Distributed Filtering With Dynamic Event-Triggered Communication for Networked Periodic Systems

Abstract

This article deals with the problem of distributed filtering of periodic systems over sensor networks under event-triggered communications, based on a piecewise impulsive method. The continuous-time periodic system is approximated by several linear time-invariant subsystems through a piecewise constant method. The sensor network deployed to the periodic system is taken into account for the actual needs of monitoring or estimation. A dynamic event-triggering mechanism is designed to reduce the unnecessary data transmission over the wireless network from the sensor to the filter as well as between each filter. Distributed peak-to-peak filters with periodic time-varying parameters are designed under the event-triggering mechanism to estimate the networked continuous-time periodic systems with the bounded and local differentiable disturbance. Then, a periodic piecewise filtering error system is modeled as an impulsive periodic piecewise system, for which sufficient conditions of a guaranteed peak-to-peak performance are proposed. The gains of the filters are obtained by the method of convex linearization. The validity of the proposed results is demonstrated by a pendulum tracking system.