Reproduce “Learning Structured Locomotion References for Optimization-Based 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
Ömer Tarık Banuş, Uluç Saranlı, M. Mert Ankaralı
The authors introduce a learned supervisor that adapts body-motion, stance-force, foothold, body-height, and gait-timing references in a quadruped controller, while model-based layers execute joint commands. In flat-ground simulation, it averages 77.29% survival across a 60-condition disturbance grid, versus 50.35% for MPC+WBIC. These results compare configured systems and do not establish an optimal action representation or transfer to hardware.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.