Research On User Clustering And Precoding In Millimeter Wave MIMO-NOMA System

As the rapid development of mobile Internet and Io T in recent years,the demand for data transmission by users and enterprises is growing exponentially.As the traditional orthogonal multiple access(OMA)technology is approaching the achievable rate limit,new multiple access technologies are attracting a lot of attention in the industry.Non-orthogonal multiple access(NOMA)schemes allow for the non-orthogonal multiplexing of multiple users in the power/code domain.Compared with OMA technology,it can increase the number of users accessed by the system,further increasing the transmission speed and system capacity.Millimeter-wave(mm Wave)communication solves the problem of insufficient spectrum resources,but it has very high path loss in the high-frequency band,which can be compensated by the high-gain multiple-input multiple-output(MIMO)technique.Besides,mm Wave communication also helps to deploy antennas in massive MIMO systems.In mm Wave large-scale MIMO systems,hybrid precoding schemes can effectively reduce the number of radio frequency(RF)chains without significant loss of power and performance of the system.The reduction in the number of RF chains implies a reduction in the number of users served in the system.Therefore,introducing NOMA technology into the system can increase the number of users that can be served in the system.This thesis focuses on user clustering,hybrid precoding,and power allocation algorithms in mm Wave massive MIMO-NOMA systems.The main points of this thesis are as follows.This thesis first investigates the basic principles of NOMA technology,compares the system capacity of NOMA and OMA,and proves the superiority of NOMA.It also introduces the basic principles of MIMO,and investigates the mm Wave transmission characteristics as well as the channel model.Then,this thesis investigates the user clustering and power allocation for mm Wave MIMO-NOMA based on machine type communications(MTC)devices.The goal is to maximize the downlink weighted sum-rate of the MTC devices by optimizing user clustering and power allocation.Firstly,the system model is introduced,and the user clustering and power allocation scheme are designed on this basis.The user clustering scheme considers mainly the quality of service(Qo S)requirements,MTC devices are assigned to different classes in the NOMA cluster they transmit over the same frequency resources.The power allocation optimization is then studied as a weighted sum rate maximization problem.To solve this non-convex problem with high-dimensional variables,we transform the it into a convex form by introducing two sets of auxiliary variables.In addition,an iterative algorithm is proposed to obtain the optimal solution for the power allocation by updating the weighted factors and auxiliary variables.Simulation results show that the proposed user clustering and power allocation method can significantly achieve weighted sum-rate in the mm Wave MIMO-NOMA system.Finally,this thesis studies the user clustering and power allocation for mm Wave MIMO-NOMA based on message passing(MP).The main focus is on user clustering,hybrid precoding,and power allocation for mm Wave MIMO-NOMA systems.In particular,downlink transmission with a hybrid precoding structure is considered.To maximize the weighted sum-rate of the system,a user clustering algorithm based on the MP strategy is proposed for a limited number of RF chains and maximum transmit power.Then,a two-stage resource allocation was designed to maximize the weighted sum-rate of the system.In the first stage,assuming that only analog precoding is available,a user clustering algorithm based on a min-sum message-passing(MSMP)strategy is proposed to maximize the weighted sum rate of the system.In the second stage,a transformation method is proposed to solve the non-convex power allocation optimization problem using the forced-zero digital precoding to maximize the weighted sum rate of the system.Based on this,a low-complexity multiple radio frequency chain NOMA(MRFC-NOMA)power allocation algorithms is proposed.Simulation results show that the proposed user clustering and power allocation algorithm outperforms the conventional scheme in terms of weighted sum rate.