The pattern recognition in analysis of vibroacoustic signal

Stanisław Radkowski , Jacck Dybała , Szymon Gontarz

Abstract

The paper addresses the problem of optimal choice of the most informative diagnostic features of a signal, particularly the vibroacoustic signal. Taking into account that the data are in the form of available attribute vectors, many methods for discriminating between the two classes have been developed. In this paper we review the results of our own research and those found in the literature that are relevant for the solution of diagnostic inference. Special attention is paid on the method of geometrical features selection and blind source separation. The first method presented here uses two criteria related to the ability to separate (isolate) classes of an object's state: the criterion of average scatters and the original criterion of the number of prototypes of classes. The method can be successfully used for initial analysis of input data to a neural network dealing with recognition of patterns of an object's state and classification of an object's state. Concerning blind separation, the method which uses the algorithm of blind equalization (BE) by iterative application of different lengths of equalizers is presented here. This approach allows us to estimate sub-signals in various frequency bands. Main features of this solution include the possibility of discovery and identification diagnostically useful information, even when it is hidden by relatively larger noise and interference.© (2006) by the International Institute of Acoustics & Vibration.

Author Stanisław Radkowski (FACME / IAE)
Stanisław Radkowski,,
- Institute of Automotive Engineering
, Jacck Dybała (FM)
Jacck Dybała,,
- Faculty of Mechatronics
, Szymon Gontarz (FACME / IAE)
Szymon Gontarz,,
- Institute of Automotive Engineering
Pages570-577
Book Eberhardsteiner J. (eds.): 13th International Congress on Sound and Vibration 2006 : (ICSV 13) ; Vienna, Austria, 2 - 6 July 2006, 2013, Curran, ISBN 3-95-01554-5-7
Languageen angielski
Score (nominal)0
Publication indicators Scopus Citations = 0
Citation count*
Cite
Share Share

Get link to the record


* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.
Back
Confirmation
Are you sure?