不确定多传感器目标识别的区间相离度法  

Method of interval deviation degree for uncertain multi-sensor target recognition

作　　者：万树平[1] 

机构地区：[1]江西财经大学信息管理学院,南昌330013

出　　处：《控制与决策》2009年第9期1306-1309,1315,共5页Control and Decision

基　　金：国家自然科学基金项目(10626029);江西省自然科学基金项目(0611082);江西省教育厅科技项目(GJJ08350)

摘　　要：针对目标特征值和测量值均以区间数表达的多传感器目标识别问题,提出了一种不确定性融合方法.该方法定义区间相离度,通过对特征值的区间聚类和诱导有序加权平均算子集结得到属性的权重,利用综合相离度给出目标识别方法.能够克服属性权重选取的主观性,提高了目标识别结果的可信度.仿真实例验证了所提出方法的有效性和实用性.For the problem of multi-sensor target recognition, in which the characteristic values of target types and observations of sensors are in the form of interval numbers, an uncertain fusion method is proposed. The method defines the interval deviation degree. The weight vector of attributes is obtained by interval cluster and integration of induced ordered weighted averaging （IOWA） operator. The method of recognition is presented according to the overall deviation degree, which can avoid the subjectivity of selecting attribute weight and improve the trustworthy degree of target recognition. The simulation example shows the effectiveness and feasibility of the method.

关 键 词：多传感器 目标识别 区间数 相离度 诱导有序加权平均算子 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]