Learning Multimodal Contact-Rich Skills from Demonstrations Without Reward Engineering

TL;DR

This paper introduces a model-free, demonstration-based learning framework enabling robots to acquire contact-rich skills using multimodal sensors without reward engineering, demonstrated on real robots with high success rates.

Contribution

A novel multimodal data representation and a generalizable learning framework that eliminates the need for explicit reward functions in contact-rich skill acquisition.

Findings

Achieved 100% success in writing and peeling tasks.

Achieved 80% success in cleaning task.

Validated on real Sawyer robot with three skills.

Abstract

Everyday contact-rich tasks, such as peeling, cleaning, and writing, demand multimodal perception for effective and precise task execution. However, these present a novel challenge to robots as they lack the ability to combine these multimodal stimuli for performing contact-rich tasks. Learning-based methods have attempted to model multi-modal contact-rich tasks, but they often require extensive training examples and task-specific reward functions which limits their practicality and scope. Hence, we propose a generalizable model-free learning-from-demonstration framework for robots to learn contact-rich skills without explicit reward engineering. We present a novel multi-modal sensor data representation which improves the learning performance for contact-rich skills. We performed training and experiments using the real-life Sawyer robot for three everyday contact-rich skills -- cleaning, writing, and peeling. Notably, the framework achieves a success rate of 100\% for the peeling and writing skill, and 80\% for the cleaning skill. Hence, this skill learning framework can be extended for learning other physical manipulation skills.