Prediction of the far field radiated by a flat antenna under test from a reduced set of near-field bi-polar measurements

In this article, a bi-polar near-field-to-far-field (NFTFF) transformation with probe compensation, which is very convenient from the data reduction standpoint when characterising a flat antenna under test (AUT), is developed. It takes advantage from the non-redundant sampling representations of the electromagnetic fields to devise a sampling representation on the scanning plane, which requires the knowledge of the bi-polar NF samples at a reduced number of sampling points. To account for the AUT geometry, it properly adopts as modelling surface a disc having diameter equal to the largest size of such an AUT. This surface allows a more effective AUT modelling from the NF data reduction standpoint than the formerly proposed modelling surfaces for quasi-planar antennas (the oblate ellipsoid or the two-bowls), because, unlike these last, it has the capability to best fit the AUT geometry, thus reducing very significantly the involved volumetric redundancy. An ad hoc developed two-dimensional optimal sampling, interpolation formula is then employed to reconstruct the plane-rectangular NF data which would be necessary to execute the classical NFTFF transformation developed by Leach and Paris. Some results of numerical simulations and experimental proofs are reported to demonstrate the efficiency of the so developed technique.