Electromagnetic (EM) simulation is ubiquitous in contemporary antenna design process. For many structures, including ultrawideband or dielectric resonator antennas, EM-simulation-driven optimization is the only way to adjust the geometry parameters so that given performance specifications are satisfied. On the other hand, accurate full-wave antenna simulation is computationally expensive so that employing the EM solver directly in the optimization loop may be impractical. In this paper, several computationally efficient simulation-driven antenna design techniques are discussed. All of the methods exploit coarse-discretization EM models of the structures under consideration. After suitable correction, these models serve as reliable prediction tools that guide the optimization process. As the coarse-discretization model is computationally much cheaper than the original, high-fidelity antenna model, the cost of the design process is greatly reduced. The specific approaches presented here include multi-fidelity optimization, adaptive design specifications and space mapping with kriging-based coarse models. Application examples are given.
展开▼
