Antennas are an important component of wireless ("without wires") communications, regardless of their use. As these systems have become increasingly complex, antenna design requirements have become more demanding. Conventional antenna design consists of selecting some canonical radiator structure described by a handful of key dimensions, and then adjusting these using an optimization algorithm that improves some performance-related objective function that is (during optimization) repeatedly evaluated via a full-wave computational electromagnetics model of the structure. This approach has been employed to great effect in the enormously successful development of wireless communications antenna technology thus far, but is limiting in the sense that the "design space" is restricted to a library of canonical (or regular near-canonical) shapes. As increased design constraints and more complicated placement requirements arise such an approach to antenna design could eventually become a bottleneck. The use of antenna shape synthesis, a process also referred to as inverse design, can widen the "design space", and include such aspects as occupancy and fabrication constraints, the presence of a platform, even weight constraints, and much more. Dielectric resonator antennas (DRAs) hold the promise of lower losses at higher frequencies. This thesis uses a three-dimensional shape optimization algorithm along with a characteristic mode analysis and a genetic algorithm to shape synthesize DRAs. Until now, a limited amount of work on such shape synthesis has been performed for single-feed fixed-beam DRAs. In this thesis we extend this approach by devising and implementing a new shaping methodology for significantly more complex problems, namely directivity-maximized multi-port fixed-beam DRAs, and multi-port DRAs capable of the beam-steering required to satisfy certain spherical coverage constraints, where the location, type and number of feed-ports need not be specified prior to shaping. The approach enables even low-profile enhanced-directivity DRAs to be shape synthesized.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/45231 |
| Date | 08 August 2023 |
| Creators | Nassor, Mohammed |
| Contributors | McNamara, Derek A., Yagoub, Mustapha |
| Publisher | Université d'Ottawa / University of Ottawa |
| Source Sets | Université d’Ottawa |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | application/pdf |
Page generated in 0.0023 seconds