H∞ State Estimation of Discrete Time Delayed Neural Networks with Multiple Missing Measurements Using Second Order Reciprocal Convex Approach

This paper focuses on H-infinity state estimation for a general class of discrete-time nonlinear systems of the neural network with time-varying delays and multiple missing measurements which is described by the unified model. The H-infinity performance for the systems described by the unified model is analyzed by using sector decomposition approach together with the Lyapunov stability theory. By constructing triple Lyapunov-Krasovskii functional, a new sufficient condition is established to ensure the asymptotic mean square stability of discrete-time delayed neural networks. Second order convex reciprocal technique is incorporated to deal with partitioned double summation terms and the conservatism of conditions for the state estimator synthesis is reduced efficiently. Finally, a numerical example is given to demonstrate the effectiveness of the proposed design method.