Accurate Wideband Localization In Massive MIMO Systems With Low-Resolution ADCs

Massive multiple-input multiple-output(MIMO)is a necessary technology for nextgeneration wireless communications,in which base station(BS)with large antenna arrays serves multiple users on the same time-frequency resources,thus achieving high spectrum and energy efficiency.The large number of antenna arrays significantly increases the hardware cost and power consumption of wireless communication systems.The desire to render massive MIMO practical prompts the research activities on quantized large-scale MIMO systems with low-resolution Analog-to-Digital Converters(ADCs).However,the low-resolution ADC introduces significant distortion to the original received signal,thus conventional channel estimation algorithms produce severe performance degradation.At the same time,the potential benefits of large-scale MIMO in terms of accurate localization are also underutilized.For this reason,the design of channel estimation and localization algorithms for quantized MIMO systems is an urgent problem to be solved.This thesis investigates channel estimation and localization of an uplink large-scale MIMO orthogonal frequency division multiplexing(OFDM)system with low-resolution ADC.On the one hand,Cramér-Rao bound(CRB)is derived for the uncertainty of channel parameter estimation and localization,taking into account the effect of nonlinear distortion due to quantization.Using the conversion matrix,the Fisher information matrix(FIM)of the channel parameters is converted into the FIM of the location parameters to obtain the corresponding CRB matrix.With the obtained CRB,one can evaluate the dependence of channel estimation and localization accuracy on the number of quantized bits.It can serve as the performance benchmark to guide the design of subsequent algorithms.Numerical results show that accurate localization can be achieved even with low-resolution ADCs.On the other hand,this thesis proposes a new two-stage joint channel estimation and localization scheme.Firstly,for a single-input multiple-output(SIMO)system,in the first stage,this thesis tries to finely estimate the channel parameters based on generalized turbo(GTurbo)and sparse Bayesian learning(SBL).In the second stage,the obtained channel parameter estimates are used to recover mobile station(MS)position,and the GaussNewton algorithm is used to minimize the nonlinear least square objective function.Simulation results show that the proposed scheme outperforms existing counterparts,and can approach the theoretical CRB at medium to high signal-to-noise ratio(SNR).Secondly,the proposed algorithm is extended to a MIMO system,where the proposed algorithm is still able to reach the corresponding theoretical CRB as the SNR increases.