基于TPE的大规模MIMO信道估计算法研究

发布时间：2018-11-19 07:58

【摘要】：经典的最小均方误差(MMSE,Minimum Mean Square Error)估计需要对协方差矩阵进行求逆,复杂度为矩阵维度的三次方,在大规模MIMO系统中用其进行信道估计的代价太高,研究如何降低此时的估计复杂度具有重要意义。本学位论文在基于导频信道估计的基础上,采用截短多项式展开(TPE,Truncated Polynomial Expansion)技术,对如何降低大规模MIMO信道估计复杂度进行深入研究。首先介绍基于Taylor级数展开的信道估计模型,该模型的复杂度远低于基于MMSE算法的信道估计模型。为了提高该模型的收敛速度,本文提出采用Kapteyn级数对协方差矩阵进行展开,从而得到基于Kapteyn级数展开的信道估计模型。仿真结果可以看出,当截短多项式阶数为10时,与Taylor-MMSE算法相比,Kapteyn-MMSE估计算法的性能有0.3dB的增益,后者的均方误差(MSE,Mean Square Error)收敛速度快于前者,但其复杂度略高。然后重点研究如何通过优化多项式系数来获得高性能的信道估计模型。在基于Kapteyn级数展开的信道估计模型基础上,通过优化Kapteyn级数展开的多项式系数来获得高性能的信道估计,提出一种基于加权Kapteyn级数展开的信道估计模型。为了解决该模型中加权系数的问题,建立了无约束的非线性优化模型。针对该模型,论文将其分解为两个子线性优化模型并利用基于坐标轮换的迭代算法来求解。仿真结果表明,随着展开多项式阶数的增加,基于加权Kapteyn级数展开的信道估计模型MSE快速收敛于基于MMSE算法的信道模型MSE,且收敛速度明显优于未加权模型。同时,在低阶多项式情况下,加权Kapteyn级数展开的估计性能优于加权Taylor级数展开的估计性能2dB。
[Abstract]:The classical least mean square error (MMSE,Minimum Mean Square Error) estimation needs to inverse the covariance matrix. The complexity is the third power of the matrix dimension, and the cost of channel estimation in large-scale MIMO systems is too high. It is of great significance to study how to reduce the estimated complexity at this time. In this thesis, based on pilot channel estimation, truncated polynomial expansion (TPE,Truncated Polynomial Expansion) technique is used to study how to reduce the complexity of large-scale MIMO channel estimation. Firstly, the channel estimation model based on Taylor series expansion is introduced. The complexity of the model is much lower than that of the channel estimation model based on MMSE algorithm. In order to improve the convergence rate of the model, the covariance matrix is expanded by Kapteyn series, and the channel estimation model based on Kapteyn series expansion is obtained. The simulation results show that when the truncated polynomial order is 10:00, compared with the Taylor-MMSE algorithm, the performance of the Kapteyn-MMSE estimation algorithm has the gain of 0.3dB, and the mean square error (MSE,Mean Square Error) of the latter algorithm is faster than the former. But its complexity is a little high. Then we focus on how to obtain high performance channel estimation model by optimizing polynomial coefficients. Based on the channel estimation model based on Kapteyn series expansion, a channel estimation model based on weighted Kapteyn series expansion is proposed by optimizing the polynomial coefficients of Kapteyn series expansion. In order to solve the problem of weighting coefficient in the model, an unconstrained nonlinear optimization model is established. For this model, it is decomposed into two sublinear optimization models and solved by iterative algorithm based on coordinate rotation. The simulation results show that the channel estimation model MSE based on weighted Kapteyn series expansion converges rapidly to the channel model MSE, based on MMSE algorithm and the convergence speed is obviously better than that of the unweighted model with the increase of the expansion polynomial order. At the same time, the estimation performance of weighted Kapteyn series expansion is better than that of weighted Taylor series expansion in the case of low order polynomial.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3

【相似文献】