摘要
针对现有方法分选准确率不高和对参数变换敏感的问题,提出一种新的雷达辐射源分选算法。将常规参数到达角、载频、脉宽和脉内特征参数Wpt6,Wpt7构成特征向量,用Kohonen神经网络实现自动分选。仿真结果表明,新方法的分选准确率较常规方法可以提高7%左右,更加适应于现代电子战环境。
Because the sorting rate of the common method is not high and senslnve to tne varieu parameters, a novel sorting algorithm for the radar emitter is proposed. The direction of ar- rival, the frequency, the pulse width parameters, and in-pulse parameters Wpt6, Wpt7 are con- structed as feature vectors. And Kohonen neural network is designed to a sort radar emitter. Simulation result shows that the sorting rate of the new method is 7% higher than the common method. Thus it is suitable for the modern electronic war.
出处
《数据采集与处理》
CSCD
北大核心
2009年第1期91-94,共4页
Journal of Data Acquisition and Processing
基金
武器装备预研基金(9140C1006090804)资助项目
