多用户大规模MIMO上行系统信道估计方法研究

发布时间：2018-04-18 08:22

本文选题：大规模MIMO + 信道估计　；参考：《大连海事大学》2015年硕士论文

【摘要】：多输入多输出(MMO)技术通过空间复用和天线分集可以在不增加额外带宽及功率的情况下提高信息传输速率和可靠性。但随着智能终端的日益普及,未来移动通信系统需要满足更高的要求：(1)在相同时频资源上服务更多的用户；(2)更高的传输速率及可靠性；(3)更低的功率消耗。为此,Marzetta在2010年提出了大规模MMO技术,可有效地满足上述要求。在大规模MMO系统中,信道估计对保障系统信号检测及基站间的协作传输等都有着至关重要的作用。因此,本文着重探讨大规模MMO系统的信道估计方法。首先,本文分析了导频污染产生的原因及其对信道估计的影响。在多个小区导频复用的场景下,导频污染是影响大规模MMO系统性能的主要因素。当用户数很大时,传统的信道估计方法(如MMSE、LS)受导频污染的影响比较严重,且频谱利用率较低。因此,研究新而有效的信道估计方法十分必要。其次,本文分析了基于矩阵特征值分解(EVD)的信道估计方法。该估计方法利用了大规模MMO系统中信道向量的渐近正交性,对发送信号的结构无任何要求；且每个用户对应的信道矩阵都可以通过接收信号的协方差矩阵估计得到,但会附带一个乘数模糊因子,该模糊因子可以通过发送一个短导频序列估计得到。为进一步降低估计误差,可以将迭代最小二乘投影算法(ILSP)应用于EVD估计方法。仿真结果表明,即使在采用很短的导频序列情况下,EVD估计方法的估计误差也远小于上述传统信道估计方法。最后,本文研究了大规模多用户MMO系统在比较符合实际场景的有限散射路径模型下信道估计问题。由于具有有限散射路径的信道矩阵的自由度远小于其自由参数,因此,本文分析了一种特别适用于该场景的信道估计方法——基于压缩感知(CS)的低秩矩阵估计方法。该方法将低秩矩阵估计问题转为核范数约束优化问题,并通过二次半定规划(SDP)来进行求解。仿真结果表明,在满足相同估计误差的情况下,该方法所需的导频长度和导频发送功率远小于LS估计方法,且对传输信道的统计分布无任何要求。
[Abstract]:Multi-Input-Multiple-Output MMOs (MMOs) can improve the transmission rate and reliability of information without adding additional bandwidth and power through spatial multiplexing and antenna diversity.However, with the increasing popularity of intelligent terminals, the future mobile communication systems need to meet higher requirements: 1) serving more users on the same time-frequency resources. (2) higher transmission rate and higher reliability.Therefore, Marzetta proposed a large scale MMO technology in 2010, which can meet the above requirements effectively.In large scale MMO systems, channel estimation plays an important role in signal detection and cooperative transmission between base stations.Therefore, this paper focuses on channel estimation methods for large scale MMO systems.Firstly, the causes of pilot pollution and its influence on channel estimation are analyzed.Pilot pollution is the main factor affecting the performance of large scale MMO systems in multiple cell pilot multiplexing scenarios.When the number of users is large, the traditional channel estimation methods, such as MMSE LSs, are seriously affected by pilot pollution, and the spectrum efficiency is low.Therefore, it is necessary to study new and effective channel estimation methods.Secondly, the channel estimation method based on matrix eigenvalue decomposition (EVD) is analyzed.The proposed method takes advantage of the asymptotic orthogonality of channel vectors in large-scale MMO systems, and has no requirement for the structure of the transmitted signal, and the channel matrix corresponding to each user can be estimated by the covariance matrix of the received signal.However, there is a multiplier ambiguity factor which can be estimated by sending a short pilot sequence.In order to further reduce the estimation error, the iterative least square projection algorithm (ILSP) can be applied to the EVD estimation method.The simulation results show that the estimation error of the EVD estimation method is much smaller than that of the traditional channel estimation method even when the pilot sequence is very short.Finally, the channel estimation problem of large scale multiuser MMO systems under the finite scattering path model is studied.Because the degree of freedom of channel matrix with finite scattering path is much less than its free parameter, this paper analyzes a channel estimation method based on compressed sensing CS-based low rank matrix, which is particularly suitable for this scenario.In this method, the low rank matrix estimation problem is transformed into a kernel norm constrained optimization problem and solved by quadratic semidefinite programming (SDP).The simulation results show that the pilot length and pilot transmission power of the proposed method are much smaller than that of the LS estimation method, and there is no requirement for the statistical distribution of the transmission channel in the case of the same estimation error.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN919.3

【相似文献】