Research On The Performance Of Massive MIMO Wireless Systems Based On Low-resolution ADC

As the date rates and bandwidths of communication systems scale up,the cost and power consumption of high-precision(8-12 bits)analog-to-digital converters(ADCs)become prohibitive.One possible approach to relieve this bottleneck is to redesign communication systems with the receiver employing ADCs with drastically reduced precision(e.g.,1-3 bits).Recent information-theoretic analysis for the AWGN channel shows that,the use of 2-3 bit ADCs leads to only a small but acceptable degradation in channel capacity.As one of the most crucial technologies of the 5th generation(5G)mobile communication systems,massive multi-input-multi-output(MIMO)technology is capable of fully exploiting the spatial resources,and allow for significant improvement in spectral and energy efficiency.This thesis focus on the application of low resolution quantization in massive MIMO communication systems,and we will investigate how low precision quantization affects the system performance and channel estimation.First,based on a simplified model of a multi-user massive MIMO system,we present an overview of prin-ciples and theory of massive MIMO,including the downlink and uplink transmission process,the strategies used in the base station for precoding and detection,channel estimation schemes and the concerns about chan-nel estimation in massive MIMO systems.Plus,by analyzing the structure of ADC,we provide an overview of the analytic methods of channel capacity,primarily including the nonlinear quantization modeling based channel capacity and additive quantization noise model(AQNM)based channel capacity.We explain the detailed elaboration and derivation of these two kind of analysis of channel capacity,which provide a sound theoretical basis for the following research on the system performance in the massive MIMO systems with low resolution receivers.Subsequently,we investigate the spectral efficiency and energy efficiency when different technologies are used in the receiver of massive MIMO systems with low precision quantization.Based on the ideal channel state information(CSI),we choose the AQNM model to analyze the effect of quantization.Then we derive an approximate expression for the uplink spectral efficiency of quantized massive MIMO systems,using maximum ratio combining(MRC)receiver and zero forcing(ZF)combing receiver,respectively.Based on the approximate expression,we analyze how quantization resolution,the number of antennas on base station,and transmit power affect the spectral efficiency of low resolution quantized massive MIMO systems.Furthermore,we compare the achievable spectral efficiency under MRC and ZF detectors,as well as illustrate the energy efficiency performance of the system.Numerical results validate the accuracy of our derived result and show that increasing the number of BS antennas can compensate the loss in system performance resulted from low-resolution ADCs,which corroborates the feasibility of installing finite precision ADCs in practical massive MIMO systems.Finally,we study the channel estimation and optimize the ADC resolution in quantized massive MIMO systems.Focusing on a quantized massive MIMO systems in time division duplex mode,in which both the minimum mean-square error channel estimator and MRC receivers are used.After a proper approximation,an explicit spectral efficiency formula is derived.Utilizing it,we analyze the effect of quantization bits and number of antennas at BS,and find that they both influence the channel estimation and system performance greatly.Based on a general power consumption model accounting for the resolution of the ADCs,we carry out an optimization problem to maximize the spectral efficiency subject to a constraint on the total power consumption.The optimization variables are the number of antennas at BS and the quantization bits.The results show the trade-off between the number of BS antennas and the quantization bits,which is important for practical hardware design.Numerical results demonstrate that the optimal level of hardware impairments yields ADCs of 4 to 5 quantization bits.