Reproduce “Online Sim-to-Real Adaptation via Closed-Loop System Modeling” 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
Yuhao Huang, Samuel A. Moore, Boyuan Chen
OSRAM improves robot tracking by adapting commands while leaving the deployed controller unchanged. It learns how the robot-controller pair responds to commands, using randomized simulation pretraining and finetuning with deployment observations, then optimizes reference trajectories. On the LimX Tron1, hardware loco-manipulation x-direction RMSE fell from 36.94 to 3.11 centimeters. Evidence covers velocity tracking and loco-manipulation in low-dimensional command spaces.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.