Performance Analysis For Grant-Free Random Access In Cell-Free Massive MIMO

Cell-free Massive MIMO technology is one of the key technologies for future wireless communication.All access points jointly provide services to users,which can meet the rapidly increasing number of users and high transmission rate requirements of beyond 5G network.Cell-free massive MIMO technology has become one of the main research hotspots of 6G technology in the future,which not only has the advantages of high spectral efficiency and large coverage of traditional massive MIMO technology,but also can solve the problem of inter-cell interference of traditional cellular cells through the cooperation between APs,and improve the array gain.In the scenario of massive machine-type communications(m MTC),this paper studies user activity detection,channel estimation and data detection based on Cell-free massive MIMO technology and two-stage grant-free random access mechanism.Firstly,for the first stage of the grant-free random access,Cell-free massive MIMO network system is adopted,in which APs send the non-orthogonal pilot sequence sent by users to the Central Processing Unit(CPU)for centralized processing to improve the robustness against shadow fading effect.In view of the large number of APs in the system,in order to reduce the complexity of calculation,this paper adopts Approximate Message Passing(AMP)algorithm based on MMSE denoiser to detect user activity and estimate channel.The iterative process of AMP algorithm,the design and derivation of MMSE denoiser are studied.The unique state evolution of AMP algorithm simplifies the detection process and effectively reduces the complexity of the algorithm.Experimental results show the feasibility of this algorithm.Compared with other algorithms,AMP algorithm has better user activity detection.At the same time,the detection results of the Cell-free massive MIMO system and the traditional cellular system were compared.The results show that Cell-free massive MIMO system can support a large number of communication devices,solve the problem of inter-cell interference in the traditional cell,and improve the accuracy of user activity detection.Secondly,for the second stage of the grant-free random access,the receiver design in the Cell-free MIMO system is optimized.This paper combines model-driven deep learning method with AMP algorithm to obtain AMP-NET algorithm,which is used for data detection.In the algorithm,the iterative process of AMP algorithm is expanded into a deep network,and a hybrid training strategy of hierarchical and global optimization is adopted to minimize the loss function by optimizing training parameters.AMP-NET algorithm not only has the parameter learning ability of deep learning,but also has the lower computational complexity of AMP algorithm.Simulation results show that AMP-NET algorithm can effectively improve the performance of data detection and reduce bit error rate.