H∞ state estimation for multi-rate artificial neural networks with integral measurements: A switched system approach

In this paper, the H-infinity state estimation problem is studied for a class of multi-rate artificial neural networks with integral measurements. A novel method, rather than the widely used lifting technique, is proposed to transform the multi-rate artificial neural networks to single-rate switched ones. The purpose of the addressed H-infinity state estimation problem is to design an estimator such that the estimation error dynamics is exponentially stable and the H-infinity performance requirement is satisfied. First, with the help of the Lyapunov-Krasovskii functional and the switched system approach, sufficient conditions are derived under which the existence of the desired estimator is ensured. Then, the characterization of the estimator gains is realized by solving certain linear matrix inequalities. Finally, two illustrative examples are given that confirm the usefulness of the developed H-infinity state estimation scheme and reveal the influence of the multi-rate sampling on the state estimation performance. (C) 2020 Elsevier Inc. All rights reserved.