Research On Channel Estimation Of Massive MIMO System With Hybrid Structure

The dramatic increase in the number of mobile devices and the development of wireless multimedia services have led to aggravation of data transmitting congestion.What's more,the user's requirements for communication quality and system capacity also increase.The fifth generation of mobile communication technology will come into commercial use in 2020.As the core technology of 5G,the application of Millimeter-Wave communication and massive MIMO is also becoming more and more extensive.Channel estimation is an important factor which can affect the communication quality of the mobile system.During the signal transmission process,multipath effects are often generated due to the refraction,scattering,and so on.At the same time,millimeter-wave communication is usually susceptible to environmental factors such as rain attenuation and the atmosphere,so the path loss is serious,and as a result,the number of paths is small.In the massive MIMO system,the traditional digital structure results in high computational complexity and hardware overhead.In the process of channel estimation,it also faces the challenge of high-dimensional channel matrix calculation overhead,training sequence load,pilot contamination generated between multi-user and multi-cell,and high estimation complexity.Therefore,parameterized channel modeling can be considered.The complete channel information can be reconstructed by estimating the number of paths in the channel,the angle information and the gain information of each channel to reduce the estimation complexity.At the same time,the hybrid structure or some novel antenna array structures can be used to reduce the number of radio frequency links and reduce hardware overhead.In this topic,the channel estimation algorithm based on analog-digital hybrid antenna array structure in Millimeter-wave communication is proposed.According to the codebook design performed by adjusting the phase shifter during analog beamforming,each sub-array generates beams in different directions to perform the multiple beam scanning at the receiver which can cover the omni-direction.The number of paths and the rough value of angle of arrival on each path are estimated by analyzing the received power on each beam.The algorithm of sum-difference beam is applied to the multi-path scenario to perform refined estimation of the incoming wave direction.Beamforming technology is used to reduce the interference between different paths by to achieve the effect of path separation.In order to improve the accuracy of the estimation,the method of obtaining the ratio of the difference beam and the sum-beam is replaced by the ratio of cross-correlation sum-beam and difference-beam and autocorrelation of sum-beam.The accuracy of the angle estimation is further improved by establishing the optimization model and updating the angle set by multiple estimation.Finally,the path gain is estimated by least square algorithm.The usage of hybrid structure can effectively reduce the number of RF chains.The modified sum-difference beam algorithm is applied in multi-path scenario which improves the accuracy of the angle estimation by establishing an optimization model and makes the channel estimation more accurate.In this paper,we also proposed the channel estimation scheme in hybrid structure based on the lens antenna in the millimeter wave background.Based on the characteristics of the lens antenna,we can plot the power scatter diagram and the number of paths is estimated by setting a threshold.We also do the further analysis on the power distribution and select the antennas for estimating the angle information on each path.The connection status between RF chains and antennas is controlled by a switch,and the signal is received in time slots.The spectral estimation method is used to estimate the information of angle of arrival accurately on each path,and the path gain is estimated by the least squares algorithm.Combining the lens antenna with the hybrid structure can reduce the number of RF links and hardware overhead,and the algorithm greatly improves the accuracy of channel estimation compared to the other simplified channel information estimation methods.