Triangulation-Based Constrained Surrogate Modeling of Antennas

Design of contemporary antenna structures is heavily based on full-wave electromagnetic (EM) simulation tools. They provide accuracy but are CPU intensive. Reduction of EM-driven design procedure cost can be achieved by using fast replacement models (surrogates). Unfortunately, standard modeling techniques are unable to ensure sufficient predictive power for real-world antenna structures (multiple parameters, wide parameter ranges, and highly nonlinear responses). Here, a design-oriented modeling technique is introduced, in which the most critical part is the domain definition. The surrogate is constructed in the region based on a set of reference designs, optimized with respect to user-selected figures of interest such as operating frequencies, bandwidth, and substrate permittivity. The domain is spanned by the simplexes, obtained by triangulation of the reference designs, further extended into their orthogonal complements. Restricting the model domain as above permits dramatic reduction of the number of training data samples necessary to build a reliable model, as compared to the conventional approach. The proposed modeling framework is demonstrated using three examples, a UWB monopole, and two dual-band antennas. Comprehensive benchmarking as well as application studies and experimental verification are also provided.