Classification of Electric Signals Based on Time-Frequency Signal Decomposition
Leonowicz, Z (Wroclaw University of Technology, Poland)
Abstract: A classification procedure based on time-frequency decomposition of the signal is presented. Parametric spectral ESPRIT method is used for estimation of relevant parameters of signal components and specific areas of the time-frequency plane are chosen, where the signal is expected to show most characteristic patterns. Classification is based on time-domain correlation of reconstructed signals. It is applied to event classification of non-stationary electric signals obtained from a simulated power converter.
Keyword(s): Signal analysis ; time-frequency analysis ; pattern classification ; correlation
Note: not published
Fulltext : http://zet10.ipee.pwr.wroc.pl/record/86/files/
Cited by: try citation search for ENY-PREPRINT-2008-003
Record created 2008-03-16, last modified 2008-03-16
|People who viewed this page also viewed:|
Be the first to review this document.
Start a discussion about any aspect of this document.