Distributed Observer Analysis and Design

A Distributed observer design is described for estimating the state of a continuous-time, input free, linear system. This thesis explains how to construct the local estimators, which comprise the observer inputs and outputs, and it is shown which are the requirements to deal with this structure. Every agent senses an output signal from the system and distributes it across a fixed-time network to its neighbors. The information flow increases the capability of each agent to estimate the state of the system and uses collaboration to improve the quality of data. The proposed solution has several positive features compared to recent results in the literature, which include milder assumptions on the network connectivity and the maximum dimension of the state of each observer does not exceed the order of the plant. The conditions are reduced to certain detectability requirements for each cluster of agents in the network, where a cluster is identified as a subset of agents that satisfy specific properties. Instead, the dimension of each observer is reduced to the number of possible observable states of the system, collected by the agent and by the neighbors.