Reaserch On The Hybrid Precoding Algorithm In Millimeter Wave Communication System

As one of the most possible technologies of the next generation communication system,millimeter wave(mmWave)communication system can provide lots of new spectrum resources and many advantages,such as large capacity and high transmission rate.In the mmWave systems,path loss is serious due to the short wave length.Although this problem can be partly resolved by beamforming using large antenna array,the traditional digital precoding method needs a RF chain implemented after each antenna.It means that large number of RF chains needed to be used in mmWave systems,which causes much hardware and power cost.The number of RF chains can be greatly reduced by the hybrid precoding method,where the precoding is performed in both digital and analog domain.Therefore,hybrid precoding algorithm of millimeter wave communication systems is very important in theoretical research and application value.Based on the new application,hybrid precoding algorithms in the millimeter wave communication systems is studied and designed in this thesis.Firstly,multiple data streams transmitted in a millimeter wave communication system leads to interference between data streams.Due to this problem,the zero-forcing hybrid precoding method based on QR-decomposition is proposed in the thesis.The proposed hybrid precoder is designed by reconstructing traditional ZF precoding matrix through QR-decomposition and using the spatial sparse characteristics of mmWave channel to process the transmitted signals.The simulation results show that the proposed method can reduce the interference between data streams,and the spectrum efficiency performance is closer to the optimal digital ZF precoding method compared to the existing hybrid precoding method.Secondly,aiming at the serious interference between users in millimeter wave multiuser MIMO system,a spatial sparse precoding method based on block diagonalization is designed by converting a hybrid precoder designing problem to a weighted sparse signal reconstruction one.Simulation results show that the proposed method can eliminate the interference between users,and the sum rate performance of the proposed algorithm is close to the full digital BD precoding method.Finally,when large number of users is involved in the multiuser system,the precoding matrix is rank deficient.Due to this problem,a spatial sparse precoding method based on user selection is designed.This algorithm aims at the optimal system capacity performance,greedy method is used to choose the subsets of users for communication and design hybrid precoder.The simulation results show that the proposed method reduces the complexity of algorithm and can obtain better sum rate performance.