Energy-Efficient Hybrid Precoding With a Grouped-Adaptive-Connected Structure

In this paper, we propose a grouped-adaptive-connected (GAC) hybrid precoding structure, where the antennas are divided into groups and the connections between each group of antennas and each RF chain are dynamically controlled by RF switches. Compared with the fully-adaptive-connected (FAC) structure, the proposed GAC structure requires fewer RF switches. The problem of maximizing energy efficiency (EE) is transformed into an adaptive spectral efficiency (SE) maximization (ASEM) problem and a combinatorial optimization problem. For the ASEM problem, we propose an equivalent baseband channel design based hybrid precoding (EBCD-HP) algorithm. Then, utilizing the solution of the ASEM problem, we propose a cross-entropy based grouped-adaptive hybrid precoding (CE-GAHP) algorithm. Simulation results reveal the advantages of the proposed hybrid precoding scheme.