A Transmitter-Based Beamforming Scheme for the MIMO Downlink Employing Adaptive Channel Decomposition

An extension to decomposition based transmit beamforming is proposed for multiple input multiple output (MIMO) multiuser systems. In contrast to conventional beamforming which is constrained towards orthogonalizing the subchannels of the different users to yield groups of correlated antennas that belong to only one user, a more flexible orthogonalization approach is proposed which is targeted directly towards the optimization of the decision variables at the multi-antenna receivers of all users. The resulting groups of correlated antennas may consist of elements belonging to multiple users. Thus, the proposed beamforming has more degrees of freedom and therefore attains a more efficient performance optimization. The trade-off to the performance improvement is an increase in the precoding complexity imposed by the adaptive nature of the proposed beamforming. A suboptimal adaptive-decomposition beamforming scheme is therefore also proposed with a reduced complexity overhead. Comparative simulations of the proposed scheme to conventional beamforming show considerable error rate reductions that justify the increased complexity and verify the superiority of the proposed scheme.