The influence of spreading sequences on iterative multiuser detectors

For asynchronous direct-sequence code-division multiple-access (DS-CDMA) communications, a number of suboptimal multiuser detectors have been proposed, most of them can either be characterized as linear multiuser detectors or as interference cancellation detectors, which are generally non-linear detectors. Previously, it was shown that linear multiuser detectors can be implemented by linear interference cancellation detectors with ideally an infinite number of stages. Thus, the complexity of one stage of both types of detectors, whether linear or non-linear, is essentially the same. The complexity of the whole detector is then given by the number of stages, which are needed to achieve convergence or to achieve a desired performance. In this paper, we investigate the influence of the spreading sequences on the convergence speed and the performance of the iterative algorithms. It is found that the convergence speed and the performance highly depend on the choice of the sequences and that both can be improved by using AIP-optimized sequences instead of Gold or random sequences.