基于深度卷积网络的PRI调制模式识别方法  

PRI modulation pattern recognition method based on deep convolution neural network

作　　者：茆旋宇 郑子扬[1] 王佩 郭涛 鲁加战 卢志龙 祁友杰[1] Mao Xuanyu;Zheng Ziyang;Wang Pei;Guo Tao;Lu Jiazhan;Lu Zhilong;Qi Youjie(No.8511 Research Institute of CASIC,Nanjing 210007,Jiangsu,China)

机构地区：[1]中国航天科工集团8511研究所

出　　处：《航天电子对抗》2019年第5期32-37,共6页Aerospace Electronic Warfare

摘　　要：基于数据驱动的思想,采用深度卷积神经网络实现对雷达信号PRI调制模式的识别。仅使用测量数据,对包含复合调制在内的8种复杂调制模式,分别针对存在不同程度干扰脉冲、丢失脉冲、测量噪声以及小样本脉冲环境的影响下的测量数据进行实验。实验结果表明,该方法在上述4种环境中仍具有良好的PRI调制模式识别能力。Based on the idea of data drive, the deep convolution neural network is used to recognize PRI modulation pattern of radar signal. Only the basic PRI data are used to carry out experiments on eight kinds of complex modulation modes, including composite modulation. The experiment is carried out in the environment of interference pulse, loss pulse, measurement noise and small sample pulse respectively. The results show that this method still has good recognition performance for data with high measurement noise, serious pulse interference and high pulse loss rate(short pulse sequence length).

关 键 词：PRI调制 卷积神经网络 模式识别 

分 类 号：TN971^+.1[电子电信—信号与信息处理] TN974[电子电信—信息与通信工程]