An intelligent self-optimization SMC-PHD filter for multitarget tracking

Abstract

The sequential Monte Carlo probability hypothesis density (SMC-PHD) filter is an effective multitarget tracking algorithm in nonlinear and non-Gaussian conditions, but its states-estimation accuracy depends on the state extraction method. In order to extract multitarget states, an intelligent self-optimization SMC-PHD filter is proposed in this paper. The main framework of the intelligent SMC-PHD filter, which transforms the clustering problem into an optimization problem, is constructed. To solve this optimization problem, a self-optimization bat algorithm (SBA) is proposed to extract multitarget states of the SMC-PHD filter, which has the ability to search for a better solution by designing global search and local search strategies. The simulation results show that the performance of the SBA is better than that of the k-means algorithm for extracting multitarget states when the targets are close to each other or even overlapping.