Impacts Of Low-resolution ADCs On Performance Of Massive MIMO Systems

Owing to increasing system capacity,massive multiple-input and multiple-output(MIMO)has extensively been studied in recent years.As the number of antennas grows larger,on one hand,channel harding and favorable propagation can be exploited to simplify the implementation of the receiver.On the other hand,from the hardware perspective,the total number of analog-to-digital converters(ADCs)should increase since each receive antenna needs a pair of ADCs.However,it has been demonstrated that the power consumption of one ADC scales exponentially in terms of the number of quantization bits.To overcome this challenge,the use of low-resolution ADCs in the upcoming massive MIMO systems is a potential solution.Therefore,the impacts of low-resolution ADCs on system performance,and how to perform channel estimation and data detection are needed to be addressed.In addition,the impact of in-phase/quadrature-phase imbalance(IQI)is also one of the main problems in quantized massive MIMO systems.Therefore,it is necessary to compensate the impacts of IQI.In summary,what we study in this thesis are given as follows.Firstly,we provide the performance analysis of uplink Massive MIMO systems with low-resolution ADCs for minimum mean square error(MMSE)and the maximum ratio combining(MRC)detections.Using perfect channel state information(CSI)and results from the random matrix theory,we derive the closed-form approximate expression of sum achievable rate for these two type detectors,respectively.Suppose that only imperfect CSI is available at the BS,We then provide a scheme to estimate the equivalent channel based on the classical LMMSE technique and derive the achievable rate using the estimated equivalent channel.Moreover,we take the impacts of channel estimation error and quantization noise into account when deriving the corresponding analytical results.Simulation results are also presented to validate our theoretical analysis.These results not only show the impacts of low-resolution ADCs on the performance of massive MIMO systems,but also testify the effectiveness of low-resolution ADCs from the view of energy efficiency(EE).Secondly,we discuss the detection techniques for quantized uplink massive MIMO systems.Based on the system model,in addition to the traditional detectors,we present an improved method based on the generalized approximate message passing(GAMP)algorithm and discuss its performance under mixed-ADC architecture.Based on the GAMP algorithm,for coded systems,a low-complexity detection method is proposed.The proposed method combines the GAMP detection with channel decoder and exchanges extrinsic information between them,by which the system performance can be improved.Contrasted to the iterative MMSE detection,our method circumvents large-scale matrix inverse operation and leverages the statistical properties of both quantization errors and transmitted symbols.Moreover,we verify that the bit error ratio(BER)performance of the proposed method is equivalent to that of iterative MMSE but with less complexity for implementation.Thirdly,for the channel estimation of quantized millimeter wave(mmWave)massive MIMO systems,we propose two methods.One of them is an improved channel estimator based on orthogonal approximate message passing(OAMP).Using channel sparsity and the orthogonality of steering matrix,the proposed method can not only improve estimate performance but also has less iterations,better convergence and inverse-free implementation.In addition,to obtain a better performance of channel estimate for the off-grid scenario,we provide a scheme to update steering matrix and channel coefficients in an alternating fashion.To further improve the estimation performance,we then propose a joint channel estimation and data detection(JCD)approach based on expectation maximization(EM).The proposed method not only uses pilots to remove phase ambiguities,but also exploits the data recovered from the previous iteration.As a result,the performance can be improved.In addition,our proposed scheme is capable of feeding back the estimated large-scale fading coefficients and channel sparsity level to the automatic gain control(AGC)to adjust the dynamic range of the received signals.Simulation results show the superiority of the proposed methods.Finally,by modeling the IQI parameters as random variables,this thesis uses the statistic property of the effective channel to simplify the receiver.To use bilinear generalized approximate message passing(BiG-AMP),we theoretically derive the probability density function(PDF)of the elements in the effective channel.Therefore,we investigate the channel estimation and IQI compensation for massive MIMO with low-resolution ADCs.For this purpose,two techniques to compensate IQI are proposed.One is referred to as combined-signal-based channel estimation and compensation(CCEC)and the other is denoted by effective channel estimation and compensation(ECEC).The CCEC remove the impacts of IQI by combining the received signal with its conjugated version.The ECEC perform channel estimation and data detection using the modified BiG-AMP.Compared with other classical methods,the proposed methods can obtain better performance based on the Monte-Carlo simulation results.