Pseudo-State Estimation for Fractional Order Neural Networks

This paper studies the pseudo-state estimation of fractional order neural networks based on the event-triggered mechanism. First, a novel Mittag–Leffler type event generator is designed and placed behind the sensor in order to save network bandwidth resources, which can be used to judge whether the measurement output of the sensor needs to be transmitted to the state estimator or not. Moreover, a new fractional order parameter Lyapunov functional is constructed. Based on the Mittag–Leffler type event trigger condition and fractional order Lyapunov functional method, a discriminant condition in the form of linear matrix inequalities for the pseudo-state estimation of fractional order neural networks is obtained. The effectiveness and feasibility of the method is illustrated by two numerical examples.