| With the popularization of mobile terminals and the emergence of mobile Internet of Things,the tra-ditional multiple-input multiple-output(MIMO)technology has been unable to meet the requirements for system capacity and transmission rate in the future mobile communication.The massive MIMO technolo-gy allocates large-scale array antennas in the base side and exploits spatial dimension resources,which can significantly improve system spectrum efficiency and power efficiency and reduce noise interference.It is believed to be a promising technology for the next generation of mobile communication systems.Massive MIMO beam division multiple access(BDMA)transmission utilizes the sparsity of the beam domain channel to allocate non-overlapping beam sets to different users for transmission,and multi-user MIMO channal are equivalent to multiple parallel single-user MIMO chananal,which reduces thetransmission complexity.How-ever,in practical mobile communication systems,radio frequency(RF)channel impaimlents are inevitable in the transmitting and receiving RF terminals.Channel impairments at the RF end deviate the amplitude-phase characteristics of each channel from the ideal one,thus destroys the channel sparseness in the beam domain and degrads the BDMA transmission performance.In view of the above problems,this thesis mainly investigates the calibration methods for RF channel impairments in massive MIMO transmission systems,and accomplishes its realization on the field programmable gate array(FPGA)platform.Firstly,we study the massive MIMO channel model and BDMA transmission method.After analyzing the wireless channel model and the characteristics of beam domain in massive MIMO system,the channel-to-beam relationship between physical channel and beam domain and the sparse characteristics of the beam domain channel are obtained.With the maximization sum rate upper bound as the goal,when the basestation knows the statistical channel information only,the optimal transmission should be for different users assigned to non-overlapping beam set.The principle and four steps of BDMA transmission method are discussed in detail.Secondly,we establish a massive MIMO RF channal impairments model and investigate the influence of RF channel impairments on the performance of BDMA transmission.Based on the massive MIMO or-thogonal fi-equency division multiplexing(OFDM)system model,we establish three kinds of RF channel impairments models of in-phase and quadrature arm(IQ)imbalance,timing offset and amplitude-frequency inconsistency respectively.Corresponding matrix representations are given and a generalized massive MI-MO RF channel error model is established.Based on the error model,the influences of IQ imbalance,timing offset and amplitude-frequency inconsistency on the transmission performance of BDMA are studied respec-tively.The simulation results show that the channel amplitude and phase rotation error caused by the three kinds of channel errors will cause beam channel energy dispersion,resulting in the degradation of the BDMA transmission performance.Relative to the user-side,the influence caused by the amplitude and phase rotation errors of the channel at the base station side is the main reason for the degradation of the BDMA transmission performance.Therefore,the channel calibration of the massive MIMO BDMA transmission system mainly calibrate the channel impairments in the base station side.Finally,we design and implement a massive MIMO RF channel wideband baseband hardware calibra-tion method based on the 5th generation(5G)massive MIMO prototype platform.The calibration system is equipped with an external calibration circuit at the base station side and adopts the channel estimation method according to the long term evolution(LTE)-like frame structure to control its operation in three dif-ferent modes to achieve the calibration of the receive channel and the transmit channel.This thesis gives the system architecture,frame structure,calibration pilots and calibration flow-process design of massive MIMO channel calibration system in detail.According to the calibration process,the calibration module is divided into several sub-modules,whose calibration scheme,functions,interfaces and timing of each sub-module are designed and the consumption of the overall module resources is also given.The module is tested on the FPGA and verified with the results showing that this method has better calibration performance. |
