Joint linear optimization with receiver constraint for MIMO systems

In this article, we propose a new linear multiple-input multiple-output (MIMO) system with the receiver matched to the transmit filter plus channel, leading to a linear transmission system which diagonalizes the MIMO channel into its eigenspace similar to joint transmit and receive filter optimization (joint TX/RX optimization). The proposed transmission scheme is shown to achieve comparable bit-error performance to joint TX/RX optimization in the case of adaptive modulation. Furthermore, the proposed scheme outperforms joint TX/RX optimization in terms of complexity: if neither the possibility of feedback, nor sufficient computational resources at both sides of the link are available, the joint TX/RX optimization approach is not applicable contrary to the proposed approach. We investigate different optimization criteria well known from linear precoder design and joint TX/RX optimization: maximizing the signal-to-noise ratio, removing the interference, and minimizing the mean square error. Closed-form solutions are derived for these cases.