Research On Channnel Reconstruction Method For Massive Mimo Over-the-air Test

Massive multiple-input multiple-output(MIMO)technology and millimeter wave communication are two key technologies in the fifth generation mobile communication.Base stations equipped with massive MIMO antenna arrays will be widely deployed to achieve high-speed transmission of the data for a large number of users in dense networks and compensate for serious transmission losses through array gains.With the commercialization of 5G,there are more and more requirements for the performance test of massive MIMO millimeter wave antenna equipments.Over-the-air(OTA)test has become the focus of performance test at present,and the channel reconstruction method for massive MIMO OTA is the object of this thesis,which is to synthesize accurate channel environment in the area to be tested through the radiation of limited probes.The purpose of OTA test is to reconstruct the required channel model in the room.The more reasonable the probe distribution and the more channel emulators are,the more accurate the channel reconstruction will be.The probe optimization problem of multiple users will be involved in the test of 5G massive MIMO antenna system,due to the expansion of the area to be tested,more probes and channel emulators will be used,which will bring high cost problems.The channel models of multiple users are added together and then optimized through the traditional method;we can get multiple probe positions by optimizing each user separately,then the union combined by the selected probes of multiple parts can be optimized through the reference method.But these two methods can not minimize the error between the real radio environment and the OTA test channel environment very well.Given this problem,a novel multi-user probe selection method is presented in this thesis,and the probe position and weight are optimized directly according to the channel characteristic of each user.The sum of the root mean square error of the correlation matrix of multi-user target channel models and the correlation matrix of channel models in OTA test is taken as the evaluation standard.It is verified that the proposed method is superior to other two methods for different scenarios.A series of characteristics of array antennas,such as beam acquisition and tracking,play an important role in massive MIMO antennas.For the dual anechoic chamber structure,the spatial correlation is mostly used as the criterion to reconstruct channels in the previous literature.Compared with the reconstruction of spatial correlation,the reconstruction of spatial spectrum can reflect the characteristics of the array more directly,so as to improve the accuracy of array performance in the OTA test channel model,such as array gain,sidelobe interference and direction-of-arrival estimation.Based on this,the spatial spectrum is proposed as the metric to determine the weight of the probe,so as to emphasize the array effect of massive MIMO antenna.According to the covariance matrix of the channel and the spatial spectrum,the weight of the probe is optimized.The error of spatial spectrum obtained by the two methods is simulated and analyzed.And the result of optimization based on spatial spectrum is better.