ImVR: Immersive VR Teleoperation System for General Purpose

Yulin Liu, Zihao He, Fanbo Xiang, Runlin Guo, Zhiao Huang, Jialin Zhang, Bo Ai, Stone Tao, Hao Su

Keywords: Teleoperation, Imitation Learning, Dexterous Manipulation

TL;DR: We present ImVR, an immersive VR-based teleoperation system that enables precise and intuitive control of both simulated and real-world robots.

Abstract: We present \textit{ImVR}, an immersive VR-based teleoperation system that places the operator in a fully 3D virtual environment for intuitive and precise robot control. Designed for general-purpose use, \textit{ImVR} supports a wide range of robot configurations—including single- and dual-arm setups, parallel grippers and multi-fingered dexterous hands—while operating seamlessly in both simulation and the real world. It ensures absolute alignment between the operator’s hand and the robot’s end-effector, offers adaptive viewpoints to enhance situational awareness and precision, and enables easy deployment across diverse robotic platforms. Experiments and user studies highlight \textit{ImVR}’s versatility, ease of use, and effectiveness in accelerating data collection, making it a powerful tool for large-scale robot learning and real-world manipulation tasks.

Submission Number: 26