面向B4G/5G移动通信系统的大规模MIMO传输方案研究及其优化设计

The project intends to systematically study the large scale MIMO transmission scheme and its optimal design in B4G/5G system. First, considering physical characteristics and specific application scenarios, with different configurations of the antenna array at the base station, the impact on the system performance in the large scale MIMO system will be studied. Secondly, theoretical research and analysis of various methods for precoding in the large scale MIMO system should be discussed. As the number of base station antennas tends to infinity, the asymptotic performance is obtained. Thus, a multi-cell coordination pilot scheduling scheme will be proposed, which can improve overall system performance. Thirdly, mean square error of channel estimation in the large scale MIMO multi-cell multi-user time division duplex system was used as a criterion to obtain a pilot power control method, which could effectively reduce pilot contamination in one-dimensional multi-cell network and could greatly improve per-cell downlink sum rate. Based on it, it could be extended to more complex cellular network. Two pilot contamination reduction methods were expected, which based on the cell classification and the channel estimation performance, respectively. And the corresponding theoretical analysis should be done. Finally, fusing large scale MIMO and Small Cell technology, as the large antenna array can make the transmitter energy concentrated to a small space region and can significantly reduce the user uplink transmit power, an enhanced interference coordination scheme will be presented, which could greatly improve the performance of the system.

本项目拟系统地研究B4G/5G系统中的大规模MIMO传输方案及其优化设计问题。首先，结合物理特性和具体应用场景，研究大规模MIMO系统中基站端天线阵列的不同配置方式对系统性能的影响。其次，对大规模MIMO下的各种预编码方法进行理论分析，获得当基站天线数趋于无穷时的渐近性能，并据此给出可以改善系统整体性能的多小区协调导频调度方案。然后，针对大规模MIMO多小区多用户时分双工系统，基于上行信道估计误差准则，获得适用于一维线性多小区网络的导频功率控制方法，有效减轻导频污染，提高每小区下行可达和速率；并在此基础上，扩展至更加复杂的蜂窝网络，提出基于小区分类和基于信道估计性能的导频污染减轻方法，并给出相应的理论分析。最后，融合大规模MIMO技术和Smallcell技术，利用大天线阵列可将发射能量集中到较小空间区域以及可大幅度降低用户上行链路发射功率的特性，给出能显著提高系统性能的增强型干扰协调方案。

项目系统研究了5G移动通信系统的大规模MIMO传输方法，及其导频污染与导频优化、能效优化联合预编码与资源分配、通用缩放特性等关键问题及对系统性能影响，同时深入探索了大规模MIMO技术与其它5G关键技术的系统融合，包括：多层异构与大规模MIMO、3D大规模MIMO、无线携能通信与大规模MIMO、毫米波大规模MIMO等，并对系统性能进行了细致理论推导和分析比较，提出若干优化系统方案。项目取得较多创新成果，在典型大规模MIMO系统关键问题及其性能分析方面包括：FDD大规模MIMO系统普适空间相关信道下的下行导频开销分析及导频设计；多小区多用户大规模MIMO系统基于缩放特性的传输性能分析；大规模MIMO多用户系统基于总速率的联合统计波束成形和用户调度；多用户大规模MIMO上行下行系统的能效优化资源分配，以及多小区协作多点系统能效优化的联合功率分配及波束成形；大规模MIMO公共信道下能效优化波束成形；基于多项式拟合的大规模MIMO信道预测器；多用户FDD大规模MIMO系统基于波束域块稀疏性和压缩反馈的信道估计；TDD大规模MIMO系统的导频污染减轻，以及多小区的预编码策略和导频调度方法；多小区广播干扰信道的分层干扰对齐传输方案。在结合5G其它关键技术的大规模MIMO传输方案优化设计与分析方面包括：双向AF中继全双工大规模MIMO系统的天线选择及性能分析；基于比例公平的大规模MIMO异构网络联合用户关联和功率分配；基于干扰消除的多层异构网络的系统性能分析；能效优化的超密集网络中基于安全的用户中心分簇方法；基于图论的超密集网络中用户中心重叠分簇和资源分配的联合设计；3D大规模MIMO系统的低复杂度双层波束成形方法，以及利用基站天线下倾角的无线携能通信；毫米波大规模MIMO中分离型子阵列的低复杂度混合波束成形；多层异构网络中基于博弈随机学习的联合基站关联和资源分配方法；具有额外窃听者及秘密消息的MISO无线携能通信安全传输方法；异构蜂窝网络中基于QoS感知的负载平衡用户关联方案和能效优化的小区选择方案。它们都可用于5G通信系统的优化设计。