基于深度学习的DRFM信号识别  

DRFM Interference Recognition Based on Deep Learning

作　　者：房津辉 宋宝军[1] 朱明哲 FANG Jinhui;SONG Baojun;ZHU Mingzhe(Air and Missile Defense College,Air Force Engineering University,Xi′an Shaanxi 710051,China;School of Electronic Engineering,Xidian University,Xi′an Shaanxi 710071,China)

机构地区：[1]空军工程大学防空反导学院,陕西西安710051 [2]西安电子科技大学电子工程学院,陕西西安710071

出　　处：《现代雷达》2024年第3期54-58,共5页Modern Radar

摘　　要：针对数字射频存储器(DRFM)产生信号与源信号之间无法有效区分的问题,运用基于小波变换的同步压缩变换将时域的雷达信号转换为时频图,运用深度学习强大的图像识别能力,实现了基于深度学习的源信号与DRFM信号识别,从而解决了在雷达信号处理中无法有效区分回波信号和DRFM欺骗信号以及在雷达干扰识别中基于DRFM的欺骗干扰难以识别的问题。为了验证深度学习过程的可靠性,通过神经网络可解释性算法对训练结果进行了验证和分析。实验结果表明,相比于识别原始信号,识别DRFM信号神经网络需要用到更多的特征,神经网络判断准确率达到了96.33%,识别精度良好。For digital radio frequency memory(DRFM)to generate signals cannot be effectively distinguished from the source signal,using synchro squeeze wavelet transform the radar signal of the time domain is converted to the time frequency diagram.Using deep learning powerful image recognition capabilities,the source signal and DRFM signal recognition based on deep learning are implemented.The problem that the echo signal cannot be effectively distinguished from the DRFM deception signal in the radar signal processing is resolved.The problem that is difficult to recognize DRFM deceptive interference in radar interference recognition is resolved also.In order to verify the reliability of the deep learning process,the training results are verified and analyzed through the explanatory algorithm of neural networks.The accuracy of the neural network judgment has reached 96.33%,and the recognition accuracy is good.

关 键 词：干扰识别 时频变换 梯度加权类激活映射 导向反向传播 深度学习 

分 类 号：TN974[电子电信—信号与信息处理]