Detection of unmodelled anomalous behaviours via localized controllers for condition systems

Jeff Ashley, Lawrence Holloway

Research output: Contribution to journalArticlepeer-review

Abstract

A method for detecting anomalous behaviours is presented for plants represented as condition systems. An anomalous behaviour can be viewed as observed behaviour of a real system (i.e. a robot, manufacturing cell, etc.) that cannot be accounted for by the set of plant models representing the system, even allowing for fault events. This work builds upon previous modular control synthesis techniques for condition systems.

Original languageEnglish
Pages (from-to)5289-5312
Number of pages24
JournalInternational Journal of Production Research
Volume46
Issue number19
DOIs
StatePublished - Oct 2008

Bibliographical note

Funding Information:
This work has been supported in part by NSF Grant ECS-0115694, the Office of Naval Research under Grant N000140110621, and the Center for Manufacturing at the University of Kentucky. J.A. is Research Scientist, Center for Manufacturing, University of Kentucky. L.H. is Professor of Electrical Engineering and Director of the Center for Manufacturing.

Keywords

  • Condition systems
  • Controller synthesis
  • Discrete event systems
  • Fault detection

ASJC Scopus subject areas

  • Strategy and Management
  • Management Science and Operations Research
  • Industrial and Manufacturing Engineering

Fingerprint

Dive into the research topics of 'Detection of unmodelled anomalous behaviours via localized controllers for condition systems'. Together they form a unique fingerprint.

Cite this