Efficient Beam Pattern Modeling for Reflector Antennas Within Multidimensional Feed Parameter Spaces Using a Reference Design

This paper presents an efficient beam pattern modeling method for reflector antennas where the performance of a specific feed antenna, as a function of frequency and geometric variations of the antenna, is known on a different (but similar) reflector antenna. Instead of modeling patterns for the desired reflector directly, the difference patterns between the desired and previously modeled reflectors are employed. Scaling and normalization measures are applied to the patterns to simplify their modeling as Characteristic Basis Function Patterns (CBFPs), which can be accurately interpolated over the desired design parameter space. The CBFP coefficients drive an adaptive sampling algorithm that aims to accurately obtain the antenna’s performance across the entire parameter space with as few electromagnetic simulations as possible. Two examples are presented using basic reflector optics and a simple choke horn and demonstrate the advantage of using a reference design compared to directly modeling the reflector patterns, showing that fewer simulations are required to obtain a certain level of accuracy.