AnyTeleop: A General Vision-Based Dexterous Robot Arm-Hand Teleoperation System

TL;DR

AnyTeleop is a versatile vision-based teleoperation system that supports multiple robot models and environments, achieving high success rates and improved imitation learning performance across various configurations.

Contribution

The paper introduces AnyTeleop, a unified system that generalizes vision-based robot teleoperation across different hardware and environments, enhancing flexibility and performance.

Findings

Outperforms previous system in real-world success rate.

Leads to better imitation learning in simulation.

Supports multiple robot models and camera configurations.

Abstract

Vision-based teleoperation offers the possibility to endow robots with human-level intelligence to physically interact with the environment, while only requiring low-cost camera sensors. However, current vision-based teleoperation systems are designed and engineered towards a particular robot model and deploy environment, which scales poorly as the pool of the robot models expands and the variety of the operating environment increases. In this paper, we propose AnyTeleop, a unified and general teleoperation system to support multiple different arms, hands, realities, and camera configurations within a single system. Although being designed to provide great flexibility to the choice of simulators and real hardware, our system can still achieve great performance. For real-world experiments, AnyTeleop can outperform a previous system that was designed for a specific robot hardware with a higher success rate, using the same robot. For teleoperation in simulation, AnyTeleop leads to better imitation learning performance, compared with a previous system that is particularly designed for that simulator. Project page: https://yzqin.github.io/anyteleop/.