基于新冲突度量的属性信息相关算法  

Attribute information correlation algorithm based on new conflict measure

作　　者：张锟[1] 张昌芳[2] 李杰[1] 

机构地区：[1]国防科技大学机电工程与自动化学院,长沙410073 [2]国防科技大学继续教育学院,长沙410073

出　　处：《控制与决策》2011年第4期601-605,共5页Control and Decision

基　　金：国家科技攀登计划基金项目(2006BAG02B05-14)

摘　　要：在传感器识别性能不理想的情况下,其给出的属性信息质量较低,传统的冲突度量难以客观地对它们之间的相异程度进行定量描述.为此,定义了一种新的冲突度量,该度量可以很好地对来自同类和异类目标的属性信息进行区分.在新冲突度量的基础上,采用多维分配算法实现了属性信息相关.理论分析和仿真实验均表明了所提出的基于新冲突度量的属性信息相关算法的有效性.The traditional conflict measure between attribute information can’t objectively give a quantitative description about their dissimilarity when the information quality is relatively low resulting from sensors’ poor identification performance.So a new conflict measure is defined,which can nicely distinguish between the attribute information from targets of the same type and that from targets of different type.Taking the new conflict measure as its basis,the correlation decision for attribute information is implemented with multi-dimensional assignment algorithm.Theoretical analysis and simulation experiment results show the effectiveness of the proposed attribute information correlation algorithm based on the new conflict measure.

关 键 词：属性信息 相关 证据理论 冲突度量 多维分配算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]