无线电工程2025，Vol.55Issue(6)：1230-1237,8.DOI:10.3969/j.issn.1003-3106.2025.06.010
一种基于异构网络的雷达辐射源型号识别方法
A Radar Emitter Type Identification Method Based on Heterogeneous Networks
蔡伟 1王宇 1刘则林 1张晓峰1
作者信息
- 1. 中国电子科技集团公司第五十一研究所,上海 201802
- 折叠
摘要
Abstract
Advanced radars with multiple functions and multiple modes have complex waveforms,flexible switching of working modes,and rapid agility of signal parameters.This leads to fuzzy or incorrect identification of radar models in the electronic reconnaissance system,posing a severe challenge to the identification of radar emitter.Regarding this issue,a radar emitter type identification method based on heterogeneous networks is proposed.The Convolutional Neural Network(CNN)is used to effectively extract radar parameter combination information and Long Short-Term Memory Network(LSTM)is good at extracting the time series features and pulse interval features of Pulse Description Word(PDW)data.Parallel structure of CNN and LSTM is used to effectively perform effective fusion of features from different dimensions and levels,and reconstruct the sample feature space,thereby achieving the type identification of advanced radar signals.Through simulation data and real measurement data,the identification scenario is constructed for nine types of multifunctional radars,and the effectiveness of the type identification ability of the proposed method is verified,the accuracy of radar type identification reaches 94.38%.This method demonstrates strong robustness in complex environments,providing a novel and effective solution for radar type identification.关键词
雷达型号识别/异构网络/卷积神经网络/长短期记忆网络/深度学习Key words
radar type identification/heterogeneous network/CNN/LSTM/deep learning分类
信息技术与安全科学引用本文复制引用
蔡伟,王宇,刘则林,张晓峰..一种基于异构网络的雷达辐射源型号识别方法[J].无线电工程,2025,55(6):1230-1237,8.
