【摘要】 缺陷接地结构 (DGS)是由光子带隙结构 (PBG)发展而来的 ,它是在微波电路的接地金属平面上人为地蚀刻出“缺陷”,改变接地电流的分布 ,从而改变传输线的频率特性。周期性缺陷接地结构 (PDGS)是一种在微波电路的接地金属平面上人为地蚀刻出周期性的“缺陷”结构 ,能够使某些频段的电磁波无法从中通过 ,存在明显的阻带特性。文中采用人工神经网络对周期性缺陷接地结构进行建模 ,将其结构尺寸和频率作为输入样本 ,传输系数参数作为输出样本 ,采用贝叶斯正则化算法对神经网络进行训练。神经网络训练完成后 ,在学习范围内将其结构尺寸和频率输入到神经网络模型 ,从输出端立即得到准确的传输系数。更多还原

【Abstract】 The defected grou nd structures (DGS) are expanded from th e photonics bandgap (PBG) structures. On the ground metallic plane, the defected units are etched artificially, the groun d current distribution can be changed, and the frequency properties of the transmission lines can be influenced. The periodic d efected ground structures (PDGS) are the structures that are etched periodically on its ground plane. PDGS can be used t o prohibit the propagation of electromag netic waves within a certain band of fre quencies. In this paper, artificial neur al network (ANN) of PDGS is developed. T he structure size of PDGS and the freque ncy are defined as the input samples of ANN model, the parameters of transmissio n coefficient are defined as the output samples. As the ANN model has been train ed with the Bayesian Regularization algo rithm, the transmission coefficient of P DGS at any arbitrary sizes and the frequ encies within the region of training can be obtained quickly from the ANN model.更多还原