A Dynamic Event-Triggered Approach to State Estimation for Switched Memristive Neural Networks With Nonhomogeneous Sojourn Probabilities
Jun Cheng, Lidan Liang, Ju H. Park, Huaicheng Yan, Kezan Li
Abstract
This paper investigates the state estimation for switched memristive neural networks with nonhomogeneous sojourn probabilities. Essentially different from most current literature, a novel switching law is developed to depict the dynamic behavior of switched memristive neural networks, in which the sojourn probabilities of each subsystem are assumed to be nonhomogeneous, and a higher-level deterministic switching signal is proposed to regulate proper feedback switching information by means of the average dwell time approach. Meanwhile, to alleviate the constraint network bandwidth resource efficiently, a dynamic event-triggered mechanism with a novel threshold parameter is proposed in determining if the current data should be released or not. By resorting to the Lyapunov functional technique and the stochastic analysis strategy, some sufficient conditions are addressed to ensure the stochastic stability of the augmented switched memristive neural networks. In the end, the effectiveness and superiority of the developed results are verified by a numerical example.