抄録
This paper describes the alcoholic gas discrimination experiment using an artificial neural network pattern recognition technique and dye absorbance spectra patterns from a 32-channel fiber-optic sensor array with different dye coatings, whose absorption spectra change with alcoholic gas exposure. The 32-channel absorption spectra were simultaneously measured by using 13 band-pass filters and 32 photodiodes, and data-processed by the pattern recognition method. The gas recognition system has been successfully applied to discrimination of alcoholic gases having similar chemical sutucture and found to discriminate methanol, ethanol, 1-propanol, and 2-propanol.
