or slipping out of the hand. In this vein, this paper proposes an integrated framework for grasping with formal safety guarantees based on Control Barrier Functions. We first design contact force and force closure constraints, which are enforced by a safety filter to accomplish safe grasping with finger force control. For sensory feedback, we develop a technique to estimate contact point, force, and torque from tactile sensors at each finger. We verify the framework with various safety filters in a numerical simulation under a two-finger grasping scenario. We then experimentally validate the framework by grasping multiple objects, including fragile lab glassware, in a real robotic setup, showing that safe grasping can be successfully achieved in the real world. We evaluate the performance of each safety filter in the context of safety violation and conservatism, and find that disturbance observer-based control barrier functions provide superior performance for safety guarantees with minimum conservatism.
comment: This paper was accepted to ECC 2025. The demonstration video is available at https://youtu.be/Cuj47mkXRdg
♻ ☆ RIZE: Adaptive Regularization for Imitation Learning
We propose a novel Inverse Reinforcement Learning (IRL) method that mitigates the rigidity of fixed reward structures and the limited flexibility of implicit reward regularization. Building on the Maximum Entropy IRL framework, our approach incorporates a squared temporal-difference (TD) regularizer with adaptive targets that evolve dynamically during training, thereby imposing adaptive bounds on recovered rewards and promoting robust decision-making. To capture richer return information, we integrate distributional RL into the learning process. Empirically, our method achieves expert-level performance on complex MuJoCo and Adroit environments, surpassing baseline methods on the Humanoid-v2 task with limited expert demonstrations. Extensive experiments and ablation studies further validate the effectiveness of the approach and provide insights into reward dynamics in imitation learning. Our source code is available at https://github.com/adibka/RIZE.
comment: Camera-ready version. Published in Transactions on Machine Learning Research (2025). Official version: https://openreview.net/forum?id=a6DWqXJZCZ
♻ ☆ U2UData+: A Scalable Swarm UAVs Autonomous Flight Dataset for Embodied Long-horizon Tasks AAAI26
Swarm UAV autonomous flight for Embodied Long-Horizon (ELH) tasks is crucial for advancing the low-altitude economy. However, existing methods focus only on specific basic tasks due to dataset limitations, failing in real-world deployment for ELH tasks. ELH tasks are not mere concatenations of basic tasks, requiring handling long-term dependencies, maintaining embodied persistent states, and adapting to dynamic goal shifts. This paper presents U2UData+, the first large-scale swarm UAV autonomous flight dataset for ELH tasks and the first scalable swarm UAV data online collection and algorithm closed-loop verification platform. The dataset is captured by 15 UAVs in autonomous collaborative flights for ELH tasks, comprising 12 scenes, 720 traces, 120 hours, 600 seconds per trajectory, 4.32M LiDAR frames, and 12.96M RGB frames. This dataset also includes brightness, temperature, humidity, smoke, and airflow values covering all flight routes. The platform supports the customization of simulators, UAVs, sensors, flight algorithms, formation modes, and ELH tasks. Through a visual control window, this platform allows users to collect customized datasets through one-click deployment online and to verify algorithms by closed-loop simulation. U2UData+ also introduces an ELH task for wildlife conservation and provides comprehensive benchmarks with 9 SOTA models. U2UData+ can be found at https://fengtt42.github.io/U2UData-2/.
comment: Accepted by AAAI26
♻ ☆ APEX: Action Priors Enable Efficient Exploration for Robust Motion Tracking on Legged Robots
Learning natural, animal-like locomotion from demonstrations has become a core paradigm in legged robotics. Despite the recent advancements in motion tracking, most existing methods demand extensive tuning and rely on reference data during deployment, limiting adaptability. We present APEX (Action Priors enable Efficient Exploration), a plug-and-play extension to state-of-the-art motion tracking algorithms that eliminates any dependence on reference data during deployment, improves sample efficiency, and reduces parameter tuning effort. APEX integrates expert demonstrations directly into reinforcement learning (RL) by incorporating decaying action priors, which initially bias exploration toward expert demonstrations but gradually allow the policy to explore independently. This is combined with a multi-critic framework that balances task performance with motion style. Moreover, APEX enables a single policy to learn diverse motions and transfer reference-like styles across different terrains and velocities, while remaining robust to variations in reward design. We validate the effectiveness of our method through extensive experiments in both simulation and on a Unitree Go2 robot. By leveraging demonstrations to guide exploration during RL training, without imposing explicit bias toward them, APEX enables legged robots to learn with greater stability, efficiency, and generalization. We believe this approach paves the way for guidance-driven RL to boost natural skill acquisition in a wide array of robotic tasks, from locomotion to manipulation. Website and code: https://marmotlab.github.io/APEX/.
comment: 9 pages; Previously this version appeared as arXiv:2511.09091, which was submitted as a new work by accident
♻ ☆ APEX: Action Priors Enable Efficient Exploration for Robust Motion Tracking on Legged Robots
Learning natural, animal-like locomotion from demonstrations has become a core paradigm in legged robotics. Despite the recent advancements in motion tracking, most existing methods demand extensive tuning and rely on reference data during deployment, limiting adaptability. We present APEX (Action Priors enable Efficient Exploration), a plug-and-play extension to state-of-the-art motion tracking algorithms that eliminates any dependence on reference data during deployment, improves sample efficiency, and reduces parameter tuning effort. APEX integrates expert demonstrations directly into reinforcement learning (RL) by incorporating decaying action priors, which initially bias exploration toward expert demonstrations but gradually allow the policy to explore independently. This is combined with a multi-critic framework that balances task performance with motion style. Moreover, APEX enables a single policy to learn diverse motions and transfer reference-like styles across different terrains and velocities, while remaining robust to variations in reward design. We validate the effectiveness of our method through extensive experiments in both simulation and on a Unitree Go2 robot. By leveraging demonstrations to guide exploration during RL training, without imposing explicit bias toward them, APEX enables legged robots to learn with greater stability, efficiency, and generalization. We believe this approach paves the way for guidance-driven RL to boost natural skill acquisition in a wide array of robotic tasks, from locomotion to manipulation. Website and code: https://marmotlab.github.io/APEX/.
comment: This work was intended as a replacement of arXiv:2505.10022 and any subsequent updates will appear there
♻ ☆ Towards High-Consistency Embodied World Model with Multi-View Trajectory Videos
Embodied world models aim to predict and interact with the physical world through visual observations and actions. However, existing models struggle to accurately translate low-level actions (e.g., joint positions) into precise robotic movements in predicted frames, leading to inconsistencies with real-world physical interactions. To address these limitations, we propose MTV-World, an embodied world model that introduces Multi-view Trajectory-Video control for precise visuomotor prediction. Specifically, instead of directly using low-level actions for control, we employ trajectory videos obtained through camera intrinsic and extrinsic parameters and Cartesian-space transformation as control signals. However, projecting 3D raw actions onto 2D images inevitably causes a loss of spatial information, making a single view insufficient for accurate interaction modeling. To overcome this, we introduce a multi-view framework that compensates for spatial information loss and ensures high-consistency with physical world. MTV-World forecasts future frames based on multi-view trajectory videos as input and conditioning on an initial frame per view. Furthermore, to systematically evaluate both robotic motion precision and object interaction accuracy, we develop an auto-evaluation pipeline leveraging multimodal large models and referring video object segmentation models. To measure spatial consistency, we formulate it as an object location matching problem and adopt the Jaccard Index as the evaluation metric. Extensive experiments demonstrate that MTV-World achieves precise control execution and accurate physical interaction modeling in complex dual-arm scenarios.
comment: 15 pages, 23 figures
♻ ☆ RL-100: Performant Robotic Manipulation with Real-World Reinforcement Learning
Real-world robotic manipulation in homes and factories demands reliability, efficiency, and robustness that approach or surpass the performance of skilled human operators. We present RL-100, a real-world reinforcement learning framework built on diffusion-based visuomotor policies. RL-100 unifies imitation and reinforcement learning under a single PPO-style objective applied within the denoising process, yielding conservative and stable policy improvements across both offline and online stages. To meet deployment latency constraints, we employ a lightweight consistency distillation procedure that compresses multi-step diffusion into a one-step controller for high-frequency control. The framework is task-, embodiment-, and representation-agnostic, and supports both single-action outputs and action-chunking control. We evaluate RL-100 on seven diverse real-robot manipulation tasks, ranging from dynamic pushing and agile bowling to pouring, cloth folding, unscrewing, and multi-stage juicing. RL-100 attains 100% success across evaluated trials, achieving 900 out of 900 successful episodes, including up to 250 out of 250 consecutive trials on one task, and matches or surpasses expert teleoperators in time-to-completion. Without retraining, a single policy attains approximately 90% zero-shot success under environmental and dynamics shifts, adapts in a few-shot regime to significant task variations (86.7%), and remains robust to aggressive human perturbations (about 95%). In a public shopping-mall deployment, the juicing robot served random customers continuously for roughly seven hours without failure. Together, these results suggest a practical path toward deployment-ready robot learning: start from human priors, align training objectives with human-grounded metrics, and reliably extend performance beyond human demonstrations.
comment: https://lei-kun.github.io/RL-100/
♻ ☆ GPA-RAM: Grasp-Pretraining Augmented Robotic Attention Mamba for Spatial Task Learning
Task failures in prior fine-grained robotic manipulation methods often stem from suboptimal initial grasping, which is critical for subsequent manipulation and reducing the requirement for complex pose adjustments. To address this, we propose Grasp-Pretraining Augmentation (GPA), a general multi-modal learning framework that enhances grasp perception without additional grasp pose data collection and labeling. GPA achieves evident enhancement on RLBench multi-task benchmark (from 79.3% to 84.2%) and ALOHA bimanual manipulation tasks (from 86%/16% to 98%/38%). Although GPA enhances fine-grained grasping performance by leveraging increased model capacity, it incurs computational latency and hinders real-time deployment. To mitigate this limitation, we propose Robotic Attention Mamba (RAM). This architecture synergizes attention mechanisms with state space models (SSMs), effectively capturing complex spatial features while maintaining superior inference efficiency. Our unified GPA-RAM framework balances model capacity with efficiency and applies to both discrete and continuous action generation. GPA-RAM demonstrates superior performance across four robotic systems with diverse camera configurations in both simulation and the real world. Compared with previous state-of-the-art methods, it improves average success rates by 8.2% over RVT2 (from 79.3% to 87.5%) and 2.6% over ARP^+ (from 84.9% to 87.5%) on the RLBench multi-task benchmark and 40% (from 16% to 56%), 12% (from 86% to 98%) on ALOHA bimanual continuous tasks, with inference speed of about 71 FPS. This work provides a framework for developing robotic systems that are simultaneously precise and responsive. The project and code are at https://gpa-ram.github.io/
♻ ☆ RoboTidy : A 3D Gaussian Splatting Household Tidying Benchmark for Embodied Navigation and Action
Household tidying is an important application area, yet current benchmarks neither model user preferences nor support mobility, and they generalize poorly, making it hard to comprehensively assess integrated language-to-action capabilities. To address this, we propose RoboTidy, a unified benchmark for language-guided household tidying that supports Vision-Language-Action (VLA) and Vision-Language-Navigation (VLN) training and evaluation. RoboTidy provides 500 photorealistic 3D Gaussian Splatting (3DGS) household scenes (covering 500 objects and containers) with collisions, formulates tidying as an "Action (Object, Container)" list, and supplies 6.4k high-quality manipulation demonstration trajectories and 1.5k naviagtion trajectories to support both few-shot and large-scale training. We also deploy RoboTidy in the real world for object tidying, establishing an end-to-end benchmark for household tidying. RoboTidy offers a scalable platform and bridges a key gap in embodied AI by enabling holistic and realistic evaluation of language-guided robots.
♻ ☆ $π^{*}_{0.6}$: a VLA That Learns From Experience
We study how vision-language-action (VLA) models can improve through real-world deployments via reinforcement learning (RL). We present a general-purpose method, RL with Experience and Corrections via Advantage-conditioned Policies (RECAP), that provides for RL training of VLAs via advantage conditioning. Our method incorporates heterogeneous data into the self-improvement process, including demonstrations, data from on-policy collection, and expert teleoperated interventions provided during autonomous execution. RECAP starts by pre-training a generalist VLA with offline RL, which we call $π^{*}_{0.6}$, that can then be specialized to attain high performance on downstream tasks through on-robot data collection. We show that the $π^{*}_{0.6}$ model trained with the full RECAP method can fold laundry in real homes, reliably assemble boxes, and make espresso drinks using a professional espresso machine. On some of the hardest tasks, RECAP more than doubles task throughput and roughly halves the task failure rate.
♻ ☆ Is Your VLM for Autonomous Driving Safety-Ready? A Comprehensive Benchmark for Evaluating External and In-Cabin Risks
Vision-Language Models (VLMs) show great promise for autonomous driving, but their suitability for safety-critical scenarios is largely unexplored, raising safety concerns. This issue arises from the lack of comprehensive benchmarks that assess both external environmental risks and in-cabin driving behavior safety simultaneously. To bridge this critical gap, we introduce DSBench, the first comprehensive Driving Safety Benchmark designed to assess a VLM's awareness of various safety risks in a unified manner. DSBench encompasses two major categories: external environmental risks and in-cabin driving behavior safety, divided into 10 key categories and a total of 28 sub-categories. This comprehensive evaluation covers a wide range of scenarios, ensuring a thorough assessment of VLMs' performance in safety-critical contexts. Extensive evaluations across various mainstream open-source and closed-source VLMs reveal significant performance degradation under complex safety-critical situations, highlighting urgent safety concerns. To address this, we constructed a large dataset of 98K instances focused on in-cabin and external safety scenarios, showing that fine-tuning on this dataset significantly enhances the safety performance of existing VLMs and paves the way for advancing autonomous driving technology. The benchmark toolkit, code, and model checkpoints will be publicly accessible.
♻ ☆ LoopSR: Looping Sim-and-Real for Lifelong Policy Adaptation of Legged Robots IROS 2025
Reinforcement Learning (RL) has shown its remarkable and generalizable capability in legged locomotion through sim-to-real transfer. However, while adaptive methods like domain randomization are expected to enhance policy robustness across diverse environments, they potentially compromise the policy's performance in any specific environment, leading to suboptimal real-world deployment due to the No Free Lunch theorem. To address this, we propose LoopSR, a lifelong policy adaptation framework that continuously refines RL policies in the post-deployment stage. LoopSR employs a transformer-based encoder to map real-world trajectories into a latent space and reconstruct a digital twin of the real world for further improvement. Autoencoder architecture and contrastive learning methods are adopted to enhance feature extraction of real-world dynamics. Simulation parameters for continual training are derived by combining predicted values from the decoder with retrieved parameters from a pre-collected simulation trajectory dataset. By leveraging simulated continual training, LoopSR achieves superior data efficiency compared with strong baselines, yielding eminent performance with limited data in both sim-to-sim and sim-to-real experiments. Please refer to https://peilinwu.site/looping-sim-and-real.github.io/ for videos and code.
comment: IROS 2025
♻ ☆ Searching in Space and Time: Unified Memory-Action Loops for Open-World Object Retrieval ICRA
Service robots must retrieve objects in dynamic, open-world settings where requests may reference attributes ("the red mug"), spatial context ("the mug on the table"), or past states ("the mug that was here yesterday"). Existing approaches capture only parts of this problem: scene graphs capture spatial relations but ignore temporal grounding, temporal reasoning methods model dynamics but do not support embodied interaction, and dynamic scene graphs handle both but remain closed-world with fixed vocabularies. We present STAR (SpatioTemporal Active Retrieval), a framework that unifies memory queries and embodied actions within a single decision loop. STAR leverages non-parametric long-term memory and a working memory to support efficient recall, and uses a vision-language model to select either temporal or spatial actions at each step. We introduce STARBench, a benchmark of spatiotemporal object search tasks across simulated and real environments. Experiments in STARBench and on a Tiago robot show that STAR consistently outperforms scene-graph and memory-only baselines, demonstrating the benefits of treating search in time and search in space as a unified problem.
comment: This paper is under review at ICRA
♻ ☆ Inference of Human-derived Specifications of Object Placement via Demonstration IJCAI'25
As robots' manipulation capabilities improve for pick-and-place tasks (e.g., object packing, sorting, and kitting), methods focused on understanding human-acceptable object configurations remain limited expressively with regard to capturing spatial relationships important to humans. To advance robotic understanding of human rules for object arrangement, we introduce positionally-augmented RCC (PARCC), a formal logic framework based on region connection calculus (RCC) for describing the relative position of objects in space. Additionally, we introduce an inference algorithm for learning PARCC specifications via demonstrations. Finally, we present the results from a human study, which demonstrate our framework's ability to capture a human's intended specification and the benefits of learning from demonstration approaches over human-provided specifications.
comment: IJCAI'25
♻ ☆ Vector Quantized-Elites: Unsupervised and Problem-Agnostic Quality-Diversity Optimization
Quality-Diversity algorithms have transformed optimization by prioritizing the discovery of diverse, high-performing solutions over a single optimal result. However, traditional Quality-Diversity methods, such as MAP-Elites, rely heavily on predefined behavior descriptors and complete prior knowledge of the task to define the behavior space grid, limiting their flexibility and applicability. In this work, we introduce Vector Quantized-Elites (VQ-Elites), a novel Quality-Diversity algorithm that autonomously constructs a structured behavior space grid using unsupervised learning, eliminating the need for prior task-specific knowledge. At the core of VQ-Elites is the integration of Vector Quantized Variational Autoencoders, which enables the dynamic learning of behavior descriptors and the generation of a structured, rather than unstructured, behavior space grid -- a significant advancement over existing unsupervised Quality-Diversity approaches. This design establishes VQ-Elites as a flexible, robust, and task-agnostic optimization framework. To further enhance the performance of unsupervised Quality-Diversity algorithms, we introduce behavior space bounding and cooperation mechanisms, which significantly improve convergence and performance, as well as the Effective Diversity Ratio and Coverage Diversity Score, two novel metrics that quantify the actual diversity in the unsupervised setting. We validate VQ-Elites on robotic arm pose-reaching, mobile robot space-covering, and MiniGrid exploration tasks. The results demonstrate its ability to efficiently generate diverse, high-quality solutions, emphasizing its adaptability, scalability, robustness to hyperparameters, and potential to extend Quality-Diversity optimization to complex, previously inaccessible domains.
comment: 15 pages (+4 supplementary), 14 (+1) figures, 1 algorithm, 1 (+8) table(s), accepted at IEEE Transactions on Evolutionary Computation
♻ ☆ RAPID: Robust and Agile Planner Using Inverse Reinforcement Learning for Vision-Based Drone Navigation
This paper introduces a learning-based visual planner for agile drone flight in cluttered environments. The proposed planner generates collision-free waypoints in milliseconds, enabling drones to perform agile maneuvers in complex environments without building separate perception, mapping, and planning modules. Learning-based methods, such as behavior cloning (BC) and reinforcement learning (RL), demonstrate promising performance in visual navigation but still face inherent limitations. BC is susceptible to compounding errors due to limited expert imitation, while RL struggles with reward function design and sample inefficiency. To address these limitations, this paper proposes an inverse reinforcement learning (IRL)-based framework for high-speed visual navigation. By leveraging IRL, it is possible to reduce the number of interactions with simulation environments and improve capability to deal with high-dimensional spaces while preserving the robustness of RL policies. A motion primitive-based path planning algorithm collects an expert dataset with privileged map data from diverse environments, ensuring comprehensive scenario coverage. By leveraging both the acquired expert and learner dataset gathered from the agent's interactions with the simulation environments, a robust reward function and policy are learned across diverse states. While the proposed method is trained in a simulation environment only, it can be directly applied to real-world scenarios without additional training or tuning. The performance of the proposed method is validated in both simulation and real-world environments, including forests and various structures. The trained policy achieves an average speed of 7 m/s and a maximum speed of 8.8 m/s in real flight experiments. To the best of our knowledge, this is the first work to successfully apply an IRL framework for high-speed visual navigation of drones.
comment: 18 pages, 11 figures, 58 references, and appendix is included
Robotics 62
☆ $π^{*}_{0.6}$: a VLA That Learns From Experience
We study how vision-language-action (VLA) models can improve through real-world deployments via reinforcement learning (RL). We present a general-purpose method, RL with Experience and Corrections via Advantage-conditioned Policies (RECAP), that provides for RL training of VLAs via advantage conditioning. Our method incorporates heterogeneous data into the self-improvement process, including demonstrations, data from on-policy collection, and expert teleoperated interventions provided during autonomous execution. RECAP starts by pre-training a generalist VLA with offline RL, which we call $π^{*}_{0.6}$, that can then be specialized to attain high performance on downstream tasks through on-robot data collection. We show that the $π^{*}_{0.6}$ model trained with the full RECAP method can fold laundry in real homes, reliably assemble boxes, and make espresso drinks using a professional espresso machine. On some of the hardest tasks, RECAP more than doubles task throughput and roughly halves the task failure rate.
☆ HMC: Learning Heterogeneous Meta-Control for Contact-Rich Loco-Manipulation
Learning from real-world robot demonstrations holds promise for interacting with complex real-world environments. However, the complexity and variability of interaction dynamics often cause purely positional controllers to struggle with contacts or varying payloads. To address this, we propose a Heterogeneous Meta-Control (HMC) framework for Loco-Manipulation that adaptively stitches multiple control modalities: position, impedance, and hybrid force-position. We first introduce an interface, HMC-Controller, for blending actions from different control profiles continuously in the torque space. HMC-Controller facilitates both teleoperation and policy deployment. Then, to learn a robust force-aware policy, we propose HMC-Policy to unify different controllers into a heterogeneous architecture. We adopt a mixture-of-experts style routing to learn from large-scale position-only data and fine-grained force-aware demonstrations. Experiments on a real humanoid robot show over 50% relative improvement vs. baselines on challenging tasks such as compliant table wiping and drawer opening, demonstrating the efficacy of HMC.
☆ Robust Verification of Controllers under State Uncertainty via Hamilton-Jacobi Reachability Analysis
As perception-based controllers for autonomous systems become increasingly popular in the real world, it is important that we can formally verify their safety and performance despite perceptual uncertainty. Unfortunately, the verification of such systems remains challenging, largely due to the complexity of the controllers, which are often nonlinear, nonconvex, learning-based, and/or black-box. Prior works propose verification algorithms that are based on approximate reachability methods, but they often restrict the class of controllers and systems that can be handled or result in overly conservative analyses. Hamilton-Jacobi (HJ) reachability analysis is a popular formal verification tool for general nonlinear systems that can compute optimal reachable sets under worst-case system uncertainties; however, its application to perception-based systems is currently underexplored. In this work, we propose RoVer-CoRe, a framework for the Robust Verification of Controllers via HJ Reachability. To the best of our knowledge, RoVer-CoRe is the first HJ reachability-based framework for the verification of perception-based systems under perceptual uncertainty. Our key insight is to concatenate the system controller, observation function, and the state estimation modules to obtain an equivalent closed-loop system that is readily compatible with existing reachability frameworks. Within RoVer-CoRe, we propose novel methods for formal safety verification and robust controller design. We demonstrate the efficacy of the framework in case studies involving aircraft taxiing and NN-based rover navigation. Code is available at the link in the footnote.
comment: Submitted to the 8th Annual Learning for Dynamics & Control Conference
☆ Co-Me: Confidence-Guided Token Merging for Visual Geometric Transformers
We propose Confidence-Guided Token Merging (Co-Me), an acceleration mechanism for visual geometric transformers without retraining or finetuning the base model. Co-Me distilled a light-weight confidence predictor to rank tokens by uncertainty and selectively merge low-confidence ones, effectively reducing computation while maintaining spatial coverage. Compared to similarity-based merging or pruning, the confidence signal in Co-Me reliably indicates regions emphasized by the transformer, enabling substantial acceleration without degrading performance. Co-Me applies seamlessly to various multi-view and streaming visual geometric transformers, achieving speedups that scale with sequence length. When applied to VGGT and MapAnything, Co-Me achieves up to $11.3\times$ and $7.2\times$ speedup, making visual geometric transformers practical for real-time 3D perception and reconstruction.
☆ NORA-1.5: A Vision-Language-Action Model Trained using World Model- and Action-based Preference Rewards
Vision--language--action (VLA) models have recently shown promising performance on a variety of embodied tasks, yet they still fall short in reliability and generalization, especially when deployed across different embodiments or real-world environments. In this work, we introduce NORA-1.5, a VLA model built from the pre-trained NORA backbone by adding to it a flow-matching-based action expert. This architectural enhancement alone yields substantial performance gains, enabling NORA-1.5 to outperform NORA and several state-of-the-art VLA models across both simulated and real-world benchmarks. To further improve robustness and task success, we develop a set of reward models for post-training VLA policies. Our rewards combine (i) an action-conditioned world model (WM) that evaluates whether generated actions lead toward the desired goal, and (ii) a deviation-from-ground-truth heuristic that distinguishes good actions from poor ones. Using these reward signals, we construct preference datasets and adapt NORA-1.5 to target embodiments through direct preference optimization (DPO). Extensive evaluations show that reward-driven post-training consistently improves performance in both simulation and real-robot settings, demonstrating significant VLA model-reliability gains through simple yet effective reward models. Our findings highlight NORA-1.5 and reward-guided post-training as a viable path toward more dependable embodied agents suitable for real-world deployment.
comment: https://declare-lab.github.io/nora-1.5
☆ Gallant: Voxel Grid-based Humanoid Locomotion and Local-navigation across 3D Constrained Terrains
Robust humanoid locomotion requires accurate and globally consistent perception of the surrounding 3D environment. However, existing perception modules, mainly based on depth images or elevation maps, offer only partial and locally flattened views of the environment, failing to capture the full 3D structure. This paper presents Gallant, a voxel-grid-based framework for humanoid locomotion and local navigation in 3D constrained terrains. It leverages voxelized LiDAR data as a lightweight and structured perceptual representation, and employs a z-grouped 2D CNN to map this representation to the control policy, enabling fully end-to-end optimization. A high-fidelity LiDAR simulation that dynamically generates realistic observations is developed to support scalable, LiDAR-based training and ensure sim-to-real consistency. Experimental results show that Gallant's broader perceptual coverage facilitates the use of a single policy that goes beyond the limitations of previous methods confined to ground-level obstacles, extending to lateral clutter, overhead constraints, multi-level structures, and narrow passages. Gallant also firstly achieves near 100% success rates in challenging scenarios such as stair climbing and stepping onto elevated platforms through improved end-to-end optimization.
☆ Active Matter as a framework for living systems-inspired Robophysics
Robophysics investigates the physical principles that govern living-like robots operating in complex, realworld environments. Despite remarkable technological advances, robots continue to face fundamental efficiency limitations. At the level of individual units, locomotion remains a challenge, while at the collective level, robot swarms struggle to achieve shared purpose, coordination, communication, and cost efficiency. This perspective article examines the key challenges faced by bio-inspired robotic collectives and highlights recent research efforts that incorporate principles from active-matter physics and biology into the modeling and design of robot swarms.
☆ Is Your VLM for Autonomous Driving Safety-Ready? A Comprehensive Benchmark for Evaluating External and In-Cabin Risks
Vision-Language Models (VLMs) show great promise for autonomous driving, but their suitability for safety-critical scenarios is largely unexplored, raising safety concerns. This issue arises from the lack of comprehensive benchmarks that assess both external environmental risks and in-cabin driving behavior safety simultaneously. To bridge this critical gap, we introduce DSBench, the first comprehensive Driving Safety Benchmark designed to assess a VLM's awareness of various safety risks in a unified manner. DSBench encompasses two major categories: external environmental risks and in-cabin driving behavior safety, divided into 10 key categories and a total of 28 sub-categories. This comprehensive evaluation covers a wide range of scenarios, ensuring a thorough assessment of VLMs' performance in safety-critical contexts. Extensive evaluations across various mainstream open-source and closed-source VLMs reveal significant performance degradation under complex safety-critical situations, highlighting urgent safety concerns. To address this, we constructed a large dataset of 98K instances focused on in-cabin and external safety scenarios, showing that fine-tuning on this dataset significantly enhances the safety performance of existing VLMs and paves the way for advancing autonomous driving technology. The benchmark toolkit, code, and model checkpoints will be publicly accessible.
☆ Masked IRL: LLM-Guided Reward Disambiguation from Demonstrations and Language
Robots can adapt to user preferences by learning reward functions from demonstrations, but with limited data, reward models often overfit to spurious correlations and fail to generalize. This happens because demonstrations show robots how to do a task but not what matters for that task, causing the model to focus on irrelevant state details. Natural language can more directly specify what the robot should focus on, and, in principle, disambiguate between many reward functions consistent with the demonstrations. However, existing language-conditioned reward learning methods typically treat instructions as simple conditioning signals, without fully exploiting their potential to resolve ambiguity. Moreover, real instructions are often ambiguous themselves, so naive conditioning is unreliable. Our key insight is that these two input types carry complementary information: demonstrations show how to act, while language specifies what is important. We propose Masked Inverse Reinforcement Learning (Masked IRL), a framework that uses large language models (LLMs) to combine the strengths of both input types. Masked IRL infers state-relevance masks from language instructions and enforces invariance to irrelevant state components. When instructions are ambiguous, it uses LLM reasoning to clarify them in the context of the demonstrations. In simulation and on a real robot, Masked IRL outperforms prior language-conditioned IRL methods by up to 15% while using up to 4.7 times less data, demonstrating improved sample-efficiency, generalization, and robustness to ambiguous language. Project page: https://MIT-CLEAR-Lab.github.io/Masked-IRL and Code: https://github.com/MIT-CLEAR-Lab/Masked-IRL
☆ Aerial Assistance System for Automated Firefighting during Turntable Ladder Operations
Fires in industrial facilities pose special challenges to firefighters, e.g., due to the sheer size and scale of the buildings. The resulting visual obstructions impair firefighting accuracy, further compounded by inaccurate assessments of the fire's location. Such imprecision simultaneously increases the overall damage and prolongs the fire-brigades operation unnecessarily. We propose an automated assistance system for firefighting using a motorized fire monitor on a turntable ladder with aerial support from an unmanned aerial vehicle (UAV). The UAV flies autonomously within an obstacle-free flight funnel derived from geodata, detecting and localizing heat sources. An operator supervises the operation on a handheld controller and selects a fire target in reach. After the selection, the UAV automatically plans and traverses between two triangulation poses for continued fire localization. Simultaneously, our system steers the fire monitor to ensure the water jet reaches the detected heat source. In preliminary tests, our assistance system successfully localized multiple heat sources and directed a water jet towards the fires.
comment: 7 pages, Presented at IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), Galway, Ireland, 2025
☆ Enhancing End-to-End Autonomous Driving with Risk Semantic Distillaion from VLM
The autonomous driving (AD) system has exhibited remarkable performance in complex driving scenarios. However, generalization is still a key limitation for the current system, which refers to the ability to handle unseen scenarios or unfamiliar sensor configurations.Related works have explored the use of Vision-Language Models (VLMs) to address few-shot or zero-shot tasks. While promising, these methods introduce a new challenge: the emergence of a hybrid AD system, where two distinct systems are used to plan a trajectory, leading to potential inconsistencies. Alternative research directions have explored Vision-Language-Action (VLA) frameworks that generate control actions from VLM directly. However, these end-to-end solutions demonstrate prohibitive computational demands. To overcome these challenges, we introduce Risk Semantic Distillation (RSD), a novel framework that leverages VLMs to enhance the training of End-to-End (E2E) AD backbones. By providing risk attention for key objects, RSD addresses the issue of generalization. Specifically, we introduce RiskHead, a plug-in module that distills causal risk estimates from Vision-Language Models into Bird's-Eye-View (BEV) features, yielding interpretable risk-attention maps.This approach allows BEV features to learn richer and more nuanced risk attention representations, which directly enhance the model's ability to handle spatial boundaries and risky objects.By focusing on risk attention, RSD aligns better with human-like driving behavior, which is essential to navigate in complex and dynamic environments. Our experiments on the Bench2Drive benchmark demonstrate the effectiveness of RSD in managing complex and unpredictable driving conditions. Due to the enhanced BEV representations enabled by RSD, we observed a significant improvement in both perception and planning capabilities.
☆ Advancing Minimally Invasive Precision Surgery in Open Cavities with Robotic Flexible Endoscopy
Flexible robots hold great promise for enhancing minimally invasive surgery (MIS) by providing superior dexterity, precise control, and safe tissue interaction. Yet, translating these advantages into endoscopic interventions within open cavities remains challenging. The lack of anatomical constraints and the inherent flexibility of such devices complicate their control, while the limited field of view of endoscopes restricts situational awareness. We present a robotic platform designed to overcome these challenges and demonstrate its potential in fetoscopic laser coagulation, a complex MIS procedure typically performed only by highly experienced surgeons. Our system combines a magnetically actuated flexible endoscope with teleoperated and semi-autonomous navigation capabilities for performing targeted laser ablations. To enhance surgical awareness, the platform reconstructs real-time mosaics of the endoscopic scene, providing an extended and continuous visual context. The ability of this system to address the key limitations of MIS in open spaces is validated in vivo in an ovine model.
☆ Towards A Catalogue of Requirement Patterns for Space Robotic Missions
In the development of safety and mission-critical systems, including autonomous space robotic missions, complex behaviour is captured during the requirements elicitation phase. Requirements are typically expressed using natural language which is ambiguous and not amenable to formal verification methods that can provide robust guarantees of system behaviour. To support the definition of formal requirements, specification patterns provide reusable, logic-based templates. A suite of robotic specification patterns, along with their formalisation in NASA's Formal Requirements Elicitation Tool (FRET) already exists. These pre-existing requirement patterns are domain agnostic and, in this paper we explore their applicability for space missions. To achieve this we carried out a literature review of existing space missions and formalised their requirements using FRET, contributing a corpus of space mission requirements. We categorised these requirements using pre-existing specification patterns which demonstrated their applicability in space missions. However, not all of the requirements that we formalised corresponded to an existing pattern so we have contributed 5 new requirement specification patterns as well as several variants of the existing and new patterns. We also conducted an expert evaluation of the new patterns, highlighting their benefits and limitations.
comment: In Proceedings FMAS 2025, arXiv:2511.13245
☆ Achieving Safe Control Online through Integration of Harmonic Control Lyapunov-Barrier Functions with Unsafe Object-Centric Action Policies
We propose a method for combining Harmonic Control Lyapunov-Barrier Functions (HCLBFs) derived from Signal Temporal Logic (STL) specifications with any given robot policy to turn an unsafe policy into a safe one with formal guarantees. The two components are combined via HCLBF-derived safety certificates, thus producing commands that preserve both safety and task-driven behavior. We demonstrate with a simple proof-of-concept implementation for an object-centric force-based policy trained through reinforcement learning for a movement task of a stationary robot arm that is able to avoid colliding with obstacles on a table top after combining the policy with the safety constraints. The proposed method can be generalized to more complex specifications and dynamic task settings.
comment: In Proceedings FMAS 2025, arXiv:2511.13245
☆ Safe-ROS: An Architecture for Autonomous Robots in Safety-Critical Domains
Deploying autonomous robots in safety-critical domains requires architectures that ensure operational effectiveness and safety compliance. In this paper, we contribute the Safe-ROS architecture for developing reliable and verifiable autonomous robots in such domains. It features two distinct subsystems: (1) an intelligent control system that is responsible for normal/routine operations, and (2) a Safety System consisting of Safety Instrumented Functions (SIFs) that provide formally verifiable independent oversight. We demonstrate Safe-ROS on an AgileX Scout Mini robot performing autonomous inspection in a nuclear environment. One safety requirement is selected and instantiated as a SIF. To support verification, we implement the SIF as a cognitive agent, programmed to stop the robot whenever it detects that it is too close to an obstacle. We verify that the agent meets the safety requirement and integrate it into the autonomous inspection. This integration is also verified, and the full deployment is validated in a Gazebo simulation, and lab testing. We evaluate this architecture in the context of the UK nuclear sector, where safety and regulation are crucial aspects of deployment. Success criteria include the development of a formal property from the safety requirement, implementation, and verification of the SIF, and the integration of the SIF into the operational robotic autonomous system. Our results demonstrate that the Safe-ROS architecture can provide safety verifiable oversight while deploying autonomous robots in safety-critical domains, offering a robust framework that can be extended to additional requirements and various applications.
comment: In Proceedings FMAS 2025, arXiv:2511.13245
☆ Mutation Testing for Industrial Robotic Systems
Industrial robotic systems (IRS) are increasingly deployed in diverse environments, where failures can result in severe accidents and costly downtime. Ensuring the reliability of the software controlling these systems is therefore critical. Mutation testing, a technique widely used in software engineering, evaluates the effectiveness of test suites by introducing small faults, or mutants, into the code. However, traditional mutation operators are poorly suited to robotic programs, which involve message-based commands and interactions with the physical world. This paper explores the adaptation of mutation testing to IRS by defining domain-specific mutation operators that capture the semantics of robot actions and sensor readings. We propose a methodology for generating meaningful mutants at the level of high-level read and write operations, including movement, gripper actions, and sensor noise injection. An empirical study on a pick-and-place scenario demonstrates that our approach produces more informative mutants and reduces the number of invalid or equivalent cases compared to conventional operators. Results highlight the potential of mutation testing to enhance test suite quality and contribute to safer, more reliable industrial robotic systems.
comment: In Proceedings FMAS 2025, arXiv:2511.13245
Self-Supervised Multisensory Pretraining for Contact-Rich Robot Reinforcement Learning
Effective contact-rich manipulation requires robots to synergistically leverage vision, force, and proprioception. However, Reinforcement Learning agents struggle to learn in such multisensory settings, especially amidst sensory noise and dynamic changes. We propose MultiSensory Dynamic Pretraining (MSDP), a novel framework for learning expressive multisensory representations tailored for task-oriented policy learning. MSDP is based on masked autoencoding and trains a transformer-based encoder by reconstructing multisensory observations from only a subset of sensor embeddings, leading to cross-modal prediction and sensor fusion. For downstream policy learning, we introduce a novel asymmetric architecture, where a cross-attention mechanism allows the critic to extract dynamic, task-specific features from the frozen embeddings, while the actor receives a stable pooled representation to guide its actions. Our method demonstrates accelerated learning and robust performance under diverse perturbations, including sensor noise, and changes in object dynamics. Evaluations in multiple challenging, contact-rich robot manipulation tasks in simulation and the real world showcase the effectiveness of MSDP. Our approach exhibits strong robustness to perturbations and achieves high success rates on the real robot with as few as 6,000 online interactions, offering a simple yet powerful solution for complex multisensory robotic control.
comment: 9 pages, 10 figures, preprint
☆ Continuous Vision-Language-Action Co-Learning with Semantic-Physical Alignment for Behavioral Cloning AAAI 2026
Language-conditioned manipulation facilitates human-robot interaction via behavioral cloning (BC), which learns control policies from human demonstrations and serves as a cornerstone of embodied AI. Overcoming compounding errors in sequential action decisions remains a central challenge to improving BC performance. Existing approaches mitigate compounding errors through data augmentation, expressive representation, or temporal abstraction. However, they suffer from physical discontinuities and semantic-physical misalignment, leading to inaccurate action cloning and intermittent execution. In this paper, we present Continuous vision-language-action Co-Learning with Semantic-Physical Alignment (CCoL), a novel BC framework that ensures temporally consistent execution and fine-grained semantic grounding. It generates robust and smooth action execution trajectories through continuous co-learning across vision, language, and proprioceptive inputs (e.g., robot internal states). Meanwhile, we anchor language semantics to visuomotor representations by a bidirectional cross-attention to learn contextual information for action generation, successfully overcoming the problem of semantic-physical misalignment. Extensive experiments show that CCoL achieves an average 8.0% relative improvement across three simulation suites, with up to 19.2% relative gain in human-demonstrated bimanual insertion tasks. Real-world tests on a 7-DoF robot further confirm CCoL's generalization under unseen and noisy object states.
comment: Accepted at AAAI 2026, the Project website is available at https://qhemu.github.io/CCoL/
☆ Perception-aware Exploration for Consumer-grade UAVs
In our work, we extend the current state-of-the-art approach for autonomous multi-UAV exploration to consumer-level UAVs, such as the DJI Mini 3 Pro. We propose a pipeline that selects viewpoint pairs from which the depth can be estimated and plans the trajectory that satisfies motion constraints necessary for odometry estimation. For the multi-UAV exploration, we propose a semi-distributed communication scheme that distributes the workload in a balanced manner. We evaluate our model performance in simulation for different numbers of UAVs and prove its ability to safely explore the environment and reconstruct the map even with the hardware limitations of consumer-grade UAVs.
☆ Identifying Time-varying Costs in Finite-horizon Linear Quadratic Gaussian Games
We address cost identification in a finite-horizon linear quadratic Gaussian game. We characterize the set of cost parameters that generate a given Nash equilibrium policy. We propose a backpropagation algorithm to identify the time-varying cost parameters. We derive a probabilistic error bound when the cost parameters are identified from finite trajectories. We test our method in numerical and driving simulations. Our algorithm identifies the cost parameters that can reproduce the Nash equilibrium policy and trajectory observations.
☆ Going Places: Place Recognition in Artificial and Natural Systems
Place recognition, the ability to identify previously visited locations, is critical for both biological navigation and autonomous systems. This review synthesizes findings from robotic systems, animal studies, and human research to explore how different systems encode and recall place. We examine the computational and representational strategies employed across artificial systems, animals, and humans, highlighting convergent solutions such as topological mapping, cue integration, and memory management. Animal systems reveal evolved mechanisms for multimodal navigation and environmental adaptation, while human studies provide unique insights into semantic place concepts, cultural influences, and introspective capabilities. Artificial systems showcase scalable architectures and data-driven models. We propose a unifying set of concepts by which to consider and develop place recognition mechanisms and identify key challenges such as generalization, robustness, and environmental variability. This review aims to foster innovations in artificial localization by connecting future developments in artificial place recognition systems to insights from both animal navigation research and human spatial cognition studies.
☆ Simultaneous Localization and 3D-Semi Dense Mapping for Micro Drones Using Monocular Camera and Inertial Sensors
Monocular simultaneous localization and mapping (SLAM) algorithms estimate drone poses and build a 3D map using a single camera. Current algorithms include sparse methods that lack detailed geometry, while learning-driven approaches produce dense maps but are computationally intensive. Monocular SLAM also faces scale ambiguities, which affect its accuracy. To address these challenges, we propose an edge-aware lightweight monocular SLAM system combining sparse keypoint-based pose estimation with dense edge reconstruction. Our method employs deep learning-based depth prediction and edge detection, followed by optimization to refine keypoints and edges for geometric consistency, without relying on global loop closure or heavy neural computations. We fuse inertial data with vision by using an extended Kalman filter to resolve scale ambiguity and improve accuracy. The system operates in real time on low-power platforms, as demonstrated on a DJI Tello drone with a monocular camera and inertial sensors. In addition, we demonstrate robust autonomous navigation and obstacle avoidance in indoor corridors and on the TUM RGBD dataset. Our approach offers an effective, practical solution to real-time mapping and navigation in resource-constrained environments.
☆ MA-SLAM: Active SLAM in Large-Scale Unknown Environment using Map Aware Deep Reinforcement Learning
Active Simultaneous Localization and Mapping (Active SLAM) involves the strategic planning and precise control of a robotic system's movement in order to construct a highly accurate and comprehensive representation of its surrounding environment, which has garnered significant attention within the research community. While the current methods demonstrate efficacy in small and controlled settings, they face challenges when applied to large-scale and diverse environments, marked by extended periods of exploration and suboptimal paths of discovery. In this paper, we propose MA-SLAM, a Map-Aware Active SLAM system based on Deep Reinforcement Learning (DRL), designed to address the challenge of efficient exploration in large-scale environments. In pursuit of this objective, we put forward a novel structured map representation. By discretizing the spatial data and integrating the boundary points and the historical trajectory, the structured map succinctly and effectively encapsulates the visited regions, thereby serving as input for the deep reinforcement learning based decision module. Instead of sequentially predicting the next action step within the decision module, we have implemented an advanced global planner to optimize the exploration path by leveraging long-range target points. We conducted experiments in three simulation environments and deployed in a real unmanned ground vehicle (UGV), the results demonstrate that our approach significantly reduces both the duration and distance of exploration compared with state-of-the-art methods.
☆ Dual-Variable Force Characterisation method for Human-Robot Interaction in Wearable Robotics
Understanding the physical interaction with wearable robots is essential to ensure safety and comfort. However, this interaction is complex in two key aspects: (1) the motion involved, and (2) the non-linear behaviour of soft tissues. Multiple approaches have been undertaken to better understand this interaction and to improve the quantitative metrics of physical interfaces or cuffs. As these two topics are closely interrelated, finite modelling and soft tissue characterisation offer valuable insights into pressure distribution and shear stress induced by the cuff. Nevertheless, current characterisation methods typically rely on a single fitting variable along one degree of freedom, which limits their applicability, given that interactions with wearable robots often involve multiple degrees of freedom. To address this limitation, this work introduces a dual-variable characterisation method, involving normal and tangential forces, aimed at identifying reliable material parameters and evaluating the impact of single-variable fitting on force and torque responses. This method demonstrates the importance of incorporating two variables into the characterisation process by analysing the normalized mean square error (NMSE) across different scenarios and material models, providing a foundation for simulation at the closest possible level, with a focus on the cuff and the human limb involved in the physical interaction between the user and the wearable robot.
comment: 36 pages, 10 figures, submitted and under-review in Journal of the Mechanical Behavior of Biomedical Materials
☆ Multi-Timescale Model Predictive Control for Slow-Fast Systems
Model Predictive Control (MPC) has established itself as the primary methodology for constrained control, enabling autonomy across diverse applications. While model fidelity is crucial in MPC, solving the corresponding optimization problem in real time remains challenging when combining long horizons with high-fidelity models that capture both short-term dynamics and long-term behavior. Motivated by results on the Exponential Decay of Sensitivities (EDS), which imply that, under certain conditions, the influence of modeling inaccuracies decreases exponentially along the prediction horizon, this paper proposes a multi-timescale MPC scheme for fast-sampled control. Tailored to systems with both fast and slow dynamics, the proposed approach improves computational efficiency by i) switching to a reduced model that captures only the slow, dominant dynamics and ii) exponentially increasing integration step sizes to progressively reduce model detail along the horizon. We evaluate the method on three practically motivated robotic control problems in simulation and observe speed-ups of up to an order of magnitude.
☆ Towards Deploying VLA without Fine-Tuning: Plug-and-Play Inference-Time VLA Policy Steering via Embodied Evolutionary Diffusion
Vision-Language-Action (VLA) models have demonstrated significant potential in real-world robotic manipulation. However, pre-trained VLA policies still suffer from substantial performance degradation during downstream deployment. Although fine-tuning can mitigate this issue, its reliance on costly demonstration collection and intensive computation makes it impractical in real-world settings. In this work, we introduce VLA-Pilot, a plug-and-play inference-time policy steering method for zero-shot deployment of pre-trained VLA without any additional fine-tuning or data collection. We evaluate VLA-Pilot on six real-world downstream manipulation tasks across two distinct robotic embodiments, encompassing both in-distribution and out-of-distribution scenarios. Experimental results demonstrate that VLA-Pilot substantially boosts the success rates of off-the-shelf pre-trained VLA policies, enabling robust zero-shot generalization to diverse tasks and embodiments. Experimental videos and code are available at: https://rip4kobe.github.io/vla-pilot/.
comment: 9 pages, 8 figures, submitted to IEEE RA-L
☆ RoboTidy : A 3D Gaussian Splatting Household Tidying Benchmark for Embodied Navigation and Action
Household tidying is an important application area, yet current benchmarks neither model user preferences nor support mobility, and they generalize poorly, making it hard to comprehensively assess integrated language-to-action capabilities. To address this, we propose RoboTidy, a unified benchmark for language-guided household tidying that supports Vision-Language-Action (VLA) and Vision-Language-Navigation (VLN) training and evaluation. RoboTidy provides 500 photorealistic 3D Gaussian Splatting (3DGS) household scenes (covering 500 objects and containers) with collisions, formulates tidying as an "Action (Object, Container)" list, and supplies 6.4k high-quality manipulation demonstration trajectories and 1.5k naviagtion trajectories to support both few-shot and large-scale training. We also deploy RoboTidy in the real world for object tidying, establishing an end-to-end benchmark for household tidying. RoboTidy offers a scalable platform and bridges a key gap in embodied AI by enabling holistic and realistic evaluation of language-guided robots.
☆ AsyncVLA: Asynchronous Flow Matching for Vision-Language-Action Models
Vision-language-action (VLA) models have recently emerged as a powerful paradigm for building generalist robots. However, traditional VLA models that generate actions through flow matching (FM) typically rely on rigid and uniform time schedules, i.e., synchronous FM (SFM). Without action context awareness and asynchronous self-correction, SFM becomes unstable in long-horizon tasks, where a single action error can cascade into failure. In this work, we propose asynchronous flow matching VLA (AsyncVLA), a novel framework that introduces temporal flexibility in asynchronous FM (AFM) and enables self-correction in action generation. AsyncVLA breaks from the vanilla SFM in VLA models by generating the action tokens in a non-uniform time schedule with action context awareness. Besides, our method introduces the confidence rater to extract confidence of the initially generated actions, enabling the model to selectively refine inaccurate action tokens before execution. Moreover, we propose a unified training procedure for SFM and AFM that endows a single model with both modes, improving KV-cache utilization. Extensive experiments on robotic manipulation benchmarks demonstrate that AsyncVLA is data-efficient and exhibits self-correction ability. AsyncVLA achieves state-of-the-art results across general embodied evaluations due to its asynchronous generation in AFM. Our code is available at https://github.com/YuhuaJiang2002/AsyncVLA.
☆ FlexiCup: Wireless Multimodal Suction Cup with Dual-Zone Vision-Tactile Sensing
Conventional suction cups lack sensing capabilities for contact-aware manipulation in unstructured environments. This paper presents FlexiCup, a fully wireless multimodal suction cup that integrates dual-zone vision-tactile sensing. The central zone dynamically switches between vision and tactile modalities via illumination control for contact detection, while the peripheral zone provides continuous spatial awareness for approach planning. FlexiCup supports both vacuum and Bernoulli suction modes through modular mechanical configurations, achieving complete wireless autonomy with onboard computation and power. We validate hardware versatility through dual control paradigms. Modular perception-driven grasping across structured surfaces with varying obstacle densities demonstrates comparable performance between vacuum (90.0% mean success) and Bernoulli (86.7% mean success) modes. Diffusion-based end-to-end learning achieves 73.3% success on inclined transport and 66.7% on orange extraction tasks. Ablation studies confirm that multi-head attention coordinating dual-zone observations provides 13% improvements for contact-aware manipulation. Hardware designs and firmware are available at https://anonymous.4open.science/api/repo/FlexiCup-DA7D/file/index.html?v=8f531b44.
☆ BIM-Discrepancy-Driven Active Sensing for Risk-Aware UAV-UGV Navigation
This paper presents a BIM-discrepancy-driven active sensing framework for cooperative navigation between unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) in dynamic construction environments. Traditional navigation approaches rely on static Building Information Modeling (BIM) priors or limited onboard perception. In contrast, our framework continuously fuses real-time LiDAR data from aerial and ground robots with BIM priors to maintain an evolving 2D occupancy map. We quantify navigation safety through a unified corridor-risk metric integrating occupancy uncertainty, BIM-map discrepancy, and clearance. When risk exceeds safety thresholds, the UAV autonomously re-scans affected regions to reduce uncertainty and enable safe replanning. Validation in PX4-Gazebo simulation with Robotec GPU LiDAR demonstrates that risk-triggered re-scanning reduces mean corridor risk by 58% and map entropy by 43% compared to static BIM navigation, while maintaining clearance margins above 0.4 m. Compared to frontier-based exploration, our approach achieves similar uncertainty reduction in half the mission time. These results demonstrate that integrating BIM priors with risk-adaptive aerial sensing enables scalable, uncertainty-aware autonomy for construction robotics.
☆ FACA: Fair and Agile Multi-Robot Collision Avoidance in Constrained Environments with Dynamic Priorities
Multi-robot systems are increasingly being used for critical applications such as rescuing injured people, delivering food and medicines, and monitoring key areas. These applications usually involve navigating at high speeds through constrained spaces such as small gaps. Navigating such constrained spaces becomes particularly challenging when the space is crowded with multiple heterogeneous agents all of which have urgent priorities. What makes the problem even harder is that during an active response situation, roles and priorities can quickly change on a dime without informing the other agents. In order to complete missions in such environments, robots must not only be safe, but also agile, able to dodge and change course at a moment's notice. In this paper, we propose FACA, a fair and agile collision avoidance approach where robots coordinate their tasks by talking to each other via natural language (just as people do). In FACA, robots balance safety with agility via a novel artificial potential field algorithm that creates an automatic ``roundabout'' effect whenever a conflict arises. Our experiments show that FACA achieves a improvement in efficiency, completing missions more than 3.5X faster than baselines with a time reduction of over 70% while maintaining robust safety margins.
☆ Searching in Space and Time: Unified Memory-Action Loops for Open-World Object Retrieval ICRA
Service robots must retrieve objects in dynamic, open-world settings where requests may reference attributes ("the red mug"), spatial context ("the mug on the table"), or past states ("the mug that was here yesterday"). Existing approaches capture only parts of this problem: scene graphs capture spatial relations but ignore temporal grounding, temporal reasoning methods model dynamics but do not support embodied interaction, and dynamic scene graphs handle both but remain closed-world with fixed vocabularies. We present STAR (SpatioTemporal Active Retrieval), a framework that unifies memory queries and embodied actions within a single decision loop. STAR leverages non-parametric long-term memory and a working memory to support efficient recall, and uses a vision-language model to select either temporal or spatial actions at each step. We introduce STARBench, a benchmark of spatiotemporal object search tasks across simulated and real environments. Experiments in STARBench and on a Tiago robot show that STAR consistently outperforms scene-graph and memory-only baselines, demonstrating the benefits of treating search in time and search in space as a unified problem.
comment: This paper is under review at ICRA
☆ SVBRD-LLM: Self-Verifying Behavioral Rule Discovery for Autonomous Vehicle Identification
As more autonomous vehicles operate on public roads, understanding real-world behavior of autonomous vehicles is critical to analyzing traffic safety, making policies, and public acceptance. This paper proposes SVBRD-LLM, a framework that automatically discovers, verifies, and applies interpretable behavioral rules from real traffic videos through zero-shot prompt engineering. The framework extracts vehicle trajectories using YOLOv8 and ByteTrack, computes kinematic features, and employs GPT-5 zero-shot prompting to compare autonomous and human-driven vehicles, generating 35 structured behavioral rule hypotheses. These rules are tested on a validation set, iteratively refined based on failure cases to filter spurious correlations, and compiled into a high-confidence rule library. The framework is evaluated on an independent test set for speed change prediction, lane change prediction, and autonomous vehicle identification tasks. Experiments on over 1500 hours of real traffic videos show that the framework achieves 90.0% accuracy and 93.3% F1-score in autonomous vehicle identification. The discovered rules clearly reveal distinctive characteristics of autonomous vehicles in speed control smoothness, lane change conservativeness, and acceleration stability, with each rule accompanied by semantic description, applicable context, and validation confidence.
☆ EGSA-PT:Edge-Guided Spatial Attention with Progressive Training for Monocular Depth Estimation and Segmentation of Transparent Objects
Transparent object perception remains a major challenge in computer vision research, as transparency confounds both depth estimation and semantic segmentation. Recent work has explored multi-task learning frameworks to improve robustness, yet negative cross-task interactions often hinder performance. In this work, we introduce Edge-Guided Spatial Attention (EGSA), a fusion mechanism designed to mitigate destructive interactions by incorporating boundary information into the fusion between semantic and geometric features. On both Syn-TODD and ClearPose benchmarks, EGSA consistently improved depth accuracy over the current state of the art method (MODEST), while preserving competitive segmentation performance, with the largest improvements appearing in transparent regions. Besides our fusion design, our second contribution is a multi-modal progressive training strategy, where learning transitions from edges derived from RGB images to edges derived from predicted depth images. This approach allows the system to bootstrap learning from the rich textures contained in RGB images, and then switch to more relevant geometric content in depth maps, while it eliminates the need for ground-truth depth at training time. Together, these contributions highlight edge-guided fusion as a robust approach capable of improving transparent object perception.
☆ Artificial intelligence approaches for energy-efficient laser cutting machines
This research addresses the significant challenges of energy consumption and environmental impact in laser cutting by proposing novel deep learning (DL) methodologies to achieve energy reduction. Recognizing the current lack of adaptive control and the open-loop nature of CO2 laser suction pumps, this study utilizes closed-loop configurations that dynamically adjust pump power based on both the material being cut and the smoke level generated. To implement this adaptive system, diverse material classification methods are introduced, including techniques leveraging lens-less speckle sensing with a customized Convolutional Neural Network (CNN) and an approach using a USB camera with transfer learning via the pre-trained VGG16 CNN model. Furthermore, a separate DL model for smoke level detection is employed to simultaneously refine the pump's power output. This integration prompts the exhaust suction pump to automatically halt during inactive times and dynamically adjust power during operation, leading to experimentally proven and remarkable energy savings, with results showing a 20% to 50% reduction in the smoke suction pump's energy consumption, thereby contributing substantially to sustainable development in the manufacturing sector.
☆ A visual study of ICP variants for Lidar Odometry
Odometry with lidar sensors is a state-of-the-art method to estimate the ego pose of a moving vehicle. Many implementations of lidar odometry use variants of the Iterative Closest Point (ICP) algorithm. Real-world effects such as dynamic objects, non-overlapping areas, and sensor noise diminish the accuracy of ICP. We build on a recently proposed method that makes these effects visible by visualizing the multidimensional objective function of ICP in two dimensions. We use this method to study different ICP variants in the context of lidar odometry. In addition, we propose a novel method to filter out dynamic objects and to address the ego blind spot problem.
comment: Iterative closest point; Registration; Odometry; Mapping
☆ Z-Merge: Multi-Agent Reinforcement Learning for On-Ramp Merging with Zone-Specific V2X Traffic Information
Ramp merging is a critical and challenging task for autonomous vehicles (AVs), particularly in mixed traffic environments with human-driven vehicles (HVs). Existing approaches typically rely on either lane-changing or inter-vehicle gap creation strategies based solely on local or neighboring information, often leading to suboptimal performance in terms of safety and traffic efficiency. In this paper, we present a V2X (vehicle-to-everything communication)-assisted Multiagent Reinforcement Learning (MARL) framework for on-ramp merging that effectively coordinates the complex interplay between lane-changing and inter-vehicle gap adaptation strategies by utilizing zone-specific global information available from a roadside unit (RSU). The merging control problem is formulated as a Multiagent Partially Observable Markov Decision Process (MA-POMDP), where agents leverage both local and global observations through V2X communication. To support both discrete and continuous control decisions, we design a hybrid action space and adopt a parameterized deep Q-learning approach. Extensive simulations, integrating the SUMO traffic simulator and the MOSAIC V2X simulator, demonstrate that our framework significantly improves merging success rate, traffic efficiency, and road safety across diverse traffic scenarios.
☆ Attacking Autonomous Driving Agents with Adversarial Machine Learning: A Holistic Evaluation with the CARLA Leaderboard
To autonomously control vehicles, driving agents use outputs from a combination of machine-learning (ML) models, controller logic, and custom modules. Although numerous prior works have shown that adversarial examples can mislead ML models used in autonomous driving contexts, it remains unclear if these attacks are effective at producing harmful driving actions for various agents, environments, and scenarios. To assess the risk of adversarial examples to autonomous driving, we evaluate attacks against a variety of driving agents, rather than against ML models in isolation. To support this evaluation, we leverage CARLA, an urban driving simulator, to create and evaluate adversarial examples. We create adversarial patches designed to stop or steer driving agents, stream them into the CARLA simulator at runtime, and evaluate them against agents from the CARLA Leaderboard, a public repository of best-performing autonomous driving agents from an annual research competition. Unlike prior work, we evaluate attacks against autonomous driving systems without creating or modifying any driving-agent code and against all parts of the agent included with the ML model. We perform a case-study investigation of two attack strategies against three open-source driving agents from the CARLA Leaderboard across multiple driving scenarios, lighting conditions, and locations. Interestingly, we show that, although some attacks can successfully mislead ML models into predicting erroneous stopping or steering commands, some driving agents use modules, such as PID control or GPS-based rules, that can overrule attacker-manipulated predictions from ML models.
comment: 12 pages
♻ ☆ OG-VLA: Orthographic Image Generation for 3D-Aware Vision-Language Action Model
We introduce OG-VLA, a novel architecture and learning framework that combines the generalization strengths of Vision Language Action models (VLAs) with the robustness of 3D-aware policies. We address the challenge of mapping natural language instructions and one or more RGBD observations to quasi-static robot actions. 3D-aware robot policies achieve state-of-the-art performance on precise robot manipulation tasks, but struggle with generalization to unseen instructions, scenes, and objects. On the other hand, VLAs excel at generalizing across instructions and scenes, but can be sensitive to camera and robot pose variations. We leverage prior knowledge embedded in language and vision foundation models to improve generalization of 3D-aware keyframe policies. OG-VLA unprojects input observations from diverse views into a point cloud which is then rendered from canonical orthographic views, ensuring input view invariance and consistency between input and output spaces. These canonical views are processed with a vision backbone, a Large Language Model (LLM), and an image diffusion model to generate images that encode the next position and orientation of the end-effector on the input scene. Evaluations on the Arnold and Colosseum benchmarks demonstrate state-of-the-art generalization to unseen environments, with over 40% relative improvements while maintaining robust performance in seen settings. We also show real-world adaption in 3 to 5 demonstrations along with strong generalization. Videos and resources at https://og-vla.github.io/
comment: 13 pages
♻ ☆ LED: Light Enhanced Depth Estimation at Night BMVC 2025
Nighttime camera-based depth estimation is a highly challenging task, especially for autonomous driving applications, where accurate depth perception is essential for ensuring safe navigation. Models trained on daytime data often fail in the absence of precise but costly LiDAR. Even vision foundation models trained on large amounts of data are unreliable in low-light conditions. In this work, we aim to improve the reliability of perception systems at night time. To this end, we introduce Light Enhanced Depth (LED), a novel, cost-effective approach that significantly improves depth estimation in low-light environments by harnessing a pattern projected by high definition headlights available in modern vehicles. LED leads to significant performance boosts across multiple depth-estimation architectures (encoder-decoder, Adabins, DepthFormer, Depth Anything V2) both on synthetic and real datasets. Furthermore, increased performances beyond illuminated areas reveal a holistic enhancement in scene understanding. Finally, we release the Nighttime Synthetic Drive Dataset, a synthetic and photo-realistic nighttime dataset, which comprises 49,990 comprehensively annotated images.
comment: BMVC 2025 (Poster). Code and dataset available on the project page : https://simondemoreau.github.io/LED/ 21 pages, 13 figures
♻ ☆ StyleDrive: Towards Driving-Style Aware Benchmarking of End-To-End Autonomous Driving
Personalization, while extensively studied in conventional autonomous driving pipelines, has been largely overlooked in the context of end-to-end autonomous driving (E2EAD), despite its critical role in fostering user trust, safety perception, and real-world adoption. A primary bottleneck is the absence of large-scale real-world datasets that systematically capture driving preferences, severely limiting the development and evaluation of personalized E2EAD models. In this work, we introduce the first large-scale real-world dataset explicitly curated for personalized E2EAD, integrating comprehensive scene topology with rich dynamic context derived from agent dynamics and semantics inferred via a fine-tuned vision-language model (VLM). We propose a hybrid annotation pipeline that combines behavioral analysis, rule-and-distribution-based heuristics, and subjective semantic modeling guided by VLM reasoning, with final refinement through human-in-the-loop verification. Building upon this dataset, we introduce the first standardized benchmark for systematically evaluating personalized E2EAD models. Empirical evaluations on state-of-the-art architectures demonstrate that incorporating personalized driving preferences significantly improves behavioral alignment with human demonstrations.
comment: 25 pages, 7 figures, 5 tables
♻ ☆ CARScenes: Semantic VLM Dataset for Safe Autonomous Driving
CAR-Scenes is a frame-level dataset for autonomous driving that enables training and evaluation of vision-language models (VLMs) for interpretable, scene-level understanding. We annotate 5,192 images drawn from Argoverse 1, Cityscapes, KITTI, and nuScenes using a 28-key category/sub-category knowledge base covering environment, road geometry, background-vehicle behavior, ego-vehicle behavior, vulnerable road users, sensor states, and a discrete severity scale (1-10), totaling 350+ leaf attributes. Labels are produced by a GPT-4o-assisted vision-language pipeline with human-in-the-loop verification; we release the exact prompts, post-processing rules, and per-field baseline model performance. CAR-Scenes also provides attribute co-occurrence graphs and JSONL records that support semantic retrieval, dataset triage, and risk-aware scenario mining across sources. To calibrate task difficulty, we include reproducible, non-benchmark baselines, notably a LoRA-tuned Qwen2-VL-2B with deterministic decoding, evaluated via scalar accuracy, micro-averaged F1 for list attributes, and severity MAE/RMSE on a fixed validation split. We publicly release the annotation and analysis scripts, including graph construction and evaluation scripts, to enable explainable, data-centric workflows for future intelligent vehicles. Dataset: https://github.com/Croquembouche/CAR-Scenes
comment: 8 pages, 6 figures, 7 tables
♻ ☆ DepthVision: Enabling Robust Vision-Language Models with GAN-Based LiDAR-to-RGB Synthesis for Autonomous Driving
Ensuring reliable autonomous operation when visual input is degraded remains a key challenge in intelligent vehicles and robotics. We present DepthVision, a multimodal framework that enables Vision--Language Models (VLMs) to exploit LiDAR data without any architectural changes or retraining. DepthVision synthesizes dense, RGB-like images from sparse LiDAR point clouds using a conditional GAN with an integrated refiner, and feeds these into off-the-shelf VLMs through their standard visual interface. A Luminance-Aware Modality Adaptation (LAMA) module fuses synthesized and real camera images by dynamically weighting each modality based on ambient lighting, compensating for degradation such as darkness or motion blur. This design turns LiDAR into a drop-in visual surrogate when RGB becomes unreliable, effectively extending the operational envelope of existing VLMs. We evaluate DepthVision on real and simulated datasets across multiple VLMs and safety-critical tasks, including vehicle-in-the-loop experiments. The results show substantial improvements in low-light scene understanding over RGB-only baselines while preserving full compatibility with frozen VLM architectures. These findings demonstrate that LiDAR-guided RGB synthesis is a practical pathway for integrating range sensing into modern vision-language systems for autonomous driving.
♻ ☆ Robust Adaptive Time-Varying Control Barrier Function with Application to Robotic Surface Treatment
Set invariance techniques such as control barrier functions (CBFs) can be used to enforce time-varying constraints such as keeping a safe distance from dynamic objects. However, existing methods for enforcing time-varying constraints often overlook model uncertainties. To address this issue, this paper proposes a CBFs-based robust adaptive controller design endowing time-varying constraints while considering parametric uncertainty and additive disturbances. To this end, we first leverage Robust adaptive Control Barrier Functions (RaCBFs) to handle model uncertainty, along with the concept of Input-to-State Safety (ISSf) to ensure robustness towards input disturbances. Furthermore, to alleviate the inherent conservatism in robustness, we also incorporate a set membership identification scheme. We demonstrate the proposed method on robotic surface treatment that requires time-varying force bounds to ensure uniform quality, in numerical simulation and real robotic setup, showing that the quality is formally guaranteed within an acceptable range.
comment: This work has been accepted to ECC 2025
♻ ☆ Tac2Motion: Contact-Aware Reinforcement Learning with Tactile Feedback for Robotic Hand Manipulation
This paper proposes Tac2Motion, a contact-aware reinforcement learning framework to facilitate the learning of contact-rich in-hand manipulation tasks, such as removing a lid. To this end, we propose tactile sensing-based reward shaping and incorporate the sensing into the observation space through embedding. The designed rewards encourage an agent to ensure firm grasping and smooth finger gaiting at the same time, leading to higher data efficiency and robust performance compared to the baseline. We verify the proposed framework on the opening a lid scenario, showing generalization of the trained policy into a couple of object types and various dynamics such as torsional friction. Lastly, the learned policy is demonstrated on the multi-fingered robot, Shadow Robot, showing that the control policy can be transferred to the real world. The video is available: https://youtu.be/poeJBPR7urQ.
comment: This paper has submitted to Dexterous Humanoid Manipulation Workshop, Humanoid 2025
♻ ☆ RynnEC: Bringing MLLMs into Embodied World
We introduce RynnEC, a video multimodal large language model designed for embodied cognition. Built upon a general-purpose vision-language foundation model, RynnEC incorporates a region encoder and a mask decoder, enabling flexible region-level video interaction. Despite its compact architecture, RynnEC achieves state-of-the-art performance in object property understanding, object segmentation, and spatial reasoning. Conceptually, it offers a region-centric video paradigm for the brain of embodied agents, providing fine-grained perception of the physical world and enabling more precise interactions. To mitigate the scarcity of annotated 3D datasets, we propose an egocentric video based pipeline for generating embodied cognition data. Furthermore, we introduce RynnEC-Bench, a region-centered benchmark for evaluating embodied cognitive capabilities. We anticipate that RynnEC will advance the development of general-purpose cognitive cores for embodied agents and facilitate generalization across diverse embodied tasks. The code, model checkpoints, and benchmark are available at: https://github.com/alibaba-damo-academy/RynnEC
comment: The technical report of RynnEC, an embodied cognition MLLM
♻ ☆ Combining High Level Scheduling and Low Level Control to Manage Fleets of Mobile Robots
The deployment of mobile robots for material handling in industrial environments requires scalable coordination of large fleets in dynamic settings. This paper presents a two-layer framework that combines high-level scheduling with low-level control. Tasks are assigned and scheduled using the compositional algorithm ComSat, which generates time-parameterized routes for each robot. These schedules are then used by a distributed Model Predictive Control (MPC) system in real time to compute local reference trajectories, accounting for static and dynamic obstacles. The approach ensures safe, collision-free operation, and supports rapid rescheduling in response to disruptions such as robot failures or environmental changes. We evaluate the method in simulated 2D environments with varying road capacities and traffic conditions, demonstrating high task completion rates and robust behavior even under congestion. The modular structure of the framework allows for computational tractability and flexibility, making it suitable for deployment in complex, real-world industrial scenarios.
♻ ☆ Benchmarking Population-Based Reinforcement Learning across Robotic Tasks with GPU-Accelerated Simulation
In recent years, deep reinforcement learning (RL) has shown its effectiveness in solving complex continuous control tasks. However, this comes at the cost of an enormous amount of experience required for training, exacerbated by the sensitivity of learning efficiency and the policy performance to hyperparameter selection, which often requires numerous trials of time-consuming experiments. This work leverages a Population-Based Reinforcement Learning (PBRL) approach and a GPU-accelerated physics simulator to enhance the exploration capabilities of RL by concurrently training multiple policies in parallel. The PBRL framework is benchmarked against three state-of-the-art RL algorithms -- PPO, SAC, and DDPG -- dynamically adjusting hyperparameters based on the performance of learning agents. The experiments are performed on four challenging tasks in Isaac Gym -- Anymal Terrain, Shadow Hand, Humanoid, Franka Nut Pick -- by analyzing the effect of population size and mutation mechanisms for hyperparameters. The results show that PBRL agents achieve superior performance, in terms of cumulative reward, compared to non-evolutionary baseline agents. Moreover, the trained agents are finally deployed in the real world for a Franka Nut Pick task. To our knowledge, this is the first sim-to-real attempt for deploying PBRL agents on real hardware. Code and videos of the learned policies are available on our project website (https://sites.google.com/view/pbrl).
comment: Accepted for publication at 2025 IEEE 21st International Conference on Automation Science and Engineering
♻ ☆ Generalizable and Fast Surrogates: Model Predictive Control of Articulated Soft Robots using Physics-Informed Neural Networks
Soft robots can revolutionize several applications with high demands on dexterity and safety. When operating these systems, real-time estimation and control require fast and accurate models. However, prediction with first-principles (FP) models is slow, and learned black-box models have poor generalizability. Physics-informed machine learning offers excellent advantages here, but it is currently limited to simple, often simulated systems without considering changes after training. We propose physics-informed neural networks (PINNs) for articulated soft robots (ASRs) with a focus on data efficiency. The amount of expensive real-world training data is reduced to a minimum -- one dataset in one system domain. Two hours of data in different domains are used for a comparison against two gold-standard approaches: In contrast to a recurrent neural network, the PINN provides a high generalizability. The prediction speed of an accurate FP model is exceeded with the PINN by up to a factor of 467 at slightly reduced accuracy. This enables nonlinear model predictive control (MPC) of a pneumatic ASR. Accurate position tracking with the MPC running at 47 Hz is achieved in six dynamic experiments.
comment: Accepted for publication in IEEE Transactions on Robotics (T-RO) 2025
♻ ☆ Spatial Policy: Guiding Visuomotor Robotic Manipulation with Spatial-Aware Modeling and Reasoning
Vision-centric hierarchical embodied models have demonstrated strong potential. However, existing methods lack spatial awareness capabilities, limiting their effectiveness in bridging visual plans to actionable control in complex environments. To address this problem, we propose Spatial Policy (SP), a unified spatial-aware visuomotor robotic manipulation framework via explicit spatial modeling and reasoning. Specifically, we first design a spatial-conditioned embodied video generation module to model spatially guided predictions through the spatial plan table. Then, we propose a flow-based action prediction module to infer executable actions with coordination. Finally, we propose a spatial reasoning feedback policy to refine the spatial plan table via dual-stage replanning. Extensive experiments show that SP substantially outperforms state-of-the-art baselines, achieving over 33% improvement on Meta-World and over 25% improvement on iTHOR, demonstrating strong effectiveness across 23 embodied control tasks. We additionally evaluate SP in real-world robotic experiments to verify its practical viability. SP enhances the practicality of embodied models for robotic control applications. Code and checkpoints are maintained at https://plantpotatoonmoon.github.io/SpatialPolicy/.
♻ ☆ MonoDream: Monocular Vision-Language Navigation with Panoramic Dreaming
Vision-Language Navigation (VLN) tasks often leverage panoramic RGB and depth inputs to provide rich spatial cues for action planning, but these sensors can be costly or less accessible in real-world deployments. Recent approaches based on Vision-Language Action (VLA) models achieve strong results with monocular input, yet they still lag behind methods using panoramic RGB-D information. We present MonoDream, a lightweight VLA framework that enables monocular agents to learn a Unified Navigation Representation (UNR). This shared feature representation jointly aligns navigation-relevant visual semantics (e.g., global layout, depth, and future cues) and language-grounded action intent, enabling more reliable action prediction. MonoDream further introduces Latent Panoramic Dreaming (LPD) tasks to supervise the UNR, which train the model to predict latent features of panoramic RGB and depth observations at both current and future steps based on only monocular input. Experiments on multiple VLN benchmarks show that MonoDream consistently improves monocular navigation performance and significantly narrows the gap with panoramic-based agents.
♻ ☆ Uni-Hand: Universal Hand Motion Forecasting in Egocentric Views IROS'25
Forecasting how human hands move in egocentric views is critical for applications like augmented reality and human-robot policy transfer. Recently, several hand trajectory prediction (HTP) methods have been developed to generate future possible hand waypoints, which still suffer from insufficient prediction targets, inherent modality gaps, entangled hand-head motion, and limited validation in downstream tasks. To address these limitations, we present a universal hand motion forecasting framework considering multi-modal input, multi-dimensional and multi-target prediction patterns, and multi-task affordances for downstream applications. We harmonize multiple modalities by vision-language fusion, global context incorporation, and task-aware text embedding injection, to forecast hand waypoints in both 2D and 3D spaces. A novel dual-branch diffusion is proposed to concurrently predict human head and hand movements, capturing their motion synergy in egocentric vision. By introducing target indicators, the prediction model can forecast the specific joint waypoints of the wrist or the fingers, besides the widely studied hand center points. In addition, we enable Uni-Hand to additionally predict hand-object interaction states (contact/separation) to facilitate downstream tasks better. As the first work to incorporate downstream task evaluation in the literature, we build novel benchmarks to assess the real-world applicability of hand motion forecasting algorithms. The experimental results on multiple publicly available datasets and our newly proposed benchmarks demonstrate that Uni-Hand achieves the state-of-the-art performance in multi-dimensional and multi-target hand motion forecasting. Extensive validation in multiple downstream tasks also presents its impressive human-robot policy transfer to enable robotic manipulation, and effective feature enhancement for action anticipation/recognition.
comment: Extended journal version of MMTwin (IROS'25)
♻ ☆ The Developments and Challenges towards Dexterous and Embodied Robotic Manipulation: A Survey
Achieving human-like dexterous robotic manipulation remains a central goal and a pivotal challenge in robotics. The development of Artificial Intelligence (AI) has allowed rapid progress in robotic manipulation. This survey summarizes the evolution of robotic manipulation from mechanical programming to embodied intelligence, alongside the transition from simple grippers to multi-fingered dexterous hands, outlining key characteristics and main challenges. Focusing on the current stage of embodied dexterous manipulation, we highlight recent advances in two critical areas: dexterous manipulation data collection (via simulation, human demonstrations, and teleoperation) and skill-learning frameworks (imitation and reinforcement learning). Then, based on the overview of the existing data collection paradigm and learning framework, three key challenges restricting the development of dexterous robotic manipulation are summarized and discussed.
♻ ☆ Large Language Models and 3D Vision for Intelligent Robotic Perception and Autonomy
With the rapid advancement of artificial intelligence and robotics, the integration of Large Language Models (LLMs) with 3D vision is emerging as a transformative approach to enhancing robotic sensing technologies. This convergence enables machines to perceive, reason and interact with complex environments through natural language and spatial understanding, bridging the gap between linguistic intelligence and spatial perception. This review provides a comprehensive analysis of state-of-the-art methodologies, applications and challenges at the intersection of LLMs and 3D vision, with a focus on next-generation robotic sensing technologies. We first introduce the foundational principles of LLMs and 3D data representations, followed by an in-depth examination of 3D sensing technologies critical for robotics. The review then explores key advancements in scene understanding, text-to-3D generation, object grounding and embodied agents, highlighting cutting-edge techniques such as zero-shot 3D segmentation, dynamic scene synthesis and language-guided manipulation. Furthermore, we discuss multimodal LLMs that integrate 3D data with touch, auditory and thermal inputs, enhancing environmental comprehension and robotic decision-making. To support future research, we catalog benchmark datasets and evaluation metrics tailored for 3D-language and vision tasks. Finally, we identify key challenges and future research directions, including adaptive model architectures, enhanced cross-modal alignment and real-time processing capabilities, which pave the way for more intelligent, context-aware and autonomous robotic sensing systems.
comment: 45 pages, 15 figures, MDPI Sensors Journal
♻ ☆ HACL: History-Aware Curriculum Learning for Fast Locomotion
We address the problem of agile and rapid locomotion, a key characteristic of quadrupedal and bipedal robots. We present a new algorithm that maintains stability and generates high-speed trajectories by considering the temporal aspect of locomotion. Our formulation takes into account past information based on a novel history-aware curriculum Learning (HACL) algorithm. We model the history of joint velocity commands with respect to the observed linear and angular rewards using a recurrent neural net (RNN). The hidden state helps the curriculum learn the relationship between the forward linear velocity and angular velocity commands and the rewards over a given time-step. We validate our approach on the MIT Mini Cheetah,Unitree Go1, and Go2 robots in a simulated environment and on a Unitree Go1 robot in real-world scenarios. In practice, HACL achieves peak forward velocity of 6.7 m/s for a given command velocity of 7m/s and outperforms prior locomotion algorithms by nearly 20%.
♻ ☆ SocialNav-Map: Dynamic Mapping with Human Trajectory Prediction for Zero-Shot Social Navigation
Social navigation in densely populated dynamic environments poses a significant challenge for autonomous mobile robots, requiring advanced strategies for safe interaction. Existing reinforcement learning (RL)-based methods require over 2000+ hours of extensive training and often struggle to generalize to unfamiliar environments without additional fine-tuning, limiting their practical application in real-world scenarios. To address these limitations, we propose SocialNav-Map, a novel zero-shot social navigation framework that combines dynamic human trajectory prediction with occupancy mapping, enabling safe and efficient navigation without the need for environment-specific training. Specifically, SocialNav-Map first transforms the task goal position into the constructed map coordinate system. Subsequently, it creates a dynamic occupancy map that incorporates predicted human movements as dynamic obstacles. The framework employs two complementary methods for human trajectory prediction: history prediction and orientation prediction. By integrating these predicted trajectories into the occupancy map, the robot can proactively avoid potential collisions with humans while efficiently navigating to its destination. Extensive experiments on the Social-HM3D and Social-MP3D datasets demonstrate that SocialNav-Map significantly outperforms state-of-the-art (SOTA) RL-based methods, which require 2,396 GPU hours of training. Notably, it reduces human collision rates by over 10% without necessitating any training in novel environments. By eliminating the need for environment-specific training, SocialNav-Map achieves superior navigation performance, paving the way for the deployment of social navigation systems in real-world environments characterized by diverse human behaviors. The code is available at: https://github.com/linglingxiansen/SocialNav-Map.
♻ ☆ SF-Loc: A Visual Mapping and Geo-Localization System based on Sparse Visual Structure Frames
For high-level geo-spatial applications and intelligent robotics, accurate global pose information is of crucial importance. Map-aided localization is a universal approach to overcome the limitations of global navigation satellite system (GNSS) in challenging environments. However, current solutions face challenges in terms of mapping flexibility, storage burden and re-localization performance. In this work, we present SF-Loc, a lightweight visual mapping and map-aided localization system, whose core idea is the map representation based on sparse frames with dense but compact depth, termed as visual structure frames. In the mapping phase, multi-sensor dense bundle adjustment (MS-DBA) is applied to construct geo-referenced visual structure frames. The local co-visbility is checked to keep the map sparsity and achieve incremental mapping. In the localization phase, coarse-to-fine vision-based localization is performed, in which multi-frame information and the map distribution are fully integrated. To be specific, the concept of spatially smoothed similarity (SSS) is proposed to overcome the place ambiguity, and pairwise frame matching is applied for efficient and robust pose estimation. Experimental results on the cross-season dataset verify the effectiveness of the system. In complex urban road scenarios, the map size is down to 3 MB per kilometer and stable decimeter-level re-localization can be achieved. The code will be made open-source soon (https://github.com/GREAT-WHU/SF-Loc).
♻ ☆ Rethinking Progression of Memory State in Robotic Manipulation: An Object-Centric Perspective AAAI 2026
As embodied agents operate in increasingly complex environments, the ability to perceive, track, and reason about individual object instances over time becomes essential, especially in tasks requiring sequenced interactions with visually similar objects. In these non-Markovian settings, key decision cues are often hidden in object-specific histories rather than the current scene. Without persistent memory of prior interactions (what has been interacted with, where it has been, or how it has changed) visuomotor policies may fail, repeat past actions, or overlook completed ones. To surface this challenge, we introduce LIBERO-Mem, a non-Markovian task suite for stress-testing robotic manipulation under object-level partial observability. It combines short- and long-horizon object tracking with temporally sequenced subgoals, requiring reasoning beyond the current frame. However, vision-language-action (VLA) models often struggle in such settings, with token scaling quickly becoming intractable even for tasks spanning just a few hundred frames. We propose Embodied-SlotSSM, a slot-centric VLA framework built for temporal scalability. It maintains spatio-temporally consistent slot identities and leverages them through two mechanisms: (1) slot-state-space modeling for reconstructing short-term history, and (2) a relational encoder to align the input tokens with action decoding. Together, these components enable temporally grounded, context-aware action prediction. Experiments show Embodied-SlotSSM's baseline performance on LIBERO-Mem and general tasks, offering a scalable solution for non-Markovian reasoning in object-centric robotic policies.
comment: Accepted at AAAI 2026
♻ ☆ iA*: Imperative Learning-based A* Search for Path Planning
Path planning, which aims to find a collision-free path between two locations, is critical for numerous applications ranging from mobile robots to self-driving vehicles. Traditional search-based methods like A* search guarantee path optimality but are often computationally expensive when handling large-scale maps. While learning-based methods alleviate this issue by incorporating learned constraints into their search procedures, they often face challenges like overfitting and reliance on extensive labeled datasets. To address these limitations, we propose Imperative A* (iA*), a novel self-supervised path planning framework leveraging bilevel optimization (BLO) and imperative learning (IL). The iA* framework integrates a neural network that predicts node costs with a differentiable A* search mechanism, enabling efficient self-supervised training via bilevel optimization. This integration significantly enhances the balance between search efficiency and path optimality while improving generalization to previously unseen maps. Extensive experiments demonstrate that iA* outperforms both classical and supervised learning-based methods, achieving an average reduction of 65.7\% in search area and 54.4\% in runtime, underscoring its effectiveness in robot path planning tasks.
♻ ☆ Bridging Language and Action: A Survey of Language-Conditioned Robot Manipulation
Language-conditioned robot manipulation is an emerging field aimed at enabling seamless communication and cooperation between humans and robotic agents by teaching robots to comprehend and execute instructions conveyed in natural language. This interdisciplinary area integrates scene understanding, language processing, and policy learning to bridge the gap between human instructions and robot actions. In this comprehensive survey, we systematically explore recent advancements in language-conditioned robot manipulation. We categorize existing methods based on the primary ways language is integrated into the robot system, namely language for state evaluation, language as a policy condition, and language for cognitive planning and reasoning. Specifically, we further analyze state-of-the-art techniques from four axes of action granularity, data and supervision regimes, system cost and latency, and environments and evaluations. Additionally, we highlight the key debates in the field. Finally, we discuss open challenges and future research directions, focusing on potentially enhancing generalization capabilities and addressing safety issues in language-conditioned robot manipulators.
♻ ☆ Natural Selection via Foundation Models for Soft Robot Evolution
Designing soft robots is a complex and iterative process that demands cross-disciplinary expertise in materials science, mechanics, and control, often relying on intuition and extensive experimentation. While foundation models, especially Large Language Models (LLMs), have demonstrated impressive reasoning abilities, their capacity to conduct embodied design remains largely unexplored. This paper introduces RoboCrafter-QA, a novel benchmark to evaluate whether LLMs can learn representations of soft robot designs that effectively bridge the gap between high-level task descriptions and low-level morphological and material choices. RoboCrafter-QA leverages the EvoGym simulator to generate a diverse set of soft robot design challenges, spanning robotic locomotion, manipulation, and balancing tasks. Our experiments with SOTA multi-modal LLMs reveal that while these models exhibit promising capabilities in learning design representations, they struggle with fine-grained distinctions between designs with subtle performance differences. To overcome these limitations, we finetune an efficient, open-source LLM that achieves SOTA performance on our benchmark, demonstrating superior capabilities in both design selection and direct generation of high-performing robot morphologies. Furthermore, we construct a physical replica of the modular soft robot and demonstrate a strong sim-to-real correlation, validating that superior benchmark performance has the potential to translate to effective real-world design selection. Our full system will be open-sourced to foster this exciting direction.
♻ ☆ Maestro: Orchestrating Robotics Modules with Vision-Language Models for Zero-Shot Generalist Robots
Today's best-explored routes towards generalist robots center on collecting ever larger "observations-in actions-out" robotics datasets to train large end-to-end models, copying a recipe that has worked for vision-language models (VLMs). We pursue a road less traveled: building generalist policies directly around VLMs by augmenting their general capabilities with specific robot capabilities encapsulated in a carefully curated set of perception, planning, and control modules. In Maestro, a VLM coding agent dynamically composes these modules into a programmatic policy for the current task and scenario. Maestro's architecture benefits from a streamlined closed-loop interface without many manually imposed structural constraints, and a comprehensive and diverse tool repertoire. As a result, it largely surpasses today's VLA models for zero-shot performance on challenging manipulation skills. Further, Maestro is easily extensible to incorporate new modules, easily editable to suit new embodiments such as a quadruped-mounted arm, and even easily adapts from minimal real-world experiences through local code edits.
comment: Plan to resubmit after significant revisions