Secure state estimation for event-triggered cyber-physical systems against deception attacks. Issue 18 (December 2022)

• Record Type: Journal Article
• Authors: Han, Zhichen; Zhang, Shengbing; Jin, Zengwang; Hu, Yanyan
• Is Part Of: Journal of the Franklin Institute. Volume 359:Issue 18(2022)
• Page Start: 11155
• Page End: 11185
• Publication Date: 2022-12
• Subjects: Science -- Periodicals; Technology -- Periodicals; Patents -- United States -- Periodicals; 505
• DOI: 10.1016/j.jfranklin.2022.10.049
• Languages: English
• ISSNs: 0016-0032

Abstract

Abstract: Nowadays, cyber-physical systems (CPSs) have been widely used in various fields due to their powerful performance and low cost. The cyber attacks will cause security risks and even huge losses according to the universality and vulnerability of CPSs. As a typical network attack, deception attacks have the features of high concealment and strong destructiveness. Compared with the traditional deception attack models with a constant value, a deception attack with random characteristics is introduced in this paper, which is difficult to identify. In order to defend against such deception attacks and overcome energy constraints in CPSs, the secure state estimation and the event-triggered communication mechanism without feedback information are co-considered to reconcile accuracy of estimation and energy consumption. Firstly, an event-triggered augmented state estimator is proposed for secure state estimation and attack identification. Then, under the ideology of equivalence, the augmented state estimator is derived as a concise two-stage estimator with reduced order. The two-stage estimator can perform the secure state estimation and attack identification respectively. The estimators ensure the accuracy of attack identification well since not treating attack information as the trigger event. Afterward, the comparison of the computational complexity of these two algorithms is analyzed. Finally, an example of a target tracking system is supplied to prove the effectiveness and efficiency of the proposed algorithm.