Research On Channel Estimation Of Cell-free Massive MIMO System

Massive multiple input multiple output(Multiple-Input Multiple-Output,MIMO)system uses the traditional cellular network,and the cell boundary of the cellular network will be disturbed,which has always limited the performance of massive MIMO.Cell free massive MIMO combines the characteristics of massive MIMO and distributed antenna,distributes a large number of access points(APS)in the wide area,and serves all users in the same time-frequency resources through time-division duplex operation.Cell free massive MIMO introduces the idea of " user-centric",which can shorten the distance between AP and users,and reduce user interference by arranging a large number of APS,so as to improve the user experience within the service range of the whole region.This thesis mainly studies the channel estimation of cell-free massive MIMO,the innovations of this thesis are as follows:(1)Aiming at the problem that cell-free massive MIMO inherits the attribute of channel hardening of massive MIMO.It is concluded that although channel hardening is applicable in cell-free massive MIMO,it has poor performance in many cases.Therefore,a blind channel estimation method is proposed in this paper.This method does not need any downlink pilot.It uses the data received by the user during the coherence interval to estimate the effective channel gain.Simulation results show that the proposed method has great advantages in the performance of blind channel estimation whether it is suitable for channel hardening or not.Through ZF processing,blind channel estimation can improve 95% possible net throughput by 53% when the channel hardening performance is poor.(2)Aiming at the problem that the increase of the pilot sequence length required to accurately estimate the channel will reduce the spectral efficiency with the increase of the number of users in time division duplex system,this thesis proposes a semi blind channel estimation method.This method is an expectation maximization algorithm,which is easy to deal with by assuming the Gaussian distribution of unknown symbols.Cell-free massive MIMO systems benefit from semi blind channel estimation in both uplink and downlink transmission.Simulation results show that using this algorithm can obtain more accurate channel estimation or use less pilot symbols to obtain the same accuracy estimation channel.In this simulation scheme,semi blind channel estimation reduces the training overhead of maximum likelihood estimation by 87.5%.Therefore,semi blind channel estimation is very attractive for cell-free massive MIMO.