Pith. sign in

REVIEW 8 cited by

Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2109.11978 v3 pith:CDLA3MQO submitted 2021-09-24 cs.RO cs.LG

classification cs.ROcs.LG
keywords trainingparallelapproachminutesterrainlearningmassivelypolicies
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

In this work, we present and study a training set-up that achieves fast policy generation for real-world robotic tasks by using massive parallelism on a single workstation GPU. We analyze and discuss the impact of different training algorithm components in the massively parallel regime on the final policy performance and training times. In addition, we present a novel game-inspired curriculum that is well suited for training with thousands of simulated robots in parallel. We evaluate the approach by training the quadrupedal robot ANYmal to walk on challenging terrain. The parallel approach allows training policies for flat terrain in under four minutes, and in twenty minutes for uneven terrain. This represents a speedup of multiple orders of magnitude compared to previous work. Finally, we transfer the policies to the real robot to validate the approach. We open-source our training code to help accelerate further research in the field of learned legged locomotion.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 101 citations worldwide. Full citation record

  1. When Is a Learned Command Adapter Worth It? Closed-Loop Identification and Counterfactual Auditing of Frozen Locomotion Policies

    cs.AI 2026-07 conditional novelty 7.0 of 10

    A counterfactual audit separates same-state headroom from recoverable state-allocation gain, returning NO-GO or ABSTAIN for learned command adapters on frozen Go2 and H1 locomotion policies at 1% thresholds.

  2. Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills

    cs.RO 2026-08 conditional novelty 6.0 of 10

    A taxonomy of robot learning on a weights-versus-skills axis, with a five-rung self-improvement ladder whose top cell (feedback plus memory plus search) holds only a few recent systems.

  3. GASP: GPU-Accelerated Safe Planner for Real-Time Collision-Aware Motion Generation with Latent Trajectory Sampling

    cs.RO 2026-08 conditional novelty 5.0 of 10

    A learned B-spline planner with latent sampling generates collision-aware joint trajectories on the GPU in near-millisecond time, matching analytical planners and outperforming a GPU optimization baseline.

  4. Towards Miniature Humanoid Tele-Loco-Manipulation Using Virtual Reality and Reinforcement Learning

    cs.RO 2026-07 conditional novelty 5.0 of 10

    A VR-teleoperated, reinforcement-learning-balanced control stack lets a miniature ROBOTIS OP3 humanoid walk and manipulate objects simultaneously.

  5. Milo, a Fully Autonomous Indoor/Outdoor Robotic Guide Dog

    cs.RO 2026-07 conditional novelty 5.0 of 10

    Milo is an open-source, fully onboard robotic guide dog that navigates unseen indoor/outdoor paths while explicitly modeling the handler's position, with preliminary real-world tests against a handler-unaware costmap ...

  6. Sim2Swim: Zero-Shot Velocity Control for Agile AUV Maneuvering in 3 Minutes

    cs.RO 2025-12 conditional novelty 5.0 of 10

    A DRL velocity controller for AUVs, trained in under 3 minutes in simulation, transfers zero-shot to a real underwater vehicle for agile 6DOF path-following.

  7. KiVi: Kinesthetic-Visuospatial Integration for Dynamic and Safe Egocentric Legged Locomotion

    cs.RO 2025-09 conditional novelty 5.0 of 10

    A quadruped locomotion controller that explicitly separates proprioceptive and visual pathways stays stable under camera occlusion and visual corruption that destabilizes fused-vision policies.

  8. HuMam: Humanoid Motion Control via End-to-End Deep Reinforcement Learning with Mamba

    cs.RO 2025-09 conditional novelty 5.0 of 10

    A single-layer Mamba encoder as the policy backbone improves learning speed, stability, and energy efficiency of an end-to-end RL humanoid walking controller in simulation compared to a feedforward baseline.

Pith tools