Jurnal Online - http://www.jurnal.lipi.go.id


Akrajas, Muhamad Mat Salleh and Muhammad Yahaya

School of Applied PhysicsFaculty of Science and TechnologyUniversiti Kebangsaan Malaysia, 43600 UKM BangiSelangor, Malaysia

Electronic nose is a device which imitates a part or a whole the biological olfactory system to detect gases or volatile compounds. The system comprises an array of sensing elements and a pattern recognition system. This paper reports the fabrication of an optical electronic nose using an array of metalloporphyrins films as sensing elements and artificial neural network (ANN) as a pattern recognition system. The detection of gases was based on the change on the optical absorption of the thin films which measured using four LED's of different colors; red, yellow, green and blue as light sources, upon exposure toward gases. The ANN has been trained to recognize several volatile chemicals such as benzene, acetone, cyclohexane, 2-propanol, methanol and ethanol. After being trained, the electronic nose was tested to identify the blind samples within the training set. Experimental results indicated that the nose was able to recognize the vapor sample with the correct score was as high as 69.83%. The score was gained to a value of 80.88% when the input vectors were preprocessed using principal component analysis. The prototype of electronic nose and some experimental results will be discussed.

Publication : Physics Journal IPS Proceeding Supplement A6 (2002) 0508
Full paper : format PDF (598.888 byte)
Contact : author(s)
Additional info : 5 pages, language English