基于联合时频辐射源信号识别方法  

Signal Recognition Method Based on Joint Time-Frequency Radiant Source

作　　者：叶文强 俞志富 YE Wen-qiang;YU Zhi-fu(Electronic Warfare Academy,National University of Defense Technology,Hefei 230037,China)

机构地区：[1]国防科技大学电子对抗学院,合肥230037

出　　处：《电子信息对抗技术》2018年第5期16-19,50,共5页Electronic Information Warfare Technology

摘　　要：针对传统基于人工判决和模块匹配识别雷达辐射源信号的不足,提出了一种基于联合时频的智能识别算法。将雷达辐射源信号进行时频变换,处理后的信号经过不同种降维方式输入到自动编码器中,通过预训练调节深度学习模型,并采用常用的softmax分类器对预训练完的模型采取有监督精校和识别,最后完成识别任务。仿真结果表明,该算法能够实现很高识别率,并且经过联合降维较其他几种方法的效果更好。A new intelligent recognition algorithm based on joint time-frequency is proposed in view of the shortcomings of traditional artificial decision and module matching recognition radar emitter signals. The radar emitter signal time-frequency transform,the processed signal through different dimension input to the automatic encoder,learning model by pretraining adjustment depth,and the softmax classifier commonly used on the pretraining end model adopts supervised sizing and recognition,finally the identification task is completed. The simulation results show that the algorithm can achieve high recognition rate,and the effect after the combined dimension reduction in any way is better.

关 键 词：雷达辐射源信号 时频变换 联合降维 编码器 识别 

分 类 号：TN971.1[电子电信—信号与信息处理]