Distributed estimation of moving targets with unknown dynamics

In this paper, we propose a distributed estimation algorithm for the detection of the behaviors of a group of moving targets in the environment. We assume that there are only a limited number of sensors available for the measurement task. Each sensor can only observe the targets in its sensing range. Sensors can exchange measurement information with their neighboring sensors in their communication ranges. The problem we studied is how a distributed estimation algorithm can be designed such that each sensor can detect the behaviors of all targets in the environment. To solve this problem, we developed a distributed Kalman filtering algorithm which only requires limited information exchange among sensors in the network. We also designed a strategy to deal with the unpredictable change of communication connectivity. Simulation results verified the effectiveness of the proposed algorithm.