Reproduce “Kimodo-Policy: Learning Kinematic Action Models for Humanoid Loco-Manipulation” 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
Yunheng Wang, Jiaxi Zhang, Wei Lu, Yuetong Fang, Taowen Wang, Yixiao Feng, Shuning Zhang, Junzhe Xu and 7 more
Kimodo-Policy uses a pretrained motion generator as the action-generation backbone for humanoid loco-manipulation. It conditions generation on visual observations and measured motion history, then repeatedly replans while a low-level tracker executes action prefixes. Across seven HumanoidArena tasks it reports 72.14% success versus 69.70% for GR00T N1.7, but trails leading baselines on HOI tasks; transfer beyond SONIC remains unestablished.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.