Learning to Coordinate Manipulation Skills via Skill Behavior Diversification

Youngwoon Lee, Jingyun Yang, Joseph J. Lim

TL;DR: We propose to tackle complex tasks of multiple agents by learning composable primitive skills and coordination of the skills.

Abstract: When mastering a complex manipulation task, humans often decompose the task into sub-skills of their body parts, practice the sub-skills independently, and then execute the sub-skills together. Similarly, a robot with multiple end-effectors can perform complex tasks by coordinating sub-skills of each end-effector. To realize temporal and behavioral coordination of skills, we propose a modular framework that first individually trains sub-skills of each end-effector with skill behavior diversification, and then learns to coordinate end-effectors using diverse behaviors of the skills. We demonstrate that our proposed framework is able to efficiently coordinate skills to solve challenging collaborative control tasks such as picking up a long bar, placing a block inside a container while pushing the container with two robot arms, and pushing a box with two ant agents. Videos and code are available at https://clvrai.com/coordination

Keywords: reinforcement learning, hierarchical reinforcement learning, modular framework, skill coordination, bimanual manipulation

Code: https://github.com/clvrai/coordination