Reproduce “Humanoid Mobility via Attention-based Contacts” 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
Junzhe He, Kehan Wen, Chong Zhang, Clemens Schwarke, Zewei Zhang, Kaixian Qu, Siyuan Hu, Matthias Müller and 2 more
The authors present a locomotion system that plans contacts rather than full-body motions. An attention-based planner selects terrain-aligned targets for nine body parts, while a controller turns them into movement; reference motions guide training but are not inputs at deployment. On eight simulation courses, it averages 92.5% success versus 58.8% for MGMT, limited to those courses.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.