Research On Technology Of Massive MIMO In Wireless Communication System

In recent years,with the huge increase in the number of terminal devices and the continuous growth of their demand for wireless communication services,new technologies are needed to overcome the bottlenecks of the spectral efficiency and energy efficiency in traditional systems.Massive multiple-input multiple-output(MIMO)technology,with a myriad of antennas,can provide a sufficiently high degree of freedom in the space domain without increasing the time-frequency resource overhead,and obtain higher diversity,multiplexing and power gains.This technology has become one of the key technologies and research topics in the physical layer over wireless communication systems.This thesis mainly focus on the technology of massive MIMO in wireless communication system.Considering massive MIMO and related technologies,which include the rate splitting(RS),millimeter wave and cell-free(CF)distributed technologies,analyses and designs of the theoretical performance,parameter optimization and transmission scheme of the systems in specific scenarios and configurations are proposed.The main contributions of this thesis are as follows:First,for the performance analysis and parameter optimization in the uplink centralized massive MIMO system,this thesis derives an asymptotic closed-form expression of the lower bound on the ergodic capacity over the K composite fading channel(closer to reality)in the massive MIMO system.Utilizing the result,this thesis elaborate on the energy efficiency(EE)optimization problem with the transmitted power constraint,and proposes an one-dimension iterative search algorithm with low complexity to obtain the optimal number of received antennas and asymptotic ergodic capacity with optimum EE.These parameters can be a reference for the design of the EE-oriented wireless communication network.Second,for the precoding issue in the downlink centralized massive MIMO system,a modified two-stage precoding based on channel extension to balance the multi-user interference and noise effect is developed.Additionally,utilizing the matrix transform method of the QR decomposition and the suboptimal construction coefficient,a low-complexity precoding algorithm is proposed with channel update into three cases of position.Without re-computing the entire matrix of QR decomposition,this construction of the precoding can be achieved using partial Givens or Householder transformation,thereby reducing the computation complexity while ensuring the performance gain of the system.Third,for the RS technology to deal with the impact of imperfect channel state information(CSI)in the centralized MIMO system,this thesis splits one symbol into a common part and a private part,and transmits signals simultaneously after superposition.With the reasonable design of the system,this impact can be well alleviated.In the scenarios of finite feedback,this thesis proposes a minimum mean square error(MMSE)precoding scheme to further suppress the imperfectness of the channel,in which a closed-form expression is derived.Then,utilizing an approximate expression of the quantization error,a modified power allocation strategy for different parts of symbols by minimizing the sum rate loss function is developed.In the scenarios of multi-user and largescale antennas,this thesis provides a regularized zero-forcing(RZF)precoding scheme based on an optimal asymptotic coefficient,and a low-complexity solution for the precoding matrix using Neumann series can be obtained,thereby reducing the cost of hardware.Using the analytical property of RS provided by large-scale antennas,a power allocation strategy based on the asymptotic sum rates is given.Effective performance gains of these proposed schemes can be demonstrated by simulation results.Fourth,for the issue of high cost and high power consumption in the centralized millimeter wave massive MIMO system,hybrid digital-analog precoding is typically employed.With the full-connected structure,this thesis constructs an adaptive low-complexity positive-skewed codebook to obtain the equivalent channel in the two-stage feedback scheme,and gets the zero-forcing(ZF)and the modified signal-to-leakage-and-noise ratio(SLNR)digital precoding method.In order to further tackle the multi-user interference and enhance the rate performance,the RS technology is introduced again,where the corresponding precoding design and the power allocation strategy are given.The above scheme also show a certain degree of robustness.With the subconnected structure,this thesis proposes a low-complexity antenna-specific transmission scheme in an adaptive selection network,where the chain selection and the phase configuration from the perspective of the antenna are performed,and the precoding vectors that meet the paired constraints are obtained.The above scheme has lower computation complexity and requires fewer hardware devices,thus maintains the advantages of the sub-connected structure over the full-connected structure.Fifth,for the Doppler shift effect caused by mobility in the CF distributed massive MIMO system,this thsis proposes a hybrid small-cell(SC)and clustered CF massive MIMO system through classifications of the user equipments(UEs)and access points(APs),and constructing the corresponding pairs to run in SC or CF mode.A joint initial AP selection of this paradigm for all the UEs is proposed,which is based on the statistics of estimated channel and the auction method.Closed-form expressions of the downlink achievable rates for both the static and moving UEs are provided under Ricean fading channel and Doppler shift effect.This thsis also develop a semi-heuristic search algorithm to deal with the AP selection for the moving UEs by maximizing the weight average achievable rate.The performance gains and effective rates balancing of the proposed system can be demonstrated by simulation results.