Dielectric Resonator Antenna Geometry-Dependent Performance Tradeoffs

Abstract

A flexible shape generation technique (Surface Contour Shape Generation) is combined with multiobjective evolutionary optimization algorithms in order to synthesize geometrically-diverse dielectric resonator antennas and explore the impact geometry has on relevant performance metrics. It is shown that noncanonical shapes can improve gain and bandwidth when compared to conventional dielectric resonator antenna geometries. Previously described free-form techniques (such as the superformula) have demonstrated improved performance over canonical shapes, but are difficult to integrate with inverse-design. In contrast, the proposed shape generation technique is coupled with multiobjective optimization in order to illustrate the solution space tradeoffs between the various design goals. This technical approach is compatible with various feed designs, antenna performance parameters, and 3D printing techniques.

Keywords

• Dielectric resonator antennas
• evolutionary algorithms
• optimization
• multiobjective
• inverse-design
• additive manufacturing