CRIL: Continual Robot Imitation Learning via Generative and Prediction Model

TL;DR

This paper introduces CRIL, a continual imitation learning framework for robots that uses generative and predictive models to learn new tasks sequentially, reducing the need for multi-task demonstrations and improving learning efficiency.

Contribution

The paper proposes a novel trajectory generation model combining GANs and dynamics-aware prediction for continual imitation learning in robots.

Findings

Effective in simulation and real-world tasks

Reduces multi-task demonstration requirements

Accelerates learning of new tasks

Abstract

Imitation learning (IL) algorithms have shown promising results for robots to learn skills from expert demonstrations. However, they need multi-task demonstrations to be provided at once for acquiring diverse skills, which is difficult in real world. In this work we study how to realize continual imitation learning ability that empowers robots to continually learn new tasks one by one, thus reducing the burden of multi-task IL and accelerating the process of new task learning at the same time. We propose a novel trajectory generation model that employs both a generative adversarial network and a dynamics-aware prediction model to generate pseudo trajectories from all learned tasks in the new task learning process. Our experiments on both simulation and real-world manipulation tasks demonstrate the effectiveness of our method.