Dynamic event-triggered security control for networked control systems with cyber-attacks : a model predictive control approach

Abstract

This article studies a dynamic event-triggered security control problem for networked control systems subject to deception attacks and packet dropouts. First, a combined cyberattack model is proposed, which utilises two sets of independent stochastic sequences to reflect randomly occurring cyber-attacks. Subsequently, a dynamic event-triggered protocol is constructed to relieve the restricted bandwidth pressure by reducing the data transmission of the communication channel from the plant to the controller. With the consideration of randomly occurring deception attacks, packet dropouts, and dynamic event-triggered protocols, an online model predictive control algorithm is established to ensure the stochastic stability of the closed-loop model with expected H2=H1 performance. Finally, two examples are simulated to interpret the validity and effectiveness of the proposed design strategy.