Precoder design for beamforming in K-User MIMO interference chennel

Beamforming in MIMO wireless networks, is a techniques used for mitigating interference and achieve the best possible throughput. Accurate channel state information is essential for efficient design of beamforming vectors required for data transmission. According to [1], if accurate channel information is available, precoding vector design problem could be reduced to a problem of Generalised Eigen Vector computation for a matrix pencil. In this thesis, it is assumed that the channel state information available is incomplete. Thus, the eigen vector computation for precoding vector design is that of a noisy matrix pencil. In [2], some SVD based techniques have been explored for eigen value and eigen vector compuatation of noisy matrix pencils, in [2], performance of these methods has been analysed assuming that the matrix pencil involved has a know structure. This thesis analyses the performance of the SVD methods for the noisy matrix pencils that are involved in K-User MIMO networks (such a matrix pencil would not have a defined structure when the channel state information available is assumed to be incomplete).