Characterising the Robustness of Reinforcement Learning for Continuous Control using Disturbance Injection
Abstract
In this study, we leverage the deliberate and systematic fault-injection capabilities of an open-source benchmark suite to perform a series of experiments on state-of-the-art deep and robust reinforcement learning algorithms. We aim to benchmark robustness in the context of continuous action spaces -- crucial for deployment in robot control. We find that robustness is more prominent for action disturbances than it is for disturbances to observations and dynamics. We also observe that state-of-the-art approaches that are not explicitly designed to improve robustness perform at a level comparable to that achieved by those that are. Our study and results are intended to provide insight into the current state of safe and robust reinforcement learning and a foundation for the advancement of the field, in particular, for deployment in robotic systems.
Cite
@article{arxiv.2210.15199,
title = {Characterising the Robustness of Reinforcement Learning for Continuous Control using Disturbance Injection},
author = {Catherine R. Glossop and Jacopo Panerati and Amrit Krishnan and Zhaocong Yuan and Angela P. Schoellig},
journal= {arXiv preprint arXiv:2210.15199},
year = {2022}
}
Comments
18 pages, 15 figures