Distributed State Estimation for Large-Scale Systems in the Presence of Data Packet Drops

Abstract

This article focuses on the distributed state estimation problem for large-scale systems in the presence of data packet drops. The large-scale system is structured in several correlated subsystems in the physical space, and each subsystem only communicates with its neighbors. In particular, the states of different subsystems are measured by different sensors, and the sensor broadcasts measurement information to the subestimator and its neighbors through the lossy communication channel. Thus, subestimators obtain different local information in the presence of data packet drops. In this article, the distributed estimator is designed and the optimal gain is obtained under the minimum mean square error (MMSE) estimation criterion by using local information set. Finally, the effectiveness of the distributed estimator is illustrated by a simulation experiment.