Self-triggered distributed filtering for sensor networks with topology switching via a multi-step ahead approach

In this paper, self-triggered filtering is developed for a class of distributed filtering networks (DFNs) with topology switching. To save unnecessary transmission and computation effort, a self-triggering policy (STP) is developed to avoid monitoring and judgments at every sampling moment, and it simultaneously decides when to get measurements and release the latest estimation to neighbor nodes. Specifically, the STP can predict the next trigger interval according to the local information and the neighbors' latest data at the triggering moment. However, under STP, a DFN will subject to asynchronous communication due to each filter triggering independently. Moreover, STP and DFN topology switching do contaminate the estimation accuracy. To deal with these challenges, a co-design of STP and distributed filters are developed to achieve good filtering performance with less design conservatism via a multi-step ahead approach. Finally, a numerical example and a smart grid distribution system are presented to demonstrate the effectiveness and applicability of the developed approach. (c) 2023 The Franklin Institute. Published by Elsevier Inc. All rights reserved.