A Measurement Set Partitioning Algorithm Based on CFSFDP for Multiple Extended Target Tracking in PHD Filter

Abstract

The extended target probability hypothesis density (ET-PHD) filter is a promising approach for multiple extended target tracking. One crucial problem of the ET-PHD filter is partitioning the measurement set. This paper proposes a partitioning algorithm based on clustering by fast search and find density peaks (CFSFDP). Firstly, we adopt CFSFDP algorithm to partition the measurement set and the field theory is introduced to determine the cutoff distance of the CFSFDP algorithm. Then, the cluster center of the CFSFDP algorithm is determined according to solved cutoff distance and measurement rate. Finally, as the CFSFDP algorithm cannot handle the case in which targets are spatially close, an improved sub-partitioning method is implemented. Simulation results show that the proposed algorithm has less computational complexity and stronger robustness than the existing algorithm without losing tracking performance.