Machine learning assisted dual port metasurface loaded MIMO antenna with linearly polarized to circularly polarized conversion features for n257 band of 5G mm-wave applications

Abstract

In this communication, a dual port printed multi-input and multi-output (MIMO) antenna with high isolation of dimension 12 × 8.5 × 0.8 mm3 is designed, analyzed, and investigated for 28-GHz 5G mm-wave applications. The metasurface layer is suspended over the MIMO antenna to convert the linearly polarized (LP) wave to a circularly polarized (CP) wave. The unit cell of the metasurface consists of a square-shaped conducting strip with one side of diagonal points interconnected. The conducting strip and rectangular gap between the strips etched on the substrate are capable of generating the π/2 phase shifted electric field components (Ex and Ey) resulting in polarization conversion from LP to CP. The machine learning concept is used to characterize the dimensional configuration of the antenna to achieve the optimum throughput. The MIMO diversity performance parameters and antenna essential results are validated by the measured counterparts.