基于贝叶斯压缩感知的毫米波MIMO信道估计  

MILLIMETER WAVE MIMO CHANNEL ESTIMATION BASED ON BAYES COMPRESSIVE SENSING

作　　者：吴贇 王萍 

机构地区：[1]数字化纺织服装技术教育部工程研究中心,上海201620 [2]东华大学信息科学与技术学院,上海201620

出　　处：《计算机应用与软件》2018年第2期126-129,182,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61571135)

摘　　要：毫米波多输入多输出(MIMO)通信系统中,获取完整的信道状态信息可使系统达到最大通信容量。利用毫米波信道的角度域稀疏特性,提出一种基于MMV模型的稀疏贝叶斯学习的(CTSBL)信道估计方法。运用信道虚实分量具有相同的稀疏结构,引入块稀疏压缩感知框架,再结合多测量信道间的时域相关性,以及超参数迭代更新算法,使毫米波信道估计性能得到了提升。算法理论分析和实验仿真结果表明,提出的CTSBL毫米波MIMO信道估计方法比传统的贪婪算法块正交匹配追踪(BOMP)信道估计方法具有更高的估计精度。In the millimeter wave multiple input multiple output( MIMO) communication system,obtaining the complete channel state information can make the system reach the maximum communication capacity. In this paper,a Sparse Bayesian Learning( CTSBL) channel estimation method based on MMV model was proposed by using the sparse features of the angular domain of millimeter wave channels. The sparse and compressive sensing block was introduced by using the same sparse structure of the channel 's real and imaginary components. Then the millimeter-wave channel estimation performance was improved by combining the time-domain correlation between multi-measurement channels and the iterative update algorithm of hyper parameter. The theoretical analysis and experimental simulation results showed that the proposed MIMO channel estimation method based on CTSBL millimeter wave had higher estimation accuracy than traditional greedy algorithm block orthogonal matching pursuit( BOMP) channel estimation.

关 键 词：毫米波MIMO 稀疏贝叶斯 学习 多测量 矢量模型 

分 类 号：TP919[自动化与计算机技术]