Reproduce “PARC: Physics-based Augmentation with Reinforcement Learning for Character Controllers” 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
Michael Xu, Yi Shi, KangKang Yin, Xue Bin Peng
A framework called PARC enables training of agile terrain traversal controllers from small motion datasets through iterative co-training between a diffusion-based motion generator and physics-based tracking controller, progressively expanding the motion repertoire while maintaining physical plausibility through simulation-based correction.
Key insights
- Physics-based correction is essential for maintaining motion quality across iterations
- Starting from small datasets, the system can progressively learn new skills beyond the initial data
- Continuous co-training allows both generator and tracker to leverage past experience
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.