Research On Probability Hypothesis Density Track Association Alogrithm Based On Measurement Partition

Based on random finite set statistics,probability hypothesis density(PHD)filtering algorithm is the main method to solve multiple target tracking problem.In order to improve the tracking performance of the probability hypothesis density(PHD)filter,a new PHD track association algorithm based on measurement partition is proposed.The traditional PHD filter cannot give the individual trajectories of each target,and the filtering efficiency will decrease drastically when there are a large number of clutters in the tracking environment.Besides,the traditional methods also have measurement mismatch problem for adjacent multi-target tracking.To solve the above problems,the measurement set at each time step can be divided into existing targets,new targets and clutter measurement subsets by the sorting ellipsoidal gate method,which targets can be updated by corresponding measurement subsets,thus reducing redundant computing time.Furthermore,the algorithm can adjust the weight of Gaussian component when targets are close to each other,by introducing a new weight distribution scheme,thus greatly reducing the state extraction error of adjacent targets.The simulation results show that the proposed method can improve the filtering efficiency and the adjacent targets tracking accuracy.