Title :
Subspace based system identification for an acoustic enclosure
Author :
McKelvey, Tomas ; Fleming, Andrew ; Moheimani, S. O Reza
Author_Institution :
Dept. of Signals & Syst., Chalmers Univ. of Technol., Goteborg, Sweden
Abstract :
This paper is aimed at identifying a dynamical model of an acoustic enclosure, a duct with rectangular cross section, closed ends, and side mounted speaker enclosures. Acoustic enclosures are known to be resonant systems of high order. In order to design a high performance feedback controller for a duct or an acoustic enclosure; one needs to have an accurate model of the system. Subspace based system identification techniques have proven to be an efficient means for identifying dynamics of high order highly resonant systems such as flexible structures. In this paper a frequency domain subspace based method together with a second iterative optimization step minimizing a frequency domain least-squares criterion is successfully employed
Keywords :
acoustic devices; controllability; frequency-domain analysis; identification; iterative methods; least squares approximations; observability; optimisation; acoustic enclosure; controllability; dynamical model; frequency domain analysis; high order resonant systems; identification; iterative optimization; least-squares; observability; Ducts; Feedback control; Frequency domain analysis; Loudspeakers; Maximum likelihood estimation; Noise cancellation; Optimization methods; Resonance; Signal processing; System identification;
Conference_Titel :
Control Applications, 2000. Proceedings of the 2000 IEEE International Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
0-7803-6562-3
DOI :
10.1109/CCA.2000.897433