Pith. sign in

REVIEW 3 major objections 4 minor 34 references

Failing Gracefully: Mitigating Impact of Inevitable Robot Failures

T0 review · 3 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A safety formulation that predicts the harm of inevitable robot failures and plans around it

desk verdict FailBench is a real asset, but the paper's central validation claim is contradicted by its own Table II, and the safety metric itself is a standard risk-weighted cost with a hand-set severity model. read the letter →

arxiv 2608.05313 v1 pith:AIM5BC45 submitted 2026-08-05 cs.RO cs.AIcs.HCcs.LG

Duc M. Nguyen , Saad A. Ghani , Andrew Marshall , Allison Andreyev , Gregory J. Stein , Xuesu Xiao This is my paper · ORCID
classification cs.ROcs.AIcs.HCcs.LG
keywords robotfailuresafetymetricimpactmotionplanningsimulationbenchmarkinjectionservicerobotspick-and-place
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 argues that since service-robot failures are inevitable, robots should plan motions by anticipating the consequences of those failures rather than only preventing them. It proposes a safety formulation in which the impact of a failure is the probability of a harmful interaction between a robot component and an environmental entity, multiplied by the severity of that interaction, summed over the trajectory and weighed against motion cost. To test this, the paper introduces FailBench, a simulation benchmark with a failure injector that produces diverse actuator, sensor, end-effector, and power failures and records contact events. Validation on four pick-and-place trajectories shows that the lowest predicted safety cost trajectory also has the lowest observed safety, with some discrepancies the paper attributes to modeling complexity. The paper's aim is to establish that failure consequences can be quantified ahead of time well enough to inform planning.

What carries the argument

The load-bearing mechanism is the interaction-probability estimator: for each robot component and entity, it takes axis-aligned bounding boxes, projects them onto the horizontal plane, computes the overlap area, and sets the interaction probability to the larger of the two overlap ratios (overlap area divided by component area or entity area), then applies a Z-axis check. This geometric proxy converts the post-failure state into a number, which is multiplied by a severity factor and summed in the planning objective. The simulator's contact-event measurements are the empirical counterpart to that number.

What would settle it

Run FailBench on a large set of randomly sampled trajectories under the same object-drop failure, and compare each trajectory's predicted safety cost with its observed safety. If the trajectory with the lowest predicted cost is not systematically among the lowest observed-safety trajectories — as already happens for trajectories 1 and 2 in Table II, where observed safety is 3.33 and 5.6 versus predictions of 0.347 and 0.430 — then the claim of predictive capability fails.

Watch

Extended reading notes

Core claim

The central claim is that the expected impact of an inevitable failure can be computed as $\sum_t \sum_{r_j} \sum_{e_i} P_t(x_t, e_i, r_j | F) \cdot S(e_i, r_j)$, with $P_t$ estimated from geometric overlap and $S$ a severity factor, and that this quantity predicts the safety outcomes measured in physics simulation. In the reported experiments, trajectory 4 has both the lowest theoretical safety cost (0.236) and the lowest observed safety (0.27), while trajectories 1 and 2 show observed safety (3.33 and 5.6) well above their predicted costs (0.347 and 0.430); the paper reads these as discrepancies that suggest refining the interaction probability model rather than as a failure of the formulation. The formulation is meant to let a robot choose plans that trade motion efficiency against failure harm.

Load-bearing premise

The whole calculation stands on the assumption that the probability a failure harms an entity can be captured by the horizontal overlap of two static bounding boxes divided by their areas, plus a height check, with no account of falling paths, speeds, or who moves.

Editorial extensions

If this is right

  • A planner using the proposed objective can rank candidate trajectories by expected failure harm before execution, without needing to run failures first.
  • The weighting parameter $w$ gives a direct dial between task efficiency and safety, allowing deployment-specific trade-offs.
  • FailBench's failure taxonomy lets researchers compare failure-handling strategies across actuator, sensor, end-effector, and power failures under controlled conditions.
  • The framework's validation on object-drop failures suggests the same formulation could be applied to other inevitable failure modes, provided the interaction probability is recomputed.
  • The benchmark's data on post-failure contacts can serve as a shared evaluation baseline for both classical planners and learned policies.

Reading between the lines

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

  • Beyond the paper, the same expected-impact objective could be used as a reward signal for learned policies, so a robot could be trained to prefer trajectories that are cheap in expected failure impact; FailBench would provide the training signal.
  • The large observed-versus-predicted gaps in trajectories 1 and 2 indicate the geometric overlap proxy is the part most worth replacing; a learned dynamics model trained on FailBench contacts is a testable upgrade.
  • Because severity values in the experiments are hand-set, an automated severity estimator could change which trajectories are preferred; this is a direct test of the framework's sensitivity.
  • The paper treats failures as undetectable; a natural extension would combine this impact cost with a detector that postpones high-impact states when a failure is likely, which the current formulation does not address.
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

3 major / 4 minor

Summary. The paper proposes a safety formulation for service robots that estimates the impact of inevitable failures by combining an interaction probability term and a severity term, and uses this in a planning objective. It also introduces FailBench, a MuJoCo-based simulation framework with a failure injector, motion planners, and contact-based data collection. The framework is evaluated on four trajectories for a tabletop pick-and-place task, comparing a theoretical safety cost with an observed safety (OBS) metric obtained from 60 simulation runs per trajectory. The paper claims that the OBS results 'generally align' with the theoretical predictions and validate the framework's predictive capability, and that the framework can inform failure-aware motion planning.

Significance. If the predictive claim were supported, the proposed formulation would be a useful step toward failure-impact-aware planning for household robots, and FailBench could serve as a useful community benchmark. The paper does provide a broad failure taxonomy, integration with MuJoCo, and a set of planners, which are positive infrastructural contributions. However, the central validation is not supported by the reported data: two of the four trajectories show order-of-magnitude discrepancies between theoretical safety cost and observed safety, and the severity values are hand-assigned and reused in both the theory and the simulation, making the validation circular with respect to the severity model. The paper's own future-work section concedes that the interaction probability model 'lacks the complexity needed for realistic failure impact prediction.' As submitted, the load-bearing claim of predictive capability is contradicted by the evidence in the manuscript.

major comments (3)
  1. [V.B.2 / Table II] The claim that the observed safety (OBS) results 'generally align' with theoretical safety cost predictions and 'validate the framework's predictive capability' is contradicted by Table II for trajectories 1 and 2. Trajectory 1 has a theoretical safety cost of 0.347 but an observed safety of 3.33, a factor of 9.6; trajectory 2 has 0.430 versus 5.6, a factor of 13. Only trajectories 3 and 4 are within a factor of about 1.1. Moreover, the theoretical ranking is Traj4 < Traj1 < Traj3 < Traj2, while the observed ranking is Traj4 < Traj3 < Traj1 < Traj2, so the ordinal agreement also fails. Since the paper's central contribution is the predictive safety metric, this discrepancy is load-bearing and is not resolved by the sentence noting 'opportunities for refining the interaction probability models.'
  2. [V.A / Algorithm 1 Step 6] The interaction probability in Algorithm 1 Step 6 is computed solely from horizontal-plane AABB overlap, P_t = max(A_overlap/A_rj, A_overlap/A_ei), plus a Z-axis check in Step 7. For object-drop failures, whether a dropped object strikes an entity depends on fall dynamics, release position, and entity motion, none of which enter the computation. The paper's Future Work section states that the model 'lacks the complexity needed for realistic failure impact prediction in dynamic environments.' Without any calibration or comparison against the simulated outcomes, the theoretical safety cost in Eq. (1) is not established as a reliable basis for planning decisions, which is the main claim of the paper.
  3. [V.B.1 / V.B.2 / Table II] The observed safety (OBS) metric is never formally defined. The text says it is 'the average measured safety cost across all simulation runs,' but it does not specify how contact events are detected, how contact forces or durations are converted to a cost, or how the severity values (10 for red, 2 for white) are used in the observed computation. If the same hand-assigned severity values are used in both the theoretical safety cost and the OBS, the comparison validates only the geometric probability term, not the severity model; if different definitions are used, the comparison is not well posed. Either way, the reported validation is insufficient to support the framework's predictive capability.
minor comments (4)
  1. [IV.C / Table I] Table I includes Power System Failures, but the overview paragraph in Section IV.C lists only actuator, sensor, and end-effector failures when describing currently supported failure types; either add power system failures to that list or remove the row from the table.
  2. [V.A] The first paragraph of Section V.A says Algorithm 1 models 'individual joint failures and complete system collapse scenarios,' but the evaluation is restricted to object-drop failures. The generality claim should be reconciled with the limited evaluation, or the text should explicitly state that other failure modes are outside the scope of the current validation.
  3. [III.D / Eq. (1)] The weighting parameter w in Eq. (1) is set to 1 in the experiments with no sensitivity analysis; a brief discussion of how w affects the trade-off between motion cost and safety cost would strengthen the planning implications.
  4. [References] Reference [10] is cited in Related Work as 'the RoboFail dataset,' but the reference entry is titled 'Reflect: Summarizing robot experiences for failure explanation and correction'; the dataset name and the reference title do not match, so the citation should be corrected or replaced.

Circularity Check

1 steps flagged · score 4.0 of 10

Partial circularity: the validation of the severity component is self-definitional, because observed and predicted safety costs use the same hand-assigned severity values; the independently testable probability part is contradicted by Table II.

  1. self definitional [Section V.B.1 (Experimental Setup) and Section V.B.2 (Analysis), Eq. (1) and Table II]
    "Object severity values are predefined in this analysis, with red objects representing high severity entities (value 10) and white objects indicating standard severity levels (value 2). ... The observed safety (OBS) represents the average measured safety cost across all simulation runs for each trajectory."

    The theoretical safety cost in Eq. (1) is a sum of w * P_t * S(e_i, r_j), where S is the severity factor. The observed safety is described as a 'measured safety cost', and the only severity values introduced in the analysis are the same predefined values (red=10, white=2). Since no independent severity estimate is defined for OBS, the comparison of predicted versus observed safety uses the identical S values on both sides; the severity component therefore cannot be tested by this validation. The validation reduces to checking the overlap-based probability P_t, which is independent. Table II then contradicts even that check for trajectories 1 and 2 (predicted 0.347 and 0.430 versus observed 3.33 and 5.6).

full rationale

The paper's core formulation is a definition (impact = probability times severity) rather than a derivation, so most of the chain is self-contained. There is no load-bearing self-citation, no imported uniqueness theorem, and no ansatz smuggled in via citation. The one genuinely circular element is the validation step: severity values are hand-assigned and reused in both the theoretical safety cost and the observed safety cost, so the severity part of the 'prediction' is equivalent to its input by construction. The overlap-based interaction probability is not forced by the simulation data, but Table II contradicts the paper's claim that observed safety 'generally align[s]' with theoretical predictions, with order-of-magnitude discrepancies on two of four trajectories. That contradiction is a correctness and validation weakness rather than a further circularity, and the paper's own future-work admission that the interaction model 'lacks the complexity needed for realistic failure impact prediction' corroborates the limitation. Overall, the central contribution retains independent content, but the validation claim is partially circular and overstated; score 4.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

No invented physical entities. FailBench is a simulation tool, not a postulated physical mechanism. All assumptions are explicit modeling choices rather than hidden entities.

free parameters (3)
  • Severity values S(e_i, r_j) = 10 (red), 2 (white)
    Assigned by hand in Section V.B.1; no calibration or estimation method is given, and the same values are used on both sides of the validation.
  • Weight w in Eq. (1) = 1
    Set to 1 in the experiments; the balance between motion cost and safety cost is arbitrary and can change the objective trade-off.
  • Failure injection probability = 25% per trajectory
    Chosen for the experiments in Section V.B.1; not a parameter of the formulation but affects the observed safety statistics.
assumptions (5)
  • domain assumption Complete environmental knowledge of object identity, material properties, and vulnerability
    Acknowledged in Section VI (Future Work) as rarely available in practice; the severity values require this knowledge.
  • domain assumption Failures are inevitable and undetectable, so failure probability itself is not modeled
    Stated in Section III.B; the framework conditions on failure and optimizes only impact, not failure likelihood.
  • ad hoc to paper Interaction probability is approximated by XY-plane AABB overlap plus a Z-axis check
    Algorithm 1, Section V.A; this ignores object dynamics, velocities, and human or pet motion, and is the main source of the validation mismatches.
  • ad hoc to paper The evaluation is restricted to object drop failures
    Section V.A states: 'in this paper we focus exclusively on object drop failures', which does not exercise the sensor, actuator, or power failure modes FailBench advertises.
  • domain assumption Severity is a static, state-independent factor
    Section III.C defines S(e_i, r_j) with no dependence on impact velocity or geometry; real harm depends on kinetic energy and impact location.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Failing Gracefully: Mitigating Impact of Inevitable Robot Failures." pith.science (2026). https://pith.science/paper/AIM5BC45

@misc{pith2026260805313,
  author       = {Pith},
  title        = {Pith review of: Failing Gracefully: Mitigating Impact of Inevitable Robot Failures},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AIM5BC45}},
  note         = {Machine review of arXiv:2608.05313}
}
read the original abstract

Service robots operate in household environments shared with humans, pets, and everyday objects, where they are highly susceptible to failures such as software crashes, hardware degradation, or unpredictable interactions. While roboticists strive to minimize failures, some remain inevitable, making it critical to mitigate their potential consequences for safe and reliable deployment. This paper introduces a novel safety formulation that evaluates both the probability of impactful interactions between robots and surrounding entities during failures, and the severity of their outcomes. By quantifying the impact of failures on different entities, our approach enables robots to make informed planning decisions that balance safety with task efficiency. To support systematic evaluation, we also present FailBench, a MuJoCo-based simulation framework for studying robot-environment interactions under diverse failure modes, including sensing issues and actuator malfunctions. Together, our safety formulation and FailBench provide a foundation for developing safer and more robust motion plans and learned policies in real-world household environments.

Figures

Figures reproduced from arXiv: 2608.05313 by the authors.

Figure 1
Figure 1. Robot failure mitigation example. Left (red): Carrying a bottle of hot water close to a human poses a higher risk of spilling and causing injury if dropped. Right (green): A safer strategy positions the bottle farther away from the human to reduce potential harm in the event of failure. essarily limits where the robot can operate and compromise task efficiency. Another major bottleneck in studying such failure conse… view at source ↗
Figure 2
Figure 2. Illustration of the components within the robot’s [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Visualization of contact detection after shoulder joint [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Trajectories carrying an object during pick-and-place [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 18 canonical work pages

  1. [1]

    Robot risk-awareness by formal risk reasoning and planning,

    X. Xiao, J. Dufek, and R. R. Murphy, “Robot risk-awareness by formal risk reasoning and planning,”IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 2856–2863, 2020

    work page 2020

  2. [2]

    A minimum risk approach for path planning of uavs,

    L. De Filippis, G. Guglieri, and F. Quagliotti, “A minimum risk approach for path planning of uavs,”Journal of Intelligent & Robotic Systems, vol. 61, pp. 203–219, 2011

    work page 2011

  3. [3]

    Zabarankin, S

    M. Zabarankin, S. Uryasev, and P. Pardalos,Optimal risk path algo- rithms. Springer, 2002

    work page 2002

  4. [4]

    An efficient motion planning algorithm for stochastic dynamic systems with constraints on probability of failure

    M. Ono and B. C. Williams, “An efficient motion planning algorithm for stochastic dynamic systems with constraints on probability of failure.” inAAAI, 2008, pp. 1376–1382

    work page 2008

  5. [5]

    Recover: A neuro-symbolic framework for failure detection and recovery,

    C. Cornelio and M. Diab, “Recover: A neuro-symbolic framework for failure detection and recovery,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024, pp. 12 435–12 442

    work page 2024

  6. [6]

    Understanding and resolving failures in human-robot interaction: Literature review and model development,

    S. Honig and T. Oron-Gilad, “Understanding and resolving failures in human-robot interaction: Literature review and model development,” Frontiers in psychology, vol. 9, p. 861, 2018

    work page 2018

  7. [7]

    Follow-up analysis of mobile robot failures,

    J. Carlson, R. R. Murphy, and A. Nelson, “Follow-up analysis of mobile robot failures,” inIEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA’04. 2004, vol. 5. IEEE, 2004, pp. 4987–4994

    work page 2004

  8. [8]

    Explicit motion risk representa- tion,

    X. Xiao, J. Dufek, and R. Murphy, “Explicit motion risk representa- tion,” in2019 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR). IEEE, 2019, pp. 278–283

    work page 2019

Show all 34 references
  1. [9]

    Errors in human-robot interaction accidents: A taxonomy and network analysis,

    B. H. GUO, Z. Yonger, Y . M. Goh, and J.-Y . Lim, “Errors in human-robot interaction accidents: A taxonomy and network analysis,” inInternational conference on construction engineering and project management. Korea Institute of Construction Engineering and Management, 2024, pp...

    2024

  2. [10]

    Reflect: Summarizing robot ex- periences for failure explanation and correction,

    Z. Liu, A. Bahety, and S. Song, “Reflect: Summarizing robot ex- periences for failure explanation and correction,”arXiv preprint arXiv:2306.15724, 2023

    2023 arXiv

  3. [11]

    Control barrier functions: Theory and applications,

    A. D. Ames, S. Coogan, M. Egerstedt, G. Notomista, K. Sreenath, and P. Tabuada, “Control barrier functions: Theory and applications,” in2019 18th European control conference (ECC). IEEE, 2019, pp. 3420–3431

    2019

  4. [12]

    Predictive control barrier functions: Enhanced safety mechanisms for learning-based control,

    K. P. Wabersich and M. N. Zeilinger, “Predictive control barrier functions: Enhanced safety mechanisms for learning-based control,” IEEE Transactions on Automatic Control, vol. 68, no. 5, pp. 2638– 2651, 2022

    2022

  5. [13]

    T-cbf: Traversability-based control barrier function to navigate vertically challenging terrain,

    M. Gupta and X. Xiao, “T-cbf: Traversability-based control barrier function to navigate vertically challenging terrain,” in2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2025

    2025

  6. [14]

    Hamilton–jacobi formulation for reach– avoid differential games,

    K. Margellos and J. Lygeros, “Hamilton–jacobi formulation for reach– avoid differential games,”IEEE Transactions on automatic control, vol. 56, no. 8, pp. 1849–1861, 2011

    2011

  7. [15]

    A time-dependent hamilton-jacobi formulation of reachable sets for continuous dynamic games,

    I. M. Mitchell, A. M. Bayen, and C. J. Tomlin, “A time-dependent hamilton-jacobi formulation of reachable sets for continuous dynamic games,”IEEE Transactions on automatic control, vol. 50, no. 7, pp. 947–957, 2005

    2005

  8. [16]

    Case study: Ros-based fault injection for risk analysis of robotic manipulator,

    Y . Ma, P. Grimmeisen, and A. Morozov, “Case study: Ros-based fault injection for risk analysis of robotic manipulator,” in2023 IEEE 19th International Conference on Automation Science and Engineering (CASE). IEEE, 2023, pp. 1–6

    2023

  9. [17]

    Fault detection in autonomous robots based on fault injection and learning,

    A. L. Christensen, R. O’Grady, M. Birattari, and M. Dorigo, “Fault detection in autonomous robots based on fault injection and learning,” Autonomous Robots, vol. 24, no. 1, pp. 49–67, 2008

    2008

  10. [18]

    Mujoco: A physics engine for model-based control,

    E. Todorov, T. Erez, and Y . Tassa, “Mujoco: A physics engine for model-based control,” in2012 IEEE/RSJ international conference on intelligent robots and systems. IEEE, 2012, pp. 5026–5033

    2012

  11. [19]

    Robocasa: Large-scale simulation of ev- eryday tasks for generalist robots,

    S. Nasiriany, A. Maddukuri, L. Zhang, A. Parikh, A. Lo, A. Joshi, A. Mandlekar, and Y . Zhu, “Robocasa: Large-scale simulation of ev- eryday tasks for generalist robots,”arXiv preprint arXiv:2406.02523, 2024

    2024 arXiv

  12. [20]

    robosuite: A modular simulation framework and benchmark for robot learning,

    Y . Zhu, J. Wong, A. Mandlekar, R. Mart´ın-Mart´ın, A. Joshi, S. Nasiri- any, and Y . Zhu, “robosuite: A modular simulation framework and benchmark for robot learning,”arXiv preprint arXiv:2009.12293, 2020

    2009 arXiv

  13. [21]

    The dynamic window approach to collision avoidance,

    D. Fox, W. Burgard, and S. Thrun, “The dynamic window approach to collision avoidance,”IEEE robotics & automation magazine, vol. 4, no. 1, pp. 23–33, 2002

    2002

  14. [22]

    Mink: Python inverse kinematics based on MuJoCo,

    K. Zakka, “Mink: Python inverse kinematics based on MuJoCo,” May 2025. [Online]. Available: https://github.com/kevinzakka/mink

    2025

  15. [23]

    Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,

    L. E. Kavraki, P. Svestka, J.-C. Latombe, and M. H. Overmars, “Prob- abilistic roadmaps for path planning in high-dimensional configuration spaces,”IEEE transactions on Robotics and Automation, vol. 12, no. 4, pp. 566–580, 2002

    2002

  16. [24]

    Rapidly-exploring random trees: Progress and prospects: Steven m. lavalle, iowa state university, a james j. kuffner, jr., university of tokyo, tokyo, japan,

    S. M. LaValle and J. J. Kuffner, “Rapidly-exploring random trees: Progress and prospects: Steven m. lavalle, iowa state university, a james j. kuffner, jr., university of tokyo, tokyo, japan,”Algorithmic and computational robotics, pp. 303–307, 2001

    2001

  17. [25]

    Adaptive rapidly-exploring-random- tree-star (rrt*)-smart: algorithm characteristics and behavior analysis in complex environments,

    J. Nasir, F. Islam, and Y . Ayaz, “Adaptive rapidly-exploring-random- tree-star (rrt*)-smart: algorithm characteristics and behavior analysis in complex environments,” 2013

    2013

  18. [26]

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

    J. J. Kuffner and S. M. LaValle, “Rrt-connect: An efficient approach to single-query path planning,” inProceedings 2000 ICRA. Millennium conference. IEEE international conference on robotics and automa- tion. Symposia proceedings (Cat. No. 00CH37065), vol. 2. IEEE, 2000, pp. 995–1001

    2000

  19. [27]

    Stomp: Stochastic trajectory optimization for motion planning,

    M. Kalakrishnan, S. Chitta, E. Theodorou, P. Pastor, and S. Schaal, “Stomp: Stochastic trajectory optimization for motion planning,” in 2011 IEEE international conference on robotics and automation. IEEE, 2011, pp. 4569–4574

    2011

  20. [28]

    Chomp: Covariant hamiltonian optimization for motion planning,

    M. Zucker, N. Ratliff, A. D. Dragan, M. Pivtoraiko, M. Klingensmith, C. M. Dellin, J. A. Bagnell, and S. S. Srinivasa, “Chomp: Covariant hamiltonian optimization for motion planning,”The International journal of robotics research, vol. 32, no. 9-10, pp. 1164–1193, 2013

    2013

  21. [29]

    Leveraging VLM- based pipelines to annotate 3d objects,

    R. Kabra, L. Matthey, A. Lerchner, and N. Mitra, “Leveraging VLM- based pipelines to annotate 3d objects,” inForty-first International Conference on Machine Learning, 2024. [Online]. Available: https://openreview.net/forum?id=5Pcl5qOOfL

    2024

  22. [30]

    Human motion trajectory prediction: A survey,

    A. Rudenko, L. Palmieri, M. Herman, K. M. Kitani, D. M. Gavrila, and K. O. Arras, “Human motion trajectory prediction: A survey,” The International Journal of Robotics Research, vol. 39, no. 8, pp. 895–935, 2020

    2020

  23. [31]

    Dino-wm: World models on pre-trained visual features enable zero-shot planning,

    G. Zhou, H. Pan, Y . LeCun, and L. Pinto, “Dino-wm: World models on pre-trained visual features enable zero-shot planning,”arXiv preprint arXiv:2411.04983, 2024

    2024 arXiv

  24. [32]

    Generalizing safety be- yond collision-avoidance via latent-space reachability analysis,

    K. Nakamura, L. Peters, and A. Bajcsy, “Generalizing safety be- yond collision-avoidance via latent-space reachability analysis,”arXiv preprint arXiv:2502.00935, 2025

    2025

  25. [33]

    Antic- ipatory task and motion planning,

    R. Dhakal, D. M. Nguyen, T. Silver, X. Xiao, and G. J. Stein, “Antic- ipatory task and motion planning,”arXiv preprint arXiv:2407.13694, 2024

    2024 arXiv

  26. [34]

    Integrated task and motion planning,

    C. R. Garrett, R. Chitnis, R. Holladay, B. Kim, T. Silver, L. P. Kael- bling, and T. Lozano-P ´erez, “Integrated task and motion planning,” Annual review of control, robotics, and autonomous systems, vol. 4, no. 1, pp. 265–293, 2021

    2021

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.