V-BLAST: Performance comparison of detectors, improved high speed detectors, and interference cancellation ordering effect for high data rate MIMO systems

Abstract

V-BLAST, a MIMO wireless communication system can achieve very high spectral efficiency in a rich multipath environment through exploiting the extra space dimension. In this thesis, firstly, various V-BLAST algorithms, such as the linear Zero Forcing (ZF) algorithm, the linear Minimum Mean Squared Error (MMSE) algorithm, the Interference Cancellation (IC) algorithm with ZF, the square-root IC (SRIC) algorithm with ZF, SRIC with MMSE, the Pre-match filtering (pre-MF) IC with ZF, are introduced and compared through computer simulations. The SRIC algorithms show very attractive property that having low complexity while keeping the good performance compared with other algorithms. Next, an improved SRIC algorithm and an improved recursive algorithm are proposed. The improved SRIC algorithm speeds up the original SRIC algorithm by 36% in the number of multiplications and additions. It maintains the advantages of the numerically stability, robustness, and hardware friendliness. Whereas the improved recursive algorithm speedups the original recursive algorithm further by 47% in the number of multiplications and 20% in the number of additions. Last the ordering effect of IC algorithms is evaluated in three different V-BLAST systems. The optimal ordering scheme outperforms the fixed ordering scheme in the conventional V-BLAST system and the system with simple antenna selection. The performance gap is undistinguishable in the V-BLAST system with both antenna selection and adaptive modulation.