Adaptive Precoding Mode Selection Based on Triple-CSI for mmWave Massive MIMO Systems

To enhance spectral efficiency, massive MIMO, hybrid precoding (HP) and non-orthogonal multiple access (NOMA), etc. are emerged. However, these enablers need to match with their appropriate application scenarios to achieve expected performances. To this end, this paper proposes an adaptive precoding mode selection (APMS) approach for millimeter wave (mmWave) systems. Firstly, we explore the criteria for precoding mode selection and adaptively group users based on proposed triple-channel state information (triple-CSI) that consists of three parameters. Then, phase-minimum-mean-square error (PMMSE) algorithm is proposed to optimize the sum-rate of HP, and equal-gain-transmission (EGT) method is used for NOMA. Simulation results show that APMS significantly outperforms HP and NOMA. At 26 dBm transmission power and 64 physical antennas, the sum-rate of AMPS is improved by 28.2% and 38.0% compared to HP and NOMA, respectively.