Pith. sign in

REVIEW 4 major objections 5 minor 43 references

DreamFLEX: Learning Fault-Aware Quadrupedal Locomotion Controller for Anomaly Situation in Rough Terrains

T0 review · 4 major / 5 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 DreamFLEX is a learned controller that explicitly estimates which quadruped leg joints are locked or weakened, and uses that estimate to modulate the policy's internal state so the robot keeps walking on rough terrain where existing…

desk verdict Solid sim-to-real fault-tolerant locomotion paper with a clear novelty, but the weakened-motor claim is trained, not actually evaluated. read the letter →

arxiv 2502.05817 v1 pith:FYFOSNNV submitted 2025-02-09 cs.RO cs.SYeess.SY

Seunghyun Lee , I Made Aswin Nahrendra , Dongkyu Lee , Byeongho Yu , Minho Oh , Hyun Myung This is my paper · ORCID
classification cs.ROcs.SYeess.SY
keywords quadrupedalrobotsfault-tolerantlocomotiondeepreinforcementlearningjointfailureestimationroughterraintraversalproprioceptivestatelatentconditioninglockedandweakenedmotor
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

DreamFLEX is a deep-reinforcement-learning controller that keeps a quadrupedal robot walking over rough terrain after a leg joint has failed. The paper's central claim is that explicitly estimating which joints are faulty, as a binary vector, and conditioning the policy's latent representation on that vector makes the controller more reliable than prior fault-tolerant locomotion approaches, both in simulation and on hardware. In the reported head-to-head simulation, the mean absolute velocity-tracking error under locked and weakened joint conditions is 0.1786 m/s, compared with 0.2848 m/s and 0.3710 m/s for the two comparison methods, and ablations that remove the fault vector or the modulation layer do worse. A sympathetic reading is that a robot can name its own hardware failure and change its gait accordingly, rather than stopping or falling when a joint locks mid-mission.

What carries the argument

The load-bearing component is the failure estimation and modulation network (FEMNet): an encoder-decoder that takes the five most recent proprioceptive observations and outputs the body linear velocity, a binary joint fault vector for all 12 joints, and a latent vector, the compressed representation of terrain and robot state. A decoder reconstructs the next observation so the latent space stays temporally consistent, with binary cross-entropy supervising fault detection and mean-squared-error supervising velocity. A modulation layer then maps the fault vector to affine parameters $\gamma_1$ and $\gamma_2$ and produces the conditioned latent $\tilde{z} = \gamma_1 \cdot z + \gamma_2$, which is what the policy network sees; this is the mechanism by which the controller changes its gait pattern depending on which joint has failed. The fault scenarios themselves are injected during training by clipping the commanded angle around a randomly sampled center, the locked joint case, or scaling the torque by a uniformly sampled factor, the weakened motor case, with a curriculum that widens the sampled range once task reward passes a threshold.

What would settle it

Inject a fault that matches neither model, such as a constant friction torque in a joint or intermittent motor dropout, then check whether the estimated fault vector stays accurate and velocity-tracking error stays below the no-modulation ablation; if estimation error rises sharply or the modulation fails to beat the ablation, the explicit-fault mechanism is not doing the claimed work.

Watch

Extended reading notes

Core claim

The discovery is that a fault-tolerant locomotion policy improves when the fault is represented explicitly and used to reshape the latent code that the policy consumes, instead of being folded implicitly into the observation stream. During training the agent is randomly assigned a normal joint or a faulty joint, locked (commanded angle clipped to a small symmetrical range) or weakened (torque scaled by $k_{\tau}$), and a failure curriculum gradually widens the severity. At deployment, the failure estimation and modulation network (FEMNet) reads the last five proprioceptive observations, estimates body velocity, a 12-dimensional binary joint fault vector, and a latent vector, then applies an affine modulation $\tilde{z} = \gamma_1 \cdot z + \gamma_2$ driven by the fault vector before the policy acts. The reported simulation benchmark places DreamFLEX's mean absolute velocity-tracking error at 0.1786 m/s under locked and weakened joints, below the 0.2848 and 0.3710 m/s of the compared approaches and below the two DreamFLEX ablations (0.2426 m/s without modulation, 0.2812 without the fault vector). Real-world trials on a quadruped platform confirm that the same controller climbs 10-15 cm steps and stairs and crosses gravel, roots, and slopes while a joint is locked or overheated, including an unplanned overheating failure.

Load-bearing premise

The argument stands on the assumption that the two simulated failure modes, a locked joint modeled as angle clipping and a weakened motor modeled as torque scaling, match the failures a quadruped actually suffers, so that the estimated binary fault vector is trustworthy in the field.

Editorial extensions

If this is right

  • A mission does not have to stop when one actuator fails: the controller detects the fault from proprioception alone and rebalances onto the healthy legs, including on slopes, stairs, and uneven ground.
  • Because the fault vector is explicit, a single trained policy covers many fault locations and severities rather than needing a separate policy per broken joint; the paper shows performance for locked or weakened hip, thigh, and calf joints in front and rear legs.
  • The policy transfers to hardware without fine-tuning, and it handled a soft, deformable terrain that was absent from the training distribution, indicating out-of-distribution terrain generalization.
  • Explicit fault estimation plus latent modulation improves command tracking more than either removing the fault vector or feeding it directly to the policy, so the conditioning step is doing measurable work.
  • A real, unplanned overheating failure during summer testing was detected and handled, indicating the estimator is not only responding to the scripted fault injection.

Reading between the lines

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

  • Beyond the paper, the same explicit-fault-plus-conditioning recipe could apply to other degradation modes, such as sensor dropout, payload shift, or limb shortening, since the estimator only needs a labeled indicator and the policy only needs a conditioning signal.
  • Beyond the paper, the binary single-fault-per-agent training leaves open the case of two simultaneous faults; extending the fault vector to multi-label or severity-valued flags would be a direct stress test of the modulation mechanism.
  • Beyond the paper, isolating the fault-estimation benefit from the base terrain-imagination benefit would be a useful experiment: run the controller on precise foot-placement tasks where a wrong fault flag should appear as a visible misstep rather than as a small velocity error.
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 DreamFLEX, a reinforcement-learning-based fault-tolerant quadruped locomotion controller built on the authors' DreamWaQ. DreamFLEX adds a failure estimation and modulation network (FEMNet) that, from a history of proprioceptive observations, jointly estimates body linear velocity, a binary joint fault vector, and a terrain latent vector; the fault vector then modulates the latent via an affine (FiLM-style) transformation before the policy consumes it. Training randomizes locked-joint (Eq. 3) and weakened-motor (Eq. 4) conditions with a failure curriculum (Eq. 6). In Isaac Gym, DreamFLEX is compared with Kim et al., FT-Net, and two ablations, reporting lower mean absolute tracking error (Table III); real-world video demonstrations on a Unitree Go1 in rough outdoor and indoor terrains are also presented.

Significance. If the quantitative claims survive scrutiny, the explicit fault-vector estimation plus modulation is a useful and interpretable design: it gives the controller an explicit indicator of where a failure has occurred and a learned mechanism for adapting the gait, and it is a natural extension of DreamWaQ. The failure curriculum is a practical contribution that could be reused. I found no circularity problem: the fault vector is supervised by ground-truth labels and the modulation parameters are learned by RL rather than fit to the evaluation metric. However, the current evaluation does not support all of the paper's claims: Table III has no variance information, the 'weakened motor' condition is never tested with partial torque loss, and the real-world evidence is qualitative. These are correctable with additional experiments, so the underlying idea merits revision rather than rejection.

major comments (4)
  1. [Table III; Fig. 5] Table III reports only point estimates of absolute tracking error, with no standard deviations, no number of seeds, and no significance tests, so the claim that DreamFLEX 'consistently outperforms' Kim et al. and FT-Net is not statistically supported. Because RL training and evaluation are stochastic, please run multiple seeds (e.g., at least five) and report mean ± std along with an appropriate significance test or effect size; the learning curves in Fig. 5 likewise need error bands before they can support the comparative claim.
  2. [Section III.B, Eq. (4), Table III] The 'Weakened Motor' entries in Table III are obtained with kτ = 0.0, which is complete torque loss, not a weakened motor in the sense of Section II.B.2 (kτ sampled from [kL, 0.25]). The real-world RR calf failure in Fig. 8 is described as a weakened motor but is also run with the fixed parameters of Section III.B, so the partial-torque regime is never demonstrated. Since f_t in Eq. (5) is binary and the modulation in Eq. (9) conditions only on f_t, there is also no evidence about how torque-loss severity would be represented; please evaluate a grid of kτ values (e.g., 0.05, 0.10, 0.20, 0.25) and report tracking error and fault-estimation accuracy for each. The 'Faulty leg joint motion' reward in Table I may additionally teach the policy to discard a partially functional leg, so this sweep is needed to verify the design in the partial-fault regime.
  3. [Section III.B, Eq. (6)] The locked-joint evaluation fixes qcen = qdef, even though training samples qcen from a range that expands through the failure curriculum (Eq. 6). A single locked angle does not test the controller's ability to cope with joints locked at different positions, which is central to the locked-joint failure mode. Please report results for several qcen values, including positions away from the default stance.
  4. [Section III.C] The real-world experiments are presented as qualitative video stills and narrative descriptions only; no quantitative metrics are reported for the physical robot, such as number of trials, success rate, velocity tracking error, terrain-completion time, or fault-vector estimation accuracy, and no baselines are run on hardware. This supports a feasibility demonstration but not the abstract's claim that DreamFLEX outperforms existing methods in real-world scenarios; please add basic logged metrics or explicitly restrict the real-world claim to feasibility.
minor comments (5)
  1. [Section II.A.2, Table I] The reward notation P(t^n_air,foot − t^n,des_air,foot) is not defined; specify whether P is a piecewise-linear operator, a projection, or a placeholder, and define v_xy,foot and the superscripts in the table.
  2. [Section II.C.1] In Eq. (7), the term 'LV AE' should be typeset as L_VAE; also define the VAE loss components explicitly rather than only by name.
  3. [Section II.C.2, Eq. (9)] Please specify the shapes of γ1 and γ2 and whether the affine transformation is applied elementwise to z_t; this matters for understanding how the modulation interacts with the latent dimension.
  4. [Section III.A] The ablation named 'DreamFLEX w/o joint fault vector' is said to be the same as DreamWaQ [10], but the name suggests only f_t is removed while the rest of FEMNet remains; clarify whether this ablation also removes the modulation layer and the fault-tolerant rewards.
  5. [References] Reference [24] lists the author as 'K. Kwamoto'; if the correct spelling is 'Kawamoto', please correct it.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: DreamFLEX's fault estimation and modulation are learned from simulation labels and evaluated against external baselines, with no derivation reducing to its own inputs.

full rationale

DreamFLEX's core claims are empirical and self-contained. The joint fault vector f_t is estimated by a supervised network trained with binary cross-entropy against ground-truth fault labels from the simulator (Eq. 8), not derived from the evaluation metric. The modulation parameters γ1 and γ2 (Eq. 9) are learned through reinforcement learning, and the policy is optimized with PPO against task and style rewards; none of these components is fitted to the reported ATE values. The comparisons in Table III are performed against external baselines (Kim et al., FT-Net) trained in the same environment, so the reported improvement is not forced by construction. The paper does rely on the authors' prior DreamWaQ and CENet work for the base architecture, but those are independently published and are used as building blocks rather than as proof of the target result. The self-citation to B-TMS is used only for terrain-map visualization in a real-world demonstration and is not load-bearing. The weakened-motor evaluation fixes kτ = 0.0 while training samples kτ in [kL, 0.25]; this is an experimental coverage gap or validity concern about generalizing to partial torque loss, not a circularity, because the performance numbers are still produced by an independent roll-out of a learned policy. No equation in the paper is equivalent to its own input by definition, and no predicted quantity is a renamed fit. Hence no circular step can be exhibited.

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

The paper introduces no new physical entities. The main assumptions are engineering: the fault models, the reward design, and the sim-to-real transfer premise. Free parameters are reward weights and curriculum constants that are hand-tuned and not given sensitivity analysis.

free parameters (5)
  • Fault-tolerant reward weights = 1.5, -0.5, -1e-5, -0.2, -0.1
    Hand-tuned weights in Table I to encode desired gait behavior; no sensitivity analysis is reported.
  • Failure curriculum step sizes = delta_q = 0.05 rad, delta_k = 0.0125
    Chosen to stabilize training; they control how quickly fault severity increases and affect the final policy robustness.
  • Fault sampling bounds = qcen in [qL, qU] via curriculum; k_tau in [kL, 0.25]
    Initial box bounds are not given; the curriculum updates qL, qU, and kL, influencing the range of failures experienced in training.
  • Fixed evaluation fault parameters = qcen = qdef, k_tau = 0.0
    The simulation and real-world evaluations use a fully locked joint or fully weakened motor; intermediate severities are not evaluated.
  • Reward reference values = not specified
    Desired feet air time, foot clearance height, and foot contact force thresholds in Table I are not given numerical values.
assumptions (5)
  • standard math PPO converges to a policy that generalizes across the fault and terrain distribution
    Standard reinforcement learning assumption used in the training procedure described in Section II.A.3.
  • domain assumption Locked joint is equivalent to clipping the desired joint angle around a center (Eq. 3)
    The paper simulates a hardware lock by constraining the position command; real locked joints may have different dynamics.
  • domain assumption Weakened motor is equivalent to scaling torque by factor k_tau (Eq. 4)
    Overheating or power loss is modeled as a constant torque efficiency factor, which may not capture time-varying degradation.
  • domain assumption The asymmetric actor-critic setup with privileged information is sufficient for sim-to-real transfer
    The critic uses privileged state during training; the paper assumes the learned policy and estimator transfer to the real robot without fine-tuning (Section III.C).
  • domain assumption Fractal noise terrain and the terrain curriculum are sufficient for rough terrain generalization
    The training distribution is assumed to cover real-world rough, slippery, stepped, and soft terrains encountered in experiments.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DreamFLEX: Learning Fault-Aware Quadrupedal Locomotion Controller for Anomaly Situation in Rough Terrains." pith.science (2026). https://pith.science/paper/FYFOSNNV

@misc{pith2026250205817,
  author       = {Pith},
  title        = {Pith review of: DreamFLEX: Learning Fault-Aware Quadrupedal Locomotion Controller for Anomaly Situation in Rough Terrains},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FYFOSNNV}},
  note         = {Machine review of arXiv:2502.05817}
}
read the original abstract

Recent advances in quadrupedal robots have demonstrated impressive agility and the ability to traverse diverse terrains. However, hardware issues, such as motor overheating or joint locking, may occur during long-distance walking or traversing through rough terrains leading to locomotion failures. Although several studies have proposed fault-tolerant control methods for quadrupedal robots, there are still challenges in traversing unstructured terrains. In this paper, we propose DreamFLEX, a robust fault-tolerant locomotion controller that enables a quadrupedal robot to traverse complex environments even under joint failure conditions. DreamFLEX integrates an explicit failure estimation and modulation network that jointly estimates the robot's joint fault vector and utilizes this information to adapt the locomotion pattern to faulty conditions in real-time, enabling quadrupedal robots to maintain stability and performance in rough terrains. Experimental results demonstrate that DreamFLEX outperforms existing methods in both simulation and real-world scenarios, effectively managing hardware failures while maintaining robust locomotion performance.

Figures

Figures reproduced from arXiv: 2502.05817 by the authors.

Figure 1
Figure 1. DreamFLEX enables the robot to autonomously detect leg joint [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our proposed training framework, DreamFLEX. The [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Examples of the fault scenarios in the front left calf joint of the [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The architecture of FEMNet. (a) FEMNet jointly estimates the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Learning curves of the (a) linear velocity tracking, (b) terrain level, [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Quantitative results of the transition from normal to fault conditions [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 8
Figure 8. Figure 8: Our controller successfully overcame the indoor complex environ [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 37 canonical work pages

  1. [1]

    ANYmal - A highly mobile and dynamic quadrupedal robot,

    M. Hutter, C. Gehring, D. Jud, A. Lauber, C. D. Bellicoso, V . Tsounis, J. Hwangbo, K. Bodie, P. Fankhauser, M. Bloesch, R. Diethelm, S. Bachmann, A. Melzer, and M. Hoepflinger, “ANYmal - A highly mobile and dynamic quadrupedal robot,” in Proc. IEEE/RSJ Int. Conf. Intell. Robot. Syst. , 2016, pp. 38–44

    work page 2016

  2. [2]

    Mini cheetah: A platform for pushing the limits of dynamic quadruped control,

    B. Katz, J. Di Carlo, and S. Kim, “Mini cheetah: A platform for pushing the limits of dynamic quadruped control,” in Proc. IEEE Int. Conf. Robot. Automat. , 2019, pp. 6295–6301

    work page 2019

  3. [3]

    Design of KAIST HOUND, a quadruped robot platform for fast and efficient locomotion with mixedinteger nonlinear optimization of a gear train,

    Y .-H. Shin, S. Hong, S. Woo, J. Choe, H. Son, G. Kim, J.-H. Kim, K. Lee, J. Hwangbo, and H.-W. Park, “Design of KAIST HOUND, a quadruped robot platform for fast and efficient locomotion with mixedinteger nonlinear optimization of a gear train,” in Proc. IEEE Int. Conf. Robot. Automat. , 2022, pp. 6614–6620

    work page 2022

  4. [4]

    ANYmal in the field: Solving industrial inspection of an offshore HVDC platform with a quadrupedal robot,

    C. Gehring, P. Fankhauser, L. Isler, R. Diethelm, S. Bachmann, M. Potz, L. Gerstenberg, and M. Hutter, “ANYmal in the field: Solving industrial inspection of an offshore HVDC platform with a quadrupedal robot,” in Field and Serv. Robot. , 2021, pp. 247–260

    work page 2021

  5. [5]

    CERBERUS in the DARPA Subterranean Challenge,

    M. Tranzatto, T. Miki, M. Dharmadhikari, L. Bernreiter, M. Kulkarni, F. Mascarich, O. Andersson, S. Khattak, M. Hutter, R. Siegwart, and K. Alexis, “CERBERUS in the DARPA Subterranean Challenge,” Sci. Robot., vol. 7, no. 666, p. eabp9742, 2022

    work page 2022

  6. [6]

    Taking the first step toward autonomous quadruped robots: The Quadruped Robot Challenge at ICRA 2023 in London,

    A. Jacoff, J. Jeon, O. Huke, D. Kanoulas, S. Ha, D. Kim, and H. Moon, “Taking the first step toward autonomous quadruped robots: The Quadruped Robot Challenge at ICRA 2023 in London,” IEEE Robot. Automat. Mag., vol. 30, no. 3, pp. 154–158, 2023

    work page 2023

  7. [7]

    RMA: Rapid motor adaptation for legged robots,

    A. Kumar, Z. Fu, D. Pathak, and J. Malik, “RMA: Rapid motor adaptation for legged robots,” in Robot. Sci. Syst. , 2021

    work page 2021

  8. [8]

    Learning robust perceptive locomotion for quadrupedal robots in the wild,

    T. Miki, J. Lee, J. Hwangbo, L. Wellhausen, V . Koltun, and M. Hutter, “Learning robust perceptive locomotion for quadrupedal robots in the wild,” Sci. Robot. , vol. 7, no. 62, p. eabk2822, 2022

    work page 2022

Show all 43 references
  1. [9]

    Fuzzy logic and neural network- based intelligent control system for quadruped robot on extreme terrain,

    J. Jeon, M. Jeong, and H. Myung, “Fuzzy logic and neural network- based intelligent control system for quadruped robot on extreme terrain,” in Proc. Int. Conf. Control, Automat. Syst. , 2023, pp. 1885– 1889

    2023

  2. [10]

    DreamWaQ: Learning robust quadrupedal locomotion with implicit terrain imagination via deep reinforcement learning,

    I. M. A. Nahrendra, B. Yu, and H. Myung, “DreamWaQ: Learning robust quadrupedal locomotion with implicit terrain imagination via deep reinforcement learning,” in Proc. IEEE Int. Conf. Robot. Au- tomat., 2023, pp. 5078–5084

    2023

  3. [11]

    Legged locomotion in challenging terrains using egocentric vision,

    A. Agarwal, A. Kumar, J. Malik, and D. Pathak, “Legged locomotion in challenging terrains using egocentric vision,” in Proc. Conf. Robot Learning, 2023, pp. 403–415

    2023

  4. [12]

    Learning quadrupedal locomotion on deformable terrain,

    S. Choi, G. Ji, J. Park, H. Kim, J. Mun, J. H. Lee, and J. Hwangbo, “Learning quadrupedal locomotion on deformable terrain,” Sci. Robot., vol. 8, no. 74, p. eade2256, 2023

    2023

  5. [13]

    MorAL: Learning morphologically adaptive locomotion controller for quadrupedal robots on challenging terrains,

    Z. Luo, Y . Dong, X. Li, R. Huang, Z. Shu, E. Xiao, and P. Lu, “MorAL: Learning morphologically adaptive locomotion controller for quadrupedal robots on challenging terrains,” IEEE Robot. Automat. Lett., vol. 9, no. 5, pp. 4019–4026, 2024

    2024

  6. [14]

    Dtc: Deep tracking control,

    F. Jenelten, J. He, F. Farshidian, and M. Hutter, “Dtc: Deep tracking control,” Sci. Robot. , vol. 9, no. 86, p. eadh5401, 2024

    2024

  7. [15]

    Robust recovery controller for a quadrupedal robot using deep reinforcement learning,

    J. Lee, J. Hwangbo, and M. Hutter, “Robust recovery controller for a quadrupedal robot using deep reinforcement learning,” arXiv preprint arXiv:1901.07517, 2019

    1901 arXiv

  8. [16]

    Multi-expert learning of adaptive legged locomotion,

    C. Yang, K. Yuan, Q. Zhu, W. Yu, and Z. Li, “Multi-expert learning of adaptive legged locomotion,” Sci. Robot., vol. 5, no. 49, p. eabb2174, 2020

    2020

  9. [17]

    Robust recovery motion control for quadrupedal robots via learned terrain imagination,

    I. M. A. Nahrendra, M. Oh, B. Yu, H. Lim, and H. Myung, “Robust recovery motion control for quadrupedal robots via learned terrain imagination,” in Robot. Sci. Syst. Workshop , 2023

    2023

  10. [18]

    Barkour: Bench- marking animal-level agility with quadruped robots,

    K. Caluwaerts, A. Iscen, J. C. Kew, W. Yu, T. Zhang, D. Freeman, K.-H. Lee, L. Lee, S. Saliceti, V . Zhuang, et al. , “Barkour: Bench- marking animal-level agility with quadruped robots,” arXiv preprint arXiv:2305.14654, 2023

    2023 arXiv

  11. [19]

    Robot parkour learning,

    Z. Zhuang, Z. Fu, J. Wang, C. Atkeson, S. Schwertfeger, C. Finn, and H. Zhao, “Robot parkour learning,” in Proc. Conf. Robot Learning , 2023, pp. 73–92

    2023

  12. [20]

    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,” Sci. Robot., vol. 9, no. 88, p. eadi7566, 2024

    2024

  13. [21]

    Extreme parkour with legged robots,

    X. Cheng, K. Shi, A. Agarwal, and D. Pathak, “Extreme parkour with legged robots,” in Proc. IEEE Int. Conf. Robot. Automat. , 2024, pp. 11 443–11 450

    2024

  14. [22]

    Fast online adaptation in robotics through meta-learning embeddings of simulated priors,

    R. Kaushik, T. Anne, and J.-B. Mouret, “Fast online adaptation in robotics through meta-learning embeddings of simulated priors,” in Proc. IEEE/RSJ Int. Conf. Intell. Robot. Syst. , 2020, pp. 5269–5276

    2020

  15. [23]

    Meta-learning for fast adaptive locomotion with uncertainties in environments and robot dynamics,

    T. Anne, J. Wilkinson, and Z. Li, “Meta-learning for fast adaptive locomotion with uncertainties in environments and robot dynamics,” in Proc. IEEE/RSJ Int. Conf. Intell. Robot. Syst. , 2021, pp. 4568–4575

    2021

  16. [24]

    Reinforcement learning with adaptive curriculum dynamics randomization for fault-tolerant robot control,

    W. Okamoto, H. Kera, and K. Kwamoto, “Reinforcement learning with adaptive curriculum dynamics randomization for fault-tolerant robot control,” arXiv preprint arXiv:2111.10005 , 2021

    2021 arXiv

  17. [25]

    Saving the limping: Fault- tolerant quadruped locomotion via reinforcement learning,

    D. Liu, T. Zhang, J. Yin, and S. See, “Saving the limping: Fault- tolerant quadruped locomotion via reinforcement learning,” arXiv preprint arXiv:2210.00474, 2021

    2021 arXiv

  18. [26]

    Learning quadrupedal locomotion with impaired joints using random joint masking,

    M. Kim, U. Shin, and J.-Y . Kim, “Learning quadrupedal locomotion with impaired joints using random joint masking,” in Proc. IEEE Int. Conf. Robot. Automat. , 2024, pp. 9751–9757

    2024

  19. [27]

    FT-Net: Learning failure recovery and fault-tolerant locomotion for quadruped robots,

    Z. Luo, E. Xiao, and P. Lu, “FT-Net: Learning failure recovery and fault-tolerant locomotion for quadruped robots,” IEEE Robot. Automat. Lett., vol. 8, no. 12, pp. 8414–8421, 2023

    2023

  20. [28]

    Unitree Go1

    “Unitree Go1.” [Online]. Available: https://www.unitree.com/go1

  21. [29]

    Asymmetric actor critic for image-based robot learning,

    L. Pinto, M. Andrychowicz, P. Welinder, W. Zaremba, and P. Abbeel, “Asymmetric actor critic for image-based robot learning,” in Robot. Sci. Syst. , 2018

    2018

  22. [30]

    Proximal policy optimization algorithms,

    J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal policy optimization algorithms,” arXiv preprint arXiv:1707.06347, 2017

    2017 arXiv

  23. [31]

    Learning to walk in minutes using massively parallel deep reinforcement learning,

    N. Rudin, D. Hoeller, P. Reist, and M. Hutter, “Learning to walk in minutes using massively parallel deep reinforcement learning,” in Proc. Conf. Robot Learning , 2022, pp. 91–100

    2022

  24. [32]

    Deep whole-body control: learning a unified policy for manipulation and locomotion,

    Z. Fu, X. Cheng, and D. Pathak, “Deep whole-body control: learning a unified policy for manipulation and locomotion,” in Proc. Conf. Robot Learning, 2022, pp. 138–149

    2022

  25. [33]

    Rapid locomotion via reinforcement learning,

    G. B. Margolis, G. Yang, K. Paigwar, T. Chen, and P. Agrawal, “Rapid locomotion via reinforcement learning,” in Robot. Sci. Syst. , 2022

    2022

  26. [34]

    Concurrent training of a control policy and a state estimator for dynamic and robust legged locomotion,

    G. Ji, J. Mun, H. Kim, and J. Hwangbo, “Concurrent training of a control policy and a state estimator for dynamic and robust legged locomotion,” IEEE Robot. Automat. Lett. , vol. 7, no. 2, pp. 4630– 4637, 2022

    2022

  27. [35]

    β-V AE: Learning basic visual concepts with a constrained variational framework,

    I. Higgins, L. Matthey, A. Pal, C. P. Burgess, X. Glorot, M. M. Botvinick, S. Mohamed, and A. Lerchner, “ β-V AE: Learning basic visual concepts with a constrained variational framework,” in Proc. Int. Conf. Learning Repre. , vol. 3, 2017

    2017

  28. [36]

    On information and sufficiency,

    S. Kullback and R. A. Leibler, “On information and sufficiency,” The Annals of Mathematical Statistics , vol. 22, no. 1, pp. 79–86, 1951

    1951

  29. [37]

    FiLM: Visual reasoning with a general conditioning layer,

    E. Perez, F. Strub, H. de Vries, V . Dumoulin, and A. Courville, “FiLM: Visual reasoning with a general conditioning layer,” inProc. Conf. Adv. Artif. Intell. , vol. 32, no. 1, 2018

    2018

  30. [38]

    User-conditioned neural control policies for mobile robotics,

    L. Bauersfeld, E. Kaufmann, and D. Scaramuzza, “User-conditioned neural control policies for mobile robotics,” in Proc. IEEE Int. Conf. Robot. Automat. , 2023, pp. 1342–1348

    2023

  31. [39]

    Isaac Gym: High performance GPU-based physics simulation for robot learning,

    V . Makoviychuk, L. Wawrzyniak, Y . Guo, M. Lu, K. Storey, M. Mack- lin, D. Hoeller, N. Rudin, A. Allshire, A. Handa, and G. State, “Isaac Gym: High performance GPU-based physics simulation for robot learning,” Adv. Neural Inf. Process. Syst. Data. Bench. , 2021

    2021

  32. [40]

    Adam: A method for stochastic gradient descent,

    D. P. Kingma and J. L. Ba, “Adam: A method for stochastic gradient descent,” in Proc. Int. Conf. Learning Repre. , 2015, pp. 1–15

    2015

  33. [41]

    Fast and accurate deep network learning by exponential linear units (ELUs),

    D.-A. Clevert, T. Unterthiner, and S. Hochreiter, “Fast and accurate deep network learning by exponential linear units (ELUs),” in Proc. Int. Conf. Learning Repre. , 2016

    2016

  34. [42]

    Design and use paradigms for Gazebo, an open-source multi-robot simulator,

    N. Koenig and A. Howard, “Design and use paradigms for Gazebo, an open-source multi-robot simulator,” in Proc. IEEE/RSJ Int. Conf. Intell. Robot. Syst. , 2004, pp. 2149–2154

    2004

  35. [43]

    B-TMS: Bayesian traversable terrain modeling and segmentation across 3D LiDAR scans and maps for enhanced off- road navigation,

    M. Oh, G. Shin, S. Jang, S. Lee, D. Lee, W. Song, B. Yu, H. Lim, J. Lee, and H. Myung, “B-TMS: Bayesian traversable terrain modeling and segmentation across 3D LiDAR scans and maps for enhanced off- road navigation,” in Proc. IEEE Intell. V eh. Symp., 2024, pp. 200–205

    2024

Pith tools

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