Secure Estimation and Control of Networked Systems under Cyber-attacks

Networked systems (NSs) are complex dynamical systems composed of a large number ofsimple systems interacting through a communication medium. These systems arise asnatural models in many areas of engineering and sciences, such as sensor networks andcyber-physical system. However, most of the existing results need continuous successfulcommunication, which bring high security requirements when under attacks. Also it isnot easy to realize real-time communication of complete information. In this project, themain objective is to develop secure estimation and control algorithms that can be usedby a NS to carry out prescribed tasks efficiently under cyber-attacks.Reliable attack detection scheme is one of the key factors for NSs to achieve desiredsecure performance. Incorrect or incomplete information caused by cyber-attacks willdegrade the system performance seriously. It is necessary to design compensationstrategy to minimize the information loss. Adaptive distributed fusion of attack detectionwill be studied in this proposal to deal with three kinds of cyber-attacks: Replayattacks, Denial of service attacks and false data injection attacks. These kinds of attacksoccuring in NSs often exhibit different dynamical characteristics, and can be found inmany cyber-physical systems.It is highly desirable to design effective information recovery schemes and event-basedswitching methods to improve security performance and enhance reliability androbustness against attacks. In particular, the hidden feature of attacks does bring somedifficulties in designing reliable distributed fusion detection schemes for cyber-attacks,and thus indirectly affects the design of optimal compensation strategy to reduceinformation loss. Cooperative design of attack detection and estimation/control will bestudied in this proposal to guarantee the secure estimation performance.This proposal has set up three new, challenging yet achievable objectives to thetheoretical investigation. The results would have both theoretical and practicalsignificance in communication-based network systems, in which actions must betransmitted to ensure a secure estimation and control performance. As an applicationexample of NSs, these adaptive attack detection, secure estimation and control schemeswill bring great benefit to the monitoring and control of power grid. Hence, it willimprove the performance, reliability and applicability. Our results are expected to providea fundamental secure estimation and control investigation for NSs under cyber-attacks.