基于可视图深度学习的雷达天线扫描样式识别  

Radar Antenna Scanning Type Recognition Based on Visual Graph Deep Learning

作　　者：葛学仕 赵雷鸣[1] 孙广海 GE Xueshi;ZHAO Leiming;SUN Guanghai(Unit 92728 of the PLA,Shanghai 200436,China;Nanjing Corad Electronic Equipment Co.,Ltd.,Nanjing 210018,China)

机构地区：[1]中国人民解放军92728部队,上海200436 [2]南京科瑞达电子装备有限责任公司,南京210018

出　　处：《电子信息对抗技术》2024年第5期44-52,共9页Electronic Information Warfare Technology

摘　　要：针对相控阵雷达扫描信号难识别的问题,提出一种基于可视图深度学习的雷达扫描样式识别方法。首先,分析了7种常见的雷达天线扫描样式。然后,将接收到的天线主瓣照射幅度与脉冲到达时间之间的变化趋势转换为可视图矩阵。最后,构建深度网络模型,验证扫描样式识别的效果。试验结果表明,基于可视图深度学习的雷达扫描样式识别方法对7种扫描样式的整体识别准确率在90%以上,并对低信噪比的信号有很好的处理适应性。To improve the recognition performance on radar antenna scan type,a visual graph based deep neural network is proposed to classify different modes of scan types.Firstly,seven kinds of commonly used radar antenna scans are analyzed.Then,the relationship of beam amplitude and pulse time of arrival is converted into binary matrix generated from visual graph.Finally,deep neural network is constructed to verify the recognition performance on antenna scan type.The experimental results indicate that the proposed method can achieve an average accuracy of above 90%on the specified 7 modes of scan types,and is robust to the scan signal with lower signal-to-noise ratio.

关 键 词：雷达天线扫描 可视图 深度神经网络 电子战 

分 类 号：TN971.1[电子电信—信号与信息处理]