节点文献

快速单边补偿逼近幂迭代子空间跟踪

Fast Single Compensation Approximated Power Iteration Subspace Tracking Algorithm

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 魏志强鲁慧丽

【Author】 WEI Zhi-qiang1,LU Hui-li2(1.College of Mathematics and Information Sciences,North China University ofWater Resources and Electric Power,Zhengzhou 450011,China;2.School of Computer Science and Technology,Nanjing University of Postsand Telecommunications,Nanjing 210046,China)

【机构】 华北水利水电学院数学与信息科学学院南京邮电大学计算机学院

【摘要】 针对子空间估计与跟踪问题,在快速逼近幂迭代子空间跟踪算法的基础上,得到了一种计算复杂度更低的快速子空间跟踪算法——快速单边补偿逼近幂迭代子空间跟踪算法.该算法采用与Badeau等人不同的逼近方式,简化了互相关矩阵的迭代计算,从而达到降低计算复杂度的目的.实验结果表明:该算法的最大主角性能远远优于PAST子空间跟踪算法,与FAPI和OPAST子空间跟踪算法相当;在正交误差度方面,其正交误差度优于FAPI子空间跟踪算法,远远优于PAST子空间跟踪算法.

【Abstract】 For the subspace estimating and tracking,based on the well-known fast approximated power iteration(FAPI) subspace tracking algorithm,a new fast subspace tracking algorithm with lower computational complexity was proposed,that is fast single compensation approximated power iteration(FSCAPI) subspace tracking algorithm.The algorithm used different approximations with Badeau et al,simplifying the iterative calculation of the correlation matrix,so as to achieve the purpose to reduce the computational complexity.The experimental results show that the maximum principal angle performance of the algorithm is far superior to the PAST subspace tracking algorithm and performances fairly the same with FAPI and OPAST subspace tracking algorithms;the deviation from orthonormality of the algorithm is better than the FAPI subspace tracking algorithm and far better than the PAST subspace tracking algorithm.

【基金】 国家自然科学基金项目(11101145);河南省基础与前沿技术研究计划项目(102300410129)
  • 【文献出处】 中北大学学报(自然科学版) ,Journal of North University of China(Natural Science Edition) , 编辑部邮箱 ,2012年04期
  • 【分类号】O177.3
  • 【被引频次】1
  • 【下载频次】99
节点文献中: 

本文链接的文献网络图示:

本文的引文网络