本文为瑞典林雪平大学（作者：Hien Quoc Ngo）的毕业论文，共67页。
在过去的数十年中，连接的无线设备数量有了巨大的增长。数以十亿计的设备通过无线网络进行连接和管理。同时，每个设备都需要高吞吐量来支持语音、实时视频、电影和游戏等应用程序，对无线吞吐量和无线设备数量的需求将不断增加。此外，人们对无线通信系统的能耗也越来越关注。因此，未来的无线系统必须满足三个主要要求：i）具有高吞吐量；ii）同时服务于多个用户；iii）具有较少的能耗。大规模多输入多输出（multiple input multiple output，MIMO）技术是在同一时频资源中为多个用户提供大量天线（并置或分布式）的基站（BS）技术，可以满足上述要求，是下一代无线系统的一种很有前途的候选技术。在BS处具有大量天线阵列时，对于大多数传播环境的信道变得有利，即用户和BS之间的信道向量是（几乎）成对正交的，因此，线性处理几乎是最优的。由于复用增益和阵列增益，可以实现巨大的吞吐量和能量效率。特别是，通过简单的功率控制方案，大规模MIMO可以为所有用户提供一致的良好服务。
本文主要研究大规模MIMO系统的性能。本文主要包括两个部分：大规模MIMO的基本原理和系统设计。在第一部分中，我们专注于在实际约束下系统性能的基本限制，例如低复杂度处理、每个相干间隔的有限长度、小区间干扰和有限维信道。我们首先研究MIMO上行链路的最大比合并（MRC）、迫零、最小均方误差接收机在完美和不完美信道中的节能问题。研究了能量和频谱效率的折衷。其次，我们考虑一个物理信道模型，其中角度域被划分成有限个不同方向。推导了信道容量的下限，分析了有限维信道模型中导频污染的影响。最后，研究了在瑞利衰落和视线（LoS）信道下大容量MIMO系统中的有利传播问题。我们发现瑞利衰落和视线环境都提供了良好的传播性能。
第二部分在第一部分基本分析的基础上，提出了大规模MIMO系统的设计方案。在大规模MIMO系统中，信道状态信息的获取是非常重要的。通常，通过上行链路训练BS处的估计信道。由于相干间隔的长度有限，系统性能受到导频污染的限制。为了降低导频污染的影响，我们提出了一种基于特征值分解的直接从接收数据中估计信道的方法。与传统的训练方案相比，由于减少了先导污染，该方案的训练效果更好。大规模MIMO中CSI捕获的另一个重要问题是如何在用户处获取CSI。为了解决这一问题，我们提出了两种针对用户的信道估计方案：i）下行波束形成训练方案，ii）一种有效下行信道增益的盲估计方案。在这两种方案中，信道估计开销与BS天线的数目无关。我们还推导出最优导频和数据功率以及训练持续时间分配，以最大化与MRC接收机共用的大规模MIMO上行链路的总频谱效率，用于在相干间隔中消耗给定总能量预算的情况。最后，提出并分析了大规模MIMO在中继信道中的应用。具体地说，我们考虑多对中继系统，其中多个源在同一时频资源中同时与多个目的地通信，从而借助于大规模MIMO中继。一个大规模的MIMO中继配备有许多并置或分布式天线。我们考虑不同的双工模式（全双工和半双工）和不同的中继协议（放大转发、解码转发、双向中继、单向中继），从频谱效率和功率效率两个方面探讨了大规模MIMO技术在这些中继系统中的潜在优势。
The last ten years have seen a massivegrowth in the number of connected wireless devices. Billions of devices areconnected and managed by wireless networks. At the same time, each device needsa high throughput to support applications such as voice, real-time video,movies, and games. Demands for wireless throughput and the number of wirelessdevices will always increase. In addition, there is a growing concern aboutenergy consumption of wireless communication systems. Thus, future wirelesssystems have to satisfy three main requirements: i) having a high throughput;ii) simultaneously serving many users; and iii) having less energy consumption.Massive multiple-input multiple-output (MIMO) technology, where a base station(BS) equipped with very large number of antennas (collocated or distributed)serves many users in the same time-frequency resource, can meet the aboverequirements, and hence, it is a promising candidate technology for nextgenerations of wireless systems. With massive antenna arrays at the BS, formost propagation environments, the channels become favorable, i.e., the channelvectors between the users and the BS are (nearly) pairwisely orthogonal, andhence, linear processing is nearly optimal. A huge throughput and energyefficiency can be achieved due to the multiplexing gain and the array gain. Inparticular, with a simple power control scheme, Massive MIMO can offer uniformlygood service for all users. In this dissertation, we focus on the performanceof Massive MIMO. The dissertation consists of two main parts: fundamentals andsystem designs of Massive MIMO. In the first part, we focus on fundamentallimits of the system performance under practical constraints such as lowcomplexity processing, limited length of each coherence interval, intercellinterference, and finite-dimensional channels. Wefirst study the potential forpower savings of the Massive MIMO uplink with maximum-ratio combining (MRC),zero-forcing, and minimum mean-square error receivers, under perfect andimperfect channels. The energy and spectral efficiency tradeoff isinvestigated. Secondly, we consider a physical channel model where the angulardomain is divided into a finite number of distinct directions. A lower bound onthe capacity is derived, and the effect of pilot contamination in this finite-dimensionalchannel model is analyzed. Finally, some aspects of favorable propagation inMassive MIMO under Rayleigh fading and line-of-sight (LoS) channels areinvestigated. We show that both Rayleigh fading and LoS environments offerfavorable propagation.
In the second part, based on thefundamental analysis in the rst part, we propose some system designs forMassive MIMO. The acquisition of channel state information (CSI) is veryimportant in Massive MIMO. Typically, the channels are estimated at the BSthrough uplink training. Owing to the limited length of the coherence interval,the system performance is limited by pilot contamination. To reduce the pilotcontamination effect, we propose an eigenvalue-decomposition based scheme toestimate the channel directly from the received data. The proposed schemeresults in better performance compared with the conventional training schemesdue to the reduced pilot contamination. Another important issue of CSIacquisition in Massive MIMO is how to acquire CSI at the users. To address thisissue, we propose two channel estimation schemes at the users: i) a downlinkbeamforming training scheme, and ii) a method for blind estimation of theeffective downlink channel gains. In both schemes, the channel estimationoverhead is independent of the number of BS antennas. We also derive theoptimal pilot and data powers as well as the training duration allocation tomaximize the sum spectral efficiency of the Massive MIMO uplink with MRCreceivers, for a given total energy budget spent in a coherence interval.Finally, applications of Massive MIMO in relay channels are proposed andanalyzed. Specifically, we consider multipair relaying systems where manysources simultaneously communicate with many destinations in the sametime-frequency resource with the help of a Massive MIMO relay. A Massive MIMOrelay is equipped with many collocated or distributed antennas. We considerdifferent duplexing modes (full-duplex and half-duplex) and different relayingprotocols (amplify-and-forward, decode-and-forward, two-way relaying, andone-way relaying) at the relay. The potential benefits of massive MIMOtechnology in these relaying systems are explored in terms of spectralefficiency and power efficiency.
I. 引言
- 研究动机
- 多用户MIMO蜂窝系统
- 大规模MIMO
- 数学基础知识
- 本文贡献总结
- 未来研究方向
II. 大规模MIMO基础
A. 超大规模多用户MIMO系统的能量与频谱效率
B. 具有超大天线阵和有限维信道的多小区多用户MIMO上行链路
C. 大规模MIMO中的有利传播问题
III. 系统设计
D. 基于EVD的超大天线阵列多小区多用户MIMO信道估计
E. 具有线性预编码和下行链路导频的大规模MU-MIMO下行链路TDD系统
F. 大规模MIMO中目标下行链路信道增益的盲估计
G. 具有最优功率和训练持续时间分配的大规模MIMO
H. 分布式AF波束形成的大规模多对双向中继网络
I. 大规模阵列多对双向中继信道的频谱利用率
J. 大规模阵列多对全双工中继及线性处理
更多精彩文章请关注公众号：
