多层异构网中基于残缺信道矩阵的鲁棒性干扰对齐问题研究

Heterogeneous networking is a promising solution to improve wireless converge, relieve over-crowded bandwidth and enhance system throughput for next generation wireless networks. Because the locations of the femtocell base stations (FBSs) are generally unknown (even can dynamic) and the mobile users are moving unpredictably. These types of interference cannot be efficiently handled by conventional global frequency resource scheduling. When femtocell network are dense, this issue becomes more severe since the large scales spectrum optimization becomes intractable. However, one of the main concerns to put through heterogeneous networks is the interference among over-laid cells. ..Recently developed interference alignment (IA) can achieve the maximum degrees-of-freedom (DoF) in the K-user multiple-input multiple-output (MIMO) interference channel, which becomes one of effective approaches to control the interference. By forcing interference signals at each receiver into a reduced-dimensional subspace of the received subspace, the receivers can observe an interference-free desired signal if it lies outside of the interference subspace. Standard IA approaches rely on perfect channel state information (CSI). Any types of errors may lead to performance degrades rapidly. However, in practical wireless systems, perfect channel estimates are not available due to limited feedback channel capacity, estimation/quantization errors or outdated channels. This has motivated the effort to develop robust algorithms against channel imperfection...In this work, we consider the heterogeneous network and develop a overhead-aware model that the uncoordinated femtos can intelligently join in and leave the existed interference channel. Considering the channel imperfection, we propose robust designs under the assumption of incomplete channel matrix. The underlying non-convex problem will be transferred into convex optimization problem and solved by standard convex programming. The proposed result could provide a theoretical basis for the IA implementation in the next generation of wireless communication systems..

由于宏蜂窝和小功率节点所构建的异构网络可以大幅度提高网络的容量，日渐成为未来移动通信的发展趋势。但是微蜂窝的引入增加了异构网中小区数量，使得小区间同道干扰问题更加突出，成为限制异构网可达容量的主要因素。近年提出的干扰对齐技术能获得干扰网络的最大自由度并可达其最优容量。然而，标准的干扰对齐技术大都基于完备信道，对信道误差十分敏感，微小的误差都可导致系统性能的衰减。因此，急需设计基于非完备信道信息的干扰对齐算法，增强算法的鲁棒性以对抗信道误差。本项目拟研究基于异构网的干扰信道，计算干扰网络的最大自由度以及可达最优容量；研究异构网中协作节点与非协作节点的工作方式，建立非协作小功率节点低复杂度的准入控制模型；针对基于残缺信道矩阵的非完备信道信息模型进行分析，设计具有鲁棒性的干扰对齐算法，为干扰对其进行在下一代无线通系统中的运用提供理论依据。

由于宏蜂窝和小功率节点所构建的异构网络可以大幅度提高网络的容量，但小区间干扰成为限制异构网容量的主要因素。申请人前期研究结果表明，基于波束赋形算法对齐算法在信道完备的情况下消除小区间同道干扰。本项目在前期工作基础上，本项目基于异构网的非完备干扰信道，创建了基于秩最小化的干扰消除模型，研究利用波束赋形对干扰进行消除。与此同时，在研究数字波束赋形算法同时，发现模型波束赋形可极大降低硬件复杂度以及天线能量消耗，可进一步提高系统能量效率。此研究为模拟波束赋形技术在毫米波大规模多入多出天线系统（mmWave Massive MIMO）使用提供了可行方案。本项目资助发表期刊论文2篇，会议论文5篇，待发表期刊论文3篇。培养硕士研究生5名，其中1名已经取得硕士学位，4名在读。