0
$\begingroup$

Continuous wavelet transformation has been quite widely used for various applications. Most of the papers that I found were using CWT for non-stationary signals. Can we use CWT for stationary signal analysis? if not what are the drawbacks in using Continuous wavelet transform?

Improve this question
$\endgroup$
3
  • $\begingroup$ The signal I'm analyzing is a very short peaky signal. I did test the signal to measure it's stationary using two ways. First one is the sliding window method where I found that statistics of the signal changes in each window. This indicates that the signal is non-stationary. but when I tried the dickey fuller test it showed me that the signal is stationary. I'm confused about which one I should really on? $\endgroup$
    – hasi
    Jun 17, 2019 at 22:47
  • $\begingroup$ is there any way that I can directly contact you? $\endgroup$
    – hasi
    Jun 17, 2019 at 23:11
  • $\begingroup$ Yes you can, yet this site gets useful when exchanges are shared openly $\endgroup$
    – Laurent Duval
    Feb 14, 2022 at 21:24

1 Answer 1

Reset to default
-1
$\begingroup$

Stationarity is a multi-fold concept in signal processing. It can denote a wide range of behavior, encompassing deterministic or stochastic aspects. Beyond that, the main question is: do you know if your signal is stationary, and how?

If you actually know how, it is probably wiser to use the generation process to build a custom, adapted model or transformation, and use it for the analysis.

Even in that case, I strongly advocate using different analysis methods in parallel, to help you detect artifacts, issues than you would not detect with a single model. For instance, let us remind that one usually observe only a few realizations of a "signal", and that acquisition issues, outliers, etc. may occur.

Finally, analyzing in first intention a signal with time-frequency or time-scale transforms is a good idea, as it can help you detect the useful scales of interest, estimate parameters of stochastic events, etc.

The drawbacks are:

  • The difficulties in choosing the appropriate wavelet (real or complex), and the associated sampling (and the resulting speed)
  • The difficulties in interpreting the scalogram, as a knowledge of the underlying processes could be useful
Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.