Switch: Learning Agile Skills Switching for Humanoid Robots

TL;DR

This paper introduces Switch, a hierarchical system enabling humanoid robots to perform seamless transitions between diverse locomotion skills using deep reinforcement learning and an online skill scheduler.

Contribution

The paper presents a novel multi-skill system with a skill graph, a trained tracking policy, and an online scheduler for real-time, stable skill switching in humanoid robots.

Findings

High success rates in agile skill transitions

Stable and real-time execution of diverse locomotion skills

Effective motion imitation performance

Abstract

Recent advancements in whole-body control through deep reinforcement learning have enabled humanoid robots to achieve remarkable progress in real-world chal lenging locomotion skills. However, existing approaches often struggle with flexible transitions between distinct skills, cre ating safety concerns and practical limitations. To address this challenge, we introduce a hierarchical multi-skill system, Switch, enabling seamless skill transitions at any moment. Our approach comprises three key components: (1) a Skill Graph (SG) that establishes potential cross-skill transitions based on kinematic similarity within multi-skill motion data, (2) a whole-body tracking policy trained on this skill graph through deep reinforcement learning, and (3) an online skill scheduler to drive the tracking policy for robust skill execution and smooth transitions. For skill switching or significant tracking deviations, the scheduler performs online graph search to find the optimal feasible path, which ensures efficient, stable, and real-time execution of diverse locomotion skills. Comprehensive experiments demonstrate that Switch empowers humanoid to execute agile skill transitions with high success rates while maintaining strong motion imitation performance.