基于平行因子分解的IRS辅助毫米波信道估计  

IRS-assisted millimeter-wave channel estimation based on parallel factorization

作　　者：杨青青[1,2] 李学文 彭艺 王健明[1,2] YANG Qingqing;LI Xuewen;PENG Yi;WANG Jianming(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China;Yunnan Key Laboratory of Computer Technologies Application,Kunming 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500 [2]云南省计算机技术应用重点实验室,云南昆明650500

出　　处：《中山大学学报（自然科学版）（中英文）》2024年第4期124-131,共8页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：云南省计算机技术应用重点实验室(2021102)。

摘　　要：提出了一种基于平行因子分解的信道估计算法。首先,根据毫米波信道固有的稀疏特性对信道进行建模,利用块衰落信道的特点将信号矩阵构建成一个3维张量,并且利用平行因子分解算法对张量进行分解。然后利用压缩感知理论将分解后的矩阵转化为稀疏信号的恢复问题。最后,利用改进的双线性交替最小二乘算法对信道进行求解。仿真结果表明,与现有的BALS算法、wBALS算法和LSKRF算法相比,本文算法估计精度较高。A channel estimation algorithm based on parallel factor(PARAFAC)decomposition is pro posed.Firstly,the channel is modeled according to the inherent sparse characteristics of the mil limeter-wave channel.Then,the signal matrix is constructed into a three-dimensional tensor by us ing the characteristics of the block fading channel,and the tensor is decomposed by the parallel factorization algorithm.Then,the compressed sensing(CS)theory is used to transform the decom posed matrix into a sparse signal recovery problem.Finally,the bilinear alternating least squares(NBALS)algorithm is improved to solve the channel.The simulation shows that compared with the existing BALS algorithm,wBALS algorithm,and LSKRF algorithm,the proposed algorithm improves the estimation accuracy.

关 键 词：可重构智能表面 毫米波通信 信道估计 张量 平行因子分解 

分 类 号：TN929.5[电子电信—通信与信息系统]