Twisting Lids Off with Two Hands

TL;DR

This paper presents a novel deep reinforcement learning approach for bimanual manipulation, specifically twisting lids on various objects, demonstrating effective sim-to-real transfer and generalization to unseen objects.

Contribution

It introduces the first sim-to-real RL system for bimanual multi-fingered hands capable of complex tasks like lid twisting with generalization.

Findings

Successful transfer of policies from simulation to real-world robots.

Generalization across diverse unseen objects.

Dynamic and dexterous manipulation behaviors achieved.

Abstract

Manipulating objects with two multi-fingered hands has been a long-standing challenge in robotics, due to the contact-rich nature of many manipulation tasks and the complexity inherent in coordinating a high-dimensional bimanual system. In this work, we share novel insights into physical modeling, real-time perception, and reward design that enable policies trained in simulation using deep reinforcement learning (RL) to be effectively and efficiently transferred to the real world. Specifically, we consider the problem of twisting lids of various bottle-like objects with two hands, demonstrating policies with generalization capabilities across a diverse set of unseen objects as well as dynamic and dexterous behaviors. To the best of our knowledge, this is the first sim-to-real RL system that enables such capabilities on bimanual multi-fingered hands.