Abstract: A new multi-target tracking (MTT) algorithm called the kernel particle probability hypothesis density filter (KP-PHDF) is proposed for MTT applications. Based on the particle probability hypothesis density filter framework, the algorithm utilizes the kernel density estimation (KDE) theory and the meanshift algorithm to further estimate the probability hypothesis density (PHD) and then to extract target state estimates. The simulation results of the proposed method show that, compared with the sequential Monte Carlo probability hypothesis density filter (SMC-PHDF), the tracking accuracy of the proposed method is increased by 30.5% in terms of miss distance.

Key words: signal processing, particle probability hypothesis density filter, kernel density estimation, simulation, multi-target tracking