Exact Performance Analysis Of MIMO And Massive MIMO Systems With MMSE Receiver

In order to meet the diversified business needs of the mobile Internet and the Internet of Things in the future,the Fifth-generation(5G)mobile communication network improves the system capacity by improving spectrum efficiency,increasing network density,and expand-ing system bandwidth.With its advantages,the large-scale Multiple-Input Multiple-Output(MIMO)technology can make full use of resources in the space dimension,significantly im-prove the system's spectrum efficiency,and improve the communication link's quality,mak-ing it a key technology for 5G communication networks.Signal detection is an essential part of multi-antenna wireless system.Using appropriate precoding and detection technology,massive MIMO systems can effectively suppress the interference and improve the system's performance.Due to the deployment of large-scale antenna arrays at the base station and the increase in the number of service users,the scale of signal processing data is too large,and the complexity of signal processing algorithms has become an essential factor restricting system performance.Therefore,low-complexity detection algorithm is an important research topic of massive MIMO technology.The linear Minimum Mean Square Error(MMSE)detection algorithm has low compu-tational complexity and can obtain good detection performance at the same time,and has received extensive attention from the academic community.However,in the past 20 years,for multi-antenna wireless communication systems,the signal-to-interference plus noise ratio(SINR)of the MMSE receiver has no accurate general distribution expression.This thesis will focus on the linear detection algorithm based on the MMSE criterion in the multi-antenna communication system,considering different scenarios:from single-cell communication system to multi-cell communication system?from uncoded communication system to convolutional coded communication system?from small-scale to large-scale fad-ing environment?from uncorrelated fading to spatially correlated fading channels,and the specific research work is as follows:First,in an independent Rayleigh fading environment,we analyze the performance of an uncoded MIMO system with a linear MMSE detector.For the output SINR of the MMSE detector,we first determine the conditional distribution for the SINR by fixing the eigenval-ues involved in SINR.In the second step,we average over the distribution of the eigenvalues,and obtain an exact closed-form expression for the distribution of the SINR,which is valid for any number of transmit and receive antennas.The derived expression for the SINR is a standard function,which is a linear combination of gamma distributions and the polynomial of 1/(1+x),where x denotes the SINR.Then,based on the SINR density,we analyze the perfor-mance of MMSE-detected MIMO systems in terms of Symbol Error Probability(SEP).Exact and closed-form expressions for the SEP are derived for Pulse Amplitude Modulation(PAM)and Quadrature Amplitude Modulation(QAM),and approximate SEP for Phase Shift Key-ing(PSK)is obtained.Compared with the current best performance analysis results[86],the newly given SINR distribution expression is general,and the SEP analytical expression does not make any approximations.In addition,using the derived SINR distribution,the MMSE MIMO system's achievable sum-rate is investigated,and the corresponding closed expression is present.Finally,Monte Carlo simulation verifies the accuracy and validity of the theoretical analysis.Second,in a single-cell large-scale multi-user MIMO system based on convolutional coding,assuming that the base station already knows the ideal channel state information(CSI),the system error performance based on MMSE detection is analyzed.To characterize the differences in the positions of users in the system,the channel model considers both small-scale fading and large-scale fading.First,an approximate SINR distribution suitable for any antenna configuration is derived.Then,using the SINR distribution,the QAM modulated massive MIMO system's pairwise error probability is further analyzed.Finally,a tight upper bound on the bit error rate performance of a convolutional coded massive MIMO system is given.The simulations in all cases show that the results obtained using the derived expressions closely match the Monte Carlo simulation results.Third,pilot pollution is fundamentally limiting the performance of multi-cell massive MIMO systems.We focus on a multi-user MIMO system with a multi-cell scenario,where the base station uses the pilot sequence sent from the terminal to estimate CSI.We adopt the multi-cell MMSE(M-MMSE)scheme that takes into account pilot pollution to suppress intra-cell and inter-cell interference.Specifically,we derive an accurate and closed-form expression of the SINR distribution for the first time,which applies to any user in the reference cell in a multi-cell system equipped with any number of antennas at the base station.Using the gen-eral probability density distribution(PDF)of the output SINR for the M-MMSE detector,we further obtained the moment generation function(MGF)and cumulative distribution function(CDF)of the output SINR.Then as a verification,based on the derived SINR distribution,we respectively analyzed the system's uplink achievable sum-rate and the outage probability that can reflect the system link reliability,and gave the corresponding accurate closed expression.Besides,we conducted an asymptotic analysis of the system's relevant performance metrics.Our accurate analysis and asymptotic analysis are verified by simulation.Fourth,under Spatial transmit-correlated Rayleigh fading conditions,the performance of the MIMO MMSE detector is investigated.Assume that the data streams on different antennas have different transmit powers,and the perfect CSI is known at the receiver.We first derived an exact expression for the statistical distribution of the output SINR for the MMSE detector.On this basis,the MIMO MMSE system's achievable sum-rate is analyzed,and an exact solution of the achievable sum-rate is given.The derived analytical expression is valid for any number of transmitting and receiving antennas.In addition,in the case of precoding optimization of the transmitted signal's spatial characteristics,the SINR distribution for MMSE MIMO system,error probability(SEP),and closed expressions of the achievable spectral efficiency are respectively derived.Finally,we conducted a progressive analysis of the system's performance metrics,and the results showed that as the number of antennas increased,the correlation of the channel gradually weakened.