Reproduce “Dexterous Robot Manipulation from Human Demonstrations via Contact-Anchored Retargeting and Residual Policy Learning” 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

Zihao Yang, Chengyuan Liu, Yu Zhou, Runze Lv, Tianyu Cui, Sheng Yi, Haohua Zhu, Irvine Lu and 1 more

The authors present a motion-capture-to-robot pipeline for dexterous manipulation that transfers finger-object contacts rather than human joint motion. It refines motion with a simulated MANO hand, retargets contacts to robot hands, and uses residual policies. Across 25,454 trajectories, average success rises from 7.26% to 59.31%; force conditioning improves results in tested setups. The pipeline uses motion capture, not arbitrary video.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.