Computational Intelligence Assisted Pattern Synthesis of Linear Antenna Arrays for 5G Applications

Abstract

newline ABSTRACT newlineThe performance of wireless communication systems is greatly influenced by newlinebeam patterns of the incorporated antenna arrays. Beam pattern synthesis is an active area newlineof interest for many researchers to improve radiation performance of antenna arrays. newlineTechnologies like 5G are based on massive Multiple-Input Multiple-Output (MIMO) and newlinebeamforming, to serve their users with dedicated targeted beams. The current newlinerequirements of 5G cellular systems demand generation of narrow adaptive and switched newlinebeams, which is not possible with existing conventional techniques. Therefore, there is a newlinenecessity to develop efficient Computational Intelligence (CI) assisted Multi-Objective newlineOptimization (MOO) techniques to achieve lower values of SLL and beam width in MUMIMO newlinebased newlinecellular newlinesystems, newlinewhich newlineare newlineconflicting newlinein newlinenature. newlineLow newlinevalue newlineof newlineSLL newlineensures newline newlineless newlineinterference newlinewith newlineusers newlinein newlinethe newlinesame newlinecell newlineas newlinewell newlineas newlinewith newlineneighboring newlinecells. newlineReduction newline newlineof newlinebeam newlinewidth newlineincreases newlinesystem newlinecapacity, newlinequality newlineand newlinecell newlinecoverage newlinearea newlineby newlineimproving newlinethe newline newlinesignal-to-noise newlineratio newline(SNR). newline newline newlineIn this work, five different techniques have been implemented and four new newlinetechniques have been developed, to optimize antenna array parameters such as array size, newlineexcitation amplitudes and elemental spacing, for obtaining low values of SLL and half newlinepower beam width (HPBW). The five techniques implemented are (i) Weighting, (ii) two newlineCI techniques namely Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), newline(iii) two MOO techniques namely Multi-Objective Genetic Algorithm (MOGA) and newlineMulti-Objective Particle Swarm Optimization (MOPSO). Further, four new hybrid newlinetechniques based on Chebyshev weighting and Surrogate optimization are proposed to newlineachieve better trade-off between SLL and HPBW. They are Chebyshev Based MultiObjective newline newlineGenetic Algorithm (CB-MOGA) and Multi-Objective Particle Swarm newlineOptimization (CB-MOPSO), Surrogate Assisted Multi-Objective Genetic Algorithm (SAMOGA) newlineand newlineMulti-Objective newlineParticle newlineSwarm newlineOptimization newline(SA-MOPSO). newline newlineThe rese