粗糙集决策系统通过分别检验特征矢量中每个属性值的符合程度进行概念型决策,并能根据暨有目标特征知识系统生成表征特征属性在分类过程中重要程度的粗糙隶属度。文中构建了一个基于粗糙集的雷达目标分类器,并以目标径向投影尺寸和航迹参数为例进行分类试验,对一个包含分属于6类的30个目标特征知识系统,在对目标特征属性值叠加雷达典型测量误差的条件下,达到了不低于90%的识别率。

Rough set makes concept based decision by examining the fitness of every attribute of condition vector separately,and output rough memberships which characterize attributes importance.With radial projection of radar target dimension,cruise velocity and height as condition attributes, a multi-feature radar target rough set classifier is proposed.Simulated classification based on an information system with 30 targets that belonging to 6 classes achieves an average accuracy not less than 90%,while attributes are subjoined with typical radar errors.

TN959.1

武器装备预研基金支持项目