一种基于任务属性的雷达辐射源智能表征方法  

A method for characterizing radar radiation sources based on task attributes

作　　者：杨佳敏 杨蔚 杨柱天[3] 陈政宇 李贵显 Yang Jiamin;Yang Wei;Yang Zhutian;Chen Zhengyu;Li Guixian(Southeast University,Nanjing 210000,Jiangsu,China;Nanjing Electronic Equipment Research Institute,Nanjing 210000,Jiangsu,China;Harbin Institute of Technology,Harbin 150000,Heilongjiang,China)

机构地区：[1]东南大学,江苏南京210000 [2]南京电子设备研究所,江苏南京210000 [3]哈尔滨工业大学,黑龙江哈尔滨150000

出　　处：《航天电子对抗》2024年第6期14-20,共7页Aerospace Electronic Warfare

摘　　要：随着新体制雷达不断向智能化方向发展,传统的PDW描述字已经不能够满足复杂电磁环境下的快速识别与认知,对雷达辐射源信号的描述需在原有的基础上增加维度,形成全新的特征描述空间。提出了面向任务属性的雷达辐射源智能表征方法,通过构建雷达辐射源目标通用特征库,提出基于迁移学习的雷达辐射源智能表征模型,为雷达辐射源智能识别与认知提供技术支撑。With the continuous development of new radar systems towards intelligence,the traditional PDW descriptors can no longer meet the needs of rapid identification and recognition of radar radiation source in com⁃plex electromagnetic environment.Therefore,the description of radar radiation sources needs to add dimensions to the original basis,forming a new feature description space.A new method for characterizing radar radiation sources based on task attributes is proposed.By constructing a universal feature library for radar radiation sourc⁃es,a transfer learning based radar radiation source characterization model is proposed to provide technical support for intelligent recognition and cognition of radar radiation sources.

关 键 词：雷达表征 深度强化学习 认知电子战 

分 类 号：TN95[电子电信—信号与信息处理]