CHORD：用人类演示的接触力旋量引导强化学习实现灵巧操作
Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstration
研究团队提出 CHORD 框架，以物体为中心的接触力旋量空间引导，将人类演示与机器人动作统一表示为对物体施加的力与力矩，从而用强化学习完成刚性与铰接物体的长程灵巧操作。团队构建了含 4,739 个双臂灵巧操作任务的仿真基准，在 1,831 个任务上平均成功率达 82.12%。CHORD 还能从单手和第三人称演示泛化到全身操作，成功率达 90.77%，学到的策略可在开环与闭环设置下迁移到真实世界。
Authors:Xinghao Zhu, Zixi Liu, Shalin Jain, Chenran Li, Milad Noori, Michael Andres Lin, Huihua Zhao, John Welsh, Mrinal Verghese, Wei Liu, Tingwu Wang, Xingye Da, Zhengyi Luo, Vishal Kulkarni, Naema Bhatti, Yuke Zhu, Linxi Fan, Bowen Wen, Danfei Xu, Soha Pouya, Yan Chang
Abstract:Dexterous robot manipulation can benefit from the abundance of human demonstrations, but transferring such demonstrations to robot policies remains challenging. We present Contact Wrench Guidance from Human Demonstration in Robotic Dexterous Manipulation (CHORD), a framework for long-horizon manipulation of rigid and articulated objects with reinforcement learning. The key idea is object-centric contact wrench space guidance: we represent human and robot motions by the forces and torques they can induce on the object, enabling similarity to be measured by the induced instantaneous motions. This guidance makes reinforcement learning more scalable for contact-rich dexterous manipulation. We further introduce a large-scale simulation benchmark with 4,739 bimanual dexterous manipulation tasks, constructed from motion-capture datasets and reconstructed in-house videos. Evaluated on 1,831 benchmark tasks, CHORD achieves an average success rate of 82.12%, demonstrating strong scalability. CHORD also generalizes to whole-body manipulation from hand-only and third-person demonstrations, achieving a 90.77% success rate, and the learned policies transfer to the real world in both open-loop and closed-loop settings.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2607.00033 [cs.RO] |
| (or arXiv:2607.00033v3 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2607.00033 arXiv-issued DOI via DataCite |
Submission history
From: Xinghao Zhu [view email]
[v1]
Mon, 22 Jun 2026 22:54:33 UTC (34,048 KB)
[v2]
Fri, 14 Aug 2026 05:48:56 UTC (34,050 KB)
[v3]
Tue, 10 Nov 2026 07:47:20 UTC (33,937 KB)
来源：arXiv cs.AI · arxiv.org