Hamiltonian deep neural network optimized with pelican optimization algorithm-fostered substrate-integrated waveguide antenna design for 5G

Abstract

Due to the growing need for higher speed data, the 5G terrestrial heterogeneous wireless network deployments are expected to happen quickly throughout the world in the next decade. In such type of networks, mm-wave small-cells overlapped the sub-6 GHz macro-cells being used to serve to population-rich areas. Subsequently, many problems appear with the antenna design technologies. The presented antenna is functioning at a frequency range from 24.8 to 31.6 GHz, with a 24% bandwidth and 8.5 dB peak gain at 27 GHz. It encompasses the complete 28 GHz frequency band utilized through 5G applications. Consequently, fifth-generation communication systems are best suited for it. The proposed Hamiltonian deep neural network optimized with pelican Optimization Algorithm-fostered Substrate-Integrated Waveguide Antenna Design for 5G (SIW-HDNN-POA-5G) is implemented, and performance of proposed technique is estimated based on several metrics, including resonant frequency (GHz), reflection coefficient (S11 in dB), mean absolute error (MAE), and root mean square error (RMSE). The proposed SIW-HDNN-POA-5G method provides 24.36%, 33.55% and 44.22% higher gain and 43.21%, 38.87% and 25.65% lesser mean absolute error comparing to the existing designs, like Design of Zero Clearance SIW End fire Antenna Array Based on Machine Learning-Assisted Optimization (SIW-MLAO-5G), SIW-Fed Wideband Filtering Antenna for Millimeter-Wave Applications (SIW-5G-MLOM), and Compact SIW Fed Dual-Port Single Element Annular Slot MIMO Antenna for 5G mm Wave Applications (SIW-FWFA-MMWA), respectively.