Pattern Classification Method for Electronic Noses Based on Olfactory Neural Network Using Time Series

Guang Li
2007-01-01
Abstract:Extracting appropriate features from the responses of chemical sensor array is the first step when applying conventional artificial neural networks to electronic noses for pattern recognition.This paper presents a novel method directly dealing with time series of the sensors' responses based on an olfactory neural network with many dynamic properties.The method not only makes the electronic nose work more similar to the way in which biological noses do,but also performs better than the conventional method with feature extraction when classifying 6 volatile organic compounds.
What problem does this paper attempt to address?