Distributed target tracking for sensor networks with only local communication

In this paper the problem of distributed target tracking is considered. A network of agents is used to observe a mobile target and, at each iteration, all the agents agree about the estimate of the target position, despite the fact that they only have local interactions and only a small percentage of them can sense the target. The proposed approach, named Consensus-based Distributed Target Tracking (CDTT), is a fully distributed iterative tracking algorithm. At each iteration our method applies two phases. During the perception phase the target position is obtained either as a measure or as a prediction; subsequently, in the consensus phase a consensus algorithm is applied in order to let all the agents agree on the target position. As a result, the estimated trajectories are identical for all the agents. Numerical simulations are carried out to show the effectiveness and feasibility of our approach. © 2011 IEEE.