Research On Massive MIMO Channel Reconstruction And Tracking Technology

With the rapid development of mobile communications,the rise of mobile multimedia services,mobile social networks and other services put forward increasingly stringent requirements for the improvement of data transmission rate and the diversification of application types.In order to meet the increasing demand of mobile communication,large-scale multiple input multiple output(MIMO)technique was applied.Large scale MIMO utilizes a large number of antennas at the base station to fully exploit the spatial degrees of freedom,enhance the spatial multiplexing and spatial diversity technology,thereby significantly improves the spectrum frequency.Although large-scale MIMO technology has been proved to be effective and has been put into primary commercial use,there are still many problems to be solved,mainly in the aspects of the high overhead of channel reconstruction,high time complexity of channel estimation algorithm,channel characteristics changes caused by extremely large antenna array and so on.Aiming at the above problems,in this thesis,we will find solutions to the existing problems of large-scale MIMO technology from the aspects of downlink channel reconstruction cost,channel estimation algorithm complexity,extremely large scale channel characteristics,and study the low-cost large-scale MIMO wireless transmission theory and key technologies.Firstly,we analyze the basic characteristics of wireless channel,the principle of orthogonal frequency division multiplexing(OFDM),the role of cyclic prefix and the principle of MIMO technology,The advantages and disadvantages of frequency division duplex system and time division duplex system are introduced.Then we introduce and analyze the current channel reconstruction methods in FDD system,and determine the research focus of this thesis.Secondly,the downlink channel reconstruction method in FDD 3D MIMO system is studied,and a new method based on uplink full dimensional channel state information(CSI)is proposed Based on the reciprocity of spatial propagation parameters of uplink and downlink channels,a newtonized orthogonal matching pursuit(NOMP)algorithm for frequency independent parameter estimation of uniform planar array(UPA)configuration at uplink base station is designed,including channel delay estimation,horizontal and vertical DOA estimation.Next,the user estimates the complex gain sparsity of the downlink channel path and feeds it back to the base station to realize the reconstruction of the downlink channel.The scheme greatly reduces the cost of pilot and feedback.Then,a channel compensation scheme is proposed for the imperfect reciprocity of channel in the real scene.Matlab simulation and measured data are used to verify the availability of the reconstruction method.The accuracy of 3d-nomp algorithm is proved by software simulation,and the applicability of the proposed scheme in different scenes is proved by the measured data.Afterwards,we study the efficient channel tracking algorithm in mobile scene,and propose a time-varying downlink channel tracking method suitable for FDD large scale antenna systems.In order to reduce the complexity of calculation,a simplified dictionary and a new algorithm are designed.This tracking method consists of two main modules.In the first detection module,the uplink channel is estimated at the initial time by using the nomp algorithm.In the second tracking module of the subsequent time,the estimation parameters of the previous time are used to track the channel.Then,the channel gain of the downlink is recalculated and feedback is achieved by applying a small amount of overhead.The uplink channel is reconstructed by using the estimated frequency independent parameters and the gain of the downlink.The simulation results show that the proposed tracking scheme can reconstruct the time-varying channel effectively with less operation time and pilot cost.Finally,for FDD extremely large-scale MIMO system,we propose a low complexity extremely large scale MIMO(XL MIMO)channel tracking algorithm which utilizes the slow variation of channel physical propagation parameters.The cost is further reduced from reducing the complexity of the algorithm and the problem of long time consuming of channel estimation algorithm is solved.One of the characteristics of the future large aperture antenna array is the spatial nonstationarity,which is also taken into account here.Simulation results show that the algorithm can accurately detect the position movement and birth and death of scatterers,ensure better channel estimation performance,obtain accurate scatterer-subarray mapping relationship,and greatly reduce the time of channel estimation to ensure real-time estimation in time-varying channel scenarios.