Design of Short-Circuited Ring-Patch Antennas Working at $hboxTM_01$ Mode Based on Neural Networks

Short-circuited ring-patch (SCRP) antennas provide some advantages derived from their property of having as fundamental mode the TM
₀₁
mode. This mode has a monopolar radiation pattern with a null in the zenith, in contrast with the radiation patterns of fundamental mode for circular and ring patches (the TM
₁₁
mode) that are broadside. The lack of a good characterization of this geometry motivates the present work. In it, an exhaustive study of SCRP antennas via simulations has been performed. The dependence between the fundamental antenna parameters: the internal radius (r), the external radius (R), the permittivity of the substrate (epsi) and its thickness (h); has been studied. Finally, a neural network has been proposed for the analysis and design of these antennas. This has demonstrated to be a useful design tool that provides advantages in computation time with a negligible error in real terms. The simulations have been validated with the manufacturing and test of an antenna