An Improved PSO Algorithm for Multi-sensor Multi-target Data Association

Abstract

Improved particle swarm optimization (PSO) algorithm is proposed to deal with the data association problem for multi-sensor multi-target tracking. The tracking gate is used to confirm the effective measurements, and the association relation between measurements and targets are described by the likelihood function of filter innovation to establish the model of the optimal combination. When solving the optimal combination problem, Lagrange relaxation technology is used to reduce the combination to two dimensions firstly, and then the improved PSO algorithm, which based on the cross and mutation rules, is used to obtain the optimal solution, and get the optimal association pairs for measurements and targets. At last, the simulation shows the superiority of the method in accuracy and speed of the data association.