Volume 4, Issue 12 (December 2017), Pages: 79-82

----------------------------------------------

 Original Research Paper

 Title: Selection of the order of autoregressive models for spectral analysis of noise corrupted signals

 Author(s):  Jakub Jeřábek *

 Affiliation(s):

 Department of Electrical Engineering, Faculty of Electrical Engineering and Informatics, Univerzita Pardubice, Pardubice, Czech Republic

 https://doi.org/10.21833/ijaas.2017.012.016

 Full Text - PDF          XML

 Abstract:

This paper presents the theoretical basis of autoregressive (AR) modelling in spectral analysis. Autoregressive modelling includes a model identification procedure based on an autocorrelation function (ACF) of the incoming signal and its statistical evaluation. This is necessary to choose the best order of an AR model that best describes the given set of data. Spectral analysis gives information about the frequency content of a signal. The AR method is an alternative to discrete Fourier transform (DFT) in the computing of a power spectrum density function of a signal, but provides a smoother power spectral density then DFT. 

 © 2017 The Authors. Published by IASE.

 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

 Keywords: ARMA, AR model, Order estimation, System identification, ACF

 Article History: Received 23 February 2017, Received in revised form 23 September 2017, Accepted 24 September 2017

 Digital Object Identifier: 

 https://doi.org/10.21833/ijaas.2017.012.016

 Citation:

 Jeřábek  J (2017). Selection of the order of autoregressive models for spectral analysis of noise corrupted signals. International Journal of Advanced and Applied Sciences, 4(12): 79-82

 Permanent Link:

 http://www.science-gate.com/IJAAS/V4I12/Jakub.html

----------------------------------------------

 References (4)

1. Porat B (1997). Course in digital signal processing. John Wiley and Sons, New York, USA. 
2. Rejfek L, Buresova D, Fiser O, and Brazda V (2015). Comparison of parametric methods for radar signal processing. In the 25th International Conference on Radioelektronika (RADIOELEKTRONIKA), IEEE, Pardubice, Czech Republic: 141-144. https://doi.org/10.1109/RADIOELEK.2015.7128986 
3. Schlindwein FS and Evans DH (1992). Autoregressive spectral analysis as an alternative to fast Fourier transform analysis of Doppler ultrasound signals. Diagnostic Vascular Ultrasound, 8: 74-84. 
4. Schwarz D (2009). Lineární systémy a modely časových řad. Institute of Biostatistics and Analyses, Brno, Czech Republic. Available online at: www.iba.muni.cz