Automated Design of Pixelated Metasurfaces using Particle Swarm Optimization for Diverse Applications

This paper introduces an automated methodology for designing pixelated metasurfaces for diverse application scenarios. Our approach dynamically adjusts pixel positions on the metaatoms of the metasurface to achieve specific functionalities. Based on Particle Swarm Optimization, the design algorithm minimizes the cost function to output optimal pixel configuration for the desired functionality. For definiteness, we focused on optimizing the Polarization Conversion Ratio and Transmission Efficiency at a target frequency $f_{0}$. We fabricated metasurface prototypes based on simulated meta-atoms and conducted experimental measurements to validate our approach. The experimental results are strongly correlated with the simulated responses, confirming the efficacy of our methodology for designing versatile metasurfaces with electromagnetic response characteristics that are desirable for specific application scenarios.