A Dynamic Response, Transformed Cluster Analysis and Radial Basis Function Neural Network Based Gases/Odors Identification Approach Using a Thick Film Gas Sensor Array

作者：Sunny^*; Kumar Ravi; Mishra V N; Das R R

来源：Journal of Computational and Theoretical Nanoscience, 2014, 11(4): 1199-1204.

DOI：10.1166/jctn.2014.3483

• Cross Validation
• Dynamic Response
• Neural Classifier
• Sensor Array
• Transformed Cluster Analysis
• Thick Film

In the present study, an experiment to identify various gases/odors has been presented. This approach is based on the dynamic responses of a thick film gas sensor array. The dynamic response of the gas sensor array for four gases viz. LPG, CCl4, CO and C3H7OH was first transformed into well separated data using modified transformed cluster analysis (MTCA) technique which is a minor variant of published basic transformed cluster analysis (TCA) data preprocessing method. Subsequently, a radial basis function neural network (RBFNN) was trained using the MTCA transformed data for identification of samples of respective gases/odors. The performance of the best trained radial basis function neural network using MTCA transformed data for 24 unknown test samples had been 100% accurate while the accuracy was 82% only when the same RBFNN was trained with respective raw form of dynamic gas sensor array response. We therefore report that superior identification could be obtained with MTCA cum RBFNN methods applied to the dynamic responses of gas sensor array.

• 出版日期2014-4