Research On Linear Precoding For Downlink Multi-User Generalized Spatial Modulation Systems

Generalized Spatial Modulation(GSM)is one of the key technologies of future ultra-high speed wireless communication system because of its high throughput and high energy efficiency.Applying GSM to Multi-User Multiple-Input Multiple-Output(MU-MIMO)systems can improve time and frequency resource utilization.However,the Multi-User Interference(MUI)will degrade the performance of Multi-User Generalized Spatial Modulation(MU-GSM)system.At present,most of the precoding schemes are dedicated to eliminating the interference among signals.This paper conducts a deep research on the linear precoding techniques in MU-GSM system by leveraging interference.The main work is as follows:1.The simulation on point-to-point GSM system shows that it can achieve higher rates than that of Vertical Bell Labs Layered Space-Time(V-BLAST)system with less RF chains.Then we analyze the Symbol Error Rate(SER)performance of GSM by applying order statistcis.It is observed that the deduced SER can give an accurate error estimation of the GSM system.2.We group the transmitting antennas to model a single-cell downlink MU-GSM system with high spectral efficiency.When the GSM system directly uses the traditional linear precoding methods to eliminate the interference signal,unique sparseness ofthe channel gain matrix and the transmitted signal matrix in GSM system will lead to the loss of sending information.Therefore,we propose the virtual transform matrix,which can extract active antennas' information,to construct an equivalent virtual MU-MIMO system for solving this problem.On this basis,two linear precoding methods which can elimit interference of MU-GSM system are proposed.Further through including the idea on interference classification and establishing interference judgement matrix,we propose two kinds of linear precoding methods which can utilize useful interference of single-cell MU-GSM system,namely Multi-User Generalized Spatial Partial Linear Precoding(MU-GSM-PLP)and Multi-User Generalized Spatial Partial Minimum Mean Square Error Partial Linear Precoding(MU-GSM-MMSE-PLP).The simulation results show that the linear precoding methods which leverages interference can provide better performance with low Bits Error Ratio(BER)and high sum-rate.3.In order to further utilize the interference,the Phase Alignment Linear Precoding(PALP)method is applied to the single-cell downlink MU-GSM system,thus we propose two linear precoding methods,the Multi-User Generalized Spatial Phase Alignment Linear Precoding(MU-GSM-PALP)and the Multi-User Generalized Spatial Minimum Mean Square Error Phase Alignment Linear Precoding(MU-GSM-MMSE-PALP).They use the phase correction matrix to make all the interference into useful interference which can provide useful energy for the target symbol.Therefore,with two linear precoding methods,we can further improve the sum-rate of the single-cell MU-GSM system and reduce BER of the single-cell MU-GSM system.Monte Carlo simulation is used to verify its superiority in this paper.