I am trying to apply Hilbert Huang transform to the following signal.
x=cos(2*pi*80*t)+0.8*sin(2*pi*50*t)+0.6*sin(2*pi*25*t)+0.4*sin(2*pi*10*t)+0.3*sin(2*pi*3*t);
As i understood the output IMFs should be five sinusoidal signals with amplitude of 1,0.8,0.6,0.4 and 0.3 with frequencies of 80,50,25,20 and 3.
However I am getting only two IMFs as outputs in my code.
Who can help me?
Mode mixing will be an issue for these frequencies and amplitudes. If two frequencies fall into one octave (05 < f1/f2 < 2), both frequencies may appear in the same IMF (to some extent depending on the ratio and the amplitudes). If the ratio is closer to 1.0, they will be completely mixed. You should be able to separate 10 and 3 well, 80 and 50 will appear partly together while the remaining part of 50 and 25 should be more or less fine. If you get only 2 IMFs, something is wrong. A paper available at research gate discussing this is:
Preprint Method for Mode Mixing Separation in Empirical Mode Decomposition
There are several codes for HHT available. The link blow is for the code of Patrick Flandrin.
http://perso.ens-lyon.fr/patrick.flandrin/emd.html
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity