Plug-and-Play Design for Linear Distributed Observers

Abstract

The distributed state estimation problem on linear time-invariant systems is addressed in this paper. We consider a scenario where the system outputs are measured via a group of sensors distributed over multiple nodes, where the joint measurements guarantee the detectability of the system. Each node is equipped with an observer exchanging its own state estimates with the neighboring nodes over a communication network, and the objective is to estimate the entire state of the system by each observer. Compared to existing designs requiring information on global graph topology or sensor placement at other nodes, our design is based only on local information. Hence, the proposed scheme is plug-and-play and scalable so that adding and removing nodes to the network or changing sensors at some existing nodes does not require observer redesign at any other node. Simulation results demonstrate the effectiveness of the proposed distributed estimation scheme.