Distributed Estimation for Discrete-Time Interconnected Systems

This paper is concerned with the distributed estimation problem for discrete-time large-scale interconnected systems. A novel decoupling strategy is developed to deal with the interconnected terms in large-scale systems. Under the fully distributed estimation structure, the local estimator gain for each subsystem is designed by solving a convex optimization problem that is constructed based on each subsystem's own information and its neighboring information. Notice that the solutions to the convex optimization problems can be easily obtained by the standard software packages, and the computational complexity of each optimization problem is low even though the scale of interconnected systems is large. Moreover, stability conditions, which are determined by the rank of each measurement matrix, are derived such that the designed distributed estimator is asymptotically bounded. Finally, an illustrative example is employed to show the effectiveness of the proposed methods.