OmniH2O: Universal and Dexterous Human-to-Humanoid Whole-Body Teleoperation and Learning

Autor:	He, Tairan, Luo, Zhengyi, He, Xialin, Xiao, Wenli, Zhang, Chong, Zhang, Weinan, Kitani, Kris, Liu, Changliu, Shi, Guanya
Rok vydání:	2024
Předmět:	Computer Science - Robotics Computer Science - Computer Vision and Pattern Recognition Computer Science - Machine Learning Electrical Engineering and Systems Science - Systems and Control
Druh dokumentu:	Working Paper
Popis:	We present OmniH2O (Omni Human-to-Humanoid), a learning-based system for whole-body humanoid teleoperation and autonomy. Using kinematic pose as a universal control interface, OmniH2O enables various ways for a human to control a full-sized humanoid with dexterous hands, including using real-time teleoperation through VR headset, verbal instruction, and RGB camera. OmniH2O also enables full autonomy by learning from teleoperated demonstrations or integrating with frontier models such as GPT-4. OmniH2O demonstrates versatility and dexterity in various real-world whole-body tasks through teleoperation or autonomy, such as playing multiple sports, moving and manipulating objects, and interacting with humans. We develop an RL-based sim-to-real pipeline, which involves large-scale retargeting and augmentation of human motion datasets, learning a real-world deployable policy with sparse sensor input by imitating a privileged teacher policy, and reward designs to enhance robustness and stability. We release the first humanoid whole-body control dataset, OmniH2O-6, containing six everyday tasks, and demonstrate humanoid whole-body skill learning from teleoperated datasets. Comment: Project page: https://omni.human2humanoid.com/
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2406.08858 Zobrazit plný text záznamu View this record from Arxiv