HandUMI: A Modular Hand-Worn Parallel-Jaw Interface for Robot-Free Bimanual Data Capture
cs.RO
Submitted: 2026-11-17
Updated: 2026-11-17
Comments: 7 pages, 2 figures
License: http://creativecommons.org/licenses/by/4.0/
The gist: HandUMI captures robot-free bimanual demonstrations with a hand-worn parallel-jaw interface.
Abstract
HandUMI captures robot-free bimanual demonstrations with a hand-worn parallel-jaw interface. Swappable fingertips match the robot's installed gripper; Quest/PICO controllers track hands and encoders measure aperture. Robot-agnostic recordings are retargeted by inverse kinematics to seven simulated bimanual arms (median position error at most 1.2 mm; at most 4.7 % of episodes rejected). On bimanual towel folding, 562 robot-free demonstrations train policies on real PiPER arms: ACT succeeds in 12/20 attempts and fine-tuned π 0.5 in 7/20; ACT on the first 101 and base π 0.5 each succeed in 0/20. Hardware and software are open source (Apache-2.0), and a pair costs USD 557 with Quest 3S, excluding the optional ZED Mini and host computer. Project website: https://murobotics.ai/research/handumi.
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation