摘要大规模MIMO系统基站配置天线数量非常大，若采用高精度的模数转换器，基站的成本将变得难以承受。接收端通过减小模数转换器（Analog-Digital Converter，ADC）的量化精度（极端情形是采用1-bit的ADC）可以显著的降低基站的能量和硬件实现开销。33514
本文主要研究大规模MIMO低精度量化系统。论文主要工作如下：
(1) 针对大规模MIMO系统模型，研究了四种线性检测算法，分别是：迫零(Zero-forcing，ZF)，最小均方误差( ，MMSE)，近似最小均方误差(Approximate ，AMMSE)，匹配滤波(Matched filter, MF)。仿真结果显示：四种检测方法误码性能排序为：MMSE>ZF>AMMSE >MF。
(2) 重点研究基于信号重建算法—广义近似消息传递算法（Generalized approximate message passing，GAMP），通过仿真和分析，我们发现：在大规模MIMO 1-bit系统，传统的线性检测器如 和 会出现误码平台，而基于GAMP的近似消息解量化算法（Message passing de-quantization，MPDQ）则能消除误码平台。
(3) 将MPDQ推广到n-bit场景，并推导了相应的迭代计算方法。仿真与分析表明：MPDQ算法在n-bit系统中仍然具有优异的误码率性能，同时3-bit量化系统性能逼近无限精度。综合考虑成本和性能，3-bit量化是一种理想的方案。
关键词  大规模MIMO  近似消息传递  数据检测  低精度量化 1-比特模数转换器
毕业论文设计说明书外文摘要
Title    Design  of  massive  MIMO  system  with   Hybrid-precision  ADC                     
Abstract
When base station is equipped with large number antenna in massive MIMO system, if using high-precision Analog-Digital Converter (ADC), the cost will become unbearable. The receiver can decrease the quantization precision of the ADC (One extreme case is the use of 1-bit ADC) to significantly reduce the cost of the energy and hardware realization.
This paper investigate the data detection techniques in massive MIMO system with low-precision ADC. The main work is as follows:
(1) Based on the system model of the massive MIMO system, we investigate four linear detection algorithms: Zero-forcing (ZF), Matched filter (MF), Minimum Mean Square Error (MMSE), and Approximate Minimum Mean Square Error (AMMSE). Simulation results show their performances have the decreasing order: MMSE>ZF>AMMSE>MF.
(2) We make an investigation of the Generalized approximate message passing (GAMP) algorithm in detail. Simulation results show that in massive MIMO 1-bit system, the curve of the MMSE detector has error floor, but the message passing de-quantization (MPDQ) based on GAMP has no error floor。
(3) The MPDQ is extended to n-bit system, and the corresponding iteration process is derived. Simulation results show that: MPDQ algorithm still has excellent error rate performance in massive MIMO n-bit system, and 3-bit quantization is very close to infinite precision. Thus, the 3-bit quantization is an ideal scheme of striking a good balance between cost and performance.
Keywords  Massive MIMO  AMP algorithm  data detection  Low-precision quantization  1-bit ADC
目  次
1  引言    1
1.1 研究背景与意义    1
1.2 大规模MIMO技术简介    1
1.3 国内外研究现状    3
1.4 论文的组织结构    4
2  大规模MIMO技术    5
2.1 大规模MIMO系统模型    5
2.2 大规模MIMO系统经典的多用户检测算法    5
2.3仿真结果与分析    6
2.4 本章小结    7
3  近似消息传递算法    8 利用混合精度模数转换器的大规模MIMO系统设计:http://www.751com.cn/tongxin/lunwen_30652.html