Abstract

This letter presents a WGAN-based inverse design method for switchable dual-band transmissive frequency selective surfaces (FSS). A simplified U-Net Wasserstein GAN is trained to map target S₂₁ responses to binary unit-cell structures, assisted by a lightweight predictor that approximates full-wave electromagnetic simulations. The generated structures are validated by full-wave simulations and experimental measurements.

Abstract

This letter presents a WGAN-based inverse design method for switchable dual-band transmissive frequency selective surfaces (FSS). A simplified U-Net Wasserstein GAN is trained to map target S₂₁ responses to binary unit-cell structures, assisted by a lightweight predictor that approximates full-wave electromagnetic simulations. The generated structures are validated by full-wave simulations and experimental measurements.

Abstract

This letter presents a WGAN-based inverse design method for switchable dual-band transmissive frequency selective surfaces (FSS). A simplified U-Net Wasserstein GAN is trained to map target S₂₁ responses to binary unit-cell structures, assisted by a lightweight predictor that approximates full-wave electromagnetic simulations. The generated structures are validated by full-wave simulations and experimental measurements.

Abstract

This letter presents a WGAN-based inverse design method for switchable dual-band transmissive frequency selective surfaces (FSS). A simplified U-Net Wasserstein GAN is trained to map target S₂₁ responses to binary unit-cell structures, assisted by a lightweight predictor that approximates full-wave electromagnetic simulations. The generated structures are validated by full-wave simulations and experimental measurements.