Performance analysis of the V-BLAST algorithm: an analytical approach

• Published in: IEEE transactions on wireless communications Vol. 3; no. 4; pp. 1326 - 1337
• Main Authors: Loyka, S., Gagnon, F.
• Format: Journal Article
• Language: English
• Published: Piscataway, NJ IEEE 01.07.2004; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithm design and analysis; Algorithms; Analytical models; Antennas; Applied sciences; Correlation; Detection, estimation, filtering, equalization, prediction; Diversity reception; Exact sciences and technology; Exact solutions; Information, signal and communications theory; Mathematical analysis; Optimization; Order disorder; Performance analysis; Receiving antennas; Signal analysis; Signal and communications theory; Signal processing; Signal to noise ratio; Signal, noise; Statistical analysis; Systems, networks and services of telecommunications; Telecommunications; Telecommunications and information theory; Transmission and modulation (techniques and equipments); Transmitting antennas; Fading channels; MIMO system; Outage; Coding; Bit error rate; Wireless telecommunication; Analytical method; Space time; Performance analysis
• ISSN: 1536-1276; 1558-2248
• DOI: 10.1109/TWC.2004.830853

A geometrically based analytical approach to the performance analysis of the V-BLAST algorithm is presented in this paper, which is based on the analytical model of the Gramm-Schmidt process. This approach presents a new geometrical view of the V-BLAST and explains some of its properties in a complete and rigorous form, including a statistical analysis of postprocessing signal-to-noise ratios for a 2/spl times/n system (where n is the number of receive antennas). Closed-form analytical expressions of the vector signal at ith processing step and its power are presented. A rigorous proof that the diversity order at ith step (without optimal ordering) is (n-m+i) is given (where m is the number of transmit antennas). It is shown that the optimal ordering is based on the least correlation criterion and that the after-processing signal power is determined by the channel correlation matrices in a fashion similar to the channel capacity. Closed-form analytical expressions are derived for outage probabilities and average BER of a 2/spl times/n system. The effect of the optimal ordering is shown to be to increase the first step SNR by 3 dB (rather than to increase the diversity order as one might intuitively expect based on the selection combining argument) and to increase the second step outage probability twice.