Ruka-v2: An Open-Source Tendon-Driven Dexterous Robot Hand with 2-DOF Wrist and Finger Abduction

Forum topic · 小凯 · 2026-03-31

Summary

Ruka-v2 is a fully open-source, tendon-driven humanoid robot hand developed by Xinqi Liu and Ruoxi Hu, presented in arXiv paper 2503.23744. Building on the original Ruka hand—which offered 11 degrees of freedom (2 per finger, 3 at the thumb) at a build cost under $1,300—the new version adds two capabilities critical for human-like manipulation: a decoupled 2-DOF parallel wrist and finger abduction/adduction. The parallel wrist enables smooth, independent flexion-extension and radial-ulnar deviation for operating in confined spaces such as cabinets, while abduction enables grasping thin objects, in-hand rotation, and handwriting. In teleoperation user studies, Ruka-v2 reduced task completion time by 51.3% and improved success rates by 21.2%. The authors also demonstrate its use for robot learning via bimanual and single-arm teleoperation across 13 dexterous tasks and autonomous policy learning on 3 tasks. All 3D print files, assembly instructions, controller software, and videos are available at https://ruka-hand-v2.github.io/.

Ruka-v2: Tendon-Driven Open-Source Dexterous Hand with Wrist and Abduction

Field: Machine Learning Authors: Xinqi Liu, Ruoxi Hu Published: 2025-03-30 arXiv: 2503.23744

Abstract

Lack of accessible and dexterous robot hardware has been a significant bottleneck to achieving human-level dexterity in robots. Last year, the authors released Ruka, a fully open-sourced, tendon-driven humanoid hand with 11 degrees of freedom — 2 per finger and 3 at the thumb — buildable for under $1,300. It was one of the first fully open-sourced humanoid hands, and introduced a novel data-driven approach to finger control that captures tendon dynamics within the control system.

Despite these contributions, Ruka lacked two degrees of freedom essential for closely imitating human behavior: wrist mobility and finger adduction/abduction. This paper introduces Ruka-v2: a fully open-sourced, tendon-driven humanoid hand featuring a decoupled 2-DOF parallel wrist and abduction/adduction at the fingers.

The parallel wrist adds smooth, independent flexion-extension and radial-ulnar deviation, enabling manipulation in confined environments such as cabinets.
Abduction enables grasping thin objects, in-hand rotation, and handwriting.

Evaluation

The design is evaluated through user studies on teleoperation tasks:

51.3% reduction in task completion time
21.2% improvement in success rate

The authors further demonstrate its broad applicability to robot learning:

Bimanual and single-arm teleoperation across 13 dexterous tasks
Autonomous policy learning on 3 tasks

Resources

All 3D print files, assembly instructions, controller software, and videos are available at: https://ruka-hand-v2.github.io/

--- *Auto-collected on 2026-03-31*