Distributed robust cooperative tracking against multiple maneuvering targets based on random finite set

Abstract

This paper considers the distributed multi-target tracking problem with mobile sensor network. Based on the random finite set and consensus methods, a distributed multi-target tracking algorithm is designed by local filtering stage and consensus fusion stage. In the local filtering stage, the state and measurement information of multi-target are described by random finite set. Each node in the sensor network obtains the state estimates of multi-target individually by applying the probability hypothesis density filtering algorithm. To reduce the computational complexity, the Jacobian matrix and Hessian matrix are used to calculate the priori probability density. With the local neighboring interactions, the estimate results of multi-target system are obtained after the consensus fusion stage. Besides, the stochastic boundedness of the estimation error of the distributed multi-target tracking algorithm is proved by constructing a stochastic process. In the final, a simulation example is given to verify the multi-target tracking performance of the proposed algorithm.