Fast structured total least squares algorithms via exploitation of the displacement structure

被引:0
|
作者
Mastronardi, N [1 ]
Lemmerling, P [1 ]
Van Huffel, S [1 ]
机构
[1] Katholieke Univ Leuven, Dept Comp Sci, B-3001 Louvain, Belgium
来源
TOTAL LEAST SQUARES AND ERRORS-IN-VARIABLES MODELING: ANALYSIS, ALGORITHMS AND APPLICATIONS | 2002年
关键词
Hankel/Toeplitz matrix; structured total least squares; displacement rank;
D O I
暂无
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
We present fast algorithms and their stability properties for Toeplitz structured total least squares (STLS) problems. The STLS problem can be formulated as the following constrained optimization problem min(DeltaA,Deltab,x) parallel to[DeltaA Deltab]parallel to(F) such that (A + DeltaA)x = b + Deltab and [DeltaA Deltab] has the same structure as [A b]. This natural extension of the TLS problem is clearly more difficult to solve than the TLS problem, because of its highly nonlinear nature and the existence of many local minima. We focus here on the frequently occurring case where [A b] is a Toeplitz matrix. The problem is solved in an iterative fashion, in which at each iteration the Kaxush-Kuhn-Tucker (KKT) equations of the locally linearized problem have to be solved. For this kernel routine we use a generalized Schur decomposition algorithm based on the low displacement rank of the KKT system matrix. By exploiting the sparsity of the associated generators, we obtain a fast algorithm that requires O(mn + n(2)) flops per iteration, where m and n axe the number of rows and the number of columns of A, respectively. We also prove the stability of the latter kernel routine. The efficiency of the proposed fast implementation is compared to the efficiency of the straightforward implementation, which does not exploit the structure of the involved matrices. The comparison is done on a recently introduced speech compression scheme in which the considered STLS problem constitutes one of the kernel problems. The numerical results confirm the high efficiency of the newly proposed fast implementation.
引用
收藏
页码:93 / 106
页数:14
相关论文
共 34 条
  • [21] Analysis of the structured total least squares problem for Hankel/Toeplitz matrices
    Lemmerling, P
    Van Huffel, S
    NUMERICAL ALGORITHMS, 2001, 27 (01) : 89 - 114
  • [22] A Structured Total Least Squares Algorithm for Spatial Straight Line Fitting
    Deng Xiao-yuan
    Lu Tie-ding
    Chang Xiao-tao
    Zhu Xiao-yong
    INTERNATIONAL CONFERENCE ON INTELLIGENT EARTH OBSERVING AND APPLICATIONS 2015, 2015, 9808
  • [23] Analysis of the Structured Total Least Squares Problem for Hankel/Toeplitz Matrices
    Philippe Lemmerling
    Sabine Van Huffel
    Numerical Algorithms, 2001, 27 : 89 - 114
  • [24] Application of structured total least squares for system identification and model reduction
    Markovsky, I
    Willems, JC
    Van Huffel, S
    De Moor, B
    Pintelon, R
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2005, 50 (10) : 1490 - 1500
  • [25] Modeling audio with damped sinusoids using total least squares algorithms
    Verhelst, W
    Hermus, K
    Lemmerling, P
    Wambacq, P
    Van Huffel, S
    TOTAL LEAST SQUARES AND ERRORS-IN-VARIABLES MODELING: ANALYSIS, ALGORITHMS AND APPLICATIONS, 2002, : 331 - 340
  • [26] A global solution for the structured total least squares problem with block circulant matrices
    Beck, A
    Ben-Tal, A
    SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2005, 27 (01) : 238 - 255
  • [27] Multiple delays estimation for chirp signals using structured total least squares
    Yeredor, A
    LINEAR ALGEBRA AND ITS APPLICATIONS, 2004, 391 : 261 - 286
  • [28] Efficient implementation of a structured total least squares based speech compression method
    Lemmerling, P
    Mastronardi, N
    Van Huffel, S
    LINEAR ALGEBRA AND ITS APPLICATIONS, 2003, 366 : 295 - 315
  • [29] On structured total least-squares for blind identification of multichannel FIR filters
    Ikram, Muhammad Z.
    2008 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, VOLS 1-12, 2008, : 2821 - 2824
  • [30] The structured total least squares algorithm research for passive location based on angle information
    Ding Wang
    Li Zhang
    Ying Wu
    Science in China Series F: Information Sciences, 2009, 52 : 1043 - 1054