Skip to content
#Research
Oct 21Wed
  1. arXiv cs.RO62

    MARE: An open-source lunar rover simulator with shadow ground truth

    The authors present MARE, an open-source lunar rover perception and autonomy simulation suite built on Unreal Engine 5 and ROS 2, comprising three parts: terrain generation, dual-physics lunar rover sensor simulation, and GNC mission control. The simulator provides stereo, lidar, event, and SPAD camera data along with pixel-level ground truth; at low solar elevation angles, the mean IoU between predicted shadows and rendered shadows is 0.79. Stereo depth analysis shows that pixel coverage in shadowed regions is lower than in sunlit regions.

  3. arXiv cs.RO67

    Idealworks and NVIDIA evaluate sim-to-real transfer of COMPASS navigation

    Idealworks and NVIDIA systematically evaluated the sim-to-real transfer of the X-Mobility and COMPASS learned navigation pipeline on industrial mobile robots. After training in a digital twin with Cosmos-generated sim-to-real data, real-robot success rates across four scenarios rose from 65.3% to 91.7%, weighted travel time nearly halved, the gap to the production-line navigation stack was 8.3 percentage points, and everything ran on the onboard Jetson AGX Xavier.

  4. arXiv cs.RO55

    TaskWAM uses stage-aware tokens to steer world-action models toward intended futures

    The paper proposes TaskWAM, a closed-loop, stage-aware framework that uses Mode Tokens to represent fine-grained task requirements and State Tokens to track execution progress, forming a structured task state that steers world-action model generation toward futures matching the current task, addressing the Stage Confusion phenomenon in long-horizon manipulation.

  5. arXiv cs.RO55

    CAPSafe benchmark evaluates safety reasoning for actions and long-horizon planning in embodied agents

    Researchers propose the CAPSafe benchmark, which evaluates the safety reasoning capabilities of large language model-driven embodied agents at two levels: an action-level dataset that judges whether a single action is safe in a given environmental context, and a plan-level dataset that assesses whether agents can generate safe long-horizon plans for complex hazardous household tasks.

  7. arXiv cs.RO46

    RILC: Reinforced Iterative Learning Control for Elastic-Joint Robotic Arms, Achieving Cross-Trajectory Generalization with Convergence Guarantees

    This study proposes Reinforced Iterative Learning Control (RILC), which couples Iterative Learning Control (ILC) with Reinforcement Learning (RL) for trajectory tracking of elastic-joint robots, addressing the problems that traditional ILC must learn each new trajectory separately and that RL has high sample complexity and limited tracking accuracy.

  10. arXiv cs.RO62

    Reasoning driving VLAs have a perception-action binding gap: they can perceive danger but do not brake

    The paper points out that reasoning-augmented vision-language-action (VLA) driving models can perceive and reason about safety-critical hazards, yet cannot translate that evidence into safe actions, which the authors call a perception-action binding gap. Using matched simulated counterfactuals and logged replay, the study finds that in-path hazards are indeed encoded in the visual representation, but longitudinal planning is dominated mainly by ego-speed history and responds only weakly to hazard distance and the model's own reasoning trace.

  11. arXiv cs.RO60

    GPL: Representing VLA reasoning state as a plan distribution

    The authors propose Generative Plan Latents (GPL), which represent the reasoning state of vision-language-action (VLA) models as a low-dimensional conditional distribution over plans, such as which object, grasp, or route to use, so that plans can be sampled, selected on request, and reweighted by deployment outcomes without retraining the VLA.

  12. arXiv cs.RO55

    From Predictive Geometry to Action: Learning Complementary Predictive Representations for Embodied Manipulation

    The paper proposes a two-stage framework that learns complementary predictive representations from multi-view video clips to enhance embodied manipulation policies. In the pretraining stage, a frozen 3D foundation model provides dense geometric features, which are compressed into predictive geometric representations and optimized to preserve scene structure and its temporal evolution; action supervision is then applied to geometric transitions to extract a motor representation that emphasizes action-relevant transitions. In the downstream post-training stage, predictive geometry is transferred through geometry queries, and the motor representation guides intermediate features along the action-generation pathway. Future observations serve only as privileged supervision during training and are not needed at inference time.

  13. arXiv cs.RO38

    Design and Implementation of Thermal Control Strategy for the Micro-Size Lunar Exploration Rover HAKUTO

    The thermal interface design of the HAKUTO Rover Flight Model (FM) was verified through thermal balance tests in a vacuum environment. The study divided the rover body into 6 radiating surfaces to build an analytical thermal model, and, combining lunar radiation, solar albedo, and lunar regolith conduction, determined the avionics mounting locations and solar panel angles, keeping mass under 4 kg; after refinement through numerical simulation, verification was completed.

  14. arXiv cs.RO62

    DirectWAM: Distilling Cascaded World-Action Models into Single-Step Action Prediction

    The paper proposes DirectWAM, which converts pretrained video and action experts into single-step predictors that require no stochastic sampling while retaining the explicit future-to-action cascade structure. After learning a single-step action model, the method trains the future predictor by decoding actions from its own predicted futures and backpropagating demonstration action error.

  16. arXiv cs.RO47

    Probabilistic Reach-Avoid Value Function (PVF): A Temporal-Difference Learning Method for Computing Probabilistic Reach-Avoid Sets

    Proposes a self-supervised method, the Probabilistic Reach-Avoid Value Function (PVF), for computing reach-avoid sets for robot tasks under stochastic, potentially unbounded disturbances, applicable to black-box systems. Compared with Hamilton-Jacobi reachability (HJR)-based baselines on systems of increasing complexity, up to the Hopper locomotion task, results show that when disturbances are non-adversarial, PVF's probabilistic reach-avoid sets are less conservative and more accurate than adversarial HJR sets. The method has been submitted to ICRA 2027.

  17. arXiv cs.RO56

    PACT-VLA: Verifying speculative action chunks with pre-commit observations

    PACT-VLA proposes a training-free framework that, in asynchronous inference, uses pre-commit observations to select the best-matching action chunk. Its scoring function, FlowRes, measures the residual between the direction of a candidate's latent flow and the velocity predicted by the frozen policy under that observation, without needing a separate verifier.

  18. arXiv cs.RO44

    Controlling Test-Time Compute for World-Model Planning with Expected Free Energy: When to Stop Thinking

    The study examines whether expected free energy (EFE) can determine how much computation to allocate to planning based on a learned world model, with an objective combining goal distance and disagreement among stochastic predictions. Experiments found that stopping when elite cost stalls reduces refinement while mean success rates stay close to fixed-depth references; with a frozen model, rewarding disagreement lowers success, while online updates partially recover losses caused by regions missing from the training data. The cost-stall rule fails when optimization exploits model errors, and predicted cost keeps falling after task success rates peak.

  21. arXiv cs.RO55

    EventSync: Event-driven memory updates for history-dependent robotic manipulation tasks

    EventSync proposes a hierarchical framework with event-driven memory updates to address history-dependent manipulation tasks where the current observation alone is insufficient to determine the next action. A vision-language model generates subgoals based on the task plan and interaction history, while a vision-language-action model handles execution. A lightweight Event Monitor combines visual observations with end-effector trajectories to determine whether the current subgoal is complete; only after obtaining completion evidence does it write the completed subgoal into the history and trigger subsequent reasoning.

  22. arXiv cs.RO57

    PointHand user study: How keypoint hand rendering affects novice teleoperation demonstrations and diffusion policy success rates

    The paper proposes the PointHand teleoperation interface, which renders a multi-fingered robotic hand as sparse keypoints at the fingertips and joints, matching the keypoints observable by the policy. In a simulated mug-in-box task, 18 novices and 4 experts collected 2,200 trials across three rendering conditions; among them, novices who saw emphasized rendering of the thumb and index finger achieved 441 successful demonstrations out of 600 trials, higher than the 320 out of the real hand-shape rendering.

  23. arXiv cs.RO47

    Hierarchical QP control and authority-guided recovery planning achieve high-precision RCM motion for ophthalmic robots

    A study proposes a high-precision remote center of motion (RCM) servo architecture for ophthalmic robots, using a hierarchical QP controller and authority-guided recovery planning to address servo oscillation caused by conflicts between RCM and tip-tracking objectives. In a 200-scenario kinematic benchmark, the approach achieved a 100% reach rate for 128 near-singular targets and a 99% reach rate for 72 near-joint-limit targets, with software RCM residual below 0.142 mm.

  24. arXiv cs.RO57

    Research on Enumeration-Guided Training of Neural Hamilton-Jacobi Safety Certificates

    This research proposes a verification-guided training framework that uses probabilistic enumeration of safety constraints to detect forward-invariance violations, generating counterexample sets that are more compact and closer to the failure boundary than those from standard adversarial attack methods. Fine-tuning a pretrained HJ value function with these counterexamples expands the certifiable region and shortens certification time across all tested benchmarks; the authors also introduce a deployment-time shielding mechanism based on the refined certificates, achieving zero safety constraint violations in Isaac Sim physics simulation and real-world autonomous racing experiments compared with a nominal controller and an LQR-based backup strategy.

  26. arXiv cs.RO57

    RoboCasa-Shift: A Stage-Aware Robustness Evaluation Benchmark for VLA Policies in Long-Horizon Household Manipulation

    The paper proposes RoboCasa-Shift, a diagnostic robustness benchmark built on RoboCasa365 that combines controlled single-factor perturbations, paired clean-perturbed rollouts, and simulator-grounded stage-aware diagnostics, completing a total of 1,040 rollouts evaluating GR00T N1.5 across four long-horizon tasks, four perturbation factors, and three severity levels.

  27. arXiv cs.RO47

    Agent-Driven Discovery of Interpretable Car-Following Laws: AERM Model Uncovers Car-Following Patterns from Multi-Source Trajectory Data

    An agent-centered framework incorporating physical constraints discovers interpretable car-following laws from multi-source trajectory data, yielding the Adaptive Equilibrium Response Model (AERM), which couples state-dependent relative-speed response with spacing-dependent equilibrium-speed restoration.

  28. arXiv cs.RO67

    Multi-View Latent Action Pretraining for Cross-Embodiment Robot Video Learning

    The paper proposes a latent action model (LAM) that can ingest synchronized multi-camera views, using cross-attention to resolve depth and occlusion issues, then compresses them into latent actions that are fed into a policy in which a latent expert and an action expert are coupled by sequential flow matching around a frozen VLM backbone.

  31. arXiv cs.RO55

    CovNav: Conformal Prediction-Guided Adaptive Inference for ObjectNav

    CovNav is a training-free framework that uses conformal prediction (CP) to turn spatial queries in ObjectNav into structured binary decisions: a local large model in fast mode judges whether the target is likely near an observed object, CP converts its token probabilities into a prediction set, and only ambiguous queries are escalated to slow chain-of-thought reasoning.

  32. arXiv cs.RO47

    PLaCA: A Robot Action Generation Framework Based on Future Latent Change

    PLaCA proposes a world-to-action framework that directly predicts spatially structured patch-level future latent change from current multimodal input and explicitly incorporates it into action generation. On LIBERO-Long, PLaCA achieves a 79.0% success rate, a 5.0 percentage point improvement over the Action-only baseline; on CALVIN, the average sequence length improves from 2.784 to 3.564. The second-stage training updates only 2.95% of the model parameters.

  33. arXiv cs.RO38

    Minimum-Time Escape from the Conic Field of View of a Reeds-Shepp Pursuer: Classification and Game Value for a Slower or Equally Fast Evader

    An arXiv paper studies how an omnidirectional evader escapes in minimum time from the forward conic field of view of a Reeds-Shepp pursuer with minimum turning radius r. For slower or equally fast evaders with speed ratio sinα<ν≤1, it constructs a feedback solution using Isaacs' characteristics and gives a scalar equation for the finite escape time.

  34. arXiv cs.RO65

    Head-Mounted Single Camera Replaces Wrist Camera and Pose Tracking in Electronics-Free UMI Glove Solution

    The paper proposes replacing the wrist camera, pose tracker, and their separate data streams in the UMI glove setup with a single monocular camera (USD 122) worn on the collector's head. It reads printed fiducial markers on an electronics-free glove (USD 22) and paper tiles in the workspace, producing the 6-DoF gripper pose and jaw width from the same frames the policy will see. At deployment, the same camera is mounted on a low-cost (USD 120) robotic arm beside the robot and mimics the collector's head motion based on the robot's joint states.

  36. arXiv cs.RO57

    Tactile-ETA Achieves Fine Manipulation Through Tactile In-Context Learning

    The Tactile-ETA system brings tactile experience into embodied task agents through in-context learning, completing fine manipulation without updating model parameters. Across four simulated manipulation tasks, adding tactile observations to the same two expert demonstrations raised average success from 15.75% to 24.00%; on a real robot task, vision- and tactile-aligned demonstrations raised success from 0% to 60%. Experiments also show that performance depends on the sensory content and representation of the demonstrations, and different agent models differ markedly on the Lift Bottle task.

  37. arXiv cs.RO47

    Orthogonal Residual Learning for Adaptive Sideslip Angle and Cornering Stiffness Estimation

    An orthogonal residual learning method constrains a model corrector via null-space parameterization to be orthogonal to cornering stiffness sensitivity, preventing the corrector from absorbing stiffness-change information and suppressing online adaptation. In full-vehicle simulation, the method reduces sideslip angle error by 12.7% to 22.0% and maintains the same adaptation speed as an uncorrected filter under step changes in stiffness. On public real-vehicle data, a 430-parameter orthogonal corrector retained more response to nominal stiffness offsets than an unconstrained corrector of the same size.

  38. arXiv cs.RO46

    ThERMO Thermal Imaging Speed Optimization Algorithm Improves UR10e Robot Autonomous Tumor Resection Accuracy

    The ThERMO closed-loop thermal feedback algorithm completed autonomous electrosurgical tests on ex vivo porcine tissue using the UR10e-based ASTR system, reducing thermal injury by up to 52% and improving margin accuracy by up to 59%, without requiring data-driven training or iterative tuning. The algorithm adjusts cutting speed via real-time thermal imaging; histological analysis and 3D photogrammetry verified thermal injury extent and margin accuracy, outperforming two fixed-speed control methods.

  39. arXiv cs.RO45

    Robot-Task Suitability in Heterogeneous Multi-Robot Task Allocation

    This study compares 10 deterministic methods and 2 LLM-based variants for heterogeneous multi-robot task allocation, covering rule-based, similarity-based, requirement-oriented, fuzzy, and semantic definitions of suitability. The results show that the methods often disagree on scoring and ranking candidate robots in the same scenario, that generic similarity measures can violate the suitability property of one-sided requirement satisfaction, and that LLM scoring introduces stochasticity and linguistic sensitivity. The paper was accepted to the FM-MRS Workshop at IROS 2026.

  40. arXiv cs.RO48

    Evaluating Social Reasoning of Gemini Robotics-ER on the SHREC Benchmark

    An evaluation tested the social reasoning capabilities of Gemini Robotics-ER on the SHREC human-robot social interaction benchmark and compared it with published general-purpose vision-language model baselines. Results show that Gemini Robotics-ER outperforms GPT-4o-mini on error/competence detection but performs worse on fine-grained social-attribute judgments; adding a visual frame helped one task while hurting the other.