Energy Efficiency Optimization for Downlink Massive MIMO With Statistical CSIT

Abstract

We investigate energy efficiency (EE) optimization for single-cell massive\nmultiple-input multiple-output (MIMO) downlink transmission with only\nstatistical channel state information (CSI) available at the base station. We\nfirst show that beam domain transmission is favorable for energy efficiency in\nthe massive MIMO downlink, by deriving a closed-form solution for the\neigenvectors of the optimal transmit covariance matrix. With this conclusion,\nthe EE optimization problem is reduced to a real-valued power allocation\nproblem, which is much easier to tackle than the original large-dimensional\ncomplex matrix-valued precoding design problem. We further propose an iterative\nwater-filling-structured beam domain power allocation algorithm with low\ncomplexity and guaranteed convergence, exploiting the techniques from\nsequential optimization, fractional optimization, and random matrix theory.\nNumerical results demonstrate the near-optimal performance of our proposed\nstatistical CSI aided EE optimization approach.\n