SciGroveBeta
Robotics

Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstration

Xinghao Zhu, Zixi Liu, Shalin Jain, Chenran Li, Milad Noori, Huihua Zhao, John Welsh, Michael Andres Lin, Wei Liu, Tingwu Wang, Xingye Da, Zhengyi Luo, Vishal Kulkarni, Naema Bhatti, Yuke Zhu, Linxi Fan, Bowen Wen, Danfei Xu, Soha Pouya, Yan Chang

Featured July 9, 2026

Original Paper

This analysis was generated by SciGrove. Upload your own PDFs or enter a DOI — and get the same AI breakdown on any paper.

Get started

AI-generated analysis — This is SciGrove's AI interpretation of the paper, not peer-reviewed content. Always refer to the original paper.

Simply

Robots learn tricky hand movements from human examples by focusing on the forces and twists they apply to objects, rather than just where they touch, making it easier to copy complex actions.

In depth
The paper introduces CHORD, a framework that addresses the challenge of transferring human dexterous manipulation skills to robots. Its core innovation is object-centric contact wrench space guidance, which represents human and robot interactions by the forces and torques they can induce on an object. This allows for a robust comparison of physically aligned contacts, enabling scalable reinforcement learning for contact-rich dexterous manipulation across diverse objects and tasks, even with morphological differences between human and robot hands.

Key Takeaways

  • 1
    Introduces CHORD, a novel framework that uses contact wrench space guidance to effectively transfer human demonstrations to robot policies for dexterous manipulation.
  • 2
    Develops a large-scale simulation benchmark of 4,739 bimanual dexterous manipulation tasks, enabling comprehensive evaluation of contact-rich manipulation algorithms.
  • 3
    Demonstrates state-of-the-art performance (82.12% success rate on 1,831 tasks) and strong generalization capabilities, including whole-body manipulation and successful real-world transfer.

Conceptual Flow

HIGH LEVEL
1
Learning Robot Dexterity from Human Examples

The robot watches a human move an object, figures out the forces the human applied, and then learns to apply similar forces to move the object itself.

Human Hand Moves Object
Extract Forces & Twists
Robot Learns to Apply Forces
2
Robots Master Complex Tasks

The new method helps robots successfully do many hard tasks, like opening boxes or stirring, even with different robot bodies, and works well in the real world.

Robot Tries Many Tasks
Achieves High Success
Works on Real Robots
Handles Many Objects