Polarization in space-time processing for propagation channel identification

We propose to estimate the multipath channel parameters using a space-time subspace-based approach with diversely polarized antennas. The diversity of the responses provided by such arrays can improve the directions and delays estimation. Our method exploits the polarized space-time steering vector. We extend the space-time MUSIC pseudo-spectrum by using the eigenvalues of a two dimensional matrix obtained by projections onto noise subspace. We first establish performances of the space-time method and compare the polarized approach with the conventional space-time MUSIC algorithm. Statistical performances are investigated and compared to the Cramer Rao bound.