Recursive least-squares algorithms with improved numerical stability and constrained least-squares algorithms for multichannel active noise control systems

Auteurs-es

  • M. Bouchard Sch. of Info. Technol./Engineering, University of Ottawa, Ottawa, Ont. K1N 6N5, Canada

Mots-clés :

Adaptive filtering, Algorithms, Computer simulation, Control systems, Convergence of numerical methods, Least squares approximations, Recursive functions, Multichannel active noise control (ANC) systems

Résumé

Recursive least squares (RLS) algorithms with improved numerical properties and constrained least-squares algorithms for multichannel active noise control (ANC) systems were discussed. The choice of the suitable algorithm for ANC, depends on the specific application. The inverse RLS algorithm has the best numerical properties, but it also has the highest computational load, while the least square lattice (LSL) algorithm has the lowest computational load, but requires the use of a forgetting factor with high value.

Fichiers supplémentaires

Publié-e

2000-09-01

Comment citer

1.
Bouchard M. Recursive least-squares algorithms with improved numerical stability and constrained least-squares algorithms for multichannel active noise control systems. Canadian Acoustics [Internet]. 1 sept. 2000 [cité 18 févr. 2025];28(3):66-7. Disponible à: https://jcaa.caa-aca.ca/index.php/jcaa/article/view/1293

Numéro

Rubrique

Actes du congrès de la Semaine canadienne d'acoustique