摘要
机载电子战系统对辐射源型号的精确识别是关乎载机空战OODA环的重要能力。基于卷积神经网络(CNN)对辐射源的多维精细特征进行离线训练学习,获取多维特征参数与辐射源型号之间的高阶非线性分类函数边界,并基于训练好的网络模型开展工程化应用验证。试验表明辐射源型号的识别正确率显著提升,迈出了机载电子战系统走向智能认知的第一步。
Radiator Identification of airborne EW system is a key issue for OODA loop in air combat.A new approach is proposed for radiator identification based on artificial intelligence,which presents an off-line training algorithm for radiator's multi-feature based on CNN model.The higher-order nonlinear function boundary can be obtained for classification of multi-type ra-diator through multi-feature.The trained model is deployed for engineering verification,and the test result indicates that using the above method can attain excellent identifying results.This in-dicates the airborne EW system takes the first key step in the field of cognitive EW.
作者
贾朝文
张学帅
鄢勃
刘翔
杨洋
张译方
杨启伦
JIA Chaowen;ZHANG Xueshuai;YAN Bo;LIU Xiang;YANG Yang;ZHANG Yifang;YANG Qilun(Science and Technology on Electronic Information Control Laboraoty,Chengdu 610036,China)
出处
《电子信息对抗技术》
2020年第5期1-5,共5页
Electronic Information Warfare Technology
