Two approaches to get efficient NF/FF transformations from inaccurately positioned planar wide-mesh scanning samples

This paper aims to give the experimental proof of two effective procedures for the compensation of severe and pessimistic positioning errors, which alter the gathered near-field samples in a near-field to far-field (NF/FF) transformation with planar wide-mesh scanning (PWMS). This scanning is so called since it has a sampling lattice whose meshes are wider and wider as their distance from the measurement plane centre grows, so that it allows a drastic reduction of the acquisition time as compared to the classical plane-rectangular (PR) scan. The here considered NF/FF transformation with PWMS relies upon a nonredundant sampling representation of the voltage detected by the scanning probe, attained by modelling the antenna under test with a double bowl, that is, a surface got by joining together two bowls with the same circular aperture, and exploits an optimal sampling interpolation formula to precisely reconstruct, from the gathered PWMS samples, the huge NF data required by the PR NF/FF transformation. The retrieval of the NF samples at the points prescribed by the representation is got from the collected PWMS samples, corrupted by known positioning errors, by employing either a singular value decomposition based procedure or an iterative algorithm. Some experimental results, which fully assess the effectiveness of both the developed procedures, are shown.