Design And Application Of Electromic Nose System

Electronic nose (EN) is a new research field which has developed rapidly in recent years, and it is a cross subject which relates to chemistry, material, electronic technology, multi-information data fusion pattern recognition, and computer. It comprises an array of odor sensors with partial specificity and an appropriate pattern recognition system, capable of recognizing simple or complex odors.This thesis showed how the electronic nose system platform was built and introduced two experiments in this platform. In addition, a few modules of the electronic nose system software portion with LabVIEW also were described in detail.On pattern resigning, a pattern recognition algorithm warehouse was built, which made EN choose the best algorithm according the characteristic of the sample. In respect, there are the disadvantages that the class number must be determined in advance, the value of learning rate is hard to fix, et al, when using traditional competitive neural network(CNN) in electronic nose(E—nose), so an optimizing method of CNN was presented, optimized network structure had a strong ability to adjust the number of clusters dynamically and good results of classification.The thesis described two experiments. They recognized the kinds of vinegars by EN and detected the quality of oil, respectively.Because there are many flaws in sensory discrimination method for species identification of vinegar, this thesis tried to apply the electronic nose system to classify for five different brands and types vinegar. Five kinds of vinegars were measured by EN, the sensor array were optimized by the minimal WilksΛstatistic value, put the data from optimized sensors into the optimized CNN.The thesis tried to identify the quality of oil by electronic nose, studied on the electro-conductivity of the different kind oils, the acid value and peroxide value of them. The experiment shows that the results of the three methods are consistent. It illuminated that electronic nose can detect the quality of oil.