Improved Box Particle CPHD Algorithm for Group Target Tracking
Xuan Cheng, Hongbing Ji, Yongquan Zhang · 2019
The existing box particle CPHD algorithm for group target tracking has large computation and poor estimation performance in strong clutter environment, an improved box particle CPHD algorithm for group target tracking is proposed to solve this problem. This algorithm utilizes the characteristic of likelihood function in box particle filter to generate an adaptive rectangular tracking threshold, which eliminates a large number of clutter measurements and reduces greatly the computational burden. In addition, the estimation performance is improved by modifying the way of box particles supplementation and k-means clustering algorithm in the state extraction. The simulation experiments show that the proposed algorithm has a higher real-time performance and better estimation performance in strong clutter environment.