March 7, 2024Open Access

Learning Human-to-Humanoid Real-Time Whole-Body Teleoperation

Key Points

Whole-body teleoperation achieved using an RGB camera, enabling dynamic humanoid robot movements like walking and kicking.
The study shows a robust real-time humanoid motion imitator trained on a large motion dataset derived from human actions.
A sim-to-data approach filtered motions effectively, utilizing reinforcement learning to enhance the humanoid robot's capabilities in real-time scenarios with zero-shot transfer to hardware from simulation setups in robotics experiments to improve interaction with environments and tasks.

Abstract

We present Human to Humanoid (H2O), a reinforcement learning (RL) based framework that enables real-time whole-body teleoperation of a full-sized humanoid robot with only an RGB camera. To create a large-scale retargeted motion dataset of human movements for humanoid robots, we propose a scalable "sim-to-data" process to filter and pick feasible motions using a privileged motion imitator. Afterwards, we train a robust real-time humanoid motion imitator in simulation using these refined motions and transfer it to the real humanoid robot in a zero-shot manner. We successfully achieve teleoperation of dynamic whole-body motions in real-world scenarios, including walking, back jumping, kicking, turning, waving, pushing, boxing, etc. To the best of our knowledge, this is the first demonstration to achieve learning-based real-time whole-body humanoid teleoperation.

Read Full Paperexternally

AI से पूछें

Bookmark

View Full Paper