摘要
提出一种基于T-S(Takagi-Sugeno)模型的模糊神经网络(Fuzzy Neural Network,FNN)降水粒子相态识别方法。该方法建立一种多层前馈的模糊神经网络,在对双线偏振气象雷达接收的偏振参量进行模糊化、规则计算、模糊推理和退模糊处理基础上,利用模糊神经网络误差反馈的思想自适应的调节不同降水类型各偏振参量隶属函数的参数,保证降水粒子相态识别的精度要求。经过实测数据的处理结果证明了该方法的有效性。
In this paper,a method based on T-S（Takagi-Sugeno）model for FNN（Fuzzy Neural Network）in identification of hydrometeors is proposed.This method establishes a fuzzy neural network with multi-layer feed forward.Based on the fuzzification,inference and defuzzification of the dual-polarization weather radar measurements,the method uses the error feedback of fuzzy neural network to adaptively adjust the membership function parameters of different measurements in different hydrometeor types.It can ensure the accuracy of identification of hydrometeors.The results of the measured data show that the method is effective.
出处
《雷达科学与技术》
北大核心
2017年第6期607-616,共10页
Radar Science and Technology
基金
国家自然科学基金(No.61471365
U1633106
61231017)
中国民航大学蓝天青年学者培养经费
中央高校基本科研业务费(No.3122017007)
关键词
相态识别
T-S模型
模糊神经网络
双线偏振气象雷达
隶属函数
identification of hydrometeors
T-S model
fuzzy neural network（FNN）
dual-polarization weather radar
membership function
