Robust transmit eigen-beamforming with imperfect knowledge of channel correlations

Transmit beamforming is a powerful approach to enhance the performance of wireless communication systems employing multiple antennas at the transmitter. Using channel covariance information, transmit beamforming has been pursued based on various criteria. A major drawback of existing techniques is that most of them require nearly perfect knowledge of the channel covariance matrix at the transmitter, which is not necessarily possible in practice. In this paper, we propose a more robust framework in which we take this issue into account and design a robust transmit beamformer to have the best performance under the worst-case mismatch. We show that, for small mismatches, eigen-beamforming along the eigenvectors of the presumed channel correlation matrix offers the best worst-case performance. We also combine eigen-beamforming with robust power loading that is achieved by spatial water-filling-type strategy in which the water level is determined in a simple form.