EVD Channel Estimation In Massive MIMO Systems

Massive MIMO technology deploys a large number of antennas at a base station,which can significantly improve system reliability,spectrum efficiency,energy efficiency,and channel capacity.It has become one of the key technologies for 5G mobile communications.However,only when the base station can acquire CSI can the advantages of the massive MIMO system be fully utilized.Due to the limited bandwidth,the strategy of reusing pilots in different cells is used,leading to pilot contamination problems.The EVD channel estimation,which utilizes the asymptotic orthogonality between channel vectors of different users and the sample covariance matrix of the received signal can effectively combat pilot contamination.But the EVD-based algorithm is sensitive to two errors: 1)the error due to non-orthogonal channel vectors caused by a finite number of BS antennas;2)the inaccuracy of the sample covariance matrix resulting from finite samples of the received signal.The thesis studies and analyzes the performance of a massive MIMO system based on the EVD algorithm,with particular emphasis on the impact of the above two errors,and introduces a shrinkage estimation method to improve the EVD channel estimation performance.The major contributions of this thesis are summarized as follows:(1)Establish a large-scale MIMO system model to explore the causes,impacts and countermeasures of pilot contamination.The principle of EVD channel estimation algorithm against pilot contamination is mainly introduced,and the main factors affecting the performance of the algorithm are analyzed.(2)In order to reduce the impact of sample covariance estimation error on EVD performance,a shrinkage estimation algorithm was introduced to improve the accuracy of the covariance matrix estimation under low-sample high-dimensional conditions.First,we improve on the Ledoit-Wolf(LW)method by conditioning on a sufficient statistic.By the Rao-Blackwell theorem,this yields a new estimator called RBLW,whose mean-squared error dominates that of LW for Gaussian variables.Then,in order to further reduce the estimation error,an iterative method is used to approximate the ideal estimation method and determine the closed expression of its limit,called Oracle Approximating Shrinkage(OAS)estimation.The RBLW and OAS estimation methods all have simple expressions and are easy to implement.Applying it to EVD channel estimation can improve system performance and reduce the requirement on the length of received signals.(3)Closed-form expressions of achievable rate in the uplink downlink transmission and the SER in downlink transmission of the massive MIMO system based on the EVD channel estimation are derived,from which the relationship with two kinds of errors is established.Discuss the closed-form expression of the rate loss and the trend of change.Based on this,derive the upper bound of the rate loss caused by the limited number of antennas.The rate loss caused by the sample covariance matrix is inversely proportional to the length of the received signal,and this loss will disappear when the length reaches infinity.We investigate the asymptotic behaviors of the SER.It is shown that the SER is upper-bounded in the regimes of high signal-to-noise ratio(SNR)and large number of received signal.Although increasing the number of BS antennas is helpful in improving the system performance,there is always a gap between the SER and that obtained without these two errors,and this gap enlarges as the number of BS antennas increases.