Modeling And Non-stationary Characteristic Analysis Of Massive MIMO Channel

Nowsday, 4G network construction is hot, the research of 5G technology has already begun.Traditional MIMO technology does not meet modern people's growing demands for high data rate,so how to improve the system capacity in the limited spectrum resources has become the crucial of the next generation communication systems. Massive MIMO technology emerges, studies show that Massive MIMO technology has the following advantages:(1) Massive MIMO technology directly improve system capacity by increasing the number of base station antennas;(2) can better reduce interference in the cell;(3) lower transmit power.since the base station has a large antenna,Massive MIMO technology can greatly improve spectrum efficiency and energy efficiency,reduce interference as 5G key technologies.Currently,research of Massive MIMO channel more in the theoretical stage, the actual measurement is small. In this paper, Starting from Massive MIMO channel modeling,introduced Massive MIMO confocal elliptical model, theoretical simulation,extract the channel parameters to research and analysis,at the same time,study the non-stationary characteristics of Massive MIMO from theoretical modeling. The main contents are as follows:1. Fisrtly, Massive MIMO channel appeared non-stationary characteristics,Such as offset angle, received power difference,Doppler frequency changes at the sending and receiving antenna elements. the tradition MIMO channel model was no longer applicable. For Massive MIMO communication system,this article described Massive MIMO confocal elliptical model. unlike the plane wave assumption of traditional MIMO channel model, confocal elliptical model used spherical wave assumption. Introduced the position of antennas and clusters to the calculation of channel impulse response. In the study of clusters, used birth-death process to simulate the dynamic characteristics of newborn and demise of clusters on antenna array. Different antenna elements observed different clusters collection in the model. In addition, it also described the random shuffling and matching of clusters.2. Secondly, In terms of theoretical simulation,theoretical simulation for Massive MIMO confocal elliptical model.When simulation,refenced the parameters of the WINNER model simulation cluster parameters.to calculate the channel impulse response.According to the channel impulse response,study the statistical properties of channel, such as spatial correlation function, power spectral on angle antenna arrays. Simulation results show,Massive MIMO confocal elliptical model can model Massive MIMO channel non-stationary characteristics. The simulation results are consistent with the literature related to the actual measurement results. In addition, sumulationed the evolution and random shuffling process of clusters on antenna array. By simulation we have deeper understand to the non-stationary characteristics of Massive MIMO.3. Finally, analyszed the emergence of non-stationary characteristics of Massive MIMO channel antenna array,Found that under normal circumstances between adjacent antenna,the number of generation cluster and demise cluster of up to 2. In this situation propose a Markov chain method to simulate the process of generating and demise of clusters,And gives the Markov chain state-space division and the calculation method of the shift probability matrix. theoretical simulation for Markov chain approach, simulation results show, compared with birth-death process, Markov chain method is more simple, less time-consuming simulation,and can model the non-stationary characteristics of clusters on antenna array.