Group-blind turbo multiuser detection for CDMA using a Gaussian approximation

We introduce a simple iterative group-blind multiuser detection algorithm in an asynchronous multipath CDMA network in the presence of multiple-access interference (MAI) and ambient channel noise. The technique is based on the observation that the unknown multiple-access interference can be approximated as a Gaussian random vector. This simple model allows us to estimate the covariance matrix of the MAI signal without having knowledge of the number of unknown users. Then, in the CDMA uplink, we employ the proposed noise variance estimator for turbo multiuser detection in the presence of unknown interfering users and also unknown ambient noise. Simulation results show that the proposed group-blind technique offers significant performance improvement over the subspace soft interference cancellation method in a CDMA uplink environment.