Massive MIMO Iterative Detection Technology Based On Low Resolution ADCs

The low-resolution ADC structure is an effective solution for reducing the energy consumption and hardware cost in massive MIMO systems.Based on this structure,this paper focus on the iterative detection algorithm to gain a good balance between complexity and performance.Different detection algorithm,two error correction codes:LDPC codes and polarization codes,iterative soft interference cancellation and how to choose low-precision ADC quantization steps in massive MIMO systems are studied.Combined with LDPC codes and polar codes,an iterative detection algorithm based on pseudo-de-quantization detection is proposed and its performance is validated in low-resolution ADC massive MU-MIMO systems.First of all,the classic detection algorithm of the uplink multi-user MIMO system is introduced.Firstly,the uplink multi-user MIMO system model is given.Then the optimal detection technique based on the minimum mean square error criterion is introduced.Since the complexity involved in the optimal detection technique is too high,some simple detection algorithms are introduced in the order of non-iterative detection and iterative detection.The advantages and disadvantages of various algorithms are summarized while describing the algorithm process.After that,the complexity comparison of various algorithms is given.Finally,simulation results of various algorithms are given in different scenarios,which lay a good foundation for the subsequent massive MIMO detection algorithm research.Secondly,three detection algorithms based on the Bayesian inference in low-resolution ADC massive MIMO systems are introduced.Firstly,the low-resolution ADC massive MIMO system model is given.Then the mathematical expression of estimated signal is obtained based on the Bayesian inference.The unified detection algorithm framework is given by combining the generalized approximation message passing technology.By means of the specific selection of the distribution of input signal and likelihood function,three detectors in the order of high complexity to low complexity are obtained: de-quantization optimal detector,pseudo-de-quantization optimal detector and linear detector.After that,how to choose the ADC quantization step to optimize the performance of the MIMO detector is explored.Combined with the simulation results,a quantization step selection criterion used for non-1-bit resolution ADCs is obtained.Finally,the performance of the pseudo-de-quantization optimal detection algorithm and the MMSE detection algorithm are compared.Then,combined with the LDPC codes,the research on iterative detection is carried out.Firstly,the basic knowledge of the LDPC codes: how to describe the LDPC codes,decoding algorithms and so on are introduced.Based on the previous research results,the pseudo-de-quantization optimal detection algorithm is modified to soft input soft output detection algorithm.After that,combined with iterative soft interference cancellation algorithm,an iterative detection algorithm based on pseudo-de-quantization optimal detection is introduced.In addition,the decoding success flag is introduced in the algorithm framework to further improve the interaction between the detector and the decoder.Finally,the ADC quantization step is further explored to optimize the performance of the proposed iterative detection algorithm,and a method calculating the quantization step based on training is proposed.The effectiveness of the proposed method is verified by simulation and the performance of the proposed algorithm is studied.Finally,the related knowledge of polar codes is introduced and the polar codes are applied to the proposed iterative detection algorithm framework based on pseudo-de-quantization detection to further explore the performance of the iterative detection algorithm.Firstly,the related knowledge of the polar codes is introduced.Then,the performance of polar codes and LDPC codes is compared in the AWGN channel.The simulation result shows that the polar codes have better performance than the LDPC codes in the short codes field.Finally,the polar codes are applied to the proposed iterative detection algorithm framework to further verify the performance of the proposed detection algorithm.