论文部分内容阅读
本文采用多层感知器建立了微带不连续性的神经网络模型。把微带不连续性尺寸和频率作为输入样本，不连续性的Ｓ参数作为输出样本，采用ＢＰ算法对多层感知器进行训练。当多层感知器训练完成，在学习范围内将微带不连续电路尺寸和频率输入到多层感知器，从输出端立即得到准确的Ｓ参数。
In this paper, a multilayer neural network model of microstrip discontinuity is established by using multilayer perceptrons. Taking the microstrip discontinuity size and frequency as the input samples and discontinuity S parameters as the output samples, the BP algorithm was used to train the multilayer perceptrons. When the multi-layer sensor training is completed, the microstrip discrete circuit size and frequency are input to the multi-layer sensor within the learning range, and the accurate S parameters are obtained immediately from the output end.