基于BP网的雷达辐射源分类器设计

Design of Radar Sortation Device Based on BP Network

为了解决侦察告警设备在强噪声背景下的雷达辐射源识别率较低的问题,设计了一种基于BP(Back Propagation)网络的雷达辐射源分类器,并就设计关键点——分类器拓扑结构、训练次数的确定以及训练集的设计进行了研究。分类器以BP网络为基础,以载频、脉宽、重频和天线扫描方式为输入特征向量,对雷达辐射源进行分类识别。仿真实验结果证明,该分类器在强噪声背景下的正确识别率可达92.4%。

In order to solve the problem of lower radar emitter recognition rate of radar warning receiver in strong noise background, we design a radar sortation device based on BP （Back Propagation） network, and research on the device structure, training degree and training data design which are the key of radar sortation device. It is based on BP network and uses RF （Radio Freqency）, PW （Pulse Width）, PRI （Pulse Repetition Interval） and scan mode of antenna as the input eigenvector of BP network to sort radar emitter. Simulation experiment resuits indicate that the recognition rate of radar sortation device can add up to 92. 4% in strong noise background.