Reproduce “Global-Local Attention Decomposition for Terrain Encoding in Humanoid Perceptive 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
Shengcheng Fu, Yang Zhang, Zhanxiang Cao, Liyun Yan, Yizhi Chen, Yunpeng Yin, Yue Gao
GLAD gives a humanoid separate global terrain context and local foothold attention from a robot-centered elevation map. A global branch summarizes terrain, while a state-conditioned query selects salient nearby features for foot placement. In simulation, it achieved 93.48±1.98% success on a winding pillar path, versus 81.13±3.70% for AME-GC; this comparison does not establish the cause of the difference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.