Event-Triggered Distributed Fusion Estimation for Clustered Sensor Networks With Multiscale Fading and Multirate Sampling Characteristics

This article addresses the distributed fusion estimation problem of clustered sensor networks with multirate sampling and multiscale fading. An event-triggered distributed two-stage fusion estimation method is proposed for the clustered sensor network. Specifically, a local fusion estimator is first constructed to fuse multirate measurements transmitted through small-scale fading channels. The parameters of the estimator are computed by solving linear matrix inequalities (LMIs). Then, the event-triggered mechanism (ETM) is utilized to adjust the transmission frequency of local estimates, aiming to conserve the energy of cluster heads. Furthermore, by solving a min–max optimization problem, a global fusion estimator is designed to fuse the local estimates in the presence of the ETM and large-scale fading channels. Sufficient conditions for the stochastic ultimate boundedness of both local and global estimation errors are derived by leveraging differential game theory and LMIs. The comparison simulation results on a dc-power network show the effectiveness and advantages of the proposed method.