Design of H∞ state estimator for delayed static neural networks under hybrid-triggered control and imperfect measurement strategy

Abstract

This paper address the problem of designing H∞ state estimator for a class of delayed static neural networks (DSNNs) under the hybrid-triggered scheme (HTS) and imperfect measurement strategy. Firstly, for reducing the redundancy of signal transmission together with saving the network resource, a HTS is introduced in DSNNs, where the HTS characterized by Bernoulli distribution and takes into account the inevitable network-induced delay and packet dropout phenomenon simultaneously. Then, without loss of generality, we extend the results to a more general dissipative property, which is proposed for the first time, and which contains strict passivity except for the traditional four cases. Further, by considering the HTS, a modified delay-product-type Lyapunov functional is proposed. Accordingly, sufficient conditions for global asymptotically stable with a prescribed level γ of augmented system are obtained in the light of a set of LMIs, and the explicit expression of expected estimator is given. Finally, the feasibility and availability of the results of development are verified by three cases as example.