On the way to electronic nose

Abstract

Two kinds of electronic nose are developed. First one is in the view of matrix of 24 gas sensors on the base of thin films of tin dioxide. The results of modeling are given. The second kind of devise uses the dynamical characteristics of heterogeneous reactions of two sensors. The recognition of reagents based on comparison of reaction of sensors with earlier formed pattern of reactions, obtained by using the correlation coefficient. The perspectives of using the neuron networks for making this analysis is reported.