Research On Downlink Precoding Technology In Massive MIMO Systems

Massive multiple input multiple output(MIMO)technology,as the core and key technology in the fifth generation wireless mobile communication system(5G),can increasing data transmission rate and improve system link reliability,so that this technology can significantly improve the performance of communication system.Precoding technology can pre-process the transmitted data stream by using the channel state information known at the base station to increase array gain and eliminate interference.It has become a key technology in massive MIMO systems.In this paper,downlink precoding technology for massive MIMO systems is studied,and hybrid precoding algorithm is designed and simulated based on millimeter-wave channel.The main contributions of this paper are as follows:1.The main basic theories of massive MIMO systems are studied,including millimeterwave channel model,channel estimation,the downlink system model,sum rate of massive MIMO system,which lays a theoretical foundation for the full study.2.The precoding technology in massive MIMO systems is studied from three aspects: digital precoding,analog precoding and hybrid precoding.The downlink system model of massive MIMO system which using hybrid precoding algorithm is studied,and the design of hybrid precoding is preliminarily studied.Several typical digital precoding algorithms are simulated and analyzed by using Matlab.3.A hybrid ZF precoding algorithm for massive MIMO systems is studied.In the radio frequency domain,equal gain transmission(EGT)algorithm and phase quantization are used to obtain large array gain;in the digital baseband domain,successive over relaxation(SOR)iteration is used to achieve zero forcing(ZF)precoding with low computational complexity.The Matlab simulation analysis is carried out and the hybrid ZF precoding is compared with full digital ZF precoding.The simulation results show that the hybrid ZF precoding algorithm can approximate the performance of full digital ZF precoding algorithm.Moreover,the complexity of the algorithm and the cost of hardware are greatly reduced.The simulation analysis also shows that the higher the precision of the hybrid ZF precoding quantization,the better the performance of the precoding algorithm.4.A hybrid block diagonalization(BD)precoding algorithm for massive MIMO systems is studied,which is more suitable for multi-antenna and multi-stream at both base station and mobile station scenarios.The precoder is designed simultaneously with the receiver combiner,and the analog precoding matrix is designed in the RF domain by EGT algorithm,and the analog combiner matrix is solved by DFT traversal.In the baseband digital domain,the equivalent baseband matrix is defined,and the equivalent baseband matrix is decomposed by two-step SVD to obtain the baseband precoding matrix and the baseband combiner for each user.Finally,the proportional water-filling method is used to distribute power among users to maximize the sum rate of system.In this chapter,the hybrid BD precoding algorithm is simulated and compared with the full-digital BD precoding algorithm.The simulation results show that the hybrid BD precoding algorithm has good performance,and slightly better than the full-digital BD precoding algorithm.