Neural Simplex Architecture

Dung T. Phan, Radu Grosu, Nils Jansen, Nicola Paoletti, Scott A. Smolka, Scott D. Stoller

Research output: Chapter in Book/Report/Conference proceedingConference contribution

57 Downloads (Pure)

Abstract

We present the Neural Simplex Architecture (NSA), a new approach to runtime assurance that provides safety guarantees for neural controllers (obtained e.g. using reinforcement learning) of autonomous and other complex systems without unduly sacrificing performance. NSA is inspired by the Simplex control architecture of Sha et al., but with some significant differences. In the traditional approach, the advanced controller (AC) is treated as a black box; when the decision module switches control to the baseline controller (BC), the BC remains in control forever. There is relatively little work on switching control back to the AC, and there are no techniques for correcting the AC’s behavior after it generates a potentially unsafe control input that causes a failover to the BC. Our NSA addresses both of these limitations. NSA not only provides safety assurances in the presence of a possibly unsafe neural controller, but can also improve the safety of such a controller in an online setting via retraining, without overly degrading its performance. To demonstrate NSA’s benefits, we have conducted several significant case studies in the continuous control domain. These include a target-seeking ground rover navigating an obstacle field, and a neural controller for an artificial pancreas system.
Original languageEnglish
Title of host publicationNASA Formal Methods - 12th International Symposium, NFM 2020, Proceedings
EditorsRitchie Lee, Susmit Jha, Anastasia Mavridou
PublisherSpringer
Pages97-114
Number of pages18
ISBN (Electronic)978-3-030-55754-6
ISBN (Print)978-3-030-55753-9
DOIs
Publication statusPublished - 2020
Event12th International Symposium on NASA Formal Methods, NFM 2020 - Moffett Field, United States
Duration: 11 May 202015 May 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12229 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Symposium on NASA Formal Methods, NFM 2020
Country/TerritoryUnited States
CityMoffett Field
Period11/05/2015/05/20

Keywords

  • Online retraining
  • Reverse switching
  • Runtime assurance
  • Safe reinforcement learning
  • Simplex architecture

Cite this