A cascaded constrained beamforming and multiuser detection using cyclostationarity

This paper addresses a simple, yet effective, cascade of the beamforming and the multiuser detection (MUD), where the beamformer (spatial filtering) first cancels out the interferences with directions of arrivals (DOAs) different from that of the desired user and the succeeding MUD (temporal filtering) eliminates the remaining interferences whose DOAs are the same as that of the desired user. Both stages are based on constrained cyclostationarity and, thus, the proposed scheme makes use of a priori information of users and, at the same time, is free of the training sequence and near-far resistant. Furnished simulations show that the new approach yields performance close to the previously addressed jointly spatial-temporal (S-T) processing structure, but with substantially reduced computational complexity.