Equivalent Circuit Theory-Assisted Deep Learning for Accelerated Generative Design of Metasurfaces

In this article, we propose an equivalent circuit theory-assisted deep learning approach to accelerate the design of metasurfaces. By combining filter theory and a sophisticated model, designers can achieve efficient metasurface designs. Compared with most existing generative methods that rely on arbitrarily generated training dataset (TDS), proposed method adaptively produce highly relevant low-noise samples under guidance theory, resulting in significantly narrowed target solution space improved model efficiency. Furthermore, select variational autoencoder (VAE) as which compress raw into lower-dimensional latent where optimization methods, such genetic algorithm, be more efficiently executed find optimal than brute-force search. To verify effectiveness method, apply it creation three examples frequency selective surfaces (FSSs), presenting wide-band, dual-band, band-stop responses. Experimental results show realize much faster stable convergence without domain knowledge.