Power Allocation for Energy Efficiency Optimization in Multi-User mmWave-NOMA System With Hybrid Precoding

Recently, the non-orthogonal multiple access (NOMA) scheme has been applied in millimeter-wave (mmWave) communication system to support more users and further improve the performance. In this paper, an energy-efficient power allocation (PA) scheme is designed for a downlink multi-user mmWave-NOMA system with hybrid precoding (HP), where two typical HP architectures, namely the fully-connected HP architecture (FHPA) and the sub-connected HP architecture (SHPA), are both considered. Firstly, we pair users in term of their channel difference and correlation. Then, analog beamforming schemes are proposed for the system with fully-connected and sub-connected HP architectures, respectively. Based on this, a two-step HP design with the proposed analog beamforming and zero-forcing precoding is presented. With these results, the optimization problem is formulated to maximize the energy efficiency (EE) of the system. This is a non-convex optimization problem, and can be approximately decomposed into independent convex sub-problems by applying the fractional programming theory. Using the coordinate descent method, the closed-form solutions of each sub-problem are derived. On these basis, an effective iterative algorithm is proposed to obtain the suboptimal power allocation. Simulation results verify the effectiveness of the proposed PA scheme, it has the same EE performance as the existing scheme with relatively low complexity, and can obtain the EE close to the exhaustive search scheme as well as the chaotic accelerated particle swarm optimization scheme. Moreover, the proposed analog beamforming obviously outperforms the conventional finite resolution analog beamforming, and the system with SHPA has higher EE than that with FHPA.