Self-Supervised Online Learning for Safety-Critical Control using Stereo Vision
Abstract
With the increasing prevalence of complex vision-based sensing methods for use in obstacle identification and state estimation, characterizing environment-dependent measurement errors has become a difficult and essential part of modern robotics. This paper presents a self-supervised learning approach to safety-critical control. In particular, the uncertainty associated with stereo vision is estimated, and adapted online to new visual environments, wherein this estimate is leveraged in a safety-critical controller in a robust fashion. To this end, we propose an algorithm that exploits the structure of stereo-vision to learn an uncertainty estimate without the need for ground-truth data. We then robustify existing Control Barrier Function-based controllers to provide safety in the presence of this uncertainty estimate. We demonstrate the efficacy of our method on a quadrupedal robot in a variety of environments. When not using our method safety is violated. With offline training alone we observe the robot is safe, but overly-conservative. With our online method the quadruped remains safe and conservatism is reduced.
Cite
@article{arxiv.2203.01404,
title = {Self-Supervised Online Learning for Safety-Critical Control using Stereo Vision},
author = {Ryan K. Cosner and Ivan D. Jimenez Rodriguez and Tamas G. Molnar and Wyatt Ubellacker and Yisong Yue and Aaron D. Ames and Katherine L. Bouman},
journal= {arXiv preprint arXiv:2203.01404},
year = {2022}
}
Comments
7 pages, 4 figures, conference publication at ICRA 2022