Search papers, labs, and topics across Lattice.
This paper introduces Switch, a hierarchical reinforcement learning system for humanoid robots that enables agile switching between locomotion skills. A Skill Graph (SG) is constructed based on kinematic similarity to define potential skill transitions, and a whole-body tracking policy is trained on this graph using deep RL. An online skill scheduler then uses graph search to find optimal transition paths, enabling real-time execution of diverse locomotion skills.
Humanoid robots can now nimbly switch between skills on the fly, thanks to a hierarchical reinforcement learning approach that leverages kinematic similarity for efficient transitions.
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.
