arXiv Robotics — research abstracts· Mike Zhang, Dongho Kang, Kevin Bergamin, Nicola Burger, Robin Deits, Jonathan Foster, Bilal Hammoud, Katie Hughes, Francesco Iacobelli, Twan Koolen, M. Eva Mungai, Zach Nobles, Shane Rozen-Levy, Jean Pierre Sleiman, Fangzhou Yu, Yunbo Zhang, Alfred Rizzi, Jessica Hodgins, Scott Kuindersma, Yeuhi Abe, Sylvain Bertrand, Farbod Farshidian·· 11 days agoEditorial score65
HuMBLE: Human Motion-Driven Behavior Learning for Embodied Locomotion
HuMBLE: Human Motion-Driven Behavior Learning for Embodied Locomotion
Summary
This research introduces HuMBLE, a learning framework that synthesizes real-time, steerable, and robust locomotion policies for humanoid robots by balancing command tracking and generalization from human motion data. The framework is validated on three humanoid robots, demonstrating robust performance in real-world scenarios and the ability to reconstruct coordinated whole-body behavior from a steering command.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org