The structure-labeled group target estimation with random finite set observation

In this paper, we combine the graphical models to the labeled random finite sets (RFS) and define the group target on a structure-labeled discrete countable space, unlike the labeled RFS without the structured information. Further, we apply it to single group tracking and propose an analytical Bayesian estimation algorithm with the RFS observation. Different from the existing group tracking approaches, we consider the dependent relationship among targets and at the same time target birth, death and spawning in the group. The final result shows that the proposed algorithm has a better performance in estimating precision than that of Gaussian-mixture CPHD filter without considering the structured information.