Overview
Field: RO (Robotics) Authors: Yunhai Feng, Natalie Leung, Jiaxuan Wang, Lujie Yang, Haozhi Qi Published: 2026-07-13 arXiv: 2607.11874
Summary
Recent work in humanoid whole-body control has found success with a simple recipe: retarget human motion to robot kinematic references, then train policies via reinforcement learning (RL) to track them. But how does this recipe transfer to dexterous manipulation? The answer is not obvious, as manipulation involves complex, contact-rich dynamics and requires delicate regulation of contact modes and forces.
Key points
- REGRIND is a minimalist retargeting-guided RL pipeline that learns dexterous manipulation policies from a single human demonstration.
- Human hand-object motion is retargeted to a robot reference that preserves hand-object spatial and contact relationships.
- A residual RL policy is trained in simulation to track object-centric keypoints along that reference.
- The learned policy is transferred zero-shot to real hardware through careful system identification.
- Results on two different multi-finger hands show fluid, human-like behaviors on contact-rich tool-use tasks, including manipulating scissors and turning a screwdriver.
- Systematic hardware experiments identify and analyze the key factors affecting sim-to-real transfer in dexterous manipulation, providing practical guidance for retargeting-based learning in contact-rich scenarios.