基于神经网络的雷达实际探测距离研究  

Research of Radar Actual Detection Range Based on Neural Network

作　　者：程子光[1] 尹康银[1] 叶全国 王前[1] CHENG Ziguang;YIN Kangyin;YE Quanguo;WANG Qian(Air Force Early Warning Academy,Wuhan 430019)

机构地区：[1]空军预警学院,武汉430019

出　　处：《舰船电子工程》2020年第2期57-61,共5页Ship Electronic Engineering

基　　金：国家自然科学基金项目(编号:61074191);空军军事理论研究项目(编号:17KJ3C1-0089R)资助。

摘　　要：雷达实际探测距离是一项十分重要的能力指标,凭经验或查阅威力图等得到的值与实际值往往存在较大偏差,不利于精细化任务行动。在量化分析影响雷达探测距离的天气、干扰、目标高度、大小等因素基础上,设计了BP神经网络框架模型,给出了样本数据获得方法,通过样本训练和实验找到了比较合理的神经网络结构参数。检验结果表明,该网络模型能准确算出各种条件下的雷达实际探测距离,具有应用简单、容错性高等特点。The actual detection range of radar is a very important capability indicator.There is a big deviation between the ac tual value and the value obtained by experience or radar-coverage-diagram,which is not conducive to the completion of refinement task.On the basis of quantitative analysis of weather,interference,target height,target size and other influencing factors,a frame work model of BP neural network is designed,a method of sample data acquisition is given.The reasonable structure parameter of BP neural network is found though sample training and experiments.The test results show that the network model can accurately cal culate the actual detection range of radar under various conditions.It has the advantages of simple application and high fault-toler ance.

关 键 词：神经网络 雷达探测距离 BP网络 影响因素 量化分析 人工智能 

分 类 号：TN95[电子电信—信号与信息处理] TP183[电子电信—信息与通信工程]