New hybrid approaches for SLL reduction and size miniaturization of planar antenna arrays based on 2D convolution and genetic algorithm

Abstract

In this paper, a new hybrid beamforming approach based on the 2D convolution and genetic algorithm (GA), denoted as C/GA, is proposed for side lobe level (SLL) reduction and size miniaturization of planar antenna arrays (PAA). The convolution algorithm determines the elements excitations, whereas the GA optimizes the element spacing to adjust the half power beamwidth (HPBW). Two beamforming approaches are derived from the C/GA approach. The first one is the single-stage C/GA that is denoted as (SC/GA), where the convolution of the excitation matrix by itself is performed just once. The SC/GA yields a twofold decrease in the SLL. The second is the double-stage C/GA, (DC/GA), which is developed for further SLL reduction, where the convolution of the excitation matrix by itself is performed twice successively, resulting in a fourfold decrease in the SLL. However, both SC/GA and DC/GA approaches provide high dynamic range ratios (DRR). This issue is addressed either by taking the middle excitations of the estimated excitation matrices or by performing array thinning by turning off the edge elements of the synthesized PAAs, which indeed minimizes the array size. The practical validation of the proposed approaches is realized using the computer simulation technology (CST) microwave package.