Robust change detection in signals using energy concentration and regression models

Reliable characterization of different signals is essential for better understanding of their generating and propagation phenomena. Many works in this area have been based on detecting special patterns or clusters in data, and event detection using parametric models. In this paper we present an appr...

Full description

Saved in:
Bibliographic Details
Main Authors: Popescu, Theodor D., Aiordachioaie, Dorel
Format: Conference Proceeding
Language:English
Subjects:
Algorithm design and analysis
Entropy
Kernel
Robustness
Signal processing
Time-frequency analysis
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Reliable characterization of different signals is essential for better understanding of their generating and propagation phenomena. Many works in this area have been based on detecting special patterns or clusters in data, and event detection using parametric models. In this paper we present an approach making use of the short-term time-frequency Renyi entropy and an algorithm to discriminate between the model parameter and noise variance changes, operating on Renyi entropy, as a new space of decision. This method enables a simpler analysis and interpretation of the signals behavior. The procedure is used, with good results, in the analysis of a seismic signal during a strong to moderate ground motion.
ISSN:2164-5221
DOI:10.1109/ICSP.2016.7878119