Research Design And Application Of Intelligent Sensory Detection System Based On Electronic Nose And Tongue

Electronic nose and electronic tongue are a kind of intelligent sensory detection system which imitated human smell and taste sensing mechanism.Therefore,this paper designs a new intelligent sensory detection system based on electronic nose and electronic tongue,and applies it to the purity identification of juice and liquor.Specific research contents are as follows:An intelligent sensory detection system is designed,which is mainly composed of electronic nose and electronic tongue detection system.The electronic nose system is mainly composed of detection chamber and gas circuit,sensor array,signal acquisition circuit,data acquisition device and software system based on LabVIEW platform.The electronic tongue system consists of multi-sensor array,signal conditioning circuit,data acquisition device and software system based on LabVIEW platform.Six kinds of adulterated juice with different purity were detected by electronic nose and electronic tongue detection system.The data of electronic nose,electronic tongue and their feature level fusion are processed by kernel principal component analysis and fuzzy support vector machine.The results showed that in the qualitative differentiation of adulterated fruit juice by nuclear principal component analysis,electronic nose or electronic tongue alone could not effectively distinguish fruit juice of different purity,but the mixed fruit juice of different purity could be well distinguished.In the quantitative prediction of adulterated juice by fuzzy support vector machine,the root mean square error of electronic nose and electronic tongue is the minimum.Six kinds of liquor with different purity were detected by electronic nose and electronic tongue detection system.The data of electronic nose,electronic tongue and their feature level fusion are processed by extreme learning machine and least squares support vector machine respectively,and the qualitative distinction and quantitative prediction of adulterated liquor with different purity are realized.The results showed that the classification accuracy of electronic nose or tongue was 93.3%,and 100% when the two were combined.The least squares support vector machine was used to predict the quantity of adulterated liquor,and the root mean square error was the smallest after the information fusion.In short,this paper will display a set of detection system based on electronic noseand electronic tongue systems.Through the information fusion,the designed system has obtained good effects,which lays a foundation for the design and application of the follow-up intelligent sensory system.