Enhancing network security in constrained switched systems: event-triggered MPC with a buffer-based anti-attack approach

Abstract

This paper explores event-triggered model predictive control (MPC) for constrained switched systems, incorporating a buffer-based anti-attack (BBAA) approach to ensure network security. Since the system state is unknown and disturbed, an observer-based $H_{\mathrm{\infty }}$ MPC approach is used to estimate the system state and suppress disturbance. Furthermore, the estimated state and control input can be optimised by minimising the performance function. Considering that the system state is subject to denial of service (DoS) attacks in the feedback channel, the BBAA approach is used to reconstruct the estimated state, thereby reducing the impact of attacks on system performance. To improve the utilisation of network resources, a set of event-triggering mechanisms (ETMs) is used. In addition, a set of matrix inequality conditions is given to constrain the gains of controllers and observers. Then, the system $H_{\mathrm{\infty }}$ performance is demonstrated by applying the average dwell time (ADT) and the Lyapunov function method, and the complex coupling between model switching, DoS attacks and event-triggering is analysed. Simulation results validate the effectiveness of the proposed approach.

Keywords:

• Switched systems
• model predictive control
• DoS attacks
• buffer-based anti-attack approach

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [grant number 61873172]; the Science Center for Gas Turbine Project [grant number P2022-B-V-001-001].