短路针加载微带天线反射损耗的神经网络模型

Artificial Neural Network Models for the Return Loss of Microstrip Patch Antennas Loaded with a Short Pin

短路针加载的微带贴片天线具有较低的谐振频率,特别适用于便携式通信设备中。作者采用多层感知器建立了短路针加载圆形贴片天线的神经网络模型,把短路针位置作为输入样本,天线的反射损耗S11参数作为输出样本,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,在学习范围内将短路针位置输入到多层感知器,从输出端立即得到准确的S11参数。

Microstrip patch antennas with a short pin, which have a lower resonant frequency, is well suited to portable communication equipment. In this paper, a multi-layer perception is introduced to characterize the circular microstrip patch antenna loaded with a short pin. The location of the short pin is defined as the input samples, and the return loss S11-parameters, the output samples. When the perception has been trained with the back propagation algorithm, the return loss S11-parameters of the circular microstrip patch antenna with the short pin at any arbitrary location can be obtained quickly from the perception.