Reproduce “Chronos Policy: Learning Robot Policies on Physical Time” 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
Changhao Wang, Jaden Clark, Shuran Song
Chronos Policy helps robots handle irregular sensing and delayed execution by using each observation’s capture time and each action’s requested execution time. Its time-aware rotary positional encoding and timing randomization expose the policy to varied schedules. In Conveyor, it achieved 32.7% success under combined timing perturbations, versus 0% for a fixed-time baseline; this result is limited to that tested condition.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.