Research On Multi-antenna Channel Estimation In Massive MIMO

With the development of wireless communication,large scale MIMO is considered as the key technology of the next generation mobile communication system.Using the large scale antenna array,which is based on the base station or the client configuration,can greatly improve the system throughput,and benefit from the statistical CSI.The performance of massive MIMO depends largely on whether the base station can accurately obtain the current downlink CSI.In practical wireless communication systems,pilot based channel estimation techniques are used to obtain CSI.The number of base station antennas in large scale MIMO system is up to tens or even hundreds of times,leading to further problems such as pilot pollution,pilot resource consumption and pilot power.In this thesis,we study the pilot design and channel estimation techniques for massive MIMO systems.Current massive MIMO technology is used in communication systems because its distribution channel has the characteristics of clusters being more obvious.And the channel matrix is sparse,so compressed sensing has obvious advantages.This thesis will focus on the use of compressed sensing to do multi-antenna system channel estimation.This article does these following work:1)The commonly used methods of channel estimation can be roughly divided into two categories: using pilot method and blind estimation method.In this paper,the traditional method of pilot channel estimation,and the compressed sensing method is introduced.2)A non-orthogonal pilot placement scheme is proposed: the spatial joint sparsity of large-scale MIMO channel is analyzed,but the traditional algorithm does not take advantage of this feature.so,we designed the corresponding pilot placement scheme under a compressed sensing framework,which greatly reduced the pilot cost compared with the traditional scheme.Finally,based on the pilot scheme,the measurement matrix used in the compressed sensing estimation algorithm is constructed,and it is proved that the proposed measurement matrix can perform reliable sparse signal recovery theoretically.3)Based on the greedy and bayesian ideas,two sparse channel estimation methods are proposed,which are based on tradional OMP and BCS algorithm.Compared withtraditional algorithms,the advantages of these two improved algorithms are the joint estimation of the the channel impulse response with common sparsity between different antennas.Then the algorithm complexity and convergence are analyzed.Finally,the simulation results show that the proposed method has a significant performance improvement compared with the traditional method.4)A CS channel estimation scheme is proposed by analyzing the common sparsity of virtual Angle domain channels between different subcarriers of FDD massive MIMO.The scheme can achieve robust and accurate channel state information acquisition in base station,greatly reducing the cost of time slots in channel estimation and feedback.In addition,this chapter summarizes the improvement of MMV to GMMV in the theory of compressed sensing,thereby the easonable non-orthogonal pilot scheme and the improved estimation algorithms are designed.The simulation results show that the scheme can achieve the performance baseline with a reasonable time-slot cost.5)finally,we summarize the full text and look forward to the unexplored area?...