Study On User Grouping Strategy For The Downlink Massive MIMO Systems

With the rapid development of wireless communication technology,large-scale MIMO technology has been widely studied as the key technology in 5G communication system.Large scale MIMO systems can theoretically deploy a large number of antennas at the base station side,allowing multiple users to reuse the same time-frequency resource to transmit data,which greatly improves the spectrum efficiency and thoughput of the system.However it is not possible to increase the number of antennas indefinitely at the base station in the actual scene,which causes the co-channel interference between users reusing the same frequency resources,and such interference will lead a poor performance of the system.Therefore,in this paper,a user grouping strategy for the downlink massive MIMO system is studied,and two improved user grouping algorithms are proposed based on the advantages of existing algorithms.The details are as follows:1.The channel model of massive MIMO system is introduced and the co-channel interference of users reusing the same frequency resources is theoretically derived.What's more,a variety of pre-coding schemes to suppress the co-channel interference are analyzed in this paper,and a lot of user grouping algorithms of the TDD system are summarized in order to provide a good idea for the design of user grouping algorithm.2.In this paper,an improved adaptive user grouping algorithm based on user correlation is proposed.This algorithm maximizes the system throughput by reasonably grouping all the users served by the base station at a certain moment by comparing the relevance of the channel between users with the threshold correlation.The simulation results show that the improved user grouping algorithm outperforms the original algorithm in the throughput performance of the system,especially in the area where the user correlation is larger,the throughput enhancement is more obvious.3.This paper presents a low complexity joint user grouping and resource allocation algorithm,which aims to maximize the system throughput.The algorithm allocates the users in the system reasonably to each RB,in order to maximize the total sum rate of users reusing the same RB.When a user reaches the expected rate,the system no longer serves this user.The simulation results show that compared with the original algorithm,the improved user grouping algorithm has little difference in system throughput,but the user fairness and algorithm complexity of the improved algorithm are better than the original algorithm.4.Finally,the total work of the whole paper is summarized,and some possible improvements are proposed as the future work.At the same time,possible research directions for user grouping algorithm in massive MIMO is expected in the end of the paper.