Blind joint equalization and multiuser detection for DS-CDMA in unknown correlated noise

The problem of blind demodulation of multiuser information symbols in an asynchronous multipath CDMA network in the presence of multiple-access interference (MAI), intersymbol interference (ISI), and ambient channel noise is considered. First, a discrete-time signal model for a multipath code-division multiple-access channel is developed, based upon which several linear receiver structures are defined. Blind channel-estimation methods exploiting the orthogonality between the signal and noise subspaces are then discussed. In particular, channel estimation in the presence of unknown correlated noise is considered. It is assumed that the receiver employs two well-separated antennas, such that the noise is spatially uncorrelated. Two subspace methods for estimating the channel response are examined, based on, respectively, the singular value decomposition (SVD) and the canonical correlation decomposition (CCD) of the cross-correlation matrix of the signals received by the two antennas. Simulation results show that the CCD-based method substantially outperforms the SVD-based method.