关键词: 扩展目标跟踪, 量测集划分, 层次划分, 椭球门限, 模糊C-均值

Abstract: In the background of clutter, the Probability Hypothesis Density（PHD） filter is used to carry out the extended target tracking where the measurement set is difficult to partition and the computational efficiency is low. A method is proposed to divide the measurements for extended target by using the Clusters Optimization based on Density of Hierarchical Division（CODHD） clustering algorithm. Firstly, the adaptive ellipsoid threshold method is used to pre-process the measurement set, the measurement division is generated by using the clusters merging method. The clustering quality for each partition are calculated so as to build the quality curve, and measurement division is obtained through calculating the gained clusters number and clustering center by Fuzzy C-Means（FCM） operation. The simulation results have shown that the method can be used to divide the measurement set while the result of accurate partitioning can be obtained, and the cost of the calculation is obviously reduced.

Key words: extended target tracking, measurement set partitioning, hierarchical division, ellipsoid threshold, Fuzzy C-Means（FCM）