Neural-network observer-based resilient finite-time H∞ asynchronous control with dual adaptive triggered protocols for singular jump systems under aperiodic DoS attacks

This work investigates the neural-network (NN) observer-based resilient finite-time H-infinity asynchronous control for a class of nonlinear discrete singular Markov jump systems (SMJSs) under the simultaneous presence of dual adaptive event-triggered protocols (AETPs) and aperiodic denial-of-service (DoS) attacks. Taking advantage of the inherent approximation capability of NN technique, an improved NN observer is devised for precisely observing the unknown states. Novel dual AETPs are introduced in sensor-to-observer (S-O) channel and controller-to-actuator (C-A) channel, respectively, with the aim of reducing consumption of communication resources and mitigating stress on network bandwidth. Then, based on hidden Markov model (HMM) mechanism, NN observer and resilient asynchronous controller are designed collaboratively under aperiodic DoS attacks with frequency and duration limits. By exploiting attack-mode dependent Lyapunov-Krasovskii (L-K) functionals, singular value decomposition method and iterative technique, fresh conditions of the regularity, causality and finite-time boundedness with desired H-infinity performance index for closed-loop discrete SMJSs are established under the framework of linear matrix inequalities (LMIs). Lastly, the oil catalytic cracking process (OCCP) serves as a means to demonstrate the validity and practicality of the offered strategies.