Blind Multiuser Detection: from MOE to Subspace Methods

ZY Xu,P Liu,XD Wang

DOI: https://doi.org/10.1109/tsp.2003.821111

IF: 4.875

2004-01-01

IEEE Transactions on Signal Processing

Abstract:The minimum output energy (MOE) multiuser receiver has been shown to approach the minimum mean-square-error (MMSE) receiver at high signal-to-noise ratio (SNR). However, performance degradation is incurred by noise induced channel estimation error. In this paper, we propose a Power of R (POR) technique to significantly improve the performance of the MOE receiver. It is shown that the new receiver asymptotically converges to the MMSE receiver without performance penalty. The convergence is established either under high SNR, with large exponent raised in the power of the covariance matrix, or with sufficiently large number of data samples. Connection between our POR method and a widely studied subspace method is investigated from the respective optimization criteria. Asymptotic equivalence between these two methods is also established. Extensive simulations based on finite data samples show that the proposed method significantly outperforms the subspace method in systems with medium to heavy loading, severe multipath distortion, or smaller processing gain. Moreover, adaptive implementation of the proposed method exhibits very robust performance in a dynamic loading environment.