Research On Techniques Of Beamspace Massive Non-orthogonal Multiple Access

With the wide applications of Internet of Things(IoT)in various social and economic fields such as industry,agriculture,transportation,and medical treatment,the number of IoT devices has experienced an explosive growth.In this context,the traditional orthogonal multiple access techniques can not meet the requirements of wireless access of a massive number of devices,and thus it is urgent to study the massive non-orthogonal multiple access techniques for the beyond fifth-generation(B5G)cellular IoT networks.However,massive non-orthogonal multiple access(NOMA)techniques face many challenging,such as superior difficulty in channel state information(CSI)acquisition,high complexity of superposition coding,and large computational amount of successive interference cancelation(SIC).In view of a series of problems existing in massive nonorthogonal multiple access techniques of B5 G cellular IoT,this paper focuses on the research of techniques of beamspace massive non-orthogonal multiple access,and provides a comprehensive solution.Firstly,this paper introduces the characteristics and requirements of B5 G cellular IoT networks,and then discusses the key techniques of massive NOMA,including CSI acquisition,devices clustering,superposition coding,and SIC.Especially,by exploiting the advantages of beamspace,this paper redesigns these four key techniques.Secondly,this paper studies the uplink of beamspace massive non-orthogonal multiple access.Based on the proposed framework for uplink massive non-orthogonal multiple access,this paper analyzes the system performance and derives a the lower bound on the weighted sum of the ergodic rates which reveals the influence of beamspace parameters on the system performance.Then,by maximizing the lower bound on the weighted sum of the ergodic rates,this paper proposes an uplink massive non-orthogonal multiple access algorithm without clustering and an uplink massive non-orthogonal multiple access algorithm with clustering,respectively.Simulation results show that the proposed uplink massive beamspace non-orthogonal multiple access algorithms can achieve a tradeoff between system performance and computational complexity.Thirdly,this paper studies the downlink of beamspace massive non-orthogonal multiple access.Based on the proposed framework for downlink massive non-orthogonal multiple access,the performance of the system is analyzed,and a upper bound on the weighted sum of the ergodic rates is obtained.Through analyzing this upper bound,it is revealed that the effective performance gain can be obtained by using a linear combination of multiple base beams of beamspace to design the transmitting beam.Motivated by this observation,three beamforming schemes are designed for beamspace massive non-orthogonal multiple access.Specifically,the first one is a full-space multiple-beam design scheme,which constructs transmit beams for each devices in the entire space;the second one is a partial-space multiple-beam design scheme,which constructs transmit beams for each device in a given subspace belonging to its cluster;the third one is a partial-space single-beam design scheme,which constructs a common transmit beam for the devices of the same cluster in a given subspace belonging to the cluster.The degrees of freedom of beam design of these three schemes are reduced in turn,and the design complexity is also gradually reduced,such that they can be selected according to the requirements of system performance and computational complexity.Finally,this paper summarizes the research on beamspace massive non-orthogonal multiple access,and provides several research directions in future works.