Distributed consensus filtering for jump Markov linear systems

This article studies the problem of distributed filtering for jump Markov linear systems in a not fully connected sensor network. A distributed consensus filter is developed by applying an improved interacting multiple model approach in which the mode-conditioned estimates are derived by the Kalman consensus filter and the mode probabilities are obtained in the sense of linear minimum variance. A numerical example is provided to demonstrate the effectiveness of the proposed algorithm for tracking a manoeuvring target in a sensor work with eight nodes.