RwoR: Generating Robot Demonstrations from Human Hand Collection for Policy Learning without Robot

TL;DR

This paper introduces a novel method to generate robot demonstrations from human hand data using a generative model, enabling scalable policy learning without requiring robot control during data collection.

Contribution

The authors propose a hand-to-gripper generative model that translates human hand demonstrations into robot gripper actions, bridging the visual gap and improving data collection efficiency.

Findings

Generated robot demonstrations are of high quality and effective for policy learning.

The method significantly reduces the need for robot control during data collection.

Experiments show improved manipulation performance and data collection efficiency.

Abstract

Recent advancements in imitation learning have shown promising results in robotic manipulation, driven by the availability of high-quality training data. To improve data collection efficiency, some approaches focus on developing specialized teleoperation devices for robot control, while others directly use human hand demonstrations to obtain training data. However, the former requires both a robotic system and a skilled operator, limiting scalability, while the latter faces challenges in aligning the visual gap between human hand demonstrations and the deployed robot observations. To address this, we propose a human hand data collection system combined with our hand-to-gripper generative model, which translates human hand demonstrations into robot gripper demonstrations, effectively bridging the observation gap. Specifically, a GoPro fisheye camera is mounted on the human wrist to capture human hand demonstrations. We then train a generative model on a self-collected dataset of paired human hand and UMI gripper demonstrations, which have been processed using a tailored data pre-processing strategy to ensure alignment in both timestamps and observations. Therefore, given only human hand demonstrations, we are able to automatically extract the corresponding SE(3) actions and integrate them with high-quality generated robot demonstrations through our generation pipeline for training robotic policy model. In experiments, the robust manipulation performance demonstrates not only the quality of the generated robot demonstrations but also the efficiency and practicality of our data collection method. More demonstrations can be found at: https://rwor.github.io/