Simulated Annealing-Driven Event-Triggered Neural Sliding Mode Control for Networked Nonlinear Markov Jump Systems

This paper presents the design of an event-triggered state estimator for neural sliding mode control (SMC) in networked nonlinear Markov jump systems (MJSs) with incomplete mode information. To improve communication efficiency, a simulated annealing-based event-triggering mechanism is introduced for observer design over the network. This mechanism is enhanced by a neural-based adaptive compensator that effectively addresses unknown nonlinearities. An integral sliding surface is then constructed in the state estimation space, serving as the foundation for deriving the sliding mode dynamics and ensuring robustness to uncertainties. In light of uncertain transition rates (TRs), a unified sliding mode controller is developed to accommodate various mode types, ensuring both the reachability condition and the maintenance of sliding motion. The stochastic stability of the system is analyzed for each transition rate scenario. Finally, simulation results are provided to validate the effectiveness and performance of the proposed approach.