Performance Analysis of Spatially Correlated MIMO-OFDM Beamforming Systems with the Maximum Eigenvalue Model from Measured MIMO Channels

The performance of MIMO-OFDM beamforming systems with maximal ratio combining receivers is considered based on the extensive wideband 2 x 8 (i.e., 2 transmitting and 8 receiving antennas) MIMO channel measurement data in non line-of-sight urban macrocell environments. Using the experimental observation that the probability density function (PDF) of the measured maximum eigenvalue for H-m(dagger) H-m, where H-m and superscript (dagger) denote a MIMO channel frequency response matrix and conjugation and transposition, respectively, matches the Gamma distribution well, we propose the maximum eigenvalue model as Gamma distributed random variables whose parameters are both transmitting and receiving correlations and the number of receiving antennas. Using the model, the performances of MIMO-OFDM beamforming systems, such as the PDF of the output signal-to-noise ratio, outage probability, symbol error rate, and achievable diversity order, are analyzed to demonstrate the effect of spatial correlation in the realistic 2 x n (downlink) and n x 2 (uplink) MIMO channels.