Random Access Methods For Massive MIMO Systems

With the popularization of intelligent terminals and the continuous growth of various new mobile services,in order to support high-capacity,high-speed,low-cost,low-energy,low-latency,high-reliability and other technical requirements,the 5th generation mobile communication systems(5G)New Radio(NR)architecture will be based on flexible parameter sets and frame structures,beamforming and large-scale antenna arrays.While greatly enhancing the user-centric experience of mobile Internet services,5G will fully support the object-oriented Internet of Things business and realize the intelligent interconnection of people,objects and things.Massive multiple-input multiple-output(MIMO)wireless transmission can significantly improve the system spectrum efficiency and transmission reliability by deeply exploiting spatial resources.Nevertheless,massive MIMO still faces several challenges before fully practical implement.We focus on the 5G-NR,and investigate random access for massive MIMO systems exploiting the characteristics of the beam domain channel.Firstly,based on the architecture of 5G-NR,we investigate the synchronization signal and uplink random access procedure in the 5G-NR system.Based on the NR series protocols,the 5G radio frame structure configuration,resource mapping location and random access procedure are discussed.Also,the definition of the physical random access channel(PRACH)signal and its resource mapping rules are presented.In order to obtain the timing adjustment of random access,user detection and timing estimation should be performed.Furthermore,based on the sparsity of the massive MIMO beam domain channels,the uplink timing estimation problem can be transformed into a compressed sensing problem.In view of this,the compressed sensing transmission model is discussed,and typical recovery algorithms under single measurement vector(SMV)and multiple measurement vectors(MMV)models are studied.Secondly,based on the detection problems for massive MIMO random access,the uplink timing estimation method for massive MIMO is studied.We introduce the massive MIMO spatial channel model,and the relationship between the spatial domain channel and the beam domain channel.With this channel model,the timing estimation algorithm under the maximum likelihood criterion is studied.Based on the two assumptions of deterministic and random channel,two timing estimation algorithms under least squares(LS)channel estimation and minimum mean square error(MMSE)channel estimation are studied,respectively.In view of the complexity of the timing estimation algorithm under MMSE channel estimation,the channel energy coupling matrix is introduced,and the beam domain approximation estimation algorithm is proposed.The algorithm can reduce the computational complexity and the performance approaches the timing estimation algorithm under MMSE channel estimation.Furthermore,the random access decision and timing estimation criteria are given.In addition,based on the sparse characteristics of the beam delay domain channel,the compressed sensing algorithm is used to estimate the timing in the case of unknown channel state information.The simulation results show that the performance of compressed sensing algorithm in the space domain is consistent with that of the timing estimation algorithm under LS channel estimation,and the performance of compressed sensing algorithm in the beam domain can approximate that of the proposed beam domain approximation estimation algorithm.Finally,We propose joint user detection and uplink timing estimation algorithm in the massive MIMO transmission systems.Based on the physical channel model,we investigate the channel sparsity in the angular delay domain for massive MIMO systems.Furthermore,the multi-user access system model is given,and the structured sparsity characteristics of the channel matrix under this model are demonstrated.In the case of unknown channel state information,joint user detection and timing estimation are performed by using the structured sparse characteristics of the channel,and a modified orthogonal matching pursuit(M-OMP)algorithm based on the beam domain channel characteristics is proposed.Simulation results show that the M-OMP algorithm can have better user detection performance and timing estimation performance than the traditional MMV algorithms such as OMP algorithm and the timing estimation algorithm under LS channel estimation,and can support more users.As the number of antennas at the base station increases,the M-OMP algorithm can achieve more significant performance gains.