Polarimetric smoothing revisited: Applicability to Randomly polarized sources and to incomplete electromagnetic vector-sensors

Randomly or partially polarized sources occur frequently in practical applications such as radar and modern communications. In this paper, we revisit the polarimetric smoothing technique, which is previously customized for decorrelating completely polarized coherent signals with identically oriented complete electromagnetic (EM) vector-sensors, on its applicability to randomly polarized signals and to incomplete EM vector-sensors instead. It is shown here that polarimetric smoothing can be extended in a rather straightforward way for decorrelating randomly polarized source signals, using either complete or incomplete EM vector-sensors. However, when utilizing incomplete EM vector-sensors for completely polarized signals, polarimetric smoothing may fail to decorrelate signals that have distinct directions of arrival (DOAs) and/or polarizations. Some typical numerical results are also presented to confirm the main points made in the paper.