大规模MIMO-OFDM系统中基于结构化压缩感知的信道估计及导频优化研究

发布时间：2018-04-09 05:26

本文选题：大规模多输入多输出正交频分复用系统　切入点：结构化压缩感知　出处：《南京邮电大学》2017年硕士论文

【摘要】：大规模多输入多输出-正交频分复用(MIMO-OFDM)系统,因其既可以获得较高的信道容量,同时又会得到较高的能量有效性,而成为未来5G技术的关键。本文研究了将结构化压缩感知理论用于该系统的稀疏信道估计。本文的主要贡献在于:(1)考虑到大规模MIMO-OFDM系统中将每个发送天线上的导频重叠放置,即每个发送天线可以在相同的时频资源块上发送导频符号,那么此时的系统稀疏信道估计问题可以建模为结构化压缩感知重建问题,从而建立了稀疏信道估计与结构化压缩感知的对应关系。(2)考虑到导频设计涉及导频位置以及符号两个关键因素,为了优化导频位置和导频符号来改进稀疏信道估计的质量,本文首先针对导频位置选取提出了与之对应的最小化完全块间相关值的导频优化准则以及基于此准则的导频搜索算法。完全块间相关值是结构化压缩感知框架下衡量恢复矩阵子块间相关程度的量值。仿真结果表明,与其他未优化导频相比,使用此优化方法获得的导频可以使信道估计误差(MSE)明显减小,信道估计性能提高约2-4dB。(3)将导频位置与导频符号这两个因素结合在一起,提出了这种情况下的基于最小化完全块间相关值的导频优化准则以及基于此准则的导频搜索算法。仿真结果同样表明,与其他导频相比,使用此优化算法获得的导频可以使信道估计的MSE明显减小,约2-5dB。同时仿真结果表明导频位置和符号联合优化方法获得的优化导频性能优于单纯优化导频位置获得的优化导频,它能使得大规模MIMO-OFDM系统的信道估计具有更低的MSE。
[Abstract]:Large-scale multi-input-multiple-output (MIMO) -OFDM (orthogonal Frequency Division Multiplexing) system is the key of 5G technology in the future because of its high channel capacity and high energy efficiency.In this paper, we study the application of structured compressed sensing theory to sparse channel estimation of the system.The main contribution of this paper is to take into account the fact that in large scale MIMO-OFDM systems the pilots on each transmit antenna are superimposed, that is, each transmission antenna can transmit pilot symbols on the same time-frequency resource block.Then the system sparse channel estimation problem can be modeled as a structured compressed perceptual reconstruction problem.Therefore, the relationship between sparse channel estimation and structured compressed sensing is established. Considering that pilot design involves two key factors, pilot position and symbol, the quality of sparse channel estimation is improved in order to optimize pilot position and pilot symbol.In this paper, a pilot optimization criterion for minimizing the correlation between complete blocks and a pilot search algorithm based on the pilot position selection are proposed.The complete block correlation is a measure of the correlation between subblocks of the recovery matrix under the framework of structured compression perception.The simulation results show that compared with other unoptimized pilots, the channel estimation error (MSE) can be significantly reduced by using this optimization method, and the channel estimation performance is improved by about 2-4dB.m3), which combines the pilot position with the pilot symbol.In this case, the pilot optimization criterion based on minimizing the correlation between complete blocks and the pilot search algorithm based on this criterion are proposed.The simulation results also show that compared with other pilot frequencies, the MSE of channel estimation can be significantly reduced by using this optimization algorithm, about 2-5 dB.The simulation results show that the optimal pilot performance obtained by the combined pilot position and symbol optimization method is better than that obtained by the simple optimization pilot position method, which can make the channel estimation of large-scale MIMO-OFDM system have lower MSE.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.53;TN919.3

【参考文献】