计算机工程

• 移动互联与通信技术 • 上一篇    下一篇

结合Tucker张量分解与交替最小二乘的ULA盲识别

胡丹 1,郭英杰  2   

  1. (1.贵州大学 大数据与信息工程学院,贵阳 550025; 2.悉尼科技大学 全球大数据技术中心,悉尼 2007)
  • 收稿日期:2016-09-23 出版日期:2017-10-15 发布日期:2017-10-15
  • 作者简介:胡丹(1979—),女,讲师、硕士,主研方向为信道编译码;郭英杰,教授、博士。
  • 基金项目:
    贵州省科技计划项目(黔科合LH字[2014]7627)。

ULA Blind Identification Combining Tucker Tensor Decomposition and Alternating Least Squares

HU Dan 1,GUO Yingjie 2   

  1. (1.College of Big Data and Information Engineering,Guizhou University,Guiyang 550025,China; 2.Globle Big Data Technologies Centre,University of Technology Sydney,Sydney 2007,Australia)
  • Received:2016-09-23 Online:2017-10-15 Published:2017-10-15

摘要: 为提高均匀线性阵列(ULA)系统盲识别过程的计算效率,提出一种改进的ULA盲识别算法。建立ULA信号传播模型,针对该传播模型给出广义生成函数的代数结构以及参数估计方式,利用交替最小二乘法对ULA广义生成函数进行求解,并在此基础上引入Tucker张量分解改进交替最小二乘法,实现广义生成函数的降维处理。实验结果表明,与经典DUET算法、欠定混叠盲辨识分解算法等相比,该算法具有更高的计算效率以及更好的ULA盲识别效果。

关键词: Tucker张量分解, 广义生成函数, 交替最小二乘, 均匀线性阵列, 盲识别

Abstract: In order to improve the computational efficiency of Blind Identification(BI) of Uniform Linear Array(ULA) system,an improved ULA BI algorithm is proposed.Firstly,the signal propagation model of ULA system is established.Then,the algebraic structure and parameter estimation method of the Generalized Generating Function(GGF) are given.Secondly,it uses alternating least squares to obtain GGF of ULA system,and then uses the Tucker tensor decomposition to improve alternating least squares,achieving dimensional reduction of GGF.Experimental results show that,compared with classic DUET algorithm and underdetermined aliasing BI decomposition algorithm,the proposed algorithm has higher computational efficiency and better ULA BI effects.

Key words: Tucker tensor decomposition, Generalized Generating Function(GGF), alternating least squares, Uniform Linear Array(ULA), Blind Identification(BI)

中图分类号: