Reproduce “PERA: Plug-in Experts for Robotic Actions in Continual Whole-Body 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
Shangchao Hu, Yadan Sun, Xibo Yu, Wei Huang
PERA extends a pretrained humanoid motion controller with a small, task-specific module while keeping the original controller frozen. The module can be enabled for a new behavior or bypassed to restore the original computation path. In simulation, it improved completion on eight untrained fast-dance references: full-reference completion rose from 59.38% to 71.88%, evidence limited to those motions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.