CaltechAUTHORS
  A Caltech Library Service

A Control Barrier Perspective on Episodic Learning via Projection-to-State Safety

Taylor, Andrew J. and Singletary, Andrew and Yue, Yisong and Ames, Aaron D. (2021) A Control Barrier Perspective on Episodic Learning via Projection-to-State Safety. IEEE Control Systems Letters, 5 (3). pp. 1019-1024. ISSN 2475-1456. https://resolver.caltech.edu/CaltechAUTHORS:20200707-104322686

[img] PDF - Submitted Version
See Usage Policy.

1530Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20200707-104322686

Abstract

In this letter we seek to quantify the ability of learning to improve safety guarantees endowed by Control Barrier Functions (CBFs). In particular, we investigate how model uncertainty in the time derivative of a CBF can be reduced via learning, and how this leads to stronger statements on the safe behavior of a system. To this end, we build upon the idea of Input-to-State Safety (ISSf) to define Projection-to-State Safety (PSSf), which characterizes degradation in safety in terms of a projected disturbance. This enables the direct quantification of both how learning can improve safety guarantees, and how bounds on learning error translate to bounds on degradation in safety. We demonstrate that a practical episodic learning approach can use PSSf to reduce uncertainty and improve safety guarantees in simulation and experimentally.


Item Type:Article
Related URLs:
URLURL TypeDescription
https://doi.org/10.1109/LCSYS.2020.3009082DOIArticle
https://arxiv.org/abs/2003.08028arXivDiscussion Paper
ORCID:
AuthorORCID
Taylor, Andrew J.0000-0002-5990-590X
Singletary, Andrew0000-0001-6635-4256
Yue, Yisong0000-0001-9127-1989
Ames, Aaron D.0000-0003-0848-3177
Additional Information:© 2020 IEEE. Manuscript received March 17, 2020; revised June 8, 2020; accepted June 29, 2020. Date of publication July 13, 2020; date of current version July 28, 2020. This work was supported by Defense Advanced Research Projects Agency under Award HR00111890035 and Award NNN12AA01C.
Funders:
Funding AgencyGrant Number
Defense Advanced Research Projects Agency (DARPA)HR00111890035
NASANNN12AA01C
Subject Keywords:Machine learning; Lyapunov methods; Uncertain system
Issue or Number:3
Record Number:CaltechAUTHORS:20200707-104322686
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20200707-104322686
Official Citation:A. J. Taylor, A. Singletary, Y. Yue and A. D. Ames, "A Control Barrier Perspective on Episodic Learning via Projection-to-State Safety," in IEEE Control Systems Letters, vol. 5, no. 3, pp. 1019-1024, July 2021, doi: 10.1109/LCSYS.2020.3009082
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:104244
Collection:CaltechAUTHORS
Deposited By: Tony Diaz
Deposited On:07 Jul 2020 17:46
Last Modified:03 Aug 2020 21:09

Repository Staff Only: item control page