Abstract
This paper investigates frequency mixing effect of empirical mode decomposition (EMD) and explores whether it can be explained by simple phase coupling between components of the input signal. The input is assumed to be a linear combination of harmonic oscillators. The hypothesis was tested assuming that phases of input signals’ components would couple according to Kuramoto’s model. Using a Kuramoto’s model with as many oscillators as the number of intrinsic mode functions (result of EMD), the model’s parameters were adjusted by a particle swarm optimization (PSO) method. The results show that our hypothesis is plausible, however, a different coupling mechanism than the simple sine-coupling Kuramoto’s model are likely to give better results.
Original language | English |
---|---|
Number of pages | 12 |
Journal | Advances in Data Science and Adaptive Analysis |
DOIs | |
Publication status | Published - 3 Aug 2016 |
Keywords
- Particle swarm optimization
- Kuramoto model
- instanteneous frequency
- Empirical mode decomposition