Vol. 129, Issue 6, June 2011, pp.57-68

Abstract: This paper assesses the potency of data denoising techniques to reduce the noise component in sensor array response, for successful recognition of chemical vapor by electronic nose. Three denoising methods specifically median filtering (MF), principal component analysis (PCA), and singular value decomposition (SVD) have been selected for study. For analysis seven data sets including both the experimental and simulated response of surface acoustic wave (SAW), metal oxide semiconductor (MOS), and conducting composite polymer (CCP) sensor array are used. After denoising, each data set is reanalyzed by PCA for identification of chemical vapor by visual discrimination in principal component (PC) space and performance assessment of denoising methods by computing class seprability measure from principal components. Outcomes from PCA analysis proved SVD as the best denoising method compare to other two referred method. Performance of MF approach is found to be superior to PCA.

Keywords: Electronic nose data denoising, Median filtering, Principal component analysis, Singular value decomposition, Chemical vapor detection

Click <here> or title of paper to download the full pages article (0.99 Mb)

Read more about Chemical Sensors

1999 - 2011 Copyright ©, International Frequency Sensor Association (IFSA). All Rights Reserved.

Home - News - Links - Archives - Tools - Standardization - Patents - Marketplace - Projects - Wish List - Subscribe - Membership - IFSA Publishing - Submit Press Release - Twitter - Search

 Members Area -Sensors Portal - Voltage-to-Frequency Converters - Training Courses - S&T Digest - For advertisers - Bookstore - Forums - Polls - Sensor Jobs - e-Shop - Site Map - Video