Vinsamlegast notið þetta auðkenni þegar þið vitnið til verksins eða tengið í það: http://hdl.handle.net/1946/28769
In this thesis, novel approaches to miniaturization of ultra-wideband antennas using trust region method are presented. We investigate the properties of design task formulation as an explicit size-reduction problem, with the footprint area minimization being a primary design objective and a penalty function introduced to control reflection characteristic of the antenna at hand. An alternative approach using the trust region framework with quadratic approximation model is demonstrated to increase control over the maximum in-band reflection level (compared to conventional benchmark method) but it does not decrease the final footprint area size. The second approach, using adaptive adjustment of the maximum acceptable in-band reflection level threshold is proposed and demonstrated to be superior over the conventional (fixed-threshold) setup. Finally, an iterative algorithm for exploring boundary of the feasible region is developed, allowing for efficient identification of a constrained optimum that ensures satisfaction of the prescribed matching requirements for the structure at hand. All methodologies are demonstrated using several case studies of ultra-wideband antennas.
|Novel Algorithmic Approaches for Simulation-Driven Miniaturization of Antenna Structures.pdf||1 MB||Opinn||Heildartexti||Skoða/Opna|