76 ( ) 43, 8, [3] [4], [5], KSS,, [6], [7-8] [9],, 1,, ,, 1 ; 8 (6,, 20 m,5 m), ( 2 ), 9 KSS 3, (Karolinskasleepinessscale), 5 KSS,, , 3

Size: px

Start display at page:

Download "76 ( ) 43, 8, [3] [4], [5], KSS,, [6], [7-8] [9],, 1,, ,, 1 ; 8 (6,, 20 m,5 m), ( 2 ), 9 KSS 3, (Karolinskasleepinessscale), 5 KSS,, , 3"

陨余国
5 years ago
Views:

1 ( ) JOURNALOFTONGJIUNIVERSITY(NATURALSCIENCE) Vol43No1 Jan2015 : X(2015) DOI: / j isn x , 1, 1 2 (1, ; 2, ) :, andvalidatedbyreplayingthevideosbasedonthosedata,the 15 decisiontree modelwasestablishedtheresultsindicatethat, themostsignificantvariablestoestimatedrowsineslevelare, percentageofeyeclosure(perclos),thestandarddeviation,, oflateralposition,thetime-spaceareaoflanecrosing, the (percentageofeyeclosure, reverserateofsteeringwheelandthosevariablesarepositively PERCLOS) correlated to drowsines level Among these variables,, ;PERCLOS PERCLOSisthe bestvariableto divide drowsineslevel, 2 : PERCLOS WhenPERCLOSislowerthan28%,therearenodriversin 28%, ; seriouslydrowsystateandwhenperclosishigherthan21 9%,therearenodriversinnon-drowsystate;thetotalcorrect PERCLOS 219%, ; 6431%, 15 predictingrateis6431%to verificationthe model,4 drivers wereselected from the 15 driversrandomlythe 4 resultsofmodelvalidationshowedthecorrectpredictingrate 6322% 2 ofthedecisiontree modelis6322%inbothexperiments, thedecisiontreemodeldoesn tmistakeseriouslydrowsystate : ; ; ; ; fornon-drowsystate ; :U4912 Driver Drowsiness Level Analysis and PredicationbasedonDecisionTree XU Chuan 1, WANG Xuesong 1, CHEN Xiaohong 1, ZHANG Hui 2 (1KeyLaboratoryofRoadandTrafficEngineeringofthe Ministry ofeducation,tongjiuniversity, Shanghai201804,China;2China FAW Group Corporation R&D Center, Vehicle Safety of Body Department,Changchun130011,China) Abstract:In ordertoimprovetheaccuracy of drowsines detection,in thisstudy multi-source data for young and middle-aged experienced drivers including vehicle lateral position,steering wheelcontroling, andeye movementare colectedinadrivingsimulatorexperimentmeanwhile,the subjectivedrowsineslevelofthedriverswerealsorecorded Key words:driving simulator;drowsy driving :A ; decision tree;drowsineslevel; p hysiologicalindex;drivingbehavior index 2007, 3349, 1768 [1], [2],,, : : ( ) : (1987 ),,, xuchuan7@gmailcom : (1977 ),,,, wangxs@tongjieducn

76 ( ) 43, 8, [3] [4], [5], KSS,, [6], [7-8] [9],, 1,, 11 2010,, 1 ; 8 (6,, 20 m,5 m), ( 2 ), 9 KSS 3, (Karolinskasleepinessscale), 5 KSS,, 250 40, 3 ;, LED(light-emitingdiode), [10]

2 76 ( ) 43, 8, [3] [4], [5], KSS,, [6], [7-8] [9],, 1,, ,, 1 ; 8 (6,, 20 m,5 m), ( 2 ), 9 KSS 3, (Karolinskasleepinessscale), 5 KSS,, , 3 ;, LED(light-emitingdiode), [10] 12 1 Fig1 Tongjiadvanceddrivingsimulator 15, 13 2, (339±49), (180±71) m 20, (75±21) km, 6 2km ( 1,, 3,5,9,11,13) 2 2km (, 7,15), 2, 1, 24h,, 12h, 1, ( 6 ( 1h), 120km h -1 ), ; ;,

3 1, : 77 ; 14 CD ; ( ); 15 2 : (,, 9, 6 2 );, 3 8:00 ; 8h, 12:00,, 13:00 2, 2 ; Fig2 Ringform highwayandsegmentid 3 Fig3 Studyprotocolfortheconditionofnightshiftandpostlunch, 5min, 1h ;, km, : 21 10Hz ; ; Smart, eye, 60 Hz, [11] KSS ( 4a ), :1,2,3, L PS,4,5, 6,7,8, S LC 4b T,9 ; KSS,, 2 L 0,, L 0 = (a-b)/2 (1) 2 :a ;b S LC 4b KSS 1,, ; 2 n 2, S LC = ( L i -L 0 t i ), L i >L 0 (2) i=1,,294% :i ;n ; KSS 1, 133% KSS L i ;t i 1,572% KSS

4 78 ( ) 43 4 Fig4 Lateralpositionindicatorsilustration 22 :fgj / =ygj u, g x j y,ygj,u, S S : ;F^gj fgj m χ 2 p, x k y p, x k ;,, S R y,,, (percentage of eye closure,,, F L PERCLOS), :KSS 1~5,F L =1; KSS 6 7,F L=2;KSS 8 PERCLOS, 9,F L=3 20% (E t ), CHAID SPSS E t N c, N, F L,L PS,S LC,S S,S R,P S,P U,B F PERCLOS : P S = N c /N (3) SPSS :, 10 ; 703,, P U ; 3, 70, B F 3 31 (chi-squared automatic interaction detection, CHAID), CHAID,,,, CHAID y,x 1,x 2, x g,,x m m, CHAID : x k, x k (1 k m) y ; χ 2 χ 2 = g j (fgj -F^gj )2 KSS 5, P S ( χ 2 = F^gj (4) 35; Pearson ; 005, 100, 0001; Bonferroni ; ( 10 )

5 1, : ), 102),4 L PS ( χ 2 =179),S LC ( χ 2 =164),S R ( χ 2 = Fig5 5 SPSS OutputofSPSSdecisiontreediagramforthedecision model 8, : (674%) ; P S (1) P S<28%, (557%) 219%, (700%),, (3) 28% <P S <86%, (2) 85% < P S <116%, (654%) ; L PS

6 80 ( ) 43,,L PS>033 1 L PS<033 Tab1 Thepredictingaccuracyofdecisiontree ( 10% ) (4) 116%<P S<219% S LC<0516, S R>177 S R<177 (5) 116%<P S<219% S LC>0516, 29%,, (575±171), 720%, ( 3), % 1min,,,, 15 (5) P FL (P FL=1,2,3), 6322%, 2 O FL (O FL=1,2,3),p(P FL ) 3, 7571%; P FL O FL, 2, 3617% p(p FL )= Z (P FL ), 3 1 (5) Z(P FL ) :Z(P FL ) P FL 2 O FL ;Z(P FL,) Tab2 Thepredictingaccuracyofdecisiontreeof P FL calibrationtest (6) : p(p FL = O FL )= Z (P FL = O FL ) Z (6) :p(p FL =O FL ) ; % 1865% % 6465% 3000% % 7000% : % 3404% 1286% % 3617% 1143% % 7571% Z(P FL=O FL ) ;Z : 643%, 4 1, 3, 70%;,, 6465% 5571% 1,, %; KSS 2, % 1670%;, :1 3, % PERCLOS,, ;2 1, 3, PERCLOS, 1, 1 PERCLOSE 2 : PERCLOS, 3 28%, 33 ; PERCLOS 219%, ;3 15 4,4, 643%, (40±476), (875±250) 3,, 70%4

7 1, : %5 3 1,, 178,,, : Steeringrelatedvariables[R]Sweden:Link pinguniversity, 2007 [1] [8] [D] : [R] :,2009,2009 TheMinistryofPublicSecurityRoadandTransportAuthority MAO Zhe Research on identification of fatigue driving Roadtraffic crash statistics of the People s Republic of China[R]Beijing : Behavior[D] Wuhan: Wuhan The Ministry of Public University of Technology, SecurityRoadandTransportAuthority, [2] RadunI,OhisaloJ, RadunJ, [9] etaldriverfatigueandthelaw MatsonKIn-vehiclepredictionoftruckdriversleepines fromtheperspectiveofpoliceofficersand prosecutors[ J ] Lane position variables [D ] Lule :Lule University of Transportation Research, Part F: Traffic Psychology and Technology, 2007 Behaviour,2013,18(3):159 [10] FeenstraP,vanderHorstR,GrácioBJC,etalEffectof [3] LiW,HeQ C,FanX M,etalEvaluationofdriverfatigue simulatormotioncuingonsteeringcontrolperformance [ J ] ontwochannelsofeegdata[ J ]NeuroscienceLeters,2012, TransportationResearchRecord,2010,2185(1):48 506(2):235 [4],, ROC drowsinesusingvehicle measures:recentinsightsandfuture [ J ],2013,48(1): chalenges[ J ]JournalofSafetyResearch,2009,40(4): ZHAO Xiaohua,XU Shili,RONGJian,etalDiscriminating thresholdofdrivingfatiguebasedontheelectroencephalography sampleentropybyreceiveroperatingcharacteristiccurveanalysis [ J ]JournalofSouthwestJiaotong University, 2013,48(1): [5] PatelM,LalS,KavanaghD,etalApplyingneuralnetwork analysisonheartratevariabilitydatatoasesdriverfatigue[ J ] ExpertSystemswithApplications,2011,38(6):7235 [6] PilutiT,Ulsoy A GIdentificationofdriverstateforlane- keepingtasks[ J ]Systems, Man and Cybernetics,Part A: SystemsandHumans,IEEETransactions,1999,29(5):486 [7] BerglundJensIn-vehiclepredictionoftruckdriversleepines [11] CharlesC L,Simon G H, MichaelG LPredicting driver 檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿檿 ( 66 ) Materials,2004,7(2):161 [8], [M] : [13] Park S, Kwon S, Jung S H,etal Modeling of water,2004 LU Jinfu, GUAN Zhi Numerical solution of partial permeabilityinearlyagedconcretewithcracksbasedon micro porestructure[ J ]ConstructionandBuilding Materials,2012, differentialequation[m]beijing : TsinghuaUniversityPres, (1):597 [14], [9],,, [ J ],2006,38(12):2191 [ J ],2010,29(2):262 WANG Chuankun,GAO Xiangjie,ZHAO Yuxi,etalPeak value distribution of surface chloride concentration and convection depth of concrete[ J ]Buletin ofthe Chinese CeramicSociety, 2010,29(2):262 CHEN Wei,XU HongfaModelingofchlorideioningresin concreteconsideringtheeffectofcyclicwetinganddrying [ J ] Journalof HarbinInstituteof Technology, 2006,38(12): 2191 [15] HobbsD WAggregateinfluenceonchlorideiondiffusioninto [10] [D] : concrete[ J ]CementandConcreteResearch,1999,29(12):,2011 XIANG Fahai Study on the transmision mechanism of chlorideions in concrete [D ]Changsha: Central South University, 2011 [11] Taylor GDispersion ofsoluble materinsolventflowing slowlythroughatube[ J ]ProceedingsoftheRoyalSocietyof London,SeriesA,MathematicalandPhysicalSciences,1953, 219(1137): [16] GB/T [S] :, 2008 Ministryof Housing and Urban-Rural Developmentofthe Reople srepublicof ChinaGB/T Codefor durabilitydesign ofconcretestructures[s]beijing : China Architecture&BuildingPres,2008 [12], [ J ] [17] derincono T,CastroP,MorenoEI,etalChlorideprofiles,2004,7(2):161 SHI Huisheng, WANG Qiong Research on service life intwo marinestructures-meaningandsomepredictions[ J ] BuildingandEnvironment,2004,39(9):1065 prediction of marine concrete [ J ] Journal of Building

标题

2012 年 6 月交通运输系统工程与信息 Journal of Transportation Systems Engineering and Information Technology Vol 郾 12 Sup. 1 June 2012 文章编号 : 1009 鄄 6744 (2012) Sup. 1 鄄 0086 鄄 05 * 张万安, 肖跃秀 ( 吉安