A Framework for Joint Attack Detection and Control Under False Data Injection

  • Luyao Niu
  • , Andrew Clark

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

In this work, we consider an LTI system with a Kalman filter, detector, and Linear Quadratic Gaussian (LQG) controller under false data injection attack. The interaction between the controller and adversary is captured by a Stackelberg game, in which the controller is the leader and the adversary is the follower. We propose a framework under which the system chooses time-varying detection thresholds to reduce the effectiveness of the attack and enhance the control performance. We model the impact of the detector as a switching signal, resulting in a switched linear system. A closed form solution for the optimal attack is first computed using the proposed framework, as the best response to any detection threshold. We then present a convex program to compute the optimal detection threshold. Our approach is evaluated using a numerical case study.

Original languageEnglish
Title of host publicationDecision and Game Theory for Security - 10th International Conference, GameSec 2019, Proceedings
EditorsTansu Alpcan, Yevgeniy Vorobeychik, John S. Baras, György Dán
PublisherSpringer
Pages352-363
Number of pages12
ISBN (Print)9783030324292
DOIs
StatePublished - 2019
Event10th International Conference on Decision and Game Theory for Security, GameSec 2019 - Stockholm, Sweden
Duration: Oct 30 2019Nov 1 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11836 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Decision and Game Theory for Security, GameSec 2019
Country/TerritorySweden
CityStockholm
Period10/30/1911/1/19

Keywords

  • Control system
  • Detection threshold
  • False data injection attacks
  • K-L divergence
  • LQG control
  • Stealthiness

Fingerprint

Dive into the research topics of 'A Framework for Joint Attack Detection and Control Under False Data Injection'. Together they form a unique fingerprint.

Cite this