Asymptotic performance analysis of V-BLAST

In this paper, we present an asymptotic analysis of the V-BLAST scheme at high signal-to-noise ratio (SNR) region. We consider point-to-point MIMO communications over an i.i.d. Rayleigh flat fading channel with n transmitting antennas and m (m ges n) receiving antennas. Both the zero-forcing V-BLAST (ZF-V-BLAST) and minimum mean-squared-error V-BLAST (MMSE-V-BLAST) are analyzed with respect to their diversity gains and BER performances. We show that the diversity gain of V-BLAST, including ZF-V-BLAST and MMSE-V-BLAST, with optimal ordering is m - n + 1, i.e. applying the optimal ordering technique does not improve the diversity gain. Contrary to the common perception that the MMSE and ZF estimators have asymptotically the same post-processing SNR for high input SNR, we show that the difference between the post-processing SNRs of the two estimators does not vanish for high SNR. We also quantify the remarkable BER performance advantage of the MMSE-V-BLAST over the ZF-V-BLAST for high SNR