Summary
International Technical Conference on Circuits/Systems, Computers and Communications
2008
Session Number:P2
Session:
Number:P2-13
Independent Component Analysis for Pattern Recognition of Gas Sensor Array Measurement Data
Kai Song, Qi Wang, Taehwan Kim, Keunsung Bae,
pp.-
Publication Date:2008/7/7
Online ISSN:2188-5079
Summary:
Commercial metal oxide semiconductor gas sensor array system is generally subject to significant cross-sensitivity due to poor selectivity for different gases. This crosstalk greatly affects the accuracy of gas recognition and concentration measurement. In this paper, we explore the independent component analysis as a novel modeling technique to process the gas sensor array response data for quantitative and qualitative analysis. The process consists of two stages: first the type of gas is classified from the independent components analysis, then gas concentration is computed using the scatter plot of adjacent independent components. Simulation results demonstrate that the ICA-based method show good performance on the qualitative and quantitative gas component analysis.