Distributed zonotopic estimation for interconnected systems: A fusing overlapping states strategy

This paper is concerned with the distributed estimation problem for a class of interconnected systems with overlapping states. Under local communication, a distributed zonotopic estimator is designed by using the information from each subsystem and its neighbors. However, the obtained state estimates are also overlapping and contain redundant information that can be shared to further improve estimation performance for subsystems. To better utilize such overlapping estimates from neighbors, zonotopic fusion criteria are proposed to combine local estimates into a better result. Notice that the proposed distributed zonotopic estimator does not require knowing the correlations of estimation errors, such as the complex calculation of cross-covariances, so it can be implemented online. Moreover, robust stability conditions are provided to ensure the boundedness of estimation errors. Finally, illustrative examples are employed to show the effectiveness of the proposed methods.