Reproduce “ReflexArc: Bottom-Up Integration of Instantaneous Tactile Reflexes into Whole-Body Humanoid Control” 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
Mingyu Wang, Si Li, Ziqing Chen, He Wang, Li Yi
ARC (Ascending Reflex Control) gives humanoid robots an immediate tactile-triggered whole-body withdrawal while a learned policy remains active to manage balance and task behavior. The system uses a model-based solver and integrates the reflex into the policy’s interfaces to limit interference. In a Unitree G1 box-contact test, immediate response reduced peak force 34.1%; this result applies to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.