A Novel Low-complexity Sub-block Nlms-drmta Algorithm for Blind Multiuser Detection in Wcdma Systems

In this paper, we present a novel adaptive beamforming algorithm for multiuser detection in wideband code division multiple access (WCDMA) systems. The algorithm is based on a new sub-block normalized least mean squares optimization in conjunction with the despread-respread technique. We denote this new algorithm using the acronym Sub-Block NLMS-DRMTA, which stands for sub-block normalized least mean squares despread-respread multitarget array. A novel adaptive step size rule has been proposed based on the variance of the array input signal, which has resulted in enhanced robustness to estimation errors and very low computational complexity. The proposed algorithm has been benchmarked against exiting blockbased and sample-based DRMTA algorithms. The bit error rate (BER) performance of our algorithm outperformed the performance of the block-based least squares despread-respread multitarget constant modulus array (LS-DRMTCMA) algorithm by a wide margin. In comparison to the sample-based (Sample NLMS-DRMTA) algorithm, the proposed algorithm achieved superior or similar BER performance at a markedly reduced computational complexity and significantly faster convergence. In addition, the proposed algorithm exhibited significantly better robustness compared to that of the sample-based recursive least squares (Sample RLS-DRMTA) algorithm at low signal-to-noise ratio (SNR) and/or high number of users. Moreover, the proposed algorithm has substantially lower complexity than the Sample RLS-DRMTA algorithm. Thus, the Sub-Block NLMS-DRMTA algorithm represents a viable approach for achieving both improved BER performance and fast convergence at very low computational complexity.