Impulse response identification by use of wavelet packets decomposition

Author

Takahashi, Masayuki ; Ohmori, Hiromitsu ; Sano, Akira

Author_Institution

Dept. of Syst. Design Eng., Keio Univ., Yokohama, Japan

Volume

1

fYear

1998

fDate

1998

Firstpage

211

Abstract

This paper is concerned with a new system identification approach based on the wavelet packets (WP) decomposition. The impulse response in the WP transform domain is estimated via the least squares method using transformed input/output data. Then the estimate of the original impulse response can be reconstructed from that in the transform domain by using the inverse WP transform. The MSE of the estimate is investigated, and it is clarified that the MSE can be minimized by an optimal choice of the threshold level, the resolution level and the packets selection. Therefore, the data adaptive choice is effective in ill-conditioned system identification using a band-limited input signal

Keywords

identification; least mean squares methods; transient response; wavelet transforms; MSE minimization; WP decomposition; WP transform domain; band-limited input signal; ill-conditioned system identification; impulse response identification; inverse WP transform; least mean squares method; least squares method; transformed I/O data; transformed input/output data; wavelet packets decomposition; Design engineering; Frequency estimation; Least squares methods; Signal processing; Signal resolution; System identification; System testing; Systems engineering and theory; Vectors; Wavelet packets;

fLanguage

English

Publisher

ieee

Conference_Titel

Decision and Control, 1998. Proceedings of the 37th IEEE Conference on

Conference_Location

Tampa, FL

ISSN

0191-2216

Print_ISBN

0-7803-4394-8

Type

conf

DOI

10.1109/CDC.1998.760670

Filename

760670