Abstract
Generalized power iteration-based precoding (GPIP) exhibits superior performance in maximizing spectral efficiency for multi-antenna communications. Increasing the number of antennas, however, makes GPIP less practical due to its cubic complexity with respect to the number of antennas in computing matrix inversion. To resolve this, we propose efficient GPIP methods with imperfect channel knowledge. Exploiting the GPIP matrix structure, we use two low-complexity matrix inversion approaches: the matrix inversion lemma-based low-dimension matrix inversion and Gauss-Seidel iteration. These approaches decrease complexity order to the square of the number of antennas. Simulations validate the computational efficiency of the proposed methods while maintaining performance.