Incorporation of sparse priors in long-code blind multiuser detection for DS-CDMA communication

This paper presents a new method of compensating for complex multipath channel fading in blind multiuser detection for long-code DS-CDMA communication. The system employs detectors that incorporate methods of independent component analysis (ICA). Such detectors are subject to performance degradation as the length of the user spreading code increases. When compensated through a class of sparse priors, the long-code ICA neural network detectors displayed significantly better bit error rate performance. Performance of the long-code detectors approached that of the short-code detectors in some cases