Research On Massive MIMO Random Access Mechanism For Massive Users

According to the requirements of ITU,5G user experience rate should reach 1Ggps and peak rate should reach 20Gbps.Meanwhile,the time delay should be reduced by 5-10 times compared to 4G system and connected user density should be increased by 10-100 times.Massive MIMO(Massive Multiple-Input Multiple-Output),by deploying hundreds of antennas and transmitting multipath signal simultaneously on the same spectrum resource,can serve multiple users at the same time and the spectrum effectiveness of the system can be improved,which has already been one of key technologies of 5G.Besides,the communication of massive devices brings about new challenges to 5G user access process:due to the limitation in the coherence time of the channel,the quantity of pilot frequency cannot satisfy simultaneous access of massive users,during which the access conflict may be caused because many users use the same pilot frequency sequence when using the same time frequency resource.For this regard,as for the problem about pilot frequency conflict due to the access of massive users in MIMO system,the main contributions of this paper are as below:Firstly,as for the possible problem about pilot frequency conflict due to the features of the centralized massive MIMO antenna array and massive users accessed,its influence on the random access process of massive users is analyzed.Besides,the improved algorithm for strongest user detection for centralized massive MIMO(C-SUCIR)algorithm based on the centralized massive MIMO is proposed.The approximate closed-form expression for solving the pilot frequency conflict through the system is deduced based on Rician channel model when the user uses C-SUCIR algorithm,50 that the average delay of user access is obtained.Through simulation verification,it is proved that the average access delay of users using C-SUCIR algorithm can be reduced by 10 times than that using the random access mode and 8%of pilot frequency conflict problems in the system could be solved when the ratio of the antenna and user quantity is lower than 1.5.Secondly,speaking of the feature that the antenna unit in distributed massive MIMO scene is approximate to the user location,the mathematical modeling of the random access process of distributed massive MIMO under Rician channel is implemented.Besides,the improved algorithm for strongest user detection for distributed massive MIMO(D-SUCIR)algorithm based on the distributed massive MIMO is proposed.The simulation result proves that 95%pilot frequency conflict problems existing during random access process of users with distributed massive MIMO still could be solved through D-SUCIR algorithm.In addition,the average delay of user access is lower by 10%than the access delay in random access algorithm.To sum up,the in-depth study on the influences of the random access process of centralized and distributed massive MIMO is implemented according to key paraneters such as antenna quantity,user quantity,user activation and competitive user interference.C-SUCIR and D-SUCIR random access algorithms are proposed based on the relationship between centralized and distributed massive MIMO antenna array and user location are proposed respectively.The study manifests that the problems about pilot frequency conflict in centralized and distributed massive MIMO system could be effectively solved and user access delay can be reduced through the random access algorithms proposed in the paper,which provides theoretical guidance for actual 5G system optimization.