基于权值改进D-S理论的对空目标识别融合  

Identifying Aerial Target Fusion Based on the Improved Weighting D- S Theory

作　　者：董杰[1] 蔡群[1] 丁锋[1] 张博[1] 

机构地区：[1]电子工程学院,安徽合肥230037

出　　处：《现代防御技术》2016年第2期185-189,196,共6页Modern Defence Technology

基　　金：安徽省自然科学基金(1308085QF105);中国博士后科学基金(2015M572673)

摘　　要：针对传统证据理论证据高冲突性问题,提出了一种基于权值改进型Dempster-Shafer(DS)理论的对空目标识别方法。在研究D-S证据理论数学模型基础上,依据泛函距离空间理论,更新计算各个传感器的信度函数可靠性权值,对各个传感器的基本概率赋值函数进行修正,最后利用D-S合成公式对各个传感器进行融合判决。仿真结果证明了该方法的有效性。A multiple sensor identifying aerial target method based on weighting D - S theory is pro- posed for solving highly conflict D-S theory By researching D - S theory mathematical model, the reliability of correct identifying probability is analyzed based on functional metric space theory. Th probability assignment （BPA） is revised by renewal weight coefficient which depends on the senso bility. Fusion decision of individual sensor is made by D - S combination rules. The example resul sensor e basic r reliats verify the validity of the method.

关 键 词：目标识别 D-S证据理论 距离空间 权值分配 冲突系数 证据源 

分 类 号：TN957.51[电子电信—信号与信息处理]