DocumentCode :
2507318
Title :
On the recurrence plot analysis method behaviour under scaling transform
Author :
Birleanu, F.-M. ; Ioana, C. ; Gervaise, C. ; Chanussot, J. ; Serbanescu, A. ; Serban, G.
Author_Institution :
GIPSA-Lab., Grenoble Inst. of Technol., Grenoble, France
fYear :
2011
fDate :
28-30 June 2011
Firstpage :
789
Lastpage :
792
Abstract :
In the last decade, the applications of the recurrence plot analysis method make it a valuable alternative to the time-frequency and time-scale tools. As it was initially developed for the study of dynamical systems, and was later used in nonlinear time series analysis, the question of using it as a signal processing tool has not been put into discussion yet. In this field the projective techniques are largely used, with good results. Nevertheless, they also have some limitations - especially regarding transient signal processing. But this kind of signals are ubiquitous in real world. In addition, propagation through various media as well as on multiple paths lead to delayed, attenuated and dilated versions of the original transients. In this paper we study the behaviour of the recurrence plot analysis method in the context of analyzing some finite duration signals being subject to rescalings of the amplitude and time axes. This study is a starting point in employing the analysis of recurrences in investigations of a large class of real world signals.
Keywords :
signal processing; time series; transforms; dynamical systems; finite duration signal analysis; nonlinear time series analysis; projective techniques; recurrence plot analysis method behaviour; scaling transform; time-frequency analysis; time-scale tools; transient signal processing; Delay effects; Extraterrestrial measurements; Shape; Signal processing; Time series analysis; Trajectory; Transient analysis; Transient signals; amplitude and time scaling; recurrence plots;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Statistical Signal Processing Workshop (SSP), 2011 IEEE
Conference_Location :
Nice
ISSN :
pending
Print_ISBN :
978-1-4577-0569-4
Type :
conf
DOI :
10.1109/SSP.2011.5967823
Filename :
5967823
Link To Document :
بازگشت