Throughput Optimization For IRS-Assisted NOMA URLLC Systems

One key technique of 5th generation(5G)mobile communication,called ultra-reliable and low-latency communication(URLLC),has been proposed for new communication scenarios such as autonomous driving and automated factories,which have strict requirements for reliability and low latency.URLLC uses the short packet communication technique,i.e.,the communication transmission uses limited-length data blocks,with each block consisting of a finite number of symbols.Compared to long packet communication,short packet communication has a more complex throughput expression and lower reliability.Thus,how to improve the communication reliability and optimize the throughput of URLLC has received a lot of attention.Meanwhile,the non-orthogonal multiple access(NOMA)technique has the advantage of high spectral efficiency and support for large-scale user access,which has become a key technique of 5G and has been widely studied by academia and industry.Recently,intelligent reflecting surface(IRS)has become a new enabling technique for wireless signal transmission.An IRS is composed of a large number of passive reflection components.By programming,IRS can independently control the phase and amplitude of the reflected signal of each component and realize intelligent enhancement of wireless channels,and the reflection process almost does not produce thermal noise.The advantages of IRS provide innovative means to solve the problem of fading loss and interference of wireless channels,and have great potential to improve the effectiveness and reliability of wireless communication.Therefore,the combined use of the IRS,NOMA,and URLLC techniques has the potential to provide lower communication latency,higher reliability,and higher throughput performance,as compared to traditional wireless communication techniques.Based on the above analysis,this thesis investigates the throughput optimization for IRS-assisted NOMA URLLC systems in the following two aspects.(1)This thesis first studies an IRS-assisted NOMA URLLC system composed of a singleantenna base station,an IRS,and multiple single-antenna users.To balance the complexity and communication performance of the successive interference cancellation technique of NOMA,the users are clustered,with two users in each cluster.The communications of intercluster users adopt the time division multiple access(TDMA)protocol,and the communications of intra-cluster users adopt the NOMA protocol.To meet the strict low latency requirement of URLLC,the system adopts the short packet communication technique.To achieve the optimal throughput performance,the joint optimization of user power allocation and IRS reflection beamforming is studied to maximize the sum throughput of all users.An alternating iteration-based algorithm is proposed to solve the resultant non-convex optimization problem,which solves the subproblems of power allocation optimization and IRS reflection beamforming optimization alternatively until achieving convergence.Simulation results show that the proposed algorithm can effectively improve the maximum throughput,display the superiority of using the NOMA technique,and reveal the throughput performance gap between short packet communication and long packet communication.(2)A multi-IRS-assisted NOMA URLLC system is then investigated,where the problem of how to pair user clusters and IRSs and how to optimize the corresponding resource allocation are considered.Specifically,the user power allocation,IRS reflection beamforming,and user-cluster-IRS pairing decision problems are jointly investigated to maximize the sum of user throughputs.To solve the resultant non-convex optimization problem,an alternating iteration-based algorithm is proposed,which solves the three subproblems of power allocation optimization,IRS reflection beamforming optimization,and user cluster IRS pairing decision optimization alternatively until achieving convergence.Simulation results show that the proposed algorithm can significantly improve maximum throughput and verify the necessity and effectiveness of user cluster IRS pairing decision optimization.