Multisensor multitarget tracking methods based on particle filter

Abstract

In order to solve the multisensor multitarget tracking problem of the non-Gaussian nonlinear systems, the paper presents a multisensor joint probabilistic data association particle (MJPDAP) algorithm. At first, the algorithm permutes and combines the measurement from each sensor using the rule of generalized S-D assignment algorithm. Then, all of measurements in each assignment are combined into one equivalent measurement and the joint likelihood function of the equivalent measurement is calculated. Finally, the particle weight is updated and the state estimation of the fusion center is obtained, using joint probability data association (JPDA) method. In this paper, some Monte Carlo simulations are used to analyze the performance of the new method. The simulation results show the MJPDAP can effectively track multitarget in the nonlinear systems, and be of much better performance than the single-sensor joint probabilistic data association particle (SJPDAP) algorithm.