基于短时频谱与自相关特征的雷达脉内调制识别  

Radar Intra-pulse Modulation Recognition Based on Features of Short Time Spectrum and Auto-correlation

作　　者：童真 姚群 高墨韵 柴恒[1] TONG Zhen;YAO Qun;GAO Mo-yun;CHAI Heng(The 723 Institute of CSIC,Yangzhou 225101,China)

机构地区：[1]中国船舶重工集团公司第七二三研究所,江苏扬州225101

出　　处：《舰船电子对抗》2021年第4期70-74,85,共6页Shipboard Electronic Countermeasure

摘　　要：针对雷达脉内调制识别问题,提出了一种基于短时频谱与自相关特征的识别方法。前端采用短时频谱变化与自相关统计作为特征输入,用以表达不同雷达调制的频率变化与相位变化特性,后端采用长短记忆(LSTM)网络作为分类输出,用以提取不同雷达调制中时序变化特性。通过雷达脉冲数据的训练与测试,表明本文所提出的方法在大范围信噪比与多种脉宽的场景下,具备较高的识别准确率与较好的泛化性能。Aiming at the problem of radar intra-pulse modulation recognition,this paper proposes a recognition method based on short-time spectrum and auto-correlation features.The front-end uses short-term spectrum variations and auto-correlation statistics as input features,which expresses the characteristics of frequency and phase variations of different radar modulations,and the back-end uses long short-term memory(LSTM)networks to generate classification outputs,which extracts the characteristics of sequence change in radar pulse of different modulations.Through the training and testing from radar pulse data,it is demonstrated that the proposed method in this paper has higher recognition accuracy and better generalization performance under the scenarios of wide range signal-to-noise ratio and various pulse widths.

关 键 词：雷达脉内调制 识别 短时频谱 自相关 长短记忆网络 

分 类 号：TN971.1[电子电信—信号与信息处理]