Reproduce “Ask1: Development and Reinforcement Learning-Based Control of a Custom Quadruped Robot” 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
Yang Zhang, Yuxing Lu, Guiyang Xin, Yufei Xue, Chenkun Qi, Kairong Qin, Yan Zhuang
Ask1’s controller trains in simulation to follow motion commands over uneven terrain, without terrain plans or reference motions. A foot-placement reward adjusts desired foot positions for gait phase and commanded velocity, replacing the authors’ earlier AMP style constraint. The authors report Ask1 climbing 18-centimeter stairs in the real world, but provide no quantitative success metric or controlled baseline comparison.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.