Booster Gym: An End-to-End Reinforcement Learning Framework for Humanoid Robot Locomotion

Feng Wu; Mingguo Zhao; Penghui Chen; Xinyu Han; Yushi Wang

REVIEW 4 major objections 6 minor 2 cited by

Booster Gym: An End-to-End Reinforcement Learning Framework for Humanoid Robot Locomotion

T0 review · 4 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 Booster Gym is an open-source, end-to-end reinforcement learning framework that claims to take humanoid locomotion from simulation training to zero-shot deployment on real hardware, demonstrated on the Booster T1 robot.

desk verdict A useful open-source humanoid RL training and deployment pipeline with real hardware validation, but the 'seamless zero-shot transfer' claim rests on qualitative evidence and one under-validated kinematic conversion module. read the letter →

arxiv 2506.15132 v1 pith:EKYWGG5W submitted 2025-06-18 cs.RO

Yushi Wang , Penghui Chen , Xinyu Han , Feng Wu , Mingguo Zhao This is my paper · ORCID

classification cs.RO

keywords humanoidrobotlocomotionsim-to-realtransferdomainrandomizationreinforcementlearningasymmetricactor-criticparallelmechanismzero-shotIsaacGym

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

Booster Gym is presented as a single open-source code framework that covers the whole humanoid-locomotion pipeline: setting up a simulation environment, training a policy with reinforcement learning, and deploying it on a physical robot without extra tuning. The authors validate it on the Booster T1 humanoid, where policies trained in Isaac Gym transfer zero-shot and produce omnidirectional walking, walking over grass, stone, soil, asphalt, concrete and tile, ascent of a 10-degree slope, and recovery after being struck by a falling 10 kg weight. The central claim is that the combination of domain randomization, a periodic gait reward, careful episode design, and a series-parallel ankle conversion module makes the sim-to-real gap small enough for direct hardware deployment. If true, this gives the robotics community a reusable starting point that removes most of the implementation burden usually associated with sim-to-real humanoid locomotion.

What carries the argument

The load-bearing mechanism is the combined training-and-deployment pipeline itself, held together by three parts: an asymmetric actor-critic PPO loop running in a GPU-parallel simulator; a domain-randomization envelope over robot dynamics, actuators, latency, terrain and disturbances; and a series-parallel ankle conversion module in the deployment SDK. The conversion module computes virtual serial-joint position and velocity feedback from the parallel ankle kinematics, then converts the policy's serial-joint targets into parallel-structure commands using a transposed Jacobian and the motor PD controller. This is what lets the policy ignore the closed kinematic chain during training while still controlling it on hardware. The gait-cycle observation and periodic gait reward carry the rhythmic stepping behavior, and the latency randomization matches the measured 9–12 ms command-to-feedback delay.

What would settle it

Run the released pre-trained policy on a Booster T1 with the series-parallel ankle conversion disabled; if the robot cannot stand or walk, the claim that the conversion module is what enables zero-shot transfer on the parallel ankle is contradicted. Also, record ten 0.5 m/s forward-walking trials in Isaac Gym and on hardware, and check whether the real joint position-velocity trajectories stay inside the envelope produced by the randomized training environments; if they leave that envelope systematically, the domain-randomization coverage claim fails.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that an end-to-end reinforcement learning pipeline, without teacher-student distillation or real-world fine-tuning, can be assembled into a software artifact that transfers zero-shot to a physical humanoid. The policy is trained with an asymmetric actor-critic and PPO, outputs joint position offsets at 50 Hz, and relies on a motor-level PD controller to produce torques. The actor observes noisy proprioception, the previous action, a velocity command, and a sinusoidal gait cycle, while the critic receives privileged simulation state. Domain randomization covers link mass and center of mass, joint stiffness, damping, friction, communication latency, contact properties, terrain, and external pushes, and the paper quantifies real-world latency at 9–12 ms round-trip with sub-1 ms inference to justify a 0–20 ms randomization range. A series-parallel ankle conversion module lets the policy be trained as if the ankle were a serial mechanism and then deploys on the physical parallel mechanism through a transposed Jacobian and PD control, which is the design that makes closed kinematic chains compatible with the GPU simulator.

Load-bearing premise

The load-bearing premise is that Isaac Gym's simulation, once augmented by the series-parallel ankle conversion and the chosen domain-randomization ranges, is faithful enough that a policy trained on a virtual serial ankle will control the physical parallel ankle correctly without any real-world adaptation.

Editorial extensions

If this is right

A new user can start from the released policy and repository rather than reimplementing environment, reward, training and deployment code, which is the practical claim behind calling Booster Gym a community resource.
The measured latency values (9–12 ms round-trip, sub-1 ms inference) give other developers concrete numbers to reproduce or tune when working on similar humanoid hardware.
Robots with parallel ankle mechanisms can train in standard GPU simulators that do not natively support closed kinematic chains, because the conversion module handles the parallel structure at deployment time.
Training on randomized terrain and disturbance conditions appears sufficient for outdoor surfaces and impact recovery without explicit terrain perception or privileged exteroception in the actor.
Cross-simulation evaluation in MuJoCo and Webots can serve as a fast validation layer that approximates real-world dynamics before hardware experiments.

Reading between the lines

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

A testable extension is to apply the same series-parallel conversion idea to other closed-chain joints, such as parallel knees or wrists; the paper only demonstrates the ankle, so this is an extrapolation.
The success of a pure randomization envelope hints that explicit adaptation mechanisms, such as a student network estimating latent dynamics, may be unnecessary for this robot class, but the paper does not run that comparison.
The community value of the released pretrained policy will depend on how easily the code retrains for a different humanoid; that portability is not tested and would be a useful follow-up benchmark.
Ablating the series-parallel conversion or the latency randomization would show which component actually closes the remaining sim-to-real gap; the paper reports no such ablation.

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.

Desk Editor's Note

read the letter

Here's the short version: this is a solid engineering contribution—an open-source end-to-end RL pipeline for humanoid locomotion, validated on the Booster T1 with real hardware results. If you work on sim-to-real for legged robots, it's worth a look. But the headline claim of 'seamless' zero-shot transfer is stronger than the measurements support.

What's actually new: the series-parallel ankle conversion module that lets you train on a virtual serial ankle and deploy on a parallel ankle via a transposed Jacobian and PD mapping, plus the Booster T1 validation. The rest is assembled from known components—PPO, asymmetric actor-critic, domain randomization, gait cycle rewards—and the authors openly acknowledge the open-source repos they drew from. That's fine; the contribution is integration and packaging. The paper does several things well: it measures communication latency (9-12 ms) and randomizes 0-20 ms in training, it tests in two additional simulators (MuJoCo, Webots), and it shows real-world robustness to pushes and terrain variation. Figure 7 shows rough overlap between Isaac Gym, MuJoCo, and real joint trajectories.

The main weakness is confirmation: no quantitative metrics like tracking error, success rates, or failure counts. The plots are illustrative, not statistical. The stress-test note focuses on the ankle conversion, and that's the right place to worry. The policy is trained entirely on a serial ankle model; every real command is routed through the conversion. The paper gives no direct validation of that module—no comparison of commanded versus achieved parallel joint trajectories, no ablation against a directly trained parallel-ankle policy. So if the Jacobian or PD mapping is off, transfer could degrade or fail even with perfect domain randomization. This is a real gap, but not a fatal one: the hardware demo works, and the conversion is a standard kinematic technique. It just needs an explicit quantitative check.

Another minor point: the reward table is dense and some weights look hand-tuned, but that's expected for this kind of framework paper. The citation pattern is honest; they cite IsaacGymEnvs, legged-gym, rsl_rl, and humanoid-gym in the acknowledgment. The code link is provided—I haven't verified the repo compiles, but the claim is checkable.

Who is this for? Practitioners who want a starting point for training humanoid locomotion policies and deploying them on parallel-ankle robots. Researchers already deep in sim-to-real won't find new learning principles, but the framework and the conversion module are useful reusable pieces. It deserves a serious referee—conditional accept with a request for quantitative transfer evaluation.

My recommendation: send it to review, but require the authors to add quantitative sim-to-real comparison (tracking error, success rates) and a direct validation of the series-parallel conversion module. If the code checks out, this becomes a worthwhile community resource.

Referee Report

4 major / 6 minor

Summary. Booster Gym proposes an end-to-end reinforcement learning framework for humanoid locomotion, spanning Isaac Gym training, cross-simulator testing in MuJoCo and Webots, and real-world deployment on the Booster T1 robot. The framework uses asymmetric actor-critic PPO with domain randomization, a gait-cycle-conditioned observation, and a series-parallel ankle conversion module in the SDK to reconcile the virtual serial ankle model with the physical parallel ankle mechanism. The paper claims zero-shot sim-to-real transfer and supports the claim with qualitative real-world demonstrations of omnidirectional walking, terrain adaptation, and disturbance recovery, together with a qualitative joint trajectory comparison across Isaac Gym, MuJoCo, and the real robot.

Significance. If the central claim holds, Booster Gym would be a useful community resource: it integrates common RL training components, domain randomization, and a practical solution for parallel ankle mechanisms, with code released as open source. The paper also contributes a concrete latency-measurement procedure and cross-simulator evaluation, and the qualitative hardware demonstrations indicate that the trained policy is at least feasible on the physical robot. However, the evidence currently falls short of supporting the stronger 'seamless transfer' claim: the real-world evaluation is almost entirely qualitative, no repeated-trial statistics or failure rates are reported, the series-parallel conversion module is not quantitatively validated, and the central transfer claim therefore rests on unmeasured assumptions about kinematic conversion fidelity and randomization coverage.

major comments (4)

[Section III-D] The zero-shot sim-to-real claim depends critically on the series-parallel ankle conversion module, but the paper provides no quantitative validation of this load-bearing component. Specifically, there is no comparison of commanded versus achieved joint trajectories at the parallel linkage, no conversion-error measurement, no ablation with a policy trained directly on a parallel-ankle model, and no success-rate or tracking-error statistics. The qualitative overlap in Fig. 7 cannot isolate the conversion error from other simulation discrepancies, and the acknowledged PhysX limitation with closed kinematic chains (Section II-B) makes this validation necessary.
[Sections IV-A, IV-B, IV-C] The experimental section is almost entirely qualitative, despite the claim in Section IV-C that the dynamics gap is 'systematically quantified.' Only the communication latency and policy inference time are quantified; the actual walking performance, terrain adaptation, push recovery, and cross-simulator fidelity are reported via photographs and single illustrative trajectories, without error bars, repeated trials, failure rates, or baseline comparisons. As a result, the statement in Section V that the framework provides 'zero-shot sim-to-real transfer capabilities' is stronger than the measured evidence.
[Fig. 7 and Section IV-C] Fig. 7 shows joint position-velocity trajectories over a single 0.5 m/s forward-walking condition across Isaac Gym, MuJoCo, and the real robot, but no quantitative similarity metric is computed and no variance information is shown for the 10 Isaac Gym environments. The claim that MuJoCo 'serves as an effective testing platform that closely approximates real-world dynamics' is therefore unsupported; a quantitative comparison over multiple gaits, speeds, and commands is needed.
[Section III-B and III-C] The framework acknowledges known simulation weaknesses, particularly the simplified collision estimation in Isaac Gym and the lack of closed-chain support in PhysX, but does not quantify their impact on the final policy. The gait reward substitutes foot height for foot contact forces, and the paper does not test whether this substitution introduces a systematic sim-to-real bias; an ablation or sensitivity study on these modeling choices would materially strengthen the transfer claim.

minor comments (6)

[Section III-A] There is a typo in the sentence describing PPO: 'trian' should be 'train'.
[Abstract and Section I] The repository URL appears inconsistently as 'booster_gym' in the abstract and 'booster gym' in the full text; please verify and unify the URL.
[Fig. 2] The label 'Parallel-Series' in the figure caption is not defined in the text; the direction of conversion between the parallel and series representations should be clarified.
[Table II] The notation '1 q>qmax + 1 q<qmin' in the joint position limit reward should be defined explicitly as indicator functions.
[Section IV-C] The phrase 'quantitatively quantify' is redundant; consider rewording to 'measure' or 'quantify'.
[Acknowledgements] The acknowledgements mention 'legged-gym' and 'rsl-rl' but do not provide explicit citations or repository URLs; these should be included for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: Booster Gym's sim-to-real claim rests on empirical hardware trials and cross-simulator comparisons, not on any fitted parameter renamed as a prediction.

full rationale

Booster Gym is an engineering and experimental framework paper rather than a formal derivation. The central claim, that policies trained in Isaac Gym transfer zero-shot to the Booster T1, is supported by reported hardware experiments such as omnidirectional walking, terrain adaptation, push recovery, and the joint trajectory comparison in Fig. 7. No equation in the paper is defined in terms of the quantity it is supposed to predict. The reward terms in Table II are hand-tuned objectives and tuning choices; the policy's success is not implied by their definitions and is tested empirically. The domain randomization ranges, including the 0-20 ms latency randomization, are calibrated to measured real-world values, but this is a standard engineering calibration step rather than a circular fit, because the later transfer test is an independent empirical outcome. The series-parallel ankle conversion in Section III-D is an engineering assumption about kinematic mapping and is not quantitatively validated in isolation, but the paper does not derive successful transfer from the conversion mechanism itself, so the lack of validation is a correctness or rigor concern rather than circularity. The only self-citations, such as reference [3], appear in related-work background and are not load-bearing for the paper's central claim. There is no imported uniqueness theorem, no ansatz smuggled in via citation, and no renaming of a known empirical result as a new framework. The main weaknesses are qualitative validation metrics and an unmeasured conversion error, both of which fall outside the definition of circular reasoning.

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

The framework depends on hand-tuned reward weights and domain randomization ranges, plus several unvalidated modeling assumptions about simulator fidelity and the series-parallel ankle conversion. No new physical entities are introduced.

free parameters (4)

Reward weights = Hand-tuned; Table II lists e.g., velocity tracking 1.0, base height -20.0
The learned gait is directly set by these weights; no tuning procedure or sensitivity analysis is reported.
Gait frequency f = Not reported
The gait cycle observation cos(2πft), sin(2πft) depends on this frequency, which is chosen but never specified.
Domain randomization ranges = Not fully reported; latency 0-20 ms is stated
Policy robustness depends on the ranges for mass, center of mass, friction, PD gains, and latency, but only the latency range is given.
PD gains k_p and k_d = Predefined but values not reported
The motor driver PD gains convert joint position targets to torques and are central to deployment, yet their values are omitted.

assumptions (4)

domain assumption PPO with GAE and the stated reward will converge to a policy that tracks velocity commands after training.
Section III-A uses the standard RL objective and relies on empirical convergence; no convergence proof or hyperparameter study is provided.
domain assumption Isaac Gym's PhysX engine, despite missing closed-chain support, accurately simulates Booster T1 dynamics closely enough for zero-shot transfer after domain randomization.
Section II and Section III-C acknowledge PhysX limitations but assert they are mitigated by design choices and randomization.
ad hoc to paper The transposed-Jacobian series-parallel conversion in the SDK faithfully maps virtual serial ankle commands to the physical parallel mechanism.
Section III-D introduces this module as the solution for closed chains, but no kinematic error analysis or independent validation is given.
domain assumption The proprioceptive observation vector, including IMU data, joint positions and velocities, previous action, gait phase, and commands, contains enough information for real-world deployment.
Section III-B defines the observation space; no ablation study shows each component is necessary or sufficient.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Booster Gym: An End-to-End Reinforcement Learning Framework for Humanoid Robot Locomotion." pith.science (2026). https://pith.science/paper/EKYWGG5W

@misc{pith2026250615132,
  author       = {Pith},
  title        = {Pith review of: Booster Gym: An End-to-End Reinforcement Learning Framework for Humanoid Robot Locomotion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EKYWGG5W}},
  note         = {Machine review of arXiv:2506.15132}
}

read the original abstract

Recent advancements in reinforcement learning (RL) have led to significant progress in humanoid robot locomotion, simplifying the design and training of motion policies in simulation. However, the numerous implementation details make transferring these policies to real-world robots a challenging task. To address this, we have developed a comprehensive code framework that covers the entire process from training to deployment, incorporating common RL training methods, domain randomization, reward function design, and solutions for handling parallel structures. This library is made available as a community resource, with detailed descriptions of its design and experimental results. We validate the framework on the Booster T1 robot, demonstrating that the trained policies seamlessly transfer to the physical platform, enabling capabilities such as omnidirectional walking, disturbance resistance, and terrain adaptability. We hope this work provides a convenient tool for the robotics community, accelerating the development of humanoid robots. The code can be found in https://github.com/BoosterRobotics/booster_gym.

Figures

Figures reproduced from arXiv: 2506.15132 by the authors.

**Figure 2.** An overview of the control architecture for training [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 5.** Walking on different types of terrains with the same [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

Figures from the paper (2 more)

**Figure 7.** Figure 7: Joint position-velocity trajectories with a 0.5 m/s forward velocity commands in Isaac Gym, MuJoCo, and real-world [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]

**Figure 8.** Figure 8: Joint position response to a 0.1 rad step command [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Bridging the Sim-to-Real Gap in Parallel-Link Leg Mechanisms via Simulator-Side Dynamics Normalization
cs.RO 2026-08 unverdicted novelty 6.0 of 10

Simulator-side normalization that adds actuator inertia redistribution and residual linkage inertia to serial-tree models of parallel-link legs reduces sim-to-real motion, torque, and force errors by 60-82%.
RAVEN: Reinforcement-Adaptive Visibility-Graph Planning for Robust Humanoid Navigation with Collision-Free MPC
cs.RO 2026-07 conditional novelty 5.0 of 10

Reinforcement learning that adjusts obstacle-inflation radii in a visibility-graph planner improves humanoid navigation robustness under delay and noise, beating a static MPC baseline and an end-to-end RL policy in a ...

Reference graph

Works this paper leans on

21 extracted references · 3 canonical work pages · cited by 2 Pith papers

[1]

An overview of multi-task control for redundant robot based on quadratic program- ming,

Q. Li, Y . Pang, W. Cai, Y . Wang, Q. Li, and M. Zhao, “An overview of multi-task control for redundant robot based on quadratic program- ming,” inChinese Intelligent Automation Conference. Springer, 2023, pp. 641–666

work page 2023
[2]

Tailoring solution accuracy for fast whole-body model predictive control of legged robots,

C. Khazoom, S. Hong, M. Chignoli, E. Stanger-Jones, and S. Kim, “Tailoring solution accuracy for fast whole-body model predictive control of legged robots,”IEEE Robotics and Automation Letters, 2024

work page 2024
[3]

Hifar: Multi-stage curriculum learning for high-dynamics humanoid fall recovery,

P. Chen, Y . Wang, C. Luo, W. Cai, and M. Zhao, “Hifar: Multi-stage curriculum learning for high-dynamics humanoid fall recovery,”arXiv preprint arXiv:2502.20061, 2025

arXiv 2025
[4]

Real-world humanoid locomotion with reinforcement learning,

I. Radosavovic, T. Xiao, B. Zhang, T. Darrell, J. Malik, and K. Sreenath, “Real-world humanoid locomotion with reinforcement learning,”Science Robotics, vol. 9, no. 89, p. eadi9579, 2024

2024
[5]

Domain randomization for transferring deep neural networks from simulation to the real world,

J. Tobin, R. Fong, A. Ray, J. Schneider, W. Zaremba, and P. Abbeel, “Domain randomization for transferring deep neural networks from simulation to the real world,” in2017 IEEE/RSJ international con- ference on intelligent robots and systems (IROS). IEEE, 2017, pp. 23–30

2017
[6]

Closing the sim-to-real loop: Adapting simula- tion randomization with real world experience,

Y . Chebotar, A. Handa, V . Makoviychuk, M. Macklin, J. Issac, N. Ratliff, and D. Fox, “Closing the sim-to-real loop: Adapting simula- tion randomization with real world experience,” in2019 International Conference on Robotics and Automation (ICRA). IEEE, 2019, pp. 8973–8979

2019
[7]

Learning quadrupedal locomotion over challenging terrain,

J. Lee, J. Hwangbo, L. Wellhausen, V . Koltun, and M. Hutter, “Learning quadrupedal locomotion over challenging terrain,”Science robotics, vol. 5, no. 47, p. eabc5986, 2020

2020
[8]

Rma: Rapid motor adaptation for legged robots,

A. Kumar, Z. Fu, D. Pathak, and J. Malik, “Rma: Rapid motor adaptation for legged robots,”arXiv preprint arXiv:2107.04034, 2021

arXiv 2021

Show all 21 references

[9]

Adapt- ing rapid motor adaptation for bipedal robots,

A. Kumar, Z. Li, J. Zeng, D. Pathak, K. Sreenath, and J. Malik, “Adapt- ing rapid motor adaptation for bipedal robots,” in2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022, pp. 1161–1168

2022
[10]

Reinforcement learning for versatile, dynamic, and robust bipedal locomotion control,

Z. Li, X. B. Peng, P. Abbeel, S. Levine, G. Berseth, and K. Sreenath, “Reinforcement learning for versatile, dynamic, and robust bipedal locomotion control,”The International Journal of Robotics Research, p. 02783649241285161, 2024

2024
[11]

Ad- vancing humanoid locomotion: Mastering challenging terrains with denoising world model learning,

X. Gu, Y .-J. Wang, X. Zhu, C. Shi, Y . Guo, Y . Liu, and J. Chen, “Ad- vancing humanoid locomotion: Mastering challenging terrains with denoising world model learning,”arXiv preprint arXiv:2408.14472, 2024

2024 arXiv
[12]

Learning agile and dynamic motor skills for legged robots,

J. Hwangbo, J. Lee, A. Dosovitskiy, D. Bellicoso, V . Tsounis, V . Koltun, and M. Hutter, “Learning agile and dynamic motor skills for legged robots,”Science Robotics, vol. 4, no. 26, p. eaau5872, 2019

2019
[13]

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. Handaet al., “Isaac gym: High performance gpu-based physics simulation for robot learning,” arXiv preprint arXiv:2108.10470, 2021

2021 arXiv
[14]

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 Conference on Robot Learning. PMLR, 2022, pp. 91–100

2022
[15]

Mujoco playground,

K. Zakka, B. Tabanpour, Q. Liao, M. Haiderbhai, S. Holt, J. Y . Luo, A. Allshire, E. Frey, K. Sreenath, L. A. Kahrset al., “Mujoco playground,”arXiv preprint arXiv:2502.08844, 2025

2025 arXiv
[16]

Genesis: A universal and generative physics engine for robotics and beyond,

G. Authors, “Genesis: A universal and generative physics engine for robotics and beyond,” December 2024. [Online]. Available: https://github.com/Genesis-Embodied-AI/Genesis

2024
[17]

Humanoid-gym: Reinforcement learning for humanoid robot with zero-shot sim2real transfer,

X. Gu, Y .-J. Wang, and J. Chen, “Humanoid-gym: Reinforcement learning for humanoid robot with zero-shot sim2real transfer,”arXiv preprint arXiv:2404.05695, 2024

2024 arXiv
[18]

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,”arXiv preprint arXiv:1710.06542, 2017

2017 arXiv
[19]

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
[20]

High- dimensional continuous control using generalized advantage estima- tion,

J. Schulman, P. Moritz, S. Levine, M. Jordan, and P. Abbeel, “High- dimensional continuous control using generalized advantage estima- tion,”arXiv preprint arXiv:1506.02438, 2015

2015 arXiv
[21]

Sim-to-real learning of all common bipedal gaits via periodic reward composition,

J. Siekmann, Y . Godse, A. Fern, and J. Hurst, “Sim-to-real learning of all common bipedal gaits via periodic reward composition,” in2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 7309–7315

2021

Pith tools

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

[1] [1]

An overview of multi-task control for redundant robot based on quadratic program- ming,

Q. Li, Y . Pang, W. Cai, Y . Wang, Q. Li, and M. Zhao, “An overview of multi-task control for redundant robot based on quadratic program- ming,” inChinese Intelligent Automation Conference. Springer, 2023, pp. 641–666

work page 2023

[2] [2]

Tailoring solution accuracy for fast whole-body model predictive control of legged robots,

C. Khazoom, S. Hong, M. Chignoli, E. Stanger-Jones, and S. Kim, “Tailoring solution accuracy for fast whole-body model predictive control of legged robots,”IEEE Robotics and Automation Letters, 2024

work page 2024

[3] [3]

Hifar: Multi-stage curriculum learning for high-dynamics humanoid fall recovery,

P. Chen, Y . Wang, C. Luo, W. Cai, and M. Zhao, “Hifar: Multi-stage curriculum learning for high-dynamics humanoid fall recovery,”arXiv preprint arXiv:2502.20061, 2025

arXiv 2025

[4] [4]

Real-world humanoid locomotion with reinforcement learning,

I. Radosavovic, T. Xiao, B. Zhang, T. Darrell, J. Malik, and K. Sreenath, “Real-world humanoid locomotion with reinforcement learning,”Science Robotics, vol. 9, no. 89, p. eadi9579, 2024

2024

[5] [5]

Domain randomization for transferring deep neural networks from simulation to the real world,

J. Tobin, R. Fong, A. Ray, J. Schneider, W. Zaremba, and P. Abbeel, “Domain randomization for transferring deep neural networks from simulation to the real world,” in2017 IEEE/RSJ international con- ference on intelligent robots and systems (IROS). IEEE, 2017, pp. 23–30

2017

[6] [6]

Closing the sim-to-real loop: Adapting simula- tion randomization with real world experience,

Y . Chebotar, A. Handa, V . Makoviychuk, M. Macklin, J. Issac, N. Ratliff, and D. Fox, “Closing the sim-to-real loop: Adapting simula- tion randomization with real world experience,” in2019 International Conference on Robotics and Automation (ICRA). IEEE, 2019, pp. 8973–8979

2019

[7] [7]

Learning quadrupedal locomotion over challenging terrain,

J. Lee, J. Hwangbo, L. Wellhausen, V . Koltun, and M. Hutter, “Learning quadrupedal locomotion over challenging terrain,”Science robotics, vol. 5, no. 47, p. eabc5986, 2020

2020

[8] [8]

Rma: Rapid motor adaptation for legged robots,

A. Kumar, Z. Fu, D. Pathak, and J. Malik, “Rma: Rapid motor adaptation for legged robots,”arXiv preprint arXiv:2107.04034, 2021

arXiv 2021