Fusion Technique Of Electronic Nose And Electronic Tongue And Its Application

Electronic nose is a new high technology product, which simulates animalolfactory organ. It can show whole information of gas characterization rapidly,namely “fingerprint information”. The working process of electronic tongue is toimitate the formation process of animal. It can analyze the whole components of themeasured liquid rapidly and make comprehensive evaluation. Many substances havegas-liquid phase in the word, so using either electronic nose or electronic tongue canonly detect one characteristic of the substances, and there will be some limitations.However, there are very few people who will fuse the technologies of electronic noseand electronic tongue.The main work and results of this research are as follows:(1) We design and build electronic nose system based on dynamic headspacesampling method, which include hardware and software. The designing of hardwareincludes sensor array, gas chamber, pneumatic conveying and so on. The software ofelectronic nose system is designed based on LabVIEW8.6, which can display, storeand invoke the signal acquisition, signal processing and pattern recognition. Inaddition, we built voltammetric electronic tongue system with three electrodes, whichincludes a working electrode, a reference electrode and an auxiliary electrode.(2) The new electronic nose and electronic tongue systems were used to detecteight different brands of liquor respectively, then the characteristic values of the twosystems were fused. We chose the principal component analysis and Bayesianclassifier to analyze the data of electronic nose, electronic tongue and fusion data ofthe two systems. The misclassification probability of samples is25%by usingelectronic nose separately, and the misclassification probability of samples is10%byusing electronic tongue separately. However, the misclassification probability ofsamples is reduced to7.5%, when fusing the two systems. (3) Nine kinds of edible oil are detected by using electronic nose and electronictongue respectively. We chose the principal component analysis and partial leastsquares to analyze the data of electronic nose, electronic tongue and fusion data of thetwo systems. Finally, prediction error about nine kinds of edible oil has been got. Ifwe just use electronic nose, the prediction error is12.099%. If using electronic tongueseparately, the prediction error is4.65%. The predictive effect has been improved byfusing electronic nose and electronic tongue for the prediction error is2.93%.In conclusion, the research has designed and constructed the electronic nosesystem. In the same time, the research has built electronic tongue experiment platform.Besides that, we have fused the two techniques and got the expected effect. Thisthesis builds a good foundation for the development of electronic nose and electronictongue fusion technique.