Nonfragile extended dissipativity state estimator design for discrete-time neural networks with time-varying delay

Abstract

This paper studies the issue of nonfragile extended dissipativity state estimator design for discrete-time neural networks with time-varying delay. Firstly, a class of nonfragile proportional-integral state estimator with exponential type linear polynomial on time-varying delay is presented, which unifies the Arcak state estimator and the Luenberger state estimator as its special cases. Since the exponential type linear polynomial on time-varying delay is considered, more information of time delay can be utilized and the estimation error convergence rate can be adjusted. Secondly, an augmented Lyapunov–Krasovskii functional tailored for delayed discrete-time neural networks is presented, in which the double summation terms of the state vector and the information of the output estimation error are fully considered in the corresponding augmented vector, then a sufficient criterion that guarantees the nonfragile extended dissipativity state estimation for delayed discrete-time neural networks is derived. Furthermore, since an extended dissipativity performance level is introduced, the issues of passitivity state estimation, H∞ state estimation, L2-L∞ state estimation, and (Q,S,R)-γ dissipativity state estimation can be solved in a unified framework. Finally, simulation results are given to demonstrate the advantage of the presented method.