Design of Zero Clearance SIW Endfire Antenna Array Using Machine Learning-Assisted Optimization

In this communication, a substrate integrated waveguide (SIW) end-fire antenna array with zero clearance is proposed for fifth-generation (5G) mobile applications using machine learning-assisted optimization. particular, novel impedance matching architecture that involves three arbitrary pad-loading metallic vias investigated and adopted the element. Due to stringent design requirements, locations sizes of pads are obtained via state-of-the-art learning assisted exploration method, parallel surrogate model-assisted hybrid differential evolution synthesis (PSADEA). Keeping very low profile, optimized by PSADEA covers an operating frequency bandwidth from 36 40 GHz. The in-band total efficiency generally better than 60% peak gain above 5 dBi. beam scanning range at 39 GHz −20° 35°.