Pith. sign in

REVIEW 4 major objections 5 minor 49 references

Learning Fast, Tool aware Collision Avoidance for Collaborative Robots

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims a tool-aware collision avoidance system that switches between precise classical control and a reactive RL policy, running in under 10 ms with sub-millimeter accuracy and about 60% lower compute than a GPU planner.

desk verdict A useful but modest extension of learned safety-critic + CMDP: tool-aware modes and randomized tool geometry are the real news; the dynamic real-world claim is not yet supported. read the letter →

arxiv 2508.20457 v1 pith:F4FY236L submitted 2025-08-28 cs.RO cs.SYeess.SY

Joonho Lee , Yunho Kim , Seokjoon Kim , Quan Nguyen , Youngjin Heo This is my paper
classification cs.ROcs.SYeess.SY
keywords collisionavoidancecollaborativerobotsconstrainedreinforcementlearningsafetycritictool-awarecontrolpointcloudperceptionpartialobservabilitydynamicobstacle
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that a collaborative robot can dodge moving, partially occluded obstacles in under 10 milliseconds while keeping sub-millimeter positioning accuracy, by pairing a learned risk estimate with a classical inverse-kinematics controller. Existing approaches are said to fall short in dynamic settings because they assume full visibility and fixed tools, while purely learned policies are too imprecise for nominal operation. The system takes the desired end-effector pose, the tool geometry, and an interaction mode (Engage, where the tool may touch, or Protective, where no contact is allowed) and switches between high-precision IK and a constrained-reinforcement-learning avoidance policy whenever a learned safety critic flags high risk. In simulated and real tests it is reported to beat APF and MPPI baselines on dynamic-obstacle collision rates and to use about 60% less computation than a GPU-based planner. The practical payoff is a modular safety layer that can be dropped into pick-and-place and other contact-rich tasks without retraining for each new tool.

What carries the argument

The load-bearing pieces: (1) the safety critic, a scalar in [0,1] estimating the discounted probability of future constraint violations (following [15]), trained by value iteration and aggregated as the max of the body and tool values; (2) the constrained-MDP policy optimized with P3O [36], with rewards for pose tracking and command smoothness and costs for collision, tool-region violation, and speed limits; (3) the encoder-decoder 3D CNN whose latent state feeds both critic and policy and whose occupancy output retains occluded areas; and (4) the switching rule, critic at or above 0.8 triggers the RL policy directly, otherwise the policy output initializes iterative IK, which balances react

What would settle it

Run the trained system on the real robot with the same dynamic-obstacle protocol as Section IV-F but a different depth camera model and different lighting, with no fine-tuning: if the 50-trial collision rate rises well above the reported 6% at 0.2 m/s obstacle speed, or the safety critic stops gating correctly, the sim-to-real transfer claim collapses. A complementary decisive test is a tool shape never sampled in training; the tool-violation rate staying near the simulated ~1.3% would support generalization, a large increase would refute it.

Watch

Extended reading notes

Core claim

Central claim: tool-aware collision avoidance can be a hybrid of a classical precise controller and a fast learned policy, gated by a learned safety critic. A 3D CNN encoder-decoder over voxelized point clouds reconstructs the collidable region, remembers occlusions, filters out the robot body and tool, and outputs safety-critic values for body and tool constraints. When the critic passes a threshold, a CMDP-trained policy (P3O) generates smooth joint residuals at 50 Hz; otherwise the policy output seeds an iterative IK solver for precision. With tool regions randomized in training and a cost checking the signed distance at the tool's eight bounding-box corners, one network adapts to tool si

Load-bearing premise

The perception, safety critic, and policy are trained only in simulation with approximate robot and workspace geometry and no explicit domain randomization, yet the real-world results depend on them transferring to the real robot and a real depth camera without sim-to-real fine-tuning.

Editorial extensions

If this is right

  • The 50 Hz loop with a ~6 ms average update means the system can run above 100 Hz on a mid-range GPU, leaving compute for higher-level planning.
  • The Engage/Protective mode input lets a single controller handle both contact-rich phases (grasping) and transport phases, changing only the collision cost applied to the tool region.
  • Learned perception removes the need for explicit robot-body filtering, mapping, and hand-crafted occlusion heuristics, simplifying integration into existing pipelines such as the pick-and-place demo with ANY-grasp.
  • Resource usage drops to 0.45 GB of GPU memory and sub-10 ms latency, versus a GPU-accelerated optimizer's 468 ms average planning time and ~4.5 GB, enabling reactive avoidance on robots that share compute with other processes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The gating pattern, a learned safety critic deciding when to hand control from a precise classical controller to a fast learned one, is a transferable template for other platforms such as mobile manipulators or humanoids, where reactive and precise modes must coexist.
  • Because training targets 2.5D scenes with a fixed side camera, extending to deep cavities or heavily occluded shapes, which the paper explicitly excludes, would likely require fusing a wrist camera or multiple views.
  • A testable extension: use the learned safety critic purely as a monitor that triggers replanning inside a global optimizer, combining long-horizon optimality with fast reactivity without the paper's IK/RL switching.
  • The absence of explicit domain randomization leaves cross-camera, cross-lighting, and novel-tool robustness as the most uncertain aspect of the real-world claims; validating on a second camera without fine-tuning would sharpen the transfer statement.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a modular, tool-aware collision-avoidance system for collaborative manipulators. A learned encoder-decoder converts a single depth-camera point cloud into an occupancy grid and a learned safety critic that estimates the discounted probability of future collisions for the robot body and tool. A constrained RL policy (P3O) trained in simulation generates reactive joint-target residuals when the safety-critic value exceeds a threshold; otherwise a classical differential-IK controller tracks the desired end-effector pose. The system takes user-specified tool bounding boxes and an interaction mode (Engage vs. Protective) as inputs. The authors report simulated and real-robot experiments claiming sub-millimeter tracking accuracy, control latency under 10 ms, lower collision rates than APF and MPPI in dynamic scenarios, and substantially lower compute and memory use than the Curobo trajectory optimizer.

Significance. If the stated results hold, the paper makes a useful practical contribution: it combines learned perception, a learned safety critic, and a traditional IK controller in a way that is modular and can be integrated into existing pick-and-place pipelines. The tool-aware Engage/Protective distinction is a sensible response to real contact-rich tasks, and the paper gives a concrete architecture, training procedure, and baseline comparisons. The work also has several concrete strengths: the system is demonstrated on a real robot for a static-obstacle task; the dynamic comparison includes two perception variants and standard baselines; and the computational-resource measurements are clearly reported. However, the strongest quantitative claims — sub-millimeter accuracy, under-10 ms latency, and superiority in dynamic environments — are weakened by an implausible reported error value, an end-to-end latency accounting that excludes point-cloud processing, and the fact that the dynamic comparison is entirely simulated with no real-robot dynamic validation or sim-to-real robustness analysis.

major comments (4)
  1. [Sec. IV-D4 / Abstract] The paper claims 'sub-millimeter accuracy' in the abstract, but Sec. IV-D4 states that 'open areas consistently exhibit errors around 0.01 mm.' An error of 0.01 mm (10 µm) is far below typical manipulator repeatability and the resolution of joint-encoder-based forward-kinematic measurements; this is almost certainly a typo for 0.01 m (10 mm) or 0.1 mm. Because the accuracy claim is a headline contribution, this must be corrected and the implications for the 'sub-millimeter' statement re-evaluated. Additionally, errors are computed from forward kinematics rather than an external ground-truth tracker, so 'tracking error' should not be equated with absolute positioning accuracy.
  2. [Sec. IV-C / Abstract] The 'under 10 ms' latency claim excludes point-cloud processing. Section IV-C reports model inference averaging 1.3 ms and other processes 3.9 ms, but point-cloud processing (voxelization, occupancy, raycasting) averages 7.9 ms with a maximum of 72.3 ms and runs 'in parallel.' Since the occupancy grid and safety critic cannot be computed without this point-cloud processing, end-to-end latency from sensor readout to command exceeds 10 ms in typical cycles and can be an order of magnitude higher in the worst case. The paper should report end-to-end latency (including sensing, point-cloud processing, inference, and communication) or clearly redefine what 'under 10 ms' means.
  3. [Sec. IV-A and Sec. III-D2] The learning-based perception and safety critic are trained entirely in simulation using approximate geometry, randomized tool boxes, and no reported domain randomization or sim-to-real fine-tuning. The real-robot experiments in Sec. IV-D use a single static obstacle and report qualitative trajectories and safety-critic values, not collision rates or perception accuracy. The quantitative dynamic-obstacle comparison in Sec. IV-F (Table II) is entirely simulated, and the simulation pauses each step until computation completes, so it does not demonstrate real-time dynamic performance. Given that the central claim is 'outperforms APF/MPPI in dynamic environments,' the absence of real-robot dynamic validation — or at least a real-world perception-accuracy evaluation — leaves the sim-to-real transfer as an unvalidated load-bearing assumption.
  4. [Sec. IV-F, Table II / Abstract] The dynamic-obstacle results show that the proposed method still has a 36% collision rate at 0.4 m/s obstacle speed and 6% at 0.2 m/s. This is better than the baselines but is a high absolute collision rate for a safety-related system; the abstract's phrase 'outperforms traditional approaches (APF, MPPI) in dynamic environments' should be qualified with these absolute numbers and the fact that no real-robot dynamic trials are presented. Also, the abstract states 'approximately 60% lower computational cost compared to a state-of-the-art GPU-based planner,' which is inconsistent with the paper's own comparison: 6 ms vs. 468 ms average for Curobo is far more than a 60% reduction, and the conclusion claims 'over 10 times more memory-efficient.' Please clarify the intended metric and make the abstract consistent with the reported measurements.
minor comments (5)
  1. [Eq. (1)] The summation notation 'TX t′>t' appears malformed; it should presumably be a standard sum over t' after t. Please fix the LaTeX/typesetting.
  2. [Fig. 6 caption] The caption lists panels '(C) RL-only approach. (D) Trajectory optimization' but the body text in Sec. IV-E refers to 'Fig. 6D' as RL alone and 'Fig. 6E' as Curobo. The panel labels in the caption and the in-text references must be reconciled.
  3. [Sec. IV-C] The relationship between the 50 Hz update rate, the 6 ms average update time, and the 7.9 ms point-cloud processing is unclear. If point-cloud processing runs asynchronously, the paper should explain how stale occupancy can affect safety-critic values and when consistency is ensured.
  4. [Sec. III-B / Sec. IV-D2] The safety-critic threshold of 0.8 is described as a design choice with no sensitivity analysis. Since the critic is a learned estimator rather than a certified bound, a sentence on how the threshold was selected and its effect on the Engage/Protective behavior would strengthen the presentation.
  5. [Sec. IV-A] The wrist camera is stated to be unused 'for collision avoidance,' but it appears in Fig. 4A. Clarifying the sensor configuration in the figure would avoid confusion.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central claims are supported by independent simulation training and external baselines; only minor, non-load-bearing self-citations exist.

full rationale

The paper's core contributions—tool-aware collision avoidance via a learned perception model, a safety critic, and a constrained RL policy—are derived from data and training processes rather than being defined in terms of the experimental outcomes. The safety critic is trained via Bellman value iteration on collisions experienced by the nominal differential IK controller in simulation, so its values are not fitted to the final experimental results. The RL policy is then optimized with the critic frozen, using an externally published P3O algorithm, and its performance is evaluated in simulation against standard baselines (APF, MPPI) under controlled conditions. Real-world experiments are qualitative and separately validate the adaptive behavior. Self-citations (e.g., Kim et al. [38], Lee et al. [39]) merely provide background for the CMDP framework and do not supply a load-bearing premise or a unique theorem; the framework itself is standard (Altman [35]) and the P3O algorithm is independently published. No equation or construction reduces to its own inputs, and no fitted parameter is renamed as a prediction. The main weakness—limited sim-to-real validation—is an empirical robustness concern, not a circularity in the derivation. Therefore the circularity score is low.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The system introduces no new physical entities or mediators. The 'safety critic' and 'RL policy' are learned functions, not new objects. The main invented abstraction is the parameterized tool bounding box with corner-based ESDF checks, which is an approximation rather than a new entity.

free parameters (5)
  • Safety critic threshold = 0.8
    Section IV-D.2: controller switches to RL when cost exceeds 0.8. Chosen by authors; no sensitivity analysis provided.
  • Tool region violation margin = 0.025 m
    Section III-C2/Table I: cost if any ESDF corner value < 0.025. Hand-set margin.
  • Reward weights (position, orientation, smoothness 1, smoothness 2) = 1.0, 1.0, 1e-3, 1e-4
    Table I: hand-tuned weighting of tracking and smoothness terms.
  • Voxel resolution = 0.05 m
    Section IV-A: workspace voxel grid resolution. Affects perception and collision checking.
  • EE speed limit cost threshold = 0.5 m/s
    Table I: speed limit cost threshold, chosen without discussion.
assumptions (4)
  • standard math MDP/CMDP framework and Bellman equations for the safety critic
    Section III-B and III-C1.
  • domain assumption 2.5D environment assumption: obstacles are primarily vertically oriented with varying heights, and deep cavities are excluded
    Section III-A explicitly excludes deep cavity insertion and complex occluded shapes, yet the abstract claims robustness under partial observability.
  • ad hoc to paper Tool shape can be approximated as a bounding box; checking ESDF at the eight corner points is sufficient for collision detection
    Section III-C3 defines the tool region as a bounding box and flags violations when any corner point is within half a voxel width.
  • domain assumption Sim-to-real transfer of the learned perception and policy without explicit domain randomization beyond random tool geometry and obstacle sampling
    Section III-D2 states all training data is generated in simulation using approximate geometry; real experiments (Section IV-A) use the same controller with no described sim-to-real adaptation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Fast, Tool aware Collision Avoidance for Collaborative Robots." pith.science (2026). https://pith.science/paper/F4FY236L

@misc{pith2026250820457,
  author       = {Pith},
  title        = {Pith review of: Learning Fast, Tool aware Collision Avoidance for Collaborative Robots},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F4FY236L}},
  note         = {Machine review of arXiv:2508.20457}
}
read the original abstract

Ensuring safe and efficient operation of collaborative robots in human environments is challenging, especially in dynamic settings where both obstacle motion and tasks change over time. Current robot controllers typically assume full visibility and fixed tools, which can lead to collisions or overly conservative behavior. In our work, we introduce a tool-aware collision avoidance system that adjusts in real time to different tool sizes and modes of tool-environment interaction. Using a learned perception model, our system filters out robot and tool components from the point cloud, reasons about occluded area, and predicts collision under partial observability. We then use a control policy trained via constrained reinforcement learning to produce smooth avoidance maneuvers in under 10 milliseconds. In simulated and real-world tests, our approach outperforms traditional approaches (APF, MPPI) in dynamic environments, while maintaining sub-millimeter accuracy. Moreover, our system operates with approximately 60% lower computational cost compared to a state-of-the-art GPU-based planner. Our approach provides modular, efficient, and effective collision avoidance for robots operating in dynamic environments. We integrate our method into a collaborative robot application and demonstrate its practical use for safe and responsive operation.

Figures

Figures reproduced from arXiv: 2508.20457 by the authors.

Figure 1
Figure 1. Pick-and-place example. Our system manages robot-environment interactions by taking an interaction mode input that determines contact permissions. (Left) The tool area is allowed to make contact with the environment during active use. (Right) At other times, contact is prohibited for both the tool and the robot. adapting trajectories in real time by measuring distances and maintaining safe separation [9], [10], [11]… view at source ↗
Figure 2
Figure 2. System Overview. The RL policy generates target joint positions, used as initial guess for IK when safe. A 3D CNN encoder processes workspace occupancy and proprioceptive inputs to estimate collision probabilities via a learned safety critic. When the risk is high, RL policy is triggered. All components are trained in simulation. We validate our system through real-world experiments and qualitative evaluations. Our … view at source ↗
Figure 3
Figure 3. Simulation Environment. Recent research frames robot navigation as a global op￾timization problem. To handle complex environments with multiple possible solutions, sampling-based approaches are employed. For example, Curobo [13] combines sampling￾based planning and TO for robotic manipulators. These model-based methods generally perform well in fully observable environments where obstacles and dynamics are predictab… view at source ↗
Figures from the paper (5 more)
Figure 5
Figure 5. Figure 5: Visualization of Point Cloud and Collidable Region Estimation. Front view snapshots at the initial configuration, at the start (i) and at a mid (ii) point of the trajectory. (A) Raw point cloud data captured during the experiment. (B) Corresponding estimated collidable…
Figure 4
Figure 4. Figure 4: Tool-Adaptive Collision Avoidance. (A) Experimental setup. (B) Motion in Engage mode, where the tool area is permitted to make contact with the environment. (C) Motion in Protective mode, where the policy and the safety critic treat the tool region as part of the robot…
Figure 6
Figure 6. Figure 6: Tracking Errors Across Different Tool Conditions and Control Approaches. Task-space position tracking error in the yz plane, with target positions in a 0.1 m grid at a fixed x-coordinate of 0.5 m. (A) Engage mode: only robot body collisions are considered. (B) Protecti…
Figure 7
Figure 7. Figure 7: Resource Usage Compared to Curobo (with NvBlox) [13] Dynamic obstacle Moving Waypoint Partially occluded Point cloud A B [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Dynamic obstacle avoidance experiment (A) Experimental setup. The obstacle and waypoint move in opposite direction. The waypoint moves at 0.25 m/s. (B) Visualization of point cloud on the obstacle time to converge to collision-free trajectories, which limits responsive…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

49 extracted references · 43 canonical work pages

  1. [1]

    Learning to open and traverse doors with a legged manipulator,

    M. Zhang, Y . Ma, T. Miki, and M. Hutter, “Learning to open and traverse doors with a legged manipulator,” in 8th Annual Conference on Robot Learning, 2024

    work page 2024

  2. [2]

    Mobile aloha: Learning bimanual mobile manipulation using low-cost whole-body teleoperation,

    Z. Fu, T. Z. Zhao, and C. Finn, “Mobile aloha: Learning bimanual mobile manipulation using low-cost whole-body teleoperation,” in 8th Annual Conference on Robot Learning , 2024

    work page 2024

  3. [3]

    Robot collisions: A survey on detection, isolation, and identification,

    S. Haddadin, A. De Luca, and A. Albu-Sch ¨affer, “Robot collisions: A survey on detection, isolation, and identification,” IEEE Transactions on Robotics, vol. 33, no. 6, pp. 1292–1312, 2017

    work page 2017

  4. [4]

    Collision detection for industrial collaborative robots: A deep learning approach,

    Y . J. Heo, D. Kim, W. Lee, H. Kim, J. Park, and W. K. Chung, “Collision detection for industrial collaborative robots: A deep learning approach,” IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 740–746, 2019

    work page 2019

  5. [5]

    Compensation modulation for tracking accuracy in free motion and compliance during interaction,

    D. Ko, D. Lee, W. Lee, K. Kim, and W. K. Chung, “Compensation modulation for tracking accuracy in free motion and compliance during interaction,” IEEE/ASME Transactions on Mechatronics , 2024

    work page 2024

  6. [6]

    Collision detection and coor- dinated compliance control for a dual-arm robot without force/torque sensing based on momentum observer,

    L. Han, W. Xu, B. Li, and P. Kang, “Collision detection and coor- dinated compliance control for a dual-arm robot without force/torque sensing based on momentum observer,” IEEE/ASME Transactions on Mechatronics, vol. 24, no. 5, pp. 2261–2272, 2019

    work page 2019

  7. [7]

    Control techniques for safe, ergonomic, and efficient human-robot collaboration in the digital industry: A survey,

    S. Proia, R. Carli, G. Cavone, and M. Dotoli, “Control techniques for safe, ergonomic, and efficient human-robot collaboration in the digital industry: A survey,” IEEE Transactions on Automation Science and Engineering, vol. 19, no. 3, pp. 1798–1819, 2021

    work page 2021

  8. [8]

    O. I. de Normalizaci ´on, ISO-TS 15066: Robots and robotic devices: Collaborative robots. ISO, 2016

    work page 2016

Show all 49 references
  1. [9]

    Human-robot coexistence and contact handling with redundant robots,

    E. Magrini and A. De Luca, “Human-robot coexistence and contact handling with redundant robots,” in 2017 IEEE/RSJ International Con- ference on Intelligent Robots and Systems (IROS) . IEEE, 2017, pp. 4611–4617

    2017

  2. [10]

    Optimal proactive path planning for collaborative robots in industrial contexts,

    A. Casalino, D. Bazzi, A. M. Zanchettin, and P. Rocco, “Optimal proactive path planning for collaborative robots in industrial contexts,” in 2019 International Conference on Robotics and Automation (ICRA) . IEEE, 2019, pp. 6540–6546

    2019

  3. [11]

    Multimodal sensor-based whole-body control for human–robot collabo- ration in industrial settings,

    J. de Gea Fern ´andez, D. Mronga, M. G ¨unther, T. Knobloch, M. Wirkus, M. Schr ¨oer, M. Trampler, S. Stiene, E. Kirchner, V . Bargsten, et al. , “Multimodal sensor-based whole-body control for human–robot collabo- ration in industrial settings,” Robotics and Autonomous System...

    2017

  4. [12]

    Computation of safe path velocity for collaborative robots,

    C. Sloth and H. G. Petersen, “Computation of safe path velocity for collaborative robots,” in 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2018, pp. 6142–6148

    2018

  5. [13]

    Curobo: Parallelized collision-free robot motion generation,

    B. Sundaralingam, S. K. S. Hari, A. Fishman, C. Garrett, K. Van Wyk, V . Blukis, A. Millane, H. Oleynikova, A. Handa, F. Ramos, et al. , “Curobo: Parallelized collision-free robot motion generation,” in 2023 IEEE International Conference on Robotics and Automation (ICRA) . IEE...

    2023

  6. [14]

    Safety-aware time-optimal motion planning with uncertain human state estimation,

    M. Faroni, M. Beschi, and N. Pedrocchi, “Safety-aware time-optimal motion planning with uncertain human state estimation,” IEEE Robotics and Automation Letters , vol. 7, no. 4, pp. 12 219–12 226, 2022

    2022

  7. [15]

    Learning to be safe: Deep rl with a safety critic,

    K. Srinivasan, B. Eysenbach, S. Ha, J. Tan, and C. Finn, “Learning to be safe: Deep rl with a safety critic,” arXiv preprint arXiv:2010.14603 , 2020

    2010 arXiv

  8. [16]

    Anymal parkour: Learning agile navigation for quadrupedal robots,

    D. Hoeller, N. Rudin, D. Sako, and M. Hutter, “Anymal parkour: Learning agile navigation for quadrupedal robots,” Science Robotics , vol. 9, no. 88, p. eadi7566, 2024

    2024

  9. [17]

    V oxblox: Incremental 3d euclidean signed distance fields for on-board mav planning,

    H. Oleynikova, Z. Taylor, M. Fehr, R. Siegwart, and J. Nieto, “V oxblox: Incremental 3d euclidean signed distance fields for on-board mav planning,” in 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2017, pp. 1366–1373

    2017

  10. [18]

    Regularized deep signed distance fields for reactive motion generation,

    P. Liu, K. Zhang, D. Tateo, S. Jauhri, J. Peters, and G. Chalvatzaki, “Regularized deep signed distance fields for reactive motion generation,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022, pp. 6673–6680

    2022

  11. [19]

    Storm: An integrated framework for fast joint- space model-predictive control for reactive manipulation,

    M. Bhardwaj, B. Sundaralingam, A. Mousavian, N. D. Ratliff, D. Fox, F. Ramos, and B. Boots, “Storm: An integrated framework for fast joint- space model-predictive control for reactive manipulation,” in Conference on Robot Learning . PMLR, 2022, pp. 750–759

    2022

  12. [20]

    Reinforcement learning for collision-free flight exploiting deep collision encoding,

    M. Kulkarni and K. Alexis, “Reinforcement learning for collision-free flight exploiting deep collision encoding,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 15 781–15 788

    2024

  13. [21]

    Learning to walk in confined spaces using 3d representation,

    T. Miki, J. Lee, L. Wellhausen, and M. Hutter, “Learning to walk in confined spaces using 3d representation,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 8649–8656

    2024

  14. [22]

    Real-time neural dense elevation mapping for urban terrain with uncertainty estimations,

    B. Yang, Q. Zhang, R. Geng, L. Wang, and M. Liu, “Real-time neural dense elevation mapping for urban terrain with uncertainty estimations,” IEEE Robotics and Automation Letters, vol. 8, no. 2, pp. 696–703, 2022

    2022

  15. [23]

    Pie: Parkour with implicit-explicit learning framework for legged robots,

    S. Luo, S. Li, R. Yu, Z. Wang, J. Wu, and Q. Zhu, “Pie: Parkour with implicit-explicit learning framework for legged robots,” IEEE Robotics and Automation Letters , 2024

    2024

  16. [24]

    Rapidly-exploring random trees: A new tool for path planning,

    S. LaValle, “Rapidly-exploring random trees: A new tool for path planning,” Research Report 9811 , 1998

    1998

  17. [25]

    Real-time obstacle avoidance for manipulators and mobile robots,

    O. Khatib, “Real-time obstacle avoidance for manipulators and mobile robots,” The international journal of robotics research, vol. 5, no. 1, pp. 90–98, 1986

    1986

  18. [26]

    Chomp: Gradient optimization techniques for efficient motion planning,

    N. Ratliff, M. Zucker, J. A. Bagnell, and S. Srinivasa, “Chomp: Gradient optimization techniques for efficient motion planning,” in 2009 IEEE international conference on robotics and automation . IEEE, 2009, pp. 489–494

    2009

  19. [27]

    Learning robust autonomous navigation and locomotion for wheeled- legged robots,

    J. Lee, M. Bjelonic, A. Reske, L. Wellhausen, T. Miki, and M. Hutter, “Learning robust autonomous navigation and locomotion for wheeled- legged robots,” Science Robotics, vol. 9, no. 89, p. eadi9641, 2024

    2024

  20. [28]

    Provably safe deep reinforcement learning for robotic manipulation in human environments,

    J. Thumm and M. Althoff, “Provably safe deep reinforcement learning for robotic manipulation in human environments,” in 2022 International Conference on Robotics and Automation (ICRA) . IEEE, 2022, pp. 6344–6350

    2022

  21. [29]

    Armor: Egocentric per- ception for humanoid robot collision avoidance and motion planning,

    D. Kim, M. Srouji, C. Chen, and J. Zhang, “Armor: Egocentric per- ception for humanoid robot collision avoidance and motion planning,” arXiv preprint arXiv:2412.00396 , 2024

    2024 arXiv

  22. [30]

    A comprehensive survey on safe reinforce- ment learning,

    J. Garcıa and F. Fern ´andez, “A comprehensive survey on safe reinforce- ment learning,” Journal of Machine Learning Research , vol. 16, no. 1, pp. 1437–1480, 2015

    2015

  23. [31]

    Consideration of risk in reinforcement learning,

    M. Heger, “Consideration of risk in reinforcement learning,” in Machine Learning Proceedings 1994 . Elsevier, 1994, pp. 105–111

    1994

  24. [32]

    Risk-sensitive reinforcement learning ap- plied to control under constraints,

    P. Geibel and F. Wysotzki, “Risk-sensitive reinforcement learning ap- plied to control under constraints,” Journal of Artificial Intelligence Research, vol. 24, pp. 81–108, 2005

    2005

  25. [33]

    Worst cases policy gradients,

    Y . C. Tang, J. Zhang, and R. Salakhutdinov, “Worst cases policy gradients,” in Conference on Robot Learning . PMLR, 2020, pp. 1078– 1093

    2020

  26. [34]

    Learning risk-aware quadrupedal locomotion using distributional reinforcement learning,

    L. Schneider, J. Frey, T. Miki, and M. Hutter, “Learning risk-aware quadrupedal locomotion using distributional reinforcement learning,” in 2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2024, pp. 11 451–11 458

    2024

  27. [35]

    Altman, Constrained Markov decision processes

    E. Altman, Constrained Markov decision processes . Routledge, 2021

    2021

  28. [36]

    Penalized proximal policy optimization for safe reinforcement learn- ing,

    L. Zhang, L. Shen, L. Yang, S. Chen, X. Wang, B. Yuan, and D. Tao, “Penalized proximal policy optimization for safe reinforcement learn- ing,” in Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, IJCAI-22 . International Joint Conferenc...

    2022

  29. [37]

    Ipo: Interior-point policy optimization under constraints,

    Y . Liu, J. Ding, and X. Liu, “Ipo: Interior-point policy optimization under constraints,” in Proceedings of the AAAI conference on artificial intelligence, vol. 34, no. 04, 2020, pp. 4940–4947

    2020

  30. [38]

    Not only rewards but also constraints: Applications on legged robot locomotion,

    Y . Kim, H. Oh, J. Lee, J. Choi, G. Ji, M. Jung, D. Youm, and J. Hwangbo, “Not only rewards but also constraints: Applications on legged robot locomotion,” IEEE Transactions on Robotics , 2024

    2024

  31. [39]

    Exploring constrained reinforcement learning algorithms for quadrupedal locomotion,

    J. Lee, L. Schroth, V . Klemm, M. Bjelonic, A. Reske, and M. Hut- ter, “Exploring constrained reinforcement learning algorithms for quadrupedal locomotion,” in 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2024, pp. 11 132– 11 138

    2024

  32. [40]

    Constrained policy optimization,

    J. Achiam, D. Held, A. Tamar, and P. Abbeel, “Constrained policy optimization,” in International conference on machine learning. PMLR, 2017, pp. 22–31

    2017

  33. [41]

    Anygrasp: Robust and efficient grasp perception in spatial and temporal domains,

    H.-S. Fang, C. Wang, H. Fang, M. Gou, J. Liu, H. Yan, W. Liu, Y . Xie, and C. Lu, “Anygrasp: Robust and efficient grasp perception in spatial and temporal domains,” IEEE Transactions on Robotics , 2023

    2023

  34. [42]

    Warp: A high-performance python framework for gpu simulation and graphics,

    M. Macklin, “Warp: A high-performance python framework for gpu simulation and graphics,” https://github.com/nvidia/warp, March 2022, nVIDIA GPU Technology Conference (GTC)

    2022

  35. [43]

    A dynamical system approach to realtime obstacle avoidance,

    S. M. Khansari-Zadeh and A. Billard, “A dynamical system approach to realtime obstacle avoidance,” Autonomous Robots, vol. 32, pp. 433–454, 2012. 9

    2012

  36. [44]

    Avoid everything: Model-free collision avoidance with expert-guided fine-tuning,

    A. Fishman, A. Walsman, M. Bhardwaj, W. Yuan, B. Sundaralingam, B. Boots, and D. Fox, “Avoid everything: Model-free collision avoidance with expert-guided fine-tuning,” in 8th Annual Conference on Robot Learning, 2024

    2024

  37. [45]

    Ramp: Hierarchical reactive motion planning for manipulation tasks using implicit signed distance functions,

    V . Vasilopoulos, S. Garg, P. Piacenza, J. Huh, and V . Isler, “Ramp: Hierarchical reactive motion planning for manipulation tasks using implicit signed distance functions,” in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 2023, pp. 10...

    2023

  38. [46]

    In- formed circular fields: a global reactive obstacle avoidance framework for robotic manipulators,

    M. Becker, P. Caspers, T. Lilge, S. Haddadin, and M. A. M ¨uller, “In- formed circular fields: a global reactive obstacle avoidance framework for robotic manipulators,” Frontiers in Robotics and AI , vol. 11, p. 1447351, 2025

    2025

  39. [47]

    Object rear- rangement using learned implicit collision functions,

    M. Danielczuk, A. Mousavian, C. Eppner, and D. Fox, “Object rear- rangement using learned implicit collision functions,” in 2021 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021, pp. 6010–6017

    2021

  40. [48]

    Collisionik: A per-instant pose optimization method for generating robot motions with environ- ment collision avoidance,

    D. Rakita, H. Shi, B. Mutlu, and M. Gleicher, “Collisionik: A per-instant pose optimization method for generating robot motions with environ- ment collision avoidance,” in 2021 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2021, pp. 9995–10 001

    2021

  41. [49]

    Aloha unleashed: A simple recipe for robot dexterity,

    T. Z. Zhao, J. Tompson, D. Driess, P. Florence, S. K. S. Ghasemipour, C. Finn, and A. Wahid, “Aloha unleashed: A simple recipe for robot dexterity,” in 8th Annual Conference on Robot Learning , 2024

    2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.