arXiv：cs.LG· An Dang, Arturo Flores Alvarez, Yu-Ming Chen, Conor Mc Gartoll, Helen Sun, Aaron Ames, Nima Fazeli, Manikantan Nambi·· 4 小时前AI 评分42
HULK：人形机器人全身强力移动作业学习框架
HULK: Learning Whole-Body Forceful Loco-Manipulation for Humanoids
AI 导读
HULK 是一个面向人形机器人全身强力移动作业（loco-manipulation）的控制框架，通过 MPC 引导强化学习、训练手臂腕力追踪与抱持大物体行走两个教师策略，并蒸馏为单一策略。
正文
Abstract:Humanoid loco-manipulation of large, heavy objects demands forceful interaction across the entire body. However, such payloads shift a humanoid's center of mass and impose sustained loads across the upper body, challenging balance and command tracking. We present HULK, a whole-body control framework for forceful loco-manipulation. Using model predictive control (MPC) to guide reinforcement learning with predictions of the loaded dynamics, we train two teachers: one tracks arm motions under wrist forces, and the other locomotes while holding large objects against the body. A capture-point control barrier function augments the wrist-force teacher during training to improve balance under load. We distill both teachers into a single policy. Evaluation spans simulation and the Unitree G1. In simulation, the teacher with the barrier function achieves the lowest forward and lateral velocity tracking errors at 10 kg per arm among evaluated controllers and reduces aggregate divergent component of motion (DCM) excursion magnitude by 35.7% relative to MPC-guided reinforcement learning alone. Our wrist-force teacher withstands torso push disturbances of up to 130 N.
| Comments: | 16 pages, 7 figures, IEEE International Conference on Robotics and Automation 2027 |
| Subjects: | Robotics (cs.RO); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.08970 [cs.RO] |
| (or arXiv:2610.08970v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08970 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: An Dang [view email]
[v1]
Tue, 6 Oct 2026 18:33:11 UTC (10,791 KB)
来源：arXiv：cs.LG · arxiv.org