Study Of Key Techniques In Non-orthogonal Multiple Access And Full-duplex Communication

Multiple access and duplex technology are important research directions in wire-less communication,and can potentially support various scenarios and applications in the next generation communication system.With the global increasing number of wire-less communication terminals and services,building orthogonal wireless data links has not been sufficient to meet the needs of network users.This dissertation summarizes the general technical requirements of 5G and the different needs corresponding to the three scenarios,analyzes the necessity of studying non-orthogonal multiple access and duplex technology,describes space division multiple access based on delayed channel state information(CSI)and statistical CSI,and emphasizes the system advantages of these two multiple access technologies in the absence of instantaneous perfect CSI.This dissertation elaborates the development history of Non-Orthogonal Multiple Ac-cess(NOMA),discriminates its difference from traditional orthogonal multiple access technology,and stresses the role of NOMA in supporting massive connections.In addi-tion this dissertation also analyzes the bottlenecks and performance limits of full-duplex(FD)communications.Faced with complex and diverse user needs and communication environments,this dissertation takes advantage of the above technologies and studies algorithms and techniques for improving system performance from the perspectives of precoding,user scheduling,and resource optimization allocation.The main work and contributions of this dissertation include the following aspectsThis dissertation studies the multi-user multiple access problem in the case where the transmitter can only obtain the statistical CSI of users.Based on the analytic ex-pression of the ergodic rate,a user scheduling algorithm named T2-MUSA is proposed with advantages in terms of both system throughput and user fairness.To this end,this dissertation compares the existing precoding design with only user statistical C-SI at the transmit side,and selects the precoding with low complexity and suboptimal performance for the single user and two user groups.The ergodic rate expressions for single-user and two-user groups are derived based on existing research.In addition,this dissertation proposes a user scheduling algorithm that ensures queue stability and has higher throughput considering users’ Quality of Service(QoS).The simulation re-sults show that the proposed two algorithms have obvious advantages compared with the traditional algorithms.This dissertation studies the multiple-input multiple-output(MIMO)system mod-el combined with the NOMA communication,involving channel estimation error and limited dynamic range(LDR)noise considerations.Firstly,the optimal receiver with the minimum mean square error(MMSE)as the optimization target is studied and its analytical expression is obtained.Then the dissertation uses D.C.programming to it-eratively search for the expression of the precoder.Hence,this dissertation proposes a whole set of transmission scheme from precoder to receiver.Simulation results support the system-level advantages of the proposed NOMA transmission scheme compared to the traditional Orthogonal Multiple Access(OMA).In addition,the simulations al-so provide the results of algorithm convergence and the system performance in terms of different numbers of antennas,hence the proposed algorithms are comprehensively evaluated and analyzed.This dissertation considers the residual self-interference issue in the full-duplex communication.The system model with channel estimation error and LDR noise is studied,and the different QoS requirements of users and the asymmetry of the link are considered.Thus,the max-min weighted rate problem is established for the case of asymmetrical links,and two algorithms,OTS-DCP and TTS-DCP,are proposed to solve the problem.The simulation results show that the proposed algorithm can obtain higher spectral efficiency.When changing the ratio of different uplink and downlink requirements,the algorithm can also obtain the rate region of MIMO FD systems.This dissertation also studies a stable matching problem in D2D cognitive radio networks based on full-duplex signaling.A joint cheating strategy is proposed from the perspective of D2D user pairs,satisfying the QoS requirements of the primary user,so that some D2D users can obtain higher data rate.Cognitive radio technology enables unlicensed secondary users to reuse the working frequency band of licensed primary users so that the spectral efficiency can be further increased.Because the reuse of fre-quency bands is self-organized by users,the topology of the network is rather dynamic and unstable,so this dissertation mainly studies the stable matching problem in this cognitive radio network.The simulation results show that the sum rate of D2D user pairs gains significant improvement under the cheating strategy.