Sorry, you need to enable JavaScript to visit this website.

Dominant Component Tracking for Empirical Mode Decomposition using a Hidden Markov Model

Primary tabs

Citation Author(s):
Steven Sandoval, Matthew Bredin, and Phillip L.~De Leon
Submitted by:
Matthew Bredin
Last updated:
26 November 2018 - 4:23pm
Document Type:
Presentation Slides
Document Year:
2018
Event:
Presenters Name:
Matthew Bredin

Abstract 

Abstract: 

It is well known that the empirical mode decomposition algorithm does not always return an appropriate decomposition due to problems like mode mixing. In this paper, we consider the problem of a component being split across several intrinsic mode functions (IMFs). We propose the use of a hidden Markov model (HMM) to track the dominant component across the set of IMFs returned by EMD. We provide an example demonstrating the proposed tracking using an acoustic recording where component splitting is present in the decomposition and compare our method to two other possible tracking approaches. We show that the proposed method provides a compromise between smoothness and energy associated with the tracked component.

up
0 users have voted:

Dataset Files

Lecture___Dominant_Component_Tracking.pdf

(195)