HISTOGRAM-BASED FEATURE EXTRACTION TECHNIQUE APPLIED FOR FAULT DIAGNOSIS OF ELECTRONIC CIRCUITS

Wojciech Toczek, Michal Kowalewski, Romuald Zielonko

Abstract

In this paper we discuss and compare two feature extraction techniques: histogram-based and Principal Component Analysis. Comparison is done on an analog filter fault diagnosis example performed in the frequency domain. Both techniques are implemented in a neural network system for the off-line diagnosis of electronic analog and mixed-signal circuits. The numerical and experimental examples of frequency domain ANN-based testing of filter are presented to demonstrate the usefulness of the histogram approach.

Keywords

analog fault diagnosis, feature extraction, histogram-based technique

Download

IMEKO-TC10-2005-06.pdf

DOI

-

IMEKO TC

TC10 - Measurement for Diagnostics, Optimization and Control

Back to the proceedings