MIMO-FNN模型的弹道导弹目标识别方法  

Ballistic Missile Target Recognition Method Based on MIMO-FNN Model

作　　者：林菡[1] 李昌玺 陈丽娟[1] LIN Han;LI Chang-xi;CHEN Li-juan(Fujian Agriculture and Forestry University,Dongfang College,Department of Information Engineering,Fujian Fuzhou 350017,China;PLA,No.66132Troop,Beijing100043,China)

机构地区：[1]福建农林大学东方学院信息工程系 [2]中国人民解放军66132部队

出　　处：《现代防御技术》2018年第6期36-43,共8页Modern Defence Technology

摘　　要：针对弹道导弹目标融合识别特点,分析了弹道导弹单传感器多特征多输入多输出模糊神经网络(multiple input multiple output fuzzy neural network,MIMO-FNN)模型、弹道导弹多传感器单特征MIMO-FNN模型,在此基础上,结合弹道导弹目标融合识别的实际流程,提出了弹道导弹目标识别多传感器多特征MIMO-FNN模型。该模型以每个单传感器多特征MIMO-FNN模型的输出为输入,并通过专家知识求取每个传感器的融合权值,采用sum-product模糊推理和加权求和法解模糊,得到模型的融合识别结果,并通过仿真实验验证了所提模型的有效性,最后从多传感器多特征优化和传感器权重2个方面对所提模型进行了可行性分析。According to the characteristics of ballistic missile target fusion recognition,analyzed the single sensor multi feature multiple input multiple output fuzzy neural network (MIMO-FNN)model and the multi sensor single feature MIMO-FNN model.On this basis and combined with the realistic process of ballistic missile target fusion recognition,a multi sensor multi feature MIMO-FNN model was proposed. This model first took the output of every single sensor multi feature MIMO-FNN model as input,and calculated the fusion weight of every sensor by expert knowledge,then used the sum product fuzzy inference and weighted summation to ambiguity resolution,and got the results of fusion recognition.The simulation showed the feasibility and effectiveness of this model.

关 键 词：模糊神经网络 弹道导弹 目标识别 多输入多输出 多传感器 融合 

分 类 号：TJ761.3[兵器科学与技术—武器系统与运用工程] TN959.17[电子电信—信号与信息处理]