摘要
为了对雷达天线扫描方式进行自动识别,改进开发了天线扫描方式模拟器,并分别研究了电子扫描和机械扫描的特征提取和识别方法。基于最大主瓣脉冲序列的特征参数实现电子扫描和机械扫描的区分,然后基于单个天线扫描周期脉冲序列的特征参数实现8种机械扫描方式的自动识别。仿真结果表明,本文方法能够区分一维电扫、二维电扫和机械扫描,并且采用支持向量机决策树对机械扫描方式的识别正确率高于决策树方法。
In order to recognize radar antenna scan type(AST) automatically,an improved antenna scan pattern simulator was developed,and features extraction and automatic recognition methods for electronic scan type(EST) and mechanical scan type(MST) were studied respectively.EST was firstly distinguished from MST based on the characteristic parameters extracted from the maximum main beam pulse sequence;then 8 MSTs were automatically recognized based on the parameters extracted from the pulse sequence in a scan period.The simulation results show that it is able to distinguish between one- dimensional EST,two-dimensional EST and MSTs by this method.Moreover,the correct recognition ratio of MSTs by support vector machine decision tree(SVMDT) is higher than that by decision tree(DT).
出处
《国防科技大学学报》
EI
CAS
CSCD
北大核心
2014年第3期156-163,共8页
Journal of National University of Defense Technology
基金
国家自然科学基金资助项目(61201336
41301490)
关键词
天线扫描方式
自动识别
电子扫描
机械扫描
支持向量机决策树
antenna scan type
automatic recognition
electronic scan type
mechanical scan type
support vector machine decision tree