| The fifth generation(5G)mobile communication system will be commercially available in 2020.The 5G system is provided for connection whenever and wherever possible of everyone and everything.Massive MIMO is one of the most important enabling technologies for 5G applications.Meanwhile,Massive MIMO introduces new propagation characteristics and puts forward specific requirements for 5G channel modeling.Channel model is essential for system design and performance evaluation,therefore it is urgent to meet the needs of accurate and efficient channel models for various 5G technologies and scenarios.As is known to all,the transmission of information is based on the propagation of electromagnetic wave in space.The performance of wireless communication system is mainly restricted by the characteristics of wireless channel.Therefore,this paper studies the new characteristics of wireless channel for 5G applications,proposes a two-dimensional(2D)stochastic channel model suitable for Massive MIMO channel propagation characteristics,and finally proposes an improved method for 3D MIMO channel modeling.Firstly,this paper introduces the background and significance of Massive MIMO channel research for 5G applications,and briefly describes the development of Massive MIMO channel modeling.Then several new channel modeling methods are introduced,and it is found that there are always two problems for Massive MIMO,namely,spherical wavefront channel modeling and non-stationary characteristics.Finally,the improved method of 3D MIMO channel modeling is introduced.Secondly,this paper aims at two drawbacks of the current theoretical framework of Massive MIMO channel modeling:(1)the spatial location of the transmitter and receiver needs to satisfy certain geometric constraints,mainly to facilitate the modeling of spherical wavefront characteristics,but this assumption is not usually satisfied in actual scenarios;(2)the key model parameters such as cluster birth and death rate are not established with real-time analysis.Because of the high complexity of simulation,it is very difficult to apply directly at present.For these two shortcomings,we propose a 2D geometry-based stochastic channel model to capture the behavior of spherical wavefront in real propagation environment.Then the N-state Markov process is used to simulate the non-stationary characteristics caused by the appearance and disappearance of clusters on the array.Finally,a method of 3D MIMO statistical channel modeling based on channel measured data is proposed for MIMO channel,which can effectively and accurately reflect the real threedimensional channel environment and improve the accuracy of channel model.The statistical characteristics of large-scale parameters were extracted through outfield measurement,and the cross-correlation matrix of large-scale parameters was generated to solve the problem of the non-positive definite matrix in the existing models.The elevation angle extension of 3D MIMO channel was statistically modeled with the linear model,and the dependence relationship between elevation angle extension and distance was increased.The interdependent sub-path azimuth angle and sub-path elevation angle are randomly generated by mixing Von Mises Fisher distribution.According to the statistical analysis of field measurements,the characterization parameters of the channel model are determined and the channel coefficients of 3D MIMO are generated. |
PDF Full Text Request