Generalized autoregressive spectral estimation

Author

Tsao, Jenho ; Shyu, Wei-Ji

Author_Institution

Dept. of Electr. Eng., Nat. Taiwan Univ., Taipei, Taiwan

Volume

5

fYear

1992

fDate

23-26 Mar 1992

Firstpage

449

Abstract

An autoregressive spectral estimation method is developed to reduce the noise effect in prediction coefficient estimation. This method solves the prediction coefficients from a generalized Yule-Walker equation which is formed by the data and its generalized autocorrelation sequence. This method provides several control parameters for the spectral estimator to combat the unmodeled additive noise in the linear least square sense. Through the efficient use of information by this method, data size will be directly helpful in noise suppression

Keywords

filtering and prediction theory; interference suppression; parameter estimation; spectral analysis; additive noise; autoregressive spectral estimation; generalized Yule-Walker equation; generalized autocorrelation sequence; noise suppression; prediction coefficients; Additive noise; Autocorrelation; Councils; Difference equations; Least squares approximation; Linear predictive coding; Noise reduction; Random processes; Spectral analysis; White noise;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1992. ICASSP-92., 1992 IEEE International Conference on

Conference_Location

San Francisco, CA

ISSN

1520-6149

Print_ISBN

0-7803-0532-9

Type

conf

DOI

10.1109/ICASSP.1992.226586

Filename

226586