Researches On Low-complexity Beamforming And Hybrid Precoding Techniques In Mmwave Massive MIMO System

The combination of millimeter-wave(mmWave)and massive multiple-input multiple-output(MIMO)systems can provide orders of magnitude increase both in available bandwidth and spectrum efficiency which is considered as a prospective technique for future wireless communication system.However,due to the high power consumption of the radio frequency link in the millimeter wave band,power consumption has become a major bottleneck for millimeter wave massive MIMO.In order to reduce the number of RF chains required for millimeter-wave massive MIMO,we focus on the structure of low-frequency transceivers in this paper.Firstly,analog beamforming can reduce the number of expensive radio frequency(RF)chains required by base stations(BS)and users,which is critical to reducing the energy consumption and hardware complexity of millimeter-wave massive MIMO systems.Among them,codebook-based analog beamforming is usually used to compensate for severe attenuation of millimeter wave signals.However,traditional beamforming schemes involve complex searches between predefined codebooks.In order to reduce the search complexity and maximize the available rate of the system,we propose an analog beamforming based on IC-PSO using the improved particle swarm optimization algorithm to more intelligently determine the analog precoder and analog combiner.Simulation results show that the proposed algorithm can achieve near-optimal performance with much lower complexity compared with the optimal full search(FS)algorithm.In order to cope with multi-stream transmission and further improve energy efficiency of the system,researchers have proposed a hybrid precoding that can reach an effective compromise between hardware complexity and system performance.Therefore,we study the hybrid precoding structure of multi-stream transmission,and further proposes a multi-user hybrid precoding structure based on MPBIL.This paper introduces the improved PBIL algorithm to determine the hybrid precoder in a new millimeter-wave massive MIMO system architecture based on switches and inverters(SI).Simulation results show that compared with other traditional hybrid precodings and other existing machine learning algorithms,the proposed MPBIL hybrid precoding with large-scale antenna array has lower computational complexity and higher energy efficiency.