Effective use of long-term transmit channel state information in multi-user MIMO communication systems

We propose a novel approach to the design of linear transmit processing for the downlink of multi-user multi-input multi-output (MU-MIMO) communication over fading wireless channels. Our purpose is to address the open problem of downlink transmit processing in MU-MIMO systems with only long-term CSI at the transmitter. In contrast to existing schemes, only long-term channel state information is required at the transmitter. For this to work, the MU-MIMO channel must experience correlated fading. The approach consists of two steps: 1) design of transmit processing which can deal with correlated fading channels but is allowed to use full CSI at first; 2) conversion of the transmit processing to use long-term CSI only. As a solution for the first step, we propose a joint receive and transmit mean square error minimization scheme, while the second step is performed by an innovative subspace-based procedure. Performance evaluation shows the significant potential of the proposed approach.