微帶線的神經(jīng)網(wǎng)絡模型
THE NEURAL NETWORK MODEL OF MICROSTRIP LINE
-
摘要: 本文采用多層感知器神經(jīng)網(wǎng)絡對微帶線進行建模。將微帶線的結(jié)構(gòu)尺寸和頻率作為神經(jīng)網(wǎng)絡的輸入樣本。微帶線的電特性參數(shù)(由電磁場數(shù)值方法獲得)作為輸出樣本,采用BP算法對多層感知器進行訓練。當對多層感知器訓練結(jié)束后,在學習范圍內(nèi)將微帶線的尺寸和頻率輸入到多層感知器,從輸出端可以立即得到準確的電特性參數(shù)。
-
關(guān)鍵詞:
- 神經(jīng)網(wǎng)絡; 多層感知器; BP算法; 微帶線
Abstract: The multi-layer perceptron is introduced to characterize the microstrip line．The size of microstrip line and the frequency are defined as the input saples of the multi-layer perceptron，and the electric characteristic parameters of the microstrip，which are obtained by the numerical method，are determined as the desired output samples．As the perceptron has been trained with the back propagation algorithm，the electric characteristic parameters of the microstrip line at any arbitrary size and frequencies can be obtained quickly from the multi-layer perceptron． -
楊詮讓，毫米波傳輸線．北京：電子工業(yè)出版社，1986，31-74．[2]Webb A R．Functional approximation by feed forward networks：A least squares approach to generation IEEE Trans．On Neural Network，1994，．NN-5(2)：363-371． -
計量
- 文章訪問數(shù): 2012
- HTML全文瀏覽量: 91
- PDF下載量: 431
- 被引次數(shù): 0
下載:
