Pith. sign in

REVIEW 2 major objections 6 minor 33 references

Multi-step manipulation task and motion planning guided by video demonstration

T0 review · 2 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Video guidance lets a multi-tree planner solve multi-step pick-and-place tasks with sequential dependencies, validated on three new benchmarks and a real robot.

desk verdict Video-guided multi-tree RRT works for the demonstrated scenes and the new generalization study is useful, but the success rates are conditioned on a resampling filter that hides how often the demonstration becomes infeasible. read the letter →

arxiv 2505.08949 v1 pith:RC5AYBEV submitted 2025-05-13 cs.RO cs.CVcs.SYeess.SY

Kateryna Zorina , David Kovar , Mederic Fourmy , Florent Lamiraux , Nicolas Mansard , Justin Carpentier , Josef Sivic , Vladimir Petrik This is my paper · ORCID
classification cs.ROcs.CVcs.SYeess.SY
keywords taskandmotionplanningvideodemonstrationmulti-stepmanipulationRRTcontactstates6Dobjectposeestimationtrajectoryoptimizationbenchmark
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

The paper claims that a human instructional video can supply enough structure to make multi-step task-and-motion planning tractable: if the planner is told when objects are grasped and released and where objects are placed at those moments, it can solve tasks with sequential dependencies that baseline TAMP planners miss. The video is processed into a sequence of contact-state changes and a 6D pose for every object at each change; these define the small set of placement states the planner is allowed to use. The planner then grows multiple rapidly-exploring random trees rooted at transitions between neighboring placement and grasp states and links them into a start-to-goal path. The approach is tested on three new benchmark tasks (shelf rearrangement, transferring an object through a narrow tunnel, and tray-based waiter transfer) across several robots, and the planned paths are refined by an optimal-control step for execution on a real arm. If correct, it implies that one video demonstration can replace hand-coded task scripts for a class of sequential pick-and-place tasks.

What carries the argument

The machinery is the admissible configuration space—the set of configurations in which every object is either stably resting on a contact surface or rigidly held by the gripper—restricted to the placement states read off the demonstration video. Each contact-state change in the video fixes the object poses for one placement state, and the planner connects consecutive placement states through grasp states, seeding a new RRT-like tree at each sampled transition. The video thus provides the discrete task skeleton; the trees fill in collision-free robot motion without having to discover the contact sequence.

What would settle it

Take the tunnel task and shift the demonstrated intermediate placement pose just outside the robot's reachable workspace while a collision-free alternative placement exists elsewhere inside the tunnel; if the planner then fails or times out instead of finding the alternative, the claim that the video-derived placement states are sufficient guidance would be falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central finding is that the hard combinatorial part of manipulation TAMP—discovering which pick-and-place sequence and which intermediate placements connect the start to the goal—does not need to be searched from scratch when a guiding video supplies the sequence. Contact-state recognizers and 6D pose estimators give the planner the times of grasp/release changes and the object poses at those times, which are turned directly into placement states. A multi-tree RRT that seeds trees at the transitions between these states then connects start and goal reliably. The paper reports that on the proposed benchmark this approach solves multi-object shelf tasks and the narrow-passage tunnel task within the time limit, where the comparison planners fail, and that the same demonstration remains usable when start poses, goal poses, object shapes, or furniture positions are varied.

Load-bearing premise

The demonstration video must already encode a complete and robot-feasible plan: every intermediate object placement and every grasp shown must be one the robot can actually reach and execute, because the planner has no mechanism to invent a different contact sequence when a demonstrated step is infeasible.

Editorial extensions

If this is right

  • Tasks that can be written as a sequence of pick-and-place actions on known objects become tractable for sampling-based planning from a single demonstration, including cases where an object must be parked at an intermediate location to be grasped later.
  • On the proposed benchmark, the video-guided planner solves the tunnel task for non-mobile robots and shelf tasks with up to three objects within the time limit, while the comparison planners fail on those cases; the waiter task is solved more reliably, though more slowly, than the symbolic TAMP baseline.
  • One demonstration can be reused after changing start object poses, goal object poses, object geometries, or the placement of furniture, because demonstration poses are stored relative to the contact-surface frames.
  • Trajectory refinement formulated as an optimal control problem shortens the final path (a 33% average reduction in the reported two-object shelf experiment) at the cost of about 60 seconds of extra computation, enabling smoother real-robot execution.

Reading between the lines

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

  • Beyond the paper: the demonstration is effectively a task skeleton, so the same multi-tree planner could accept the skeleton from any source—language instructions, a programmatic goal specification, or a symbolic planner—as long as grasp/release states and intermediate object poses are provided.
  • Because the planner never searches over alternative contact sequences, solution quality is inherited from the demonstration; comparing plans generated from two different videos of the same task would reveal how much of the reported path length and grasp count comes from the video itself rather than from the planner.
  • Beyond the paper: the generalization experiments vary furniture translations but keep rotations fixed, so a natural stress test would rotate the shelf or table to see whether the frame-relative pose representation keeps the same demonstration usable.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The paper proposes a video-guided task-and-motion planning (TAMP) method. From an instructional video of a human manipulating known objects, the method extracts hand-object contact states and 6D object poses at contact changes (Sec. III-A). These poses define placement states in an admissible configuration space, and grasp states are connected to neighboring placement states through transitions (Sec. III-B). The planner grows multiple RRT trees rooted at these transitions, linking trees until the start and goal configurations are connected (Sec. III-C). The resulting path is simplified with random shortcuts and then refined by solving an optimal control problem to obtain smoother trajectories (Secs. III-E, III-F). The authors introduce a benchmark with three tasks (shelf rearrangement, tunnel transfer, waiter tray transfer) and compare against HPP RRT-connect and PDDLStream on the Franka Panda, UR5, and KMR iiwa robots (Sec. IV). They further evaluate generalization to new start/goal poses, objects, environments, and combined variations, and demonstrate real-robot execution (Sec. IV-C).

Significance. If the reported results hold, the paper makes a useful contribution by showing that a classical sampling-based planner can exploit a single video demonstration to solve multi-step manipulation tasks with sequential dependencies, without learning a policy. The method is geometrically transparent, re-plans when the scene changes, and is evaluated on a new benchmark with several robots. The paper also ships a project page with code and video, which supports reproducibility. The strongest point is that the planner must still discover collision-free configurations, select grasp handles, and connect the demonstrated states, so the output trajectories are not trivially determined by the input poses. The main weakness is that the generalization evaluation filters out cases where the video-demonstrated grasp sequence is infeasible, which narrows the scope of the generalization claim.

major comments (2)
  1. [Sec. IV-C, Fig. 7] The generalization results are conditional on a resampling filter that is not accounted for in the reported success rates. The text states: 'The parameters were resampled if all objects were not graspable by the same handle at the consecutive placement states.' This removes exactly the scenarios in which the video-demonstrated placement sequence is not robot-feasible in the novel scene. The number or fraction of discarded samples is never reported, and the Limitations section (Sec. V) does not mention this filter. As a result, the claim that 'guidance from one video demonstration allows us to solve the given tasks for novel scenes' is stronger than the data support. Please report the rejection rate for each generalization scenario and, ideally, evaluate or discuss what happens when the demonstrated grasp sequence is infeasible, since this is a central limitation of the approach.
  2. [Sec. IV-B, Sec. IV-C, Fig. 6, Fig. 7] The success rates and average metrics are reported without any measure of variance, confidence interval, or number of successful runs underlying each mean. Since the planner is stochastic (random tree sampling, random configurations, random robot base poses) and some differences between methods are small, the reader cannot assess whether the reported improvements are statistically meaningful. Please report error bars, confidence intervals, or at least the per-run success counts for the 10 and 100 runs.
minor comments (6)
  1. [Sec. III-E] The section heading 'Path optimization sing random shortcut' contains a typo; it should read 'Path optimization using random shortcut'.
  2. [Table I and Sec. IV-C] The text states that trajectory refinement decreases path length by 33% on average, but the table shows a decrease from 32.6 rad to 25.7 rad, which is about 21%. Please correct the percentage or clarify the calculation.
  3. [Sec. II and Sec. IV-B] The baseline planner is referred to as both 'PDDLStreams' and 'PDDLStream'; please use one consistent name throughout.
  4. [Sec. IV-B] The names 'KUKA IIW A arm' and 'KMR IIW A robot' appear to contain typos; they should likely be 'KUKA iiwa' and 'KMR iiwa'.
  5. [Sec. III-F, Eq. (1)] In the OCP formulation, the objective includes a sum over keyframe indices i of w_d d(q_i, F_i), but the optimization variables are the control torques u_t and the state x_t = (q_t, v_t). Please clarify how q_i is obtained from x_t for a keyframe index i and how the keyframe set K is defined in practice.
  6. [Sec. III-F] The OCP weights (w_d, w_x, w_u, w_c, w_b) are not specified. Please state whether these were fixed across all tasks or tuned per task, since this affects reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the video-derived contact states and object poses are declared inputs, and the planner still performs genuine search over robot configurations, grasp handles, and collision-free connections.

full rationale

The paper's derivation chain is not circular. The method takes as its explicit input the contact states c_k and object 6D poses A_k extracted from the demonstration video (Sec. III-A). These poses are external observations, not functions of the planner's output. In Sec. III-B the placement states P_k are constructed directly from A_k, and the planner's task is to connect those states through grasp states by sampling robot configurations, solving numerical constraints, checking collisions, and linking trees (Alg. 1 and 2). The output trajectory is therefore not equivalent to the input poses by construction: the planner must still find feasible robot joint configurations, select among multiple grasp handles, and verify collision-free transitions between the demonstrated placement states. The central claim, that video guidance enables solving multi-step sequential-dependency tasks, is supported by the planner's ability to use the extracted states as waypoints while the search itself remains non-trivial. Self-citations are present but not load-bearing in a circular way. Reference [5] (HPP, with a co-author) supplies the admissible-configuration-space modeling formalism and is also used as a baseline; it is an independently published software framework rather than an unverified assertion. Reference [20] (CosyPose, with co-authors) is an off-the-shelf pose estimator used as a component, and reference [6] is the prior conference version of this work. None of these is invoked as a uniqueness theorem or to forbid alternative explanations. The closest concern is the generalization protocol's filtering statement in Sec. IV-C: 'The parameters were resampled if all objects were not graspable by the same handle at the consecutive placement states.' This is an experimental feasibility filter, and the paper separately acknowledges in Sec. V that 'The start and goal object poses should be reachable by the robot and suitable for grasping.' This narrows the scope of the generalization claims and may affect how strongly the results transfer, but it does not make the planner's output definitionally equal to its input. Filtering out instances that violate the method's stated preconditions is a benchmark-design limitation, not a circular derivation. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the assumption that a reliable video-to-state extraction pipeline (contact recognition plus 6D pose estimation) exists and that the demonstrated task sequence is a feasible blueprint for the robot. The method introduces no physical entities but does rely on several hand-chosen algorithmic parameters whose values are not reported.

free parameters (4)
  • eta_sample_tree (probability of sampling a new tree) = not reported
    Gating probability in Alg. 1 line 4 controls the balance between growing existing trees and seeding new trees at video-derived transitions; its value is not given, affecting how the results depend on this tuning.
  • delta (tree step size) = not reported
    Step size used for tree extension and linking in Alg. 1 and Alg. 2; no value is reported.
  • OCP cost weights (w_d, w_x, w_u, w_c, w_b) = not reported
    Weights in Eqs. (1)-(2) determine the trade-off between keyframe pose tracking, state regularization, torque cost, collision avoidance, and joint limits; values and any scheduling ('large weight w_x' relaxed during optimization) are not quantified.
  • Generalization sampling bounds = e.g., shelf x in [-0.5, 0.5], y in [0.0, 0.3], z in [-0.5, 0.5]; table x in [-1.0, 1.0], y in [-2.0, 0.0]
    Manually defined bounds in Sec. IV-C define the generalization test distribution; the reported success rates and planning times are specific to these ranges.
assumptions (6)
  • domain assumption The 3D models of all manipulated objects are known in advance (Sec. III-A).
    The pose estimator [20] requires known CAD models; the method cannot handle unseen object geometries.
  • domain assumption The contact recognizer [19] and CosyPose [20] produce accurate contact-state change times and 6D poses at those times (Sec. III-A).
    If either estimator fails under occlusion or lighting, the constructed placement states are wrong and the planner is misled.
  • domain assumption The demonstration shows one object manipulated at a time using pick-and-place motions without pushing or sliding (Sec. III-A and Limitations).
    The admissible configuration space and the placement-state sequence assume this restricted interaction class.
  • domain assumption A single-arm robot is used, and the demonstration's start and goal object poses are reachable and graspable by the robot (Sec. III-A, Limitations).
    The paper states that unreachable or too-close objects cause planning failure.
  • domain assumption The admissible configuration space formulation with placement and grasp constraints from HPP [5] is taken as given (Sec. III-B).
    The paper builds on HPP's state/transition modeling and its numerical constraint solver [26].
  • domain assumption Robot dynamics f is modeled by the Articulated Body Algorithm with gravity set to zero due to internal gravity compensation (Sec. III-F).
    The OCP refinement assumes this dynamics model; the zero-gravity assumption is specific to the experimental platform.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-step manipulation task and motion planning guided by video demonstration." pith.science (2026). https://pith.science/paper/RC5AYBEV

@misc{pith2026250508949,
  author       = {Pith},
  title        = {Pith review of: Multi-step manipulation task and motion planning guided by video demonstration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RC5AYBEV}},
  note         = {Machine review of arXiv:2505.08949}
}
read the original abstract

This work aims to leverage instructional video to solve complex multi-step task-and-motion planning tasks in robotics. Towards this goal, we propose an extension of the well-established Rapidly-Exploring Random Tree (RRT) planner, which simultaneously grows multiple trees around grasp and release states extracted from the guiding video. Our key novelty lies in combining contact states and 3D object poses extracted from the guiding video with a traditional planning algorithm that allows us to solve tasks with sequential dependencies, for example, if an object needs to be placed at a specific location to be grasped later. We also investigate the generalization capabilities of our approach to go beyond the scene depicted in the instructional video. To demonstrate the benefits of the proposed video-guided planning approach, we design a new benchmark with three challenging tasks: (I) 3D re-arrangement of multiple objects between a table and a shelf, (ii) multi-step transfer of an object through a tunnel, and (iii) transferring objects using a tray similar to a waiter transfers dishes. We demonstrate the effectiveness of our planning algorithm on several robots, including the Franka Emika Panda and the KUKA KMR iiwa. For a seamless transfer of the obtained plans to the real robot, we develop a trajectory refinement approach formulated as an optimal control problem (OCP).

Figures

Figures reproduced from arXiv: 2505.08949 by the authors.

Figure 1
Figure 1. The proposed planning approach is guided by the demonstration [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Approach overview. (i) First, we extract contact states and 6D object poses from the input instructional video, as described in Sec. III-A. (ii) Next, we grow multiple trees in the admissible configuration space until we find a path between the start and goal configurations. More details on the state space are in Sec. III-B, and more details on planning the path in Sec. III-C. (iii) This path is then further refined… view at source ↗
Figure 3
Figure 3. The admissible configuration space is a set of (i) placement states Pi where objects rest on the contact surfaces at specific poses and (ii) grasp states Gj where one of the objects is grasped by the robot gripper. To transit to another placement state, i.e. to change the pose of one object, the robot needs to grasp the object and release it at the new location. Since we define several ways to grasp each object, we … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Generalization scenarios include variation of: start object poses (4a), goal object poses (4b), object type (4c) and environment (4d). 4a, 4b: start and goal object pose variation. We sample randomly x and y displacements for object start (goal) poses and a rotation ar…
Figure 5
Figure 5. Figure 5: The proposed benchmark includes three tasks: the shelf task (a) with the goal of moving several objects to a predefined pose on the table or on the shelf; the tunnel task (b) with the goal of transferring an object through the tunnel; the waiter task (c) with the goal …
Figure 6
Figure 6. Figure 6: Results reported for different robots (rows) and different metrics (columns). We report (from left): the success rate, the planning time [s], the path length, and the number of grasps. For the success rate, the higher number the better (↑); for the other metrics lower …
Figure 7
Figure 7. Figure 7: Quantitative evaluation of generalization capabilities. We study performance for the following generalization scenarios (see IV-C for more details): (i) start object pose variation: random modification of start object poses; (ii) goal object pose variation: random modi…
Figure 8
Figure 8. Figure 8: Planning time per object pose. We show two heatmaps that show planning time (in seconds) when (A) the object start pose is varied, the goal pose is fixed on the shelf (highlighted with green), (B) the object goal pose is varied, the start pose is fixed on the table (hi…
Figure 9
Figure 9. Figure 9: Real robot experiment. The first line (A) shows the demonstration video with the recognized object pose and state of contact between the hand and the object (blue and red rectangles). The second line (B) shows the visualization of the planned trajectory. Here starting …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 28 canonical work pages

  1. [1]

    K. M. Lynch and F. C. Park,Modern robotics. Cambridge University Press, 2017

    2017

  2. [2]

    Probabilis- tic roadmaps for fast path planning in high dimensional configuration spaces,

    L. E. Kavraki, P. ˇSvestka, J.-C. Latombe, and M. Overmars, “Probabilis- tic roadmaps for fast path planning in high dimensional configuration spaces,”IEEE Transactions on Robotics and Automation, vol. 12, no. 4, pp. 566–580, 1996

    work page 1996

  3. [3]

    Rrt-connect: An efficient approach to single- query path planning,

    J. Kuffner and S. LaValle, “Rrt-connect: An efficient approach to single- query path planning,” inProceedings of 2000 ICRA. Millennium Con- ference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065), vol. 2, 2000, pp. 995– 1001 vol.2

    work page 2000

  4. [4]

    Motion planning for humanoid robots,

    J. Kuffner, K. Nishiwaki, S. Kagami, M. Inaba, and H. Inoue, “Motion planning for humanoid robots,” inRobotics research. The eleventh international symposium. Springer, 2005, pp. 365–374

    work page 2005

  5. [5]

    Prehensile manipulation planning: Mod- eling, algorithms and implementation,

    F. Lamiraux and J. Mirabel, “Prehensile manipulation planning: Mod- eling, algorithms and implementation,”IEEE Transactions on Robotics, vol. 38, no. 4, pp. 2370–2388, 2022

    work page 2022

  6. [6]

    Multi-contact task and motion planning guided by video demonstration,

    K. Zorina, D. Kovar, F. Lamiraux, N. Mansard, J. Carpentier, J. Sivic, and V . Petrik, “Multi-contact task and motion planning guided by video demonstration,” in2023 IEEE International Conference on Robotics and Automation (ICRA), 2023, pp. 3764–3770

    work page 2023

  7. [7]

    [Online]

    Guided TAMP Project Webpage. [Online]. Available: https://data.ciirc. cvut.cz/public/projects/2023VideoGuidedTAMP

    work page

  8. [8]

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

    S. M. LaValleet al., “Rapidly-exploring random trees: A new tool for path planning,”The annual research report, 1998

    work page 1998

Show all 33 references
  1. [9]

    Augmenting rrt-planners with local trees,

    M. Strandberg, “Augmenting rrt-planners with local trees,” inIEEE In- ternational Conference on Robotics and Automation, 2004. Proceedings. ICRA ’04. 2004, vol. 4, 2004, pp. 3258–3262 V ol.4

    2004

  2. [10]

    Multi-goal path planning using multiple random trees,

    J. Jano ˇs, V . V on´asek, and R. P ˇeniˇcka, “Multi-goal path planning using multiple random trees,”IEEE Robotics and Automation Letters, vol. 6, no. 2, pp. 4201–4208, 2021

    2021

  3. [11]

    A multi-tree extension of the transition-based rrt: Application to ordering-and-pathfinding problems in continuous cost spaces,

    D. Devaurs, T. Sim ´eon, and J. Cort ´es, “A multi-tree extension of the transition-based rrt: Application to ordering-and-pathfinding problems in continuous cost spaces,” in2014 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2014, pp. 2991–2996

    2014

  4. [12]

    Pddlstream: Integrating symbolic planners and blackbox samplers via optimistic adaptive planning,

    C. R. Garrett, T. Lozano-P ´erez, and L. P. Kaelbling, “Pddlstream: Integrating symbolic planners and blackbox samplers via optimistic adaptive planning,”Proceedings of the International Conference on Automated Planning and Scheduling, vol. 30, no. 1, pp. 440–448, Jun

  5. [13]

    Dealing with difficult instances of object rearrangement

    A. Krontiris and K. E. Bekris, “Dealing with difficult instances of object rearrangement.” inRobotics: Science and Systems, vol. 1123, 2015

    2015

  6. [14]

    High-quality tabletop rearrangement with overhand grasps: Hardness results and fast methods,

    H. Shuai, N. Stiffler, A. Krontiris, K. E. Bekris, and J. Yu, “High-quality tabletop rearrangement with overhand grasps: Hardness results and fast methods,” inRobotics: Science and Systems (RSS), Cambridge, MA, 07/2017 2017. [Online]. Available: https://arxiv.org/pdf/1705.09180.pdf

    2017 arXiv

  7. [15]

    Monte-carlo tree search for efficient visually guided rearrangement planning,

    Y . Labb ´e, S. Zagoruyko, I. Kalevatykh, I. Laptev, J. Carpentier, M. Aubry, and J. Sivic, “Monte-carlo tree search for efficient visually guided rearrangement planning,”IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 3715–3722, 2020

    2020

  8. [16]

    Guided motion planning,

    G. Ye and R. Alterovitz, “Guided motion planning,” inRobotics re- search. Springer, 2017, pp. 291–307

    2017

  9. [17]

    Part-based robot grasp planning from human demonstration,

    J. Aleotti and S. Caselli, “Part-based robot grasp planning from human demonstration,” in2011 IEEE International Conference on Robotics and Automation, 2011, pp. 4554–4560

    2011

  10. [18]

    Translating videos to commands for robotic manipulation with deep recurrent neural networks,

    A. Nguyen, D. Kanoulas, L. Muratore, D. G. Caldwell, and N. G. Tsagarakis, “Translating videos to commands for robotic manipulation with deep recurrent neural networks,” in2018 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2018, pp. 3782–3788. 11 A B C...

    2018

  11. [19]

    Understanding human hands in contact at internet scale,

    D. Shan, J. Geng, M. Shu, and D. F. Fouhey, “Understanding human hands in contact at internet scale,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2020

    2020

  12. [20]

    Cosypose: Consistent multi-view multi-object 6d pose estimation,

    Y . Labb ´e, J. Carpentier, M. Aubry, and J. Sivic, “Cosypose: Consistent multi-view multi-object 6d pose estimation,” inProceedings of the Euro- pean Conference on Computer Vision (ECCV), A. Vedaldi, H. Bischof, T. Brox, and J.-M. Frahm, Eds. Cham: Springer International Publ...

    2020

  13. [21]

    Reconstructing hand-object interactions in the wild,

    Z. Cao, I. Radosavovic, A. Kanazawa, and J. Malik, “Reconstructing hand-object interactions in the wild,” inProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2021, pp. 12 417–12 426

    2021

  14. [22]

    Learning joint reconstruction of hands and manipulated objects,

    Y . Hasson, G. Varol, D. Tzionas, I. Kalevatykh, M. J. Black, I. Laptev, and C. Schmid, “Learning joint reconstruction of hands and manipulated objects,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2019

    2019

  15. [23]

    End-to-end training of deep visuomotor policies,

    S. Levine, C. Finn, T. Darrell, and P. Abbeel, “End-to-end training of deep visuomotor policies,”The Journal of Machine Learning Research, vol. 17, no. 1, pp. 1334–1373, 2016

    2016

  16. [24]

    Learning to control a low-cost manipulator using data-efficient reinforcement learning,

    M. P. Deisenroth, C. E. Rasmussen, and D. Fox, “Learning to control a low-cost manipulator using data-efficient reinforcement learning,” Robotics: Science and Systems VII, vol. 7, pp. 57–64, 2011

    2011

  17. [25]

    Learning robust manipulation skills with guided policy search via generative motor reflexes,

    P. Ennen, P. Bresenitz, R. V ossen, and F. Hees, “Learning robust manipulation skills with guided policy search via generative motor reflexes,” in2019 International Conference on Robotics and Automation (ICRA). IEEE, 2019, pp. 7851–7857

    2019

  18. [26]

    Nocedal and S

    J. Nocedal and S. J. Wright,Numerical Optimization, 2nd ed. New York, NY , USA: Springer, 2006

    2006

  19. [27]

    On delaying collision checking in prm planning: Application to multi-robot coordination,

    G. S ´anchez and J.-C. Latombe, “On delaying collision checking in prm planning: Application to multi-robot coordination,”The International Journal of Robotics Research, vol. 21, no. 1, pp. 5–26, 2002. [Online]. Available: https://doi.org/10.1177/027836402320556458

    2002 doi

  20. [28]

    aligator

    W. Jallet, A. Bambade, S. El Kazdadi, C. Justin, and M. Nicolas, “aligator.” [Online]. Available: https://github.com/Simple-Robotics/ aligator

  21. [29]

    PROXDDP: Proximal Constrained Trajectory Optimiza- tion,

    W. Jallet, A. Bambade, E. Arlaud, S. El-Kazdadi, N. Mansard, and J. Carpentier, “PROXDDP: Proximal Constrained Trajectory Optimiza- tion,” 2023, https://inria.hal.science/hal-04332348v1

    2023

  22. [30]

    The pinocchio c++ library – a fast and flexible implementation of rigid body dynamics algorithms and their analytical derivatives,

    J. Carpentier, G. Saurel, G. Buondonno, J. Mirabel, F. Lamiraux, O. Stasse, and N. Mansard, “The pinocchio c++ library – a fast and flexible implementation of rigid body dynamics algorithms and their analytical derivatives,” inIEEE International Symposium on System Integration...

    2019

  23. [31]

    A micro lie theory for state estimation in robotics,

    J. Sola, J. Deray, and D. Atchuthan, “A micro lie theory for state estimation in robotics,”arXiv preprint arXiv:1812.01537, 2018

    2018 arXiv

  24. [32]

    Posecnn: A convolu- tional neural network for 6d object pose estimation in cluttered scenes,

    Y . Xiang, T. Schmidt, V . Narayanan, and D. Fox, “Posecnn: A convolu- tional neural network for 6d object pose estimation in cluttered scenes,” Robotics: Science and Systems (RSS), 2018

    2018

  25. [2020]

    Available: https://ojs.aaai.org/index.php/ICAPS/article/ view/6739

    [Online]. Available: https://ojs.aaai.org/index.php/ICAPS/article/ view/6739

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.