Performance of a polarization isolation enhancement method in the presence of errors

A radar measurement system designed to obtain complete backscatter polarization matrix data requires an antenna with feed structures that excite two orthogonal polarizations (or two orthogonally polarized antennas for quasi-backscatter measurement). Unfortunately, due to construction tolerances and the laws of electromagnetics, there will be coupling between the polarization excitation ports. The result is that the actual polarizations produced are really not orthogonal. However, it is possible to characterize the coupling in such a way that it can be cancelled to a degree through a calibration procedure. Such a procedure has been developed for a linearly polarized system. The author examines the effect of measurement error on the successful application of that method. The results show that while the technique will certainly provide degraded performance in the presence of errors, useful improvement in polarization isolation can be obtained under practical measurement conditions.<>