• DocumentCode
    394709
  • Title

    Convergence analysis of multichannel ANC system using FXLMS algorithm with inexact secondary path model

  • Author

    Yeung, Tuk Keung ; Yau, Sze Fong

  • Author_Institution
    Dept. of Electr. & Electron. Eng., Hong Kong Univ. of Sci. & Technol., China
  • Volume
    5
  • fYear
    2003
  • fDate
    6-10 April 2003
  • Abstract
    The paper conducts the convergence analysis of multichannel an active noise control (ANC) system using a filtered-X least-mean-square (FXLMS) algorithm with inexact secondary path modeling. An update equation of the excess mean-square error (MSE) is developed first. A criterion on secondary path modeling for the convergence of the excess MSE is then derived, which reveals that the convergence rate can be enhanced by choosing the secondary path models appropriately. Next, an upper bound on the step size of the update equation is derived. Furthermore, the steady-state excess MSE is expressed explicitly for a thorough insight into the steady-state response of the system. Computer simulations verify that our theoretical analysis can predict accurately the performance of a feedforward multichannel ANC system with inexact secondary path modeling.
  • Keywords
    acoustic noise; acoustic signal processing; active noise control; convergence of numerical methods; feedforward; least mean squares methods; transfer functions; convergence rate; feedforward system; filtered-X LMS algorithm; filtered-X least-mean-square algorithm; multichannel ANC system; secondary path modeling; steady-state excess MSE; steady-state excess mean-square error; transfer functions; update equation; Active noise reduction; Algorithm design and analysis; Computer simulation; Control system synthesis; Convergence; Equations; Performance analysis; Predictive models; Steady-state; Upper bound;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). 2003 IEEE International Conference on
  • ISSN
    1520-6149
  • Print_ISBN
    0-7803-7663-3
  • Type

    conf

  • DOI
    10.1109/ICASSP.2003.1200035
  • Filename
    1200035