Power Optimization Of Cell-Free Massive MIMO Based On Bayesian Estimation

As one of the core technologies of 5G,massive multiple-input multiple-output has made great progress in recent years.Its basic architecture is divided into centralized system and distributed system.One of the most important architectures in distributed system is the Cell-Free massive multiple-input multiple-output.It contains a large number of distributed access points which provide services to a small number of users under the same time-frequency resources.This architecture has a strong ability to resist shadow fading and can greatly improve the fairness of user quality of service.Most of the existing studies of power optimization and other signal processing problems under the conditions of known large-scale fading.However,the channel hardening phenomenon of Cell-Free massive multiple-input multiple-output is not significant compared with that of centralized massive multiple-input multiple-output due to the small number of antennas assembled at access points.So signal processing under the condition of unknown large-scale fading is one of the problems that need to be studied in Cell-Free massive multiple-input multiple-output.Firstly,this paper introduces the basic architecture of Cell-Free massive multiple-input multiple-output,including its transmission protocol and key technology,briefly describes the method of estimating channel information in uplink training when the large-scale fading information is known,and gives the closed expression of data transmission rate of user's uplink and downlink when the access point is equipped with one single antenna.Then the basic knowledge of Bayesian estimation is introduced,and the particle filter algorithm is introduced in detail.The possibility of applying particle filter to downlink power optimization is discussed.Finally,the scheme of downlink data transmission for Cell-Free massive multiple-input multiple-output is given,and the quality of service of users in small-cell scenario is compared.Secondly,a channel estimation scheme is proposed in the case of unknown large-scale fading.Based on this,the estimation of channel state information is obtained.The closed expression of downlink data transmission rate based on the large-scale fading estimation is derived when multiple antennas are assembled at the access point.Based on this expression,a downlink power optimization scheme using beamforming precoding is presented.Then a power optimization algorithm based on Bayesian filtering is proposed by analyzing the correlation of large-scale fading in adjacent coherence intervals.Finally,the performance of the algorithm to guarantee the fairness of user's quality of service is verified by simulation.Finally,based on the previous analysis of the correlation of large-scale fading in adjacent coherent intervals,a channel state information estimation and tracking algorithm based on particle filter is proposed.In the power optimization algorithm which maximizes the minimum user signal-to-interference plus noise ratio,the above scheme is used to obtain large-scale fading and improve the estimation accuracy of the optimal power coefficient.The simulation results show that the proposed algorithm guarantees the fairness of the user's quality of service and is robust to the user's mobile speed.At the end of this paper,some problems to be solved and research directions for power optimization of Cell-Free Massive MIMO are presented.