H∞ performance state estimation of delayed static neural networks based on an improved proportional-integral estimator

Abstract

In this paper, an improved proportional-integral (PI) estimator is presented to analyze the problem of H∞ performance state estimation of static neural networks with disturbance. An exponential gain term is added to the PI estimator, which leads to the convenience of analysis and design. In order to guarantee the H∞ performance state estimation, a less conservative delay-dependent criterion is derived by using an improved reciprocally convex inequality. Finally, simulation results are given to verify the advantage of the presented approach.