Reproduce “TMA: Sim-to-Real Forceful Humanoid Manipulation via Tactile-Driven Motor Adaptation” 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
Yiting Chen, Kelin Yu, Sarthak Kaingade, Bryan Trinh, Kaiyu Hang, Roberto Martin-martin, Nima Fazeli, X. Alice Wu
TMA uses tactile history to adapt a humanoid while turning door knobs and opening doors. Its four-fingertip representation encodes normal load and shear, and a history encoder adjusts a frozen manipulation policy. In trials, it opened 27 of 40 doors, versus 0 for RMA and 6 for a reactive tactile baseline, across four knobs using pose estimation and friction-increasing tape.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.