Automated Design of Polarization Converter Based on Pixel Metasurface

Conventional design methods for metasurface-based polarization converters (PCs) inevitably involve extensive theoretical analysis and parameter investigation, making the design process lengthy. In this communication, an automated network-based fusion design concept for PCs is proposed, which combines numerical precision with rapid optimization. Following this concept, a linear-to-circular polarization converter (LCPC) has been automatically designed, successfully integrating dual-band, dual-sense, angular-stable conversion characteristics into a pixel metasurface. Prototype validation was carried out, with the measured results aligning well with the simulated ones. The proposed efficient optimization scheme can be applied to reflective as well as transmissive polarization converters, significantly reducing the human effort and computational resources in the design process.