Slowing Manhattan MIMO Channels

This paper discusses the application of a beam forming algorithm that slows channel variations to measured MIMO channels. A measurement campaign is described which surveyed a section of Manhattan from a low base. The data were processed to obtain angular spectra. The spectra were then used to generate synthetic channels for algorithm evaluation. With 4 antennas the beam forming approach was found to improve channel coherence time by a factor of 2. Receiver correlation was increased, and the channel capacity reduced slightly. The slowed channel may mitigate the effect of feedback delay or compensate for the sparseness of pilots in cellular systems. Introduction