Reproduce “RoboPace: Contact-Aware Time-Optimal Retiming for Action-Chunk Policies” 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
Mimo Shirasaka, Takehiko Ohkawa, Takuya Okubo, Nicola Scianca, Tatsuya Matsushima, Kei Ota
RoboPace retimes policy-predicted robot paths online, keeping routes while changing speed near predicted contact. It combines geometric proximity and wrist-camera contact cues with robot joint velocity and acceleration limits. On one dual-arm robot and policy family, it succeeded in 32/50 trials, versus 22/50 for uniform slow execution, and took 45–59% of slow execution’s successful-trial time on four of five commands.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.