第19卷第6期2006年12月传感技术学报CHⅡ铘ESEJC^HtNAI.0FSENSORSANDACllIAl’ORSV01.19No.6Dec.2006odourReco印itionAlgorithmsforMachineolfacti伽System+LJUHo咒g—zi“‘,LU-0D矿砌以,ZHANGZ旷yo,zg(Sf^Dozo,h,0r7m£io前函gi揪一前g,G懈订gDong砺i口.o,n娩,G蚍站g孙D“510643,吼i舰)AbStr舵t:Thedevelopingprocessofmachineolfactionisreviewed.Thesignalmodelofsensorarraysasso—ciatedwithspecificodorclassesispresentedbasedonhumanchemosensoryperceptionprincipleofolfactioncells.Themultivariatepatternanalysistechniquesappliedtomachineolfactionsystemareintroduced.Thestatisticalpatternanalysis(1inearandnonlinear)andintelligentpatternanalysis(artificialneuralnetworksandothers)aredescribed.Itisconstructiveformachineolfactionsystemtoputintopractice.1墨|t孑words:machine01faction;sensorarrays;patternanalysis;recognitionalgorithmsEEACC:7230J机器嗅觉系统气味识别算法*刘红秀”,骆德汉,张泽勇(广