Research On Pilot Contamination Reduction Method Based On Spatial Domain In Massive MIMO Systems

Massive multiple input multiple output (Massive MIMO or Very Large MIMO) technology is a key technology of International Mobile Telecom System-2020 (IMT 2020,5G). Interference and noise will be negligible when the number of antennas is far greater than the number of users. However, pilot contamination (PC) becomes a major bottleneck of the system performance. Therefore, it is of important practical significance to reduce or eliminate the pilot contamination in 5G communications.A pilot contamination reduction method based on spatial domain is proposed for multi-cell multi-user massive MIMO systems in this dissertation. In order to obtain the target user’s direction of arrival (DOA) in massive MIMO systems, DOA estimation methods based on compressive sensing (CS) are proposed for two kinds of antenna shapes (linear antenna array and square array antenna).The main works and contributions are as follows:Firstly, the channel model, link model and pilot contamination model are analyzed for two kinds of antenna shapes in massive MIMO systems. One PC reduction method is proposed, which using the information of the target user’s DOA. The performance indicators for describing the reduction of PC is discussed. The feasibility and effectiveness of the proposed method are analyzed and numerically simulated. Meanwhile, the accuracy requirement on DOA estimation is determined through computer simulations.Secondly, the derivation and analysis of CS model are provided for two kinds of antenna arrays in massive MIMO scenarios. Two computation-efficiency improved matching pursuit (MP) methods are proposed, which can estimate target user’s DOA more efficiently than the original MP algorithm. When performing two dimensional (2D) DOA estimation, the synthetic-angle construction algorithm is provided to resolve the coupling between azimuth and elevation. Then the summary and comparison of the performance of these DOA algorithms, including computational complexity and estimation accuracy, are given by simulations.Finally, the proposed DOA estimation algorithms based on CS theory are applied to the proposed pilot contamination reduction method and some comprehensive simulations and analysis are performed. By adjusting the different parameters, the factors affecting the method’s performance are investigated. The advantages and disadvantages between different algorithms are summed up. The proposed DOA estimation algorithms are verified that can meet the demand, and it is testified the feasibility and effectiveness of the proposed method based on spatial domain.