Reproduce “Learning Torque Control for Quadrupedal Locomotion” 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
Shuxiao Chen, Bike Zhang, Mark W. Mueller, Akshara Rai, Koushil Sreenath
The authors present a sim-to-real controller that directly commands quadruped motor torques, instead of sending joint-position targets to a separate proportional-derivative controller. They train it in simulation and test it on a Unitree A1 across terrain and disturbance conditions. In a standing kick test, it withstands three kicks while the position-policy robot falls after two; this comparison is limited.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.