Reproduce “Learning Force-Reactive Humanoid Loco-Manipulation” 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
Sihang Li, Xiaofeng Guo, Jinkun Cao, Charlie Cheng, Ruihan Yang, Chen Feng, Guanya Shi, Haozhi Qi
FLORE connects force-annotated demonstrations to force-aware control for humanoid contact tasks. Its diffusion policy uses wrist images, robot state, and hand wrenches to predict motion and per-hand force commands, which a controller tracks. On Unitree G1, FLORE succeeds in 24 of 30 rollouts across three tasks, versus 8 of 30 for a position-only policy; evidence is limited to those tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.