Improving the Robustness of Reinforcement Learning Policies with $\mathcal{L}_{1}$ Adaptive Control

TL;DR

This paper introduces an add-on $ ext{L}_1$ adaptive controller to enhance the robustness of pre-trained reinforcement learning policies against dynamic variations in new or perturbed environments, validated through numerical and real-world experiments.

Contribution

The paper proposes a novel method combining RL with an $ ext{L}_1$ adaptive controller to improve robustness without retraining in varied environments.

Findings

Improved robustness of RL policies in perturbed environments.

Effective in both model-free and model-based RL methods.

Validated through numerical and real-world experiments.

Abstract

A reinforcement learning (RL) control policy could fail in a new/perturbed environment that is different from the training environment, due to the presence of dynamic variations. For controlling systems with continuous state and action spaces, we propose an add-on approach to robustifying a pre-trained RL policy by augmenting it with an $\mathcal{L}_{1}$ adaptive controller ($\mathcal{L}_{1}$AC). Leveraging the capability of an $\mathcal{L}_{1}$AC for fast estimation and active compensation of dynamic variations, the proposed approach can improve the robustness of an RL policy which is trained either in a simulator or in the real world without consideration of a broad class of dynamic variations. Numerical and real-world experiments empirically demonstrate the efficacy of the proposed approach in robustifying RL policies trained using both model-free and model-based methods.