eduzhai > Applied Sciences > Engineering >

Privacy-Preserving Resilience of Cyber-Physical Systems to Adversaries

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: A cyber-physical system (CPS) is expected to be resilient to more than onetype of adversary. In this paper, we consider a CPS that has to satisfy alinear temporal logic (LTL) objective in the presence of two kinds ofadversaries. The first adversary has the ability to tamper with inputs to theCPS to influence satisfaction of the LTL objective. The interaction of the CPSwith this adversary is modeled as a stochastic game. We synthesize a controllerfor the CPS to maximize the probability of satisfying the LTL objective underany policy of this adversary. The second adversary is an eavesdropper who canobserve labeled trajectories of the CPS generated from the previous step. Itcould then use this information to launch other kinds of attacks. A labeledtrajectory is a sequence of labels, where a label is associated to a state andis linked to the satisfaction of the LTL objective at that state. We usedifferential privacy to quantify the indistinguishability between states thatare related to each other when the eavesdropper sees a labeled trajectory. Twotrajectories of equal length will be differentially private if they aredifferentially private at each state along the respective trajectories. We usea skewed Kantorovich metric to compute distances between probabilitydistributions over states resulting from actions chosen according to policiesfrom related states in order to quantify differential privacy. Moreover, we dothis in a manner that does not affect the satisfaction probability of the LTLobjective. We validate our approach on a simulation of a UAV that has tosatisfy an LTL objective in an adversarial environment.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...