5G大规模MIMO在非理想系统约束下保密传输技术研究

The fifth generation cellular communication technology (5G) is going to introduce disruptive changes in the cellular architecture, and in the newly introduced application scenarios and services. As a consequence, security strategies for mobile networking must be substantially redesigned, which is very challenging task. It is noted that standardization organizations such as the IMT-2020(5G) in China and the METIS in Europe have not specified design principles for 5G security in their proposed architectures. Massive multiple-input-multiple-output (MIMO) technology, also known as large-scale MIMO, is one of the most important base station (BS) technological breakthroughs for 5G air interface. This technology is critical to the realization of service-specific security and edge-network security in 5G through physical layer (PHY) security. However, due to cost considerations, commercial massive MIMO BS equipments will introduce to the system non-ideal constraints such as hardware impairments and limited number of RF chains. There are still many open problems about the impact of those non-ideal constraints on PHY security of massive MIMO. This project will be based on modeling of the non-ideal system constraints in practical commercial massive MIMO systems, and use secrecy capacity as the fundamental analytical tool in studying massive MIMO secure transmission in the presence of non-ideal system constraints. In particular, the use of artificial noise (AN) and joint information-AN precoding for improved secrecy performance is considered as the security enhancement scheme for massive MIMO by exploiting its excessive redundant spatial degrees of freedom. Three non-ideal system constraints in massive MIMO, namely transceiver hardware impairments, per-antenna constant envelope non-linear precoding that is insensitive to hardware imperfections, hybrid precoding for limited RF chains constraints are investigated, and AN-aided secure transmission of massive MIMO under these constraints studied. The project is of significance to both theoretical studies and practical system design of 5G, especially in this year when the standardization process of 5G security specifications has been officially launched by ITU-T.

第五代移动通信技术(5G)在网络架构、应用场景和业务类型等方面引入的颠覆性变化，为移动网络安全设计带来巨大挑战，国内外标准化组织对5G架构的描述，都尚未明确其安全设计路线。大规模MIMO技术作为5G宏基站空口技术的核心，是应用物理层安全方法实现5G业务定制安全和边缘网络安全的有效手段。然而，商用大规模MIMO系统由于成本制约将存在硬件缺陷、有限射频链路数等非理想系统约束，其对物理层保密传输的影响尚有诸多开放性问题有待研究。项目以非理想系统约束建模为研究立足点，以系统保密容量推导为基本研究手段，提出非理想约束下大规模MIMO保密传输的理论分析与设计架构。具体研究收发信机硬件缺陷、对硬件精度不敏感的预编码、降低射频链路数等约束模型，结合利用大规模MIMO冗余空间自由度注入人工噪声、抬升窃听接收底噪的保密传输策略，研究保密传输性能与系统设计，从物理层安全角度对5G大规模MIMO基站设计提供指导。

大规模MIMO技术，是5G移动通信及其演进中宏基站空口物理层的核心，是应用物理层安全方法实现业务定制安全和边缘网络安全的有效手段。商用大规模MIMO系统由于成本制约而存在硬件缺陷、有限射频通道数、非线性预编码等非理想系统约束，其对传输性能特别是物理层安全传输的影响尚有待深入研究。项目以非理想系统约束建模为研究立足点，以系统保密容量分析推导为基本研究手段，提出非理想约束下大规模MIMO保密传输的理论分析与设计架构，主要在以下三个方面取得了创新性的研究结果。.1. 针对I/Q通道不平衡和低精度DAC两种硬件缺陷对大规模MIMO预编码的影响进行建模，在不同场景下研究其对预编码方案和保密传输性能的影响。通过研究发现接收机端存在I/Q通道不平衡对系统性能的影响远大于发射机端，而低精度DAC的使用在信道空间相关性高的情况下，有利于提高基于固定功率人工噪声注入的安全传输预编码性能。.2. 研究了对硬件精度不敏感的每天线恒包络（PACE）非线性预编码方案。针对文献中对PACE预编码解读的不准确性和传统实现算法复杂度高的问题，从理论上提出了PACE预编码“信号偏离度”和“中断性”的性能分析思想；基于该问题与MPSK多用户信号检测的相似性，提出了基于SDR松弛和高斯随机化平均的低复杂度优化预编码算法，并发现了该方法在智能反射面（RIS）相控阵控制中的应用前景。.3. 毫米波射频器件昂贵使得毫米波大规模MIMO须采用射频通道数远小于天线数的射频结构和模拟/数字混合处理架构，项目聚焦毫米波大规模MIMO混合架构下的保密预编码设计问题，结合信能同传研究高能效且具有保密传输性能保障的混合预编码方案，通过将原非凸问题分解为两层优化问题，提出了基于序贯参数凸逼近的低计算复杂度高鲁棒性预编码算法，并基于最小化混合预编码与全数字预编码欧氏距离的思想提出了求解模拟预编码的相位匹配算法。