Optimal Linear Precoding with the Extreme Bounds of Eigenmode SNR Variation

Abstract

A linear precoding for block transmission system with multi transmit/receive antennas is considered. In the environment that exact channel state information can be available at the transmitter, optimal transmission can be achieved by updating the linear precoder at each transmission block. However, since a power loading algorithm based on water-filling requires the eigenvalues of channel matrix, the computational cost to obtain the linear precoder increases. In this paper, we propose a method of optimal linear precoding with the extreme bounds of the eigenmode signal-to-noise ratio (SNR) variation to reduce a computational cost. The applicability of this method is demonstrated in computational simulation.