无源定位系统的多目标数据关联算法  

Data Association Algorithm for Multi-target in Passive Location System

作　　者：窦丽华[1] 刘航[1] 陈杰[1] 辛斌[1] 

机构地区：[1]北京理工大学信息科学技术学院,北京100081

出　　处：《兵工学报》2008年第2期217-220,共4页Acta Armamentarii

基　　金：北京市教育委员会基建重点实验室资助项目(CSYS100070417)

摘　　要：主要研究多目标无源定位系统的数据关联方法。数据关联实质上是求解后验概率的极大值,首先建立两传感器三维空间联合关联概率数学模型,将求解后验概率的极大值问题转化为求解特殊的指派问题。由于漏检和虚警的存在,导致指派个数不确定,提出了一种新的改进单亲遗传算法来求解该问题,该遗传算法将解阵看作一个个体,将矩阵的列看作个体的基因,始终对一个个体进行交叉和变异操作。仿真结果表明,提出的求解方法收敛速度快、关联正确率高,验证了模型和求解方法的正确性和可行性。The data association in passive location system was proposed, which is essentially searching the maximum of the posterior probability. The mathematical model of joint data association for two passive sensors in 3-D was established, the solving of maximum of the posterior probability was transformed into a particular assignment problem. The number of assignment is uncertain due to the missed detection and false alarm, so a new method based on the improved partheno-genetic algorithm was proposed. The solution matrix and its column are respectively regarded as the individual and the gene, the crossover, and the mutation were performed to the individual. The simulated results show that the proposed method has the character of fast convergence and high association accuracy, so the model and the solving process are right.

关 键 词：运筹学 数据关联 无源定位 单亲遗传算法 指派问题 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]