Fast Antenna Diagnosis Algorithm Using Oblate Spheroidal Nonuniform Grids

A fast algorithm for the diagnosis of large antennas is developed. The algorithm is based on the Rayleigh-Sommerfeld (RS) backpropagation method, accelerated by a multilevel nonuniform grid algorithm. Comparison between the field distributions reconstructed from measurements and the desired ones can be used for the localization of antenna defects and geometrical distortions. The optimal performance is obtained by using oblate spheroidal grids, which are tailored to both the measurement and antenna surfaces. Further acceleration can be achieved, for particular cases, via optimal sampling on the antenna surface. The algorithm's performance is demonstrated for the example of a parabolic reflector with a number of surface distortions, and compared to that achieved with spherical grids. The accuracy and error controllability are studied in comparison to the direct RS integral computation.