State Estimation for Discrete Neural Networks with Randomly Occurring Uncertainties and Missing Measurements

In this paper, the state estimation problem is investigated for a class of discrete uncertain neural networks subject to time delays and missing measurements. Several mutually independent sets of Bernoulli-distributed white sequences are employed to describe the phenomena of randomly occurring uncertainties (ROUs) and missing measurements. We aim to design a state estimator such that, in the presence of all admissible parameter variations, the dynamic of the estimation error is asymptotically stable and satisfies the performance constraint. By adopting the Lyapunov-Krasovakii functional and the stochastic analysis theory, sufficient conditions are established to ensure the existence of the desired state estimator. The explicit expression of such estimators is parameterized by solving a linear matrix inequality (LMI) problem. A numerical simulation example is provided to verify the usefulness of the proposed approach.