Reproduce “HumanoidNDM: Closing the Training-Deployment Gap of Motion Tracking via Humanoid Neural Dynamics Model” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chao Chen, Zhenrong Wang, Jilong Wang, Yusen Feng, Jiangran Lyu, Li Yi, He Wang
HumanoidNDM adapts simulation-trained humanoid motion-tracking policies using robot data. It learns a differentiable dynamics model from proprioceptive histories and fine-tunes through model rollouts without motion capture. On a Unitree G1 walking with a 2 kg water bucket, path deviation fell from 38 ± 17° with direct transfer to 5 ± 3° after fine-tuning. Tests cover six tasks; the method is limited to time-invariant disturbances.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.