Learning Dexterous Bimanual Catch Skills through Adversarial-Cooperative Heterogeneous-Agent Reinforcement Learning

TL;DR

This paper presents a novel reinforcement learning framework enabling robots to learn dexterous bimanual catching skills by training heterogeneous agents with an adversarial reward scheme, improving coordination and robustness across diverse objects.

Contribution

It introduces a new HARL-based approach with adversarial rewards for bimanual catching, advancing robotic dexterity and coordination in complex object handling.

Findings

Approximately 2x increase in catching reward over baselines

Robustness demonstrated across 15 diverse objects

Effective coordination learned in simulated environments

Abstract

Robotic catching has traditionally focused on single-handed systems, which are limited in their ability to handle larger or more complex objects. In contrast, bimanual catching offers significant potential for improved dexterity and object handling but introduces new challenges in coordination and control. In this paper, we propose a novel framework for learning dexterous bimanual catching skills using Heterogeneous-Agent Reinforcement Learning (HARL). Our approach introduces an adversarial reward scheme, where a throw agent increases the difficulty of throws-adjusting speed-while a catch agent learns to coordinate both hands to catch objects under these evolving conditions. We evaluate the framework in simulated environments using 15 different objects, demonstrating robustness and versatility in handling diverse objects. Our method achieved approximately a 2x increase in catching reward compared to single-agent baselines across 15 diverse objects.