Visible to the public Adversarial Regression for Detecting Attacks in Cyber-Physical SystemsConflict Detection Enabled

TitleAdversarial Regression for Detecting Attacks in Cyber-Physical Systems
Publication TypeConference Proceedings
Year of Publication2018
AuthorsAmin Ghafouri, Xenofon Koutsoukos, Yevgeniy Vorobeychik
Conference NameTwenty-Seventh International Joint Conference on Artificial Intelligence
Conference LocationStokholm, Sweden
KeywordsRobust monitoring diagnosis and network control, Vanderbilt
Abstract

Attacks in cyber-physical systems (CPS) which manipulate sensor readings can cause enormous physical damage if undetected. Detection of attacks on sensors is crucial to mitigate this issue. We study supervised regression as a means to detect anoma- lous sensor readings, where each sensor's measure- ment is predicted as a function of other sensors. We show that several common learning approaches in this context are still vulnerable to stealthy at- tacks, which carefully modify readings of compro- mised sensors to cause desired damage while re- maining undetected. Next, we model the interac- tion between the CPS defender and attacker as a Stackelberg game in which the defender chooses detection thresholds, while the attacker deploys a stealthy attack in response. We present a heuris- tic algorithm for finding an approximately optimal threshold for the defender in this game, and show that it increases system resilience to attacks without significantly increasing the false alarm rate.

URLhttp://www.vuse.vanderbilt.edu/~koutsoxd/www/Publications/0524.pdf
Citation Keynode-60981