Abstract
For nonlinear cyber-physical systems (CPSs) with input saturation and denial-of-service (DoS) attacks, an adaptive secure control strategy is investigated in this paper. To estimate the unmeasurable states, a novel switched neural network (NN) observer is designed, where two sub-observers can be freely switched from each other. Considering unnecessary packet transmissions, an improved event-triggered mechanism (ETM) is introduced to reduce the communication burden in the controller-to-actuator channel. In addition, in order to simultaneously solve the input saturation and unknown control direction problems, a Nussbaum-type function is introduced to the control design process. Finally, it is proven that all closed-loop signals are the semiglobal uniformly ultimately bounded, and simulation results demonstrate the effectiveness of the proposed secure control scheme.
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.