Conditional Generative Adversarial Networks for Inverse Design of Multifunctional Metasurfaces

Abstract

Electromagnetic (EM) metasurfaces can present a versatile platform for realization of multiple diverse EM functionalities with incident wave frequency, polarization, or propagation direction through appropriate choice of unit cells structures. However, the inverse design of multifunctional metasurfaces relies on massive full‐wave EM numerical simulations to obtain an optimized solution. Herein, a step‐by‐step procedure based on conditional generative adversarial networks (cGANs) integrated with Gramian angular fields (GAFs) to reduce the computational time required for the EM simulations in the inverse design of multifunctional microwave metasurfaces is proposed. The proposed procedure initially implements GAFs to encode the desired multiobjective scattering parameters (SPs) to images and then passes them through the cGAN model to map them to three‐layer metasurfaces. The present study uses a robust dataset, including 54 000 metasurface structures and corresponding SPs to train and validate the cGAN model. This article also presents two case study examples using two multifunctional metasurfaces with different independent functionalities and full‐space coverage to justify the performance of the proposed procedure in the inverse design of multifunctional microwave metasurfaces. The case studies demonstrate that, despite the random nature of the training data samples, the cGAN reliably predicts the corresponding metasurfaces of the desired multiobjective SPs.

Keywords

• conditional generative adversarial networks
• convolutional neural networks
• multifunctional metasurface inverse designs