Spatial Diversity and Eigenspectra

We analyse the spatial diversity of a multipath fading process for a general geometrical base of a dense antenna array in the plane and characterise it by the eigenvalue spectrum of the spatial autocorrelation kernel. We show how this diversity spectrum can be calculated for any spatial fading statistics represented by the power azimuth spectrum (PAS). Numerically calculated diversity spectra help to optimise the geometry of the antenna array. They demonstrate that the diversity gain of a onedimensional uniform linear array may be very small for a PAS with small angular spread. Circular arrays turn out to be more robust. For a channel coded system, diversity spectra give useful hints to design the time interleaving depth which is necessary to exploit the diversity gain of the code.