基于动态粒子群优化的目标跟踪算法

Target Tracking Algorithm Based on Dynamic Particle Swarm Optimizer

目标跟踪问题的关键在于如何寻找与目标运动状态匹配的运动模型;交互式多模型算法的模型集是根据先验信息确定的,它不随时间变化而变化,并且要求在模型集中任意时刻都存在描述目标运动模型;在实际中需要大量模型来描述运动;将粒子群优化和变结构多模型算法相结合,不仅能充分利用系统的实时量测信息,还能根据其先验信息调节优化算法结构;仿真表明,运用动态自适应粒子群优化算法实现模型集自适应,可以提高目标跟踪的精度和实时性。

The key to target tracking problem is how to find and match the motion model of target motion state.Interactive multiple model algorithm of the model set is set according to the prior information,it does not changes over time,and requires concentration at any time in the model are described target motion model.In practice,need a lot of model to describe the motion.Particle swarm optimization combined with variable structure multiple model algorithm,can not only make full use of the information system of real-time measurement, can also according to the prior information structure optimization algorithm.Simulation shows that the use of dynamic adaptive particle swarm optimization algorithm adaptive implementation model set,can improve the accuracy and real-time performance of target tracking.