Energy-Efficiency Optimization for Multi-User Multi-stream Massive MIMO Hybrid Precoding

Millimeter-wave (mmWave) massive multi-input multi-output (MIMO) has attracted significant attention for 5G communications. In this paper, a fully-connected hybrid architecture supporting multiple streams per user is considered using a single cell downlink multi-user massive MIMO system. To effectively suppress the inter-user interference and inter-stream interference, the block diagonalization algorithm (BD) is combined with phase quantization to solve the optimal analog precoding and analog combining. Meanwhile, to avoid the high computational cost, the Dinkelbach method and weighted minimum mean square error (WMMSE) are adopted to solve optimal baseband precoding and combining. Simulation results show that the proposed EE model is capable of minimizing the bit error rate (BER) and improving the spectrum efficiency and energy efficiency (EE) of the mmWave massive MIMO system.

Automated summary

The paper introduces a fully-connected hybrid architecture for a single-cell downlink multi-user massive MIMO system, which effectively suppresses inter-user interference and inter-stream interference. Simulation results demonstrate the effectiveness of the proposed method in minimizing bit error rate and improving spectrum efficiency and energy efficiency in mmWave massive MIMO systems.