Transmission optimization over flat Rayleigh fading channel with multiple antennas

Spatial diversity can be exploited to increase the point-to-point information capacity enormously by deploying multiple antennas for both the transmitter and receiver in a rich scattering fading channel. Assuming that perfect channel state information (CSI) is available at both ends of such a communication link, we derive a spatial shaping matrix by maximizing mutual information of the multi-antenna fading channel subject to a power constraint on each transmission antenna. The maximization is carried out using a polynomial time interior-point algorithm with linear matrix inequality constraint (LMI). The spatially shaped matrix channel is further converted into a set of independent fading subchannels through an eigenvector driving transmission. The proposed scheme facilitates the use of 1-dimensional code to achieve diversity coding gain.