【电信学】【2018.03】大规模多用户MIMO检测算法与架构

本文为黎巴嫩贝鲁特美国大学（作者：HADI AKRAM SARIEDDEEN）的博士论文，共199页。

在对多输入多输出（MIMO）技术进行了几十年的研究之后，包括从点到点到多用户MIMO（MUMIMO）的模式转换，大量的文献研究了利用空间维度来提高无线通信系统链路吞吐量和网络容量的技术。大规模MIMO支持基站（BS）的数百个天线，被誉为即将到来的第五代（5G）无线通信标准的关键技术。然而，未来使用大型MIMO系统也是必不可少的，特别是对于高速无线回程连接。大型MIMO系统在通信终端中使用数十根天线，在发射端和接收端都能够提供大量天线。虽然在大规模MIMO中的良好传播确保了通过简单的线性处理实现可靠的性能，但是在大规模MIMO中固有的对称性使得接近最佳的信号处理方案的计算复杂度随天线数目呈现指数变化。

本文主要研究大型MIMO和高阶MU-MIMO系统中的有效数据检测问题。首先，针对常规MIMO系统提出了一种近似最优的低复杂度检测算法。然后，提出了一种基于大型MIMO系统的信道矩阵删失的低复杂度硬输出和软输出检测方案。对这些方案的性能进行了数学描述和分析，给出了容量、分集增益和误码概率的边界。在联合调制分类和子空间检测的基础上，提出了一种高效的高阶MU-MIMO检测器，该检测器估计干扰源的调制类型，同时单独检测多个分离数据流。设计了该算法的硬件结构，并通过仿真验证了算法的预期效果。最后，我们将所研究的基于搜索的检测方案映射到大规模MIMO中发射机侧的低分辨率预编码，并给出了性能与复杂度折衷的分析。

After decades of research on multiple-inputmultiple-output (MIMO) technology, including paradigm shifts frompoint-to-point to multiuser MIMO (MUMIMO), an ample literature exists ontechniques to exploit the spatial dimension to increase link throughput andnetwork capacity of wireless communication systems. Massive MIMO, which supportshundreds of antennas at the base station (BS), is celebrated as the keyenabling technology of the upcoming fifth generation (5G) wirelesscommunication standard. However, the use of large MIMO systems in the future isalso indispensable, especially for high-speed wireless backhaul connectivity.Large MIMO systems use tens of antennas in communication terminals, and canafford a large number of antennas on both the transmitter and the receiversides. While favorable propagation in massive MIMO ensures that reliableperformance can be achieved by simple linear processing, the inherent symmetryin large MIMO renders the computational complexity of near-optimal signalprocessing schemes exponential in the number of antennas. In this thesis, weinvestigate the problem of efficient data detection in large MIMO and highorder MU-MIMO systems. First, near-optimal low-complexity detection algorithmsare proposed for regular MIMO systems. Then, a family of low-complexityhard-output and soft-output detection schemes based on channel matrixpuncturing targeted for large MIMO systems is proposed. The performance ofthese schemes is characterized and analyzed mathematically, and bounds oncapacity, diversity gain, and probability of bit error are derived. After that,efficient high order MU-MIMO detectors are proposed, based on joint modulationclassification and subspace detection, where the modulation type of theinterferer is estimated, while multiple decoupled streams are individuallydetected. Hardware architectures are designed for the proposed algorithms, andthe promised gains are verified via simulations. Finally, we map the studiedsearchbased detection schemes to low-resolution precoding at the transmitterside in massive MIMO and report the performance-complexity tradeoffs.

1. 引言
2. 系统模型与参考检测器
3. 基于大型星座的迭代MIMO检测
4. 降复杂度的QRD检测
5. 基于信道删失的大型MIMO检测
6. 通信容量、分集与BER分析
7. 双层多用户MIMO检测
8. 高阶多用户MIMO检测
9. 自适应MIMO系统的逐层MC
10. 大规模MIMO的1bit预编码
11. 结论与未来工作展望

更多精彩文章请关注公众号：