Exploring the linkage between partition-based distributed extended Kalman filter and distributed full-information estimation

In this paper, we revisit a partition-based distributed extended Kalman filter (DEKF) method proposed in [1] for continuous-time nonlinear systems. Our objective is to offer a comprehensive perspective on the development of this DEKF method, elucidating its relationship with partition-based distributed full-information estimation within a discrete-time linear framework. Specifically, we present a partition-based distributed full-information estimation formulation for discrete-time linear systems. We derive an analytical solution for this full-information estimation problem which is in the form of a partition-based distributed Kalman filter (DKF). The DKF approach is extended to address nonlinear systems through successive linearization of nonlinear subsystem models, and a discrete-time distributed extended Kalman filter (DEKF) approach is derived. We compare the derived discrete-time DEKF with the continuous-time DEKF approach in [1] to reveal the connection between the DEKF approach in [1] and the distributed full-information estimation in a discrete-time context. A simulated process is utilized to verify the effectiveness and assess the performance of the distributed extended Kalman filter.