Secure Control for Switched Nonlinear Systems With DoS Attacks: A Switching Event-Triggered Adaptive Output-Feedback Control Method
Hongzhen Xie, Guangdeng Zong, Dong Yang, Ying Guo, Xudong Zhao
Abstract
The security problem for a kind of switched nonlinear systems with denial-of-service (DoS) attacks is addressed here with a novel event-triggered neural network (NN) adaptive control technique. The provided event-triggered control algorithm, compared to the current output feedback control schemes on continuous-time systems, not only conserves communication resources but also defends against DoS attacks throughout the system stabilization process. First, a novel adaptive NN switching observer is proposed by examining the stability of each subsystem in the sleep and active intervals of the attack, which makes the standard backstepping controller workable even in the DoS attacks scenario. Further, an event-triggered secure control strategy is established by creating a common coordinate transformation, which ensures the bound of all the signals in the resultant system under arbitrary switching and prevents the occurrence of Zeno behavior. A first-order filter is developed to fix the complexity explosion problem in backstepping design. Finally, a continuous stirred tank reactor example is presented to support the established method.