Low complexity channel decomposition for spatial multiplexing MIMO systems

Spatial multiplexing is a common technique for multiple-input multiple-output (MIMO) communications that independent information streams are sent over different antennas. The precoder design based on the channel state information (CSI) can process the transmitting symbol before transmission. However, the spatial multiplexing is still sensitive to the ill-conditioning of the channel matrix. The conventional beam directions in the precoder are designed via singular value decomposition (SVD) which diagonalize the MIMO channel matrix H into eigenmodes. In this paper, we present a different approach using LU decomposition (LUD) to obtain the beam directions. We show that the LUD yields better channel condition and hence better error performance. The LUD-based precoder also has less feedback channel bandwidth requirement and lower complexity than the conventional SVD-based precoder.