基于深度学习的双通带频率选择表面的设计研究  

Research on the Design of Double Passband Frequency-selective Surface Based on Deep Learning

作　　者：王毅 眭世荣 柳亚男 谢雨 WANG Yi;SUI Shirong;LIU Yanan;XIE Yu(Guangdong Provincial Institute of High-tech Emerging Industry Development,Foshan 528226,China;Jinling Institute of Technology,Nanjing 211169,China)

机构地区：[1]广东省高智新兴产业发展研究院,广东佛山528226 [2]金陵科技学院网络安全学院,江苏南京211169

出　　处：《金陵科技学院学报》2024年第4期15-21,共7页Journal of Jinling Institute of Technology

基　　金：国家自然科学基金青年项目(42101428)。

摘　　要：频率选择表面(FSS)的设计方法高度依赖于全波数值模拟或等效电路模型(ECM)辅助分析。提出了一种基于深度学习的FSS结构智能设计方法,以提高其设计效率。首先,选择变分自编码器(VAE)将数据集样本压缩到低维潜在空间中;然后,将这些样本及其相应的标签转换并输入到多层感知机网络中进行训练;最后,将训练好的网络模型在输入所需的S曲线后预测结构参数。全波分析仿真结果表明,设计的方法达到了预期效果。The design method of frequency-selective surface(FSS)is highly dependent on full-wave numerical simulation or equivalent circuit model(ECM)assisted analysis.In this paper,we propose a deep learning-based intelligent design method for FSS structures to improve its design efficiency.Firstly,a variational autoencoder(VAE)is selected to compress the dataset samples into the low-dimensional latent space.Then,these samples and their corresponding labels are transformed and inputted into a multi-layer perceptron network for training.Finally,the trained network model is used to predict the structural parameters after inputting the required S-curve.The simulation results of full-wave analysis show that the proposed design method in this paper has achieved the expected effect.

关 键 词：变分自编码器 双通带频率选择表面 逆向设计 无监督学习 

分 类 号：TN603[电子电信—电路与系统]