Analysis And Design Research Of Cell-free Massive MIMO Systems

As one of the key technologies of 5th Generation(5G),massive multiple-input multiple-output(MIMO)provides tremendous array gain,diversity gain and spatial multiplexing gain by deploying tens to hundreds of antennas at the base station(BS),which can serve many mobile terminals(MTs)in the same time-frequency blocks and substantially improve system spectral efficiency.However,with the further densification of cellular cells,the inter-cell interference has become the major bottleneck restricting the performance of the massive MIMO system,which severely affects the cell edge MT performance.To address this limitation and provide uniformly good service for all MTs,the cell-free massive MIMO system came into being.In such a system,the BS is replaced by many distributed access points(APs)and an MT can be served by many APs in the same time-frequency resource blocks.It enables MTs to break through the cell boundary restrictions,and all MTs are located in the center of the cell,thus promising immense macro-diversity and coverage,high channel capacity and link reliability,low interference and path loss.Focusing on the high economic cost and hardware energy consumption issue in the actual deployment of cell-free massive MIMO systems,this dissertation studies the impact of non-ideal hardware on the performance of cell-free massive MIMO systems,and uses random matrix theories to derive the underlying performance closed-form expressions,and theoretically explores the feasibility of equipping non-ideal hardware.The main contributions of this dissertation are summarized as follows:1.Focusing on the cell-free massive MIMO system that is equipped with low-resolution analog-to-digital converters(ADCs)and multi-antenna MTs,the impacts of reducing the ADC resolution and increasing the MT antenna number on spectral efficiency and energy efficiency are investi-gated.Specifically,leveraging on the additive quantization noise model(AQNM),closed-form achievable spectral efficiency expressions for maximum ratio combining(MRC)receiver are derived.Besides,the total power consumption of the considered system is presented by in-volving the power cost of low-resolution ADCs,and then the energy efficiency expression is exhibited.Based on these closed-form results,the impacts of key system parameters on spec-tral efficiency and energy efficiency are analyzed.Moreover,the trade-offs between spectral efficiency and energy efficiency are also explored.2.In the uplink cell-free massive MIMO system with low-resolution ADCs at the APs,low-quality radio frequency(RF)chains are considered to further reduce the hardware cost and energy consumption.Leveraging on the AQNM and Gaussian RF impairment model(GRFIM),closed-form minimum mean square error(MMSE)channel estimation and MRC-based achievable rate expressions are derived,which facilitate establishing the RF scaling law when the total number of AP antennas tends to infinity.In addition,the compensation strategy between RF and ADC is explored.The simulation results indicate that the high-quality RF chains can be utilized to compensate for the rate degradation that arises from low-resolution ADCs,and vice versa.3.To mitigate the rate loss that arises from low-resolution ADCs and digital-to-analog converters(DACs)and reduce the hardware cost of high-resolution ADCs and DACs,an uplink cell-free massive MIMO system with mixed-ADC/DAC architecture is considered.To be specific,the mixed architecture permits some antennas in each AP to be linked with high-resolution ADCs/-DACs,while the remaining antennas are paired with low-resolution ADCs/DACs.Leveraging on the AQNM,closed-form achievable rate expressions with MRC receiver are first derived,and then the energy efficiency expression is also presented by considering the power consump-tion of ADCs and DACs with different resolutions.These closed-form expressions facilitate analyzing the impact of system parameters(such as the ADC/DAC resolution,the proportion ofideal antenna components in the mixed-ADC/DAC structure,and the total number of AP anten-nas)on spectral efficiency and energy efficiency.Additionally,the trade-offs between spectral efficiency and energy efficiency under different antenna architectures are explored.Finally,a max-min fairness power control algorithm is designed to provide uniformly excellent service for all MTs.4.Considering the Rician fading scenario that is more in line with the actual communication en-vironment,the achievable rates for both uplink and downlink cell-free massive MIMO systems with low-resolution ADCs at the APs are investigated.Specifically,the system model over Rician fading channels is first presented,and then the closed-form uplink and downlink rate expressions for MRC receiver and conjugate beamforming(CB)precoder are respectively de-rived by resorting to the AQNM,which reveal the impacts of the ADC resolution,the Rician K factor,and the total number of AP antennas on achievable rates.Besides,based on the de- rived closed-form rate expressions,the uplink and downlink weighted max-min fairness power control algorithms are proposed to address the rate differentiation issue caused by the uneven distributions of APs and MTs and the different priorities of MTs.