A Compensation Method for Group Target State and Structure Estimation based on GLMB

To overcome the coupling problem of structure and state estimation in cluster target tracking, this paper introduces a compensation algorithm for group target estimation founded on Generalized Labeled Multi-Bernoull filter. Firstly, based on the dynamic model of the cluster target, the collaborative noise model of the cluster target is described and established. Secondly, the structure of the cluster is estimated to obtain the cooperative noise compensation term and compensate the group target prediction state under the cooperative condition. Finally, under the framework of GLMB filter, the update status and estimated structure of the cluster are obtained, and the prediction structure of the cluster is further compensated.