面向5G的毫米波FDD大规模MIMO系统信道估计方法和导频序列研究

Massive multiple-input multiple-output (MIMO) is a key technique for the fifth generation of broad mobile communication networks (5G)，in which the number of antennas at base station is much larger than the conventional MIMO. At the same time, as most of the frequency bands bellows 3GHz have been occupied, the attention on acquiring new spectrum for 5G has shifted to millimeter wave (mmWave). These promising techniques work well only based on the reliable channel state information (CSI) estimation and the acquirement of the mmWave channel propagation characteristics. However, it is difficult to achieve the reliable CSI in massive MIMO due to the heavy CSI feedback load and pilot overhead. Therefore, in this project the channel estimation and pilot design for mmWave massive FDD MIMO-OFDM systems are investigated from three perspectives.1) The wireless transmit characteristics of mmWave massive MIMO are investigated as well as the mmWave FDD massive MIMO-OFDM system model that is used for conducting the open/closed loop channel estimation and the best pilot training. The mmWave massive MIMO channels are sparse in nature due to the limited local scatters and the highly spatial correlation since a practical antenna array is usually installed in a physically fixed space. By exploiting the inherent channel sparsity, a structured block-sparse channel scheme would be modeled. 2) In conventional OFDM, the orthogonal pilot paradigms are employed to prevent the mutual interference among different antennas. However, the orthogonal pilot schemes result in a heavy pilot overhead in FDD massive MIMO-OFDM. Therefore, a non-orthogonal pilot paradigm would be proposed. Then, based on the non-orthogonal pilot design, the downlink channel estimation would be investigated by using the sparsity support within individual user channel matrix. At the same time, based on two-dimension CS techniques, one novel feedback scheme for channel estimation would be presented. Finally, in this section the block-sparse recovery algorithm would be investigated. 3) By exploiting the partially shared support between different user channel matrices, we would present a closed-loop channel estimation and pilot training paradigm as well as the best training sequence selection strategies. In the closed loop scheme, the users feed back the pilot measurement to base station. After the base station collects all the pilot measurements, the channel estimation and the best pilot selection are performed by combining the current measurements and the previous ones stored in based station. The project research results not only can provide technical support for massive MIMO application, will also lay the theoretical foundation for the development of mmWave FDD massive MIMO-OFDM.

课题拟从三个方面研究毫米波FDD Massive MIMO-OFDM系统信道估计和导频设计问题，提出可大幅度降低反馈开销和导频开销的导频规划和信道估计反馈方案，促进5G研究和发展。1）在毫米波，网络覆盖和天线间距变小，用户共享小区散射体，课题先研究毫米波Massive MIMO信道特征，建立结构化块稀疏信道模型，并设计用于信道估计和导频训练的毫米波FDD Massive MIMO-OFDM系统模型；2）应用单用户信道脉冲响应共同支撑集概念，建立基于压缩感知（CS）的单用户开环下行信道估计和非正交导频设计方案，建立基于两维CS技术的信道估计反馈方案，研究结构化块稀疏CS恢复算法；3）为了进一步降低导频开销和用户负担，利用多用户毫米波Massive MIMO-OFDM信道共同支撑集模型，研究基于CS技术的多用户协同闭环信道估计和导频训练方案，获得最佳导频序列选择准则和多用户支撑集CS恢复算法。

课题研究了毫米波Massive MIMO-OFDM系统信道估计、导频设计及其应用，提出了可大幅度降低系统开销的导频规划和信道估计反馈方案。. 1、毫米波Massive MIMO信道特征及其估计. 课题先研究了毫米波Massive MIMO信道特征，建立结构化块稀疏信道模型，设计出了可用于信道估计和导频训练的毫米波FDD Massive MIMO-OFDM系统模型。应用单用户信道脉冲响应共同支撑集概念，建立了基于压缩感知(CS)的单用户开环下行信道估计和非正交导频设计方案，建立了基于两维CS技术的信道估计反馈方案，研究了结构化块稀疏CS恢复算法。利用多用户毫米波Massive MIMO-OFDM信道共同支撑集模型，研究了多用户协同闭环信道估计和导频训练方案，获得了最佳导频序列选择准则和多用户支撑集CS恢复算法。. 2、基于位置协助的空-地毫米波自适应信道估计. 未来的5G/B5G网络将可能突破传统的单一的地面网络配置模式，将通过配置空中无人机(UAV)基站、联合卫星通信构建空、天、地一体化通信网络。为了支持基站与无人机之间稳定的高传输速率通信及快速接入，项目提出了一种基于位置信息协助的无人机毫米波通信信道估计方法，构建了无人机和毫米波基站的三维位置模型，使用分层多分辨率码本和自适应信道估计方法，利用来自全球卫星导航系统的侧面信息辅助信道估计。. 3、MIMO OFDM系统中基于导频的FBMC/OQAM双选择信道估计. 课题提出了一种基于导频的FBMC/OQAM双选择信道估计算法。该方案采用Turbo编码辅助导频的方法来消除导频处的干扰，计算导频处的信道估计信息。通过导频处的估计信息来更新信道传输矩阵，然后，采用最小均方误差加权法对传输矩阵进行估计。仿真结果证明，所提出的估计方案在误码率和信号干扰比方面优于传统方案。. 4、基于课题研究成果，开展了相关的应用研究。研究了联合大规模MIMO和D2D技术的混合网络，设计了无线充电方案；为了提高大规模MIMO系统的能量效率，研究了基于低分辨率模数转换的全双工大规模MIMO中继系统、混合分辨率大规模MIMO系统等；为了降低传输延迟，研究了基于全毫米波MIMO基站的异构网络混合回程及缓存协助内容传递方案；为了提高网络的安全性，研究大规模MIMO网络的物理层安全问题。

内容获取失败，请点击重试