Title :
Linear-predictive data extrapolation narrow-band spectral estimation
Author :
Swingler, D.N. ; Walker, R.S.
Author_Institution :
Div. of Eng., Saint Mary´´s Univ., Halifax, NS, Canada
fDate :
9/1/1988 12:00:00 AM
Abstract :
It is demonstrated that T.J. Ulrych et al.´s (J. Geophys., Res., vol.78, p.4959-64, 1973) hybrid Fourier-Burg method of narrowband spectral estimation has significant advantages which may have been overlooked. Apart from improved resolution in comparison to the conventional Fourier approach, it also exhibits reduced sensitivity to model order compared to the Burg maximum-entropy-method algorithm. Additionally, the hybrid method has a frequency-estimation variance which is less than that of the Burg technique, approaching that of the conventional (unwindowed) Fourier method for low signal/noise ratio. Finally, it can provide for enhanced tonal detection in some instances
Keywords :
filtering and prediction theory; signal processing; spectral analysis; Burg maximum-entropy-method algorithm; hybrid Fourier-Burg method; linear-predictive data extrapolation; narrowband spectral estimation; signal/noise ratio; tonal detection enhancement; Explosions; Extrapolation; Filters; Frequency estimation; Hybrid power systems; Mouth; Narrowband; Signal detection; Signal resolution; Sonar measurements;
Journal_Title :
Proceedings of the IEEE
