摘要
针对传统使用脉间参数难以识别低信噪比条件下的复杂体制雷达信号问题,提出了一种利用深度学习模型辅助训练并对雷达辐射源进行识别的方法。首先利用时频分析的方法产生雷达信号的时频图像作为训练集1。接着利用深度卷积生成对抗网络的样本学习能力在训练集1的基础上二次生成时频图像作为训练集2,训练集2相对于1拥有着去噪和数据增强的效果。最后利用训练集2辅助视觉几何组在训练集1上的训练进行雷达辐射源识别。对5种常见的雷达信号进行了仿真实验,实验结果验证了该方法的有效性。
To deal with the problem that using traditional inter⁃pulse parameters is difficult to identify complex radar signals under low signal to noise ratio conditions,a method of using deep learning model to assist training and identifying radar emitter is proposed.First,the time⁃frequency image of the radar signal is generated as the training set 1 by the method of time⁃frequency analysis.Then,using the sample learning ability of the deep convolutional generative adversarial network,the time⁃frequency image is generated on the basis of the training set 1,and taken as the training set 2,which has the effect of de⁃noising and data enhancement compared withtraining set 1.Finally,the training set 2 is used to assist the visual geometry group training on the training set 1 to perform radar emitter identification.Simulation experiments on five common radar signals are carried out.The experimental results demonstrate the effectiveness of the proposed method.
作者
李昆
朱卫纲
LI Kun;ZHU Wei-gang(Graduate School,Space Engineering University,Beijing 101416,China;Department of Electronic and Optical Engineering,Space Engineering University,Beijing 101416,China)
出处
《电子设计工程》
2020年第12期99-104,共6页
Electronic Design Engineering
基金
CEMEE国家重点实验室项目(2018Z0202B)。