Adaptive Frequency Sampling for Fast and Accurate Simulation of the Active Reflection Coefficient

The Active Reflection Coefficient (ARC) is a key factor in designing electrically small array antennas, as it accounts for the antenna's performance in a use-case. Accurate modeling of the ARC requires the complete set of scattering parameters, which requires a full-wave simulation. In this paper, we investigate how to accelerate full-wave frequency domain simulations by using the Loewner framework with adaptive frequency sampling, to accurately model the ARC. Here we compare interpolation on the ARC data directly with interpolation on the scattering parameters from which the corresponding ARC is determined, and interpolation on the scattering parameters, but with adaptive frequency sampling according to the estimated ARC error. The three strategies are tested on two cases, one <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$2 \times 2$</tex> IoT array, and one <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$10 \times 10$</tex> BoR Array. In both antenna cases, the latter strategy for accurately modeling the ARC yields the most accurate model for the least number of interpolation points used.