Robust distributed observer for uncertain systems

Abstract

This paper proposes a robust distributed observer for uncertain systems where some unknown, unstructured, and bounded uncertainties exist in system matrices. In distributed observer, each node includes an observer which directly estimates a part of the system states by its limited measurements and further, the estimation of other states indirectly obtains via exchanging information with neighbouring nodes. It is shown that whole states convergence achieves if the LTI system is observable as well as the network graph is strongly connected. Also, in the proposed distributed observer, by applying a new term in dynamic equation, the robust performance is achieved in dealing with uncertainties in system matrices. The observer gains synthesis is then formulated in terms of linear matrix inequalities. Finally, the simulation results are presented to illustrate the effectiveness of the proposed distributed observer to handle system uncertainties.