资源描述：
《correlation analysis between international roughness index 国际平整度指数的相关性分析》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、Linetal.CorrelationAnalysisBetweenInternationalRoughnessIndex(IRI)andPavementDistressbyNeuralNetworkByJyh-DongLinProfessorjyhdongl@cc.ncu.edu.twJyh-TyngYauGraduateStudents8322019@cc.ncu.edu.twLiang-HaoHsiaoGraduateStudentLhhsiao@cyit.edu.twDepartmentofCivilEngineeringNationalCentralUnivers
2、ity,TaiwanTaoyuan,Chungli,32054,TaiwanTel:886-3-4581080Fax:886-3-4227183PaperSubmittedforPresentationandPublicationatthe82thAnnualMeetinginJanuary2003oftheTransportationResearchBoard,Washington,D.C.21Linetal.ABSTRACTIn1999,theTaiwanHighwayBureauimportedARAN(AutomaticRoadAnalyzer),aspeciall
3、yequippedcarthatprovidesspeedyandaccuratepavementinspection,andoutputsthemeasuredpavementroughnessasIRI(InternationalRoughnessIndex).Sincethen,thetraditionalroad-inspectorvisualinspectionperformedatnetworklevelpavementinspectionhadthetendencytobereplacedbyautomaticinspectiondevices.Enginee
4、rsalsohavegraduallystartedtopayattentiontoIRIindex.However,somepavementdistressdeteriorationconditionsareuniquetoTaiwanduetotheislandclimateandspecialcharacteristicsofthetrafficandaxleload.Inthispaper,analysisoftherelationshipbetweenIRIandpavementdistressesisperformedbasedonaback-propagati
5、onneuralnetworkmethodologytoevaluatetheapplicabilityofIRItobetreatedasacriticalrepresentationindexofpavementperformance.Moreover,theeffectsofdistinctdistresstypesandtheirextentonIRIcanalsobeanalyzedtoinferroadworkbenefit.Basedonaback-propagationneuralnetworkweconstructinthisresearch,IRImay
6、bepredictedaccuratelyfromdistressratingresultsobtainedfrompavementvideoimages,whichisrecordedbyacameramountedonARAN.ThecorrelationcoefficientbetweenIRIandthedistressvariablesreaches0.944,whichshowsthatIRImaycompletelyreflectpavementdistressconditions.Thus,itisfeasibletouseIRIasapavementper
7、formanceindex.UsingIRImaysimplifyinspectionworks.Obviously,differentdistresstypesdifferentlyimpactIRI.WefindthedefiniteeffectofeachdistresstypeonIRIbysensitivityanalysis.Thus,wemayinfertheimprovementintheIRIresultedfromcertainextentoramountofrepairforeachdistr
