Parametric Modeling of UWB Antennas Using Convolutional Neural Networks

A convolutional neural network (CNN) model is proposed for parametric modeling of ultra-wideband (UWB) antennas. Given the image of an antenna, CNN fast predicts its electromagnetic (EM) responses and speeds up the design process. The validity and efficiency of the proposed model is confirmed by a parametric modeling example of a UWB band-notched antenna. The results also demonstrate that this model has potential use in mapping the relationship between the antenna structure and corresponding EM responses.