Learning Multimodal Contact-Rich Skills from Demonstrations Without Reward Engineering

TL;DR

This work proposes a generalizable model-free learning-from-demonstration framework for robots to learn contact-rich skills without explicit reward engineering and presents a novel multi-modal sensor data representation which improves the learning performance for contact- rich 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.

Figures (8)

• Figure 1: The task setup with human demonstrations to the robot for cleaning
• Figure 2: The Learning from Demonstration (LfD) Framework
• Figure 3: The skill learning framework
• Figure 4: The skill learning framework
• Figure 5: Execution Controller Model. Here $Q$ is the controller input and gain parameters, $\tau_m$ is the joint torques after clamping to motor limits.