Efficient Optimization of the Blended Rolled Edge of a Rectangular Single Offset-Fed Compact Antenna Test Range Reflector Using Genetic Evolution

The physical and electrical size of the parabolic reflector largely determines the usable size of the Compact Antenna Test Range (CATR) quiet zone (QZ), with the reflector edge treatment being hugely influential in terms of determining the overall quality and uniformity of the collimated, pseudo plane wave. This is especially true at mm-wave frequencies where the edge illumination combined with the edge treatment can result in QZ ripple that quickly exceeds the widely used 1 dB and 10 degree peak-to-peak amplitude and phase ripple requirements. Recently, it has been demonstrated that modern powerful digital computational simulation technology in combination with evolutionary optimization strategies can be successfully harnessed to optimize the serrated edge treatment of an offset-fed CATR as part of the design process. This is crucial as it simultaneously attempts to maximize the performance of a given solution while ensuring the minimum use of space. In this paper, a further enhancement is presented in which the evolutionary computing technique is applied to the more geometrically complex blended rolled edge (BRE) treatment of a rectangular CATR reflector. Results of the validation of this novel technique are presented together with optimized QZ performance that highlight the successful use of this design procedure in the development of a new testbed for verifying base station system performance for a mm-wave, FR2, 5G New Radio applications.

I. Introduction

Electromagnetic (EM) optimization problems generally involve the examination and refinement of several different parameters. These can be either continuous or discrete valued and are in many cases bounded. Furthermore, the parameters that are involved within EM optimization problems are frequently complex, nonlinear, multiextremal, and nondifferentiable. In demanding applications such as these, genetic optimization algorithms (GOAs) have found great utility as they are generally robust, stochastic-based, methods which can handle EM optimization problems that involve many parameters that are not easily tractable by other techniques such as linear parameter sweeps, or a fully random optimization strategies. GOAs have been employed in many areas of application, e.g. shaped reflector antenna design, array antenna design and excitation optimization. However, these applications tend to be primarily concerned with the optimization of fields radiated within the Fraunhofer region, whereas here the requirement is to optimize the fields within the Fresnel region. In principle this is a closely related problem, although here; far greater importance is attached to the phase function than is typical when far-field performance alone is the primary performance metric. Thus, with the use of a GOA, a broad parameter-search may be conducted during the optimization of the edge treatment that specifically focuses on the CATR quiet zone (QZ) electromagnetic performance, allowing the optimization goal to be tailored to the requirements of the specific application including the facility layout.