)DENTIlCATIONOF$RIVERS&ITNESSUSING6IDEO)MAGESAND3TEERING"ASED&EATURES
2UZELITA.GADIRAN-OHD*AILANI-OHD.OR-OHD(ANIF-D3AADAND9OHAN+URNIAWAN
-%-32ESEARCH'ROUP
$EPARTMENTOF-ECHANICALAND-ATERIALS%NGINEERING 5NIVERSITI+EBANGSAAN-ALAYSIA
5+-"ANGI3ELANGOR -ALAYSIA
2ECEIVED$ATETH!UGUST !CCEPTED$ATETH&EBRUARY
!"342!#4
$RIVERS &ITNESS IS DElNED AS A MEASURE OF A PERSONS PHYSICAL STRENGTH mEXIBILITY AND ENDURANCE TO DRIVE4WOMAINFACTORSTHATLEADTOUNlTDRIVERAREDROWSINESSANDFATIGUE4HISPAPERDISCUSSESFEATURES EXTRACTEDFROMLIVEVIDEOOFDRIVERSANDSTEERINGWHEELDISPLACEMENTTOIDENTIFYTHERELATIVESTATEOF DRIVERSlTNESS!SOFTWAREBASEDSYSTEM$RIVERS&ITNESS-ONITORINGAND4RAINING3YSTEM$&-43 WAS DEVELOPEDFORTOACQUIRETHEREQUIREDDATAEXTRACTTHESELECTEDFEATURESANDIDENTIFYDRIVERSlTNESS
&IFTEENPARTICIPANTSWERETESTEDINANAUTOMOTIVESIMULATOR2ESULTSOBTAINEDSHOWEDTHATTHEREARE IDENTICALPATTERNSOFTHESELECTEDFEATURESFOUNDAMONGTHEUNlTDRIVERS
+EYWORDS$RIVERS&ITNESSSTEERINGWHEELDISPLACEMENT$RIVERS&ITNESS-ONITORINGAND4RAINING3YSTEM
$&-43 AUTOMOTIVESIMULATOR
!"342!+
+ECERGASANPEMANDUDIDElNISIKANSEBAGAIUKURANKEKUATANlZIKALKEmEKSIBELANDANDAYATAHANSESEORANG UNTUKMEMANDU$UAFAKTORYANGMENYEBABKANPEMANDUTIDAKSESUAIUNTUKMEMANDUADALAHKELETIHAN DANMENGANTUK+ERTASINIAKANMEMBINCANGKANPENCIRIANYANGDIEKSTRAKDARIVIDEONYATAPEMANDUDAN ANJAKANRODASTERENGUNTUKMENGENALPASTIPARASRELATIFKECERGASANPEMANDU3ISTEMBERDASARKANPERISIAN 3ISTEM0EMANTAUAN+ECERGASAN0EMANDUDAN,ATIHAN$&-43 DIBANGUNKANUNTUKMENDAPATKANDATA MENGEKSTRAK PENCIRIAN YANG DIPILIH DAN MENGENALPASTI KECERGASAN PEMANDU ,IMA BELAS PENYERTAAN TELAHDIPILIHSECARARAWAKUNTUKPENGUJIANDALAMPENYELAKUAUTOMOTIF+EPUTUSANAWALYANGDIPEROLEHI MENUNJUKKAN TERDAPAT CORAK YANG HAMPIR SAMA BAGI CIRI YANG DIPILIH DI ANTARA PEMANDU YANG TIDAK SESUAI
+ATAKUNCI+ECERGASANPEMANDUANJAKANRODASTERENG3ISTEM0EMANTAUAN+ECERGASAN0EMANDUDAN ,ATIHAN$&-43 PENYELAKUAUTOMOTIF
).42/$5#4)/.
!lTDRIVERISDElNEDASADRIVERWHOISALERTAND ABLETOREACTTOANYGIVENSITUATION4HElTNESS LEVEL IS INVERSELY RELATED TO DRIVERS DROWSINESS AND FATIGUE !CCIDENTS DUE TO DROWSINESS OR FATIGUE CAN CAUSE UNNECESSARY LOSS OF MONEY AND LIVES 3UCH SITUATION INCREASES THE INTEREST TOOFFERINVEHICLETECHNIQUETHATMAKESDRIVING SAFER HENCE THE NEED FOR A RELIABLE INVEHICLE DROWSINESSDETECTIONSYSTEM!NONLINEDRIVERS lTNESSIDENTIlCATIONSYSTEMSHOULDPREFERABLYBE ABLETOWARNDRIVERBEFORETHEDRIVERSBECOMES UNFIT TO DRIVE 2ELIABLE DETECTION SHOULD NOT ONLY DEPEND TO SOPHISTICATED SENSOR THAT MAY DISTURB DRIVER /NE FEATURE THAT MAKES SUCH SYSTEMAPPLICABLETOREALWORLDSITUATIONISTHAT THESENSINGMECHANISMMUSTBEUNOBTRUSIVEIN DRIVINGSCENARIOSUCHTHATITWOULDNOTDISTURB THEDRIVERORCAUSEUNNECESSARYDISCOMFORT -ULTIFEATURESDETECTIONMETHODSALLOWSTHE DETECTIONPROCESSTOBECOMEMOREACCURATEAND RELIABLEASCOMPAREDTOSINGLEFEATUREDROWSINESS DETECTION)NTHISRESEARCHTHEFEATURESUSEDWERE REVIEWEDFROMOTHERRESEARCHESBYRESEARCHERS AND ORGANIZATION SUCH AS .4(3! .ATIONAL 4RAFlC(IGHWAYAND3AFETY!DMINISTRATION 64) 3WEDISH.ATIONAL2OADAND4RANSPORT2ESEARCH )NSTITUTE .ATIONAL 2OAD4RANSPORT #OMMISSION .24# ANDSEVERALCARMANUFACTURERRESEARCHES SUCH AS BY4OYOTA AND .ISSAN .(43!
+IRCHERETAL
"ASEDONTHEIRlNDINGSDRIVERSPERFORMANCE DATA3TEERING2EVERSAL2ATE0%2#,/3ANDHEAD MOVEMENT ALONG YAXIS AND TWO EXOGENOUS VARIABLETIMEOFDRIVINGANDPERIODOFDRIVING FOR FUSED TOGETHER FOR THE DETECTION PURPOSES 4HESEFEATURESWERESELECTEDPARTLYBECAUSETHEIR ACQUISITIONISNONOBTRUSIVEANDTHERESPECTIVE COMPUTATIONALCOSTISMINIMAL
/"*%#4)6%
4HE OBJECTIVE OF THIS RESEARCH IS TO DETERMINE WHETHERTHEREARERELATIONSHIPSBETWEENDRIVERS lTNESSANDIMAGEANDSTEERINGBASEDFEATURES 5LTIMATELY THESE FEATURES WILL BE USED IN THE
$RIVERS &ITNESS -ONITORING 3YSTEM $&-3 AN INVEHICLEONLINEDRIVERSlTNESSMONITORINGSYSTEM
*AILANI ET AL 4HE WORKING FRAMEWORK OF
$&-3ISSHOWNIN&IGURE
3934%-$%3)'.
3TEERING7HEEL"ASED&EATURES
4HERE ARE FEW TYPES OF FEATURES WHICH CAN BE OBTAINED FROM STEERING WHEEL MOVEMENT AND THESE FEATURES WERE SHOWN BY THEIR RESPECTIVE RESEARCHERS TO HAVE A CERTAIN DEGREE OF RELATIONSHIPTODRIVERSDROWSINESSANDFATIGUE 4HE FEATURES ARE 3TEERING REVERSAL RATE "ELZ 3- 3TEERINGWHEELMOVEMENT3TEERING VELOCITY3HERMANETAL 3TANDARDDEVIATION OFSTEERINGWHEELMOVEMENT"ITTNERETAL AND3TEERINGADJUSTMENTMETHOD&UKUDAETAL
"ITTNERETAL USES6(!,INDEXARATIO OFFASTMOVEMENTSTEERINGANDSLOWMOVEMENT STEERING TO DETECT DROWSINESS4HEY ASSUMED THAT DROWSY DRIVER SELECT EASY DRIVING STRATEGY ANDCOMPENSATELARGEDEVIATIONSWHICHCANBE DETECTED FROM STEERING WHEEL MOVEMENT4HE 6(!, INDEX VALUE DECREASED WITH INCREASING FATIGUE "ITTNER ET AL 3TEERING WHEEL ADJUSTMENTINTERVALWASUSEDBY&UKUDAETALTO DETECT DROWSINESS4HEY ESTIMATED DROWSINESS ACCORDING TO THE CHANGE OF STEERING INTERVAL BYEXTRACTINGDATAFROMSTEERINGANGLE&UKUDA ETALALSOIMPLEMENTINDIVIDUALDIFFERENCEASA COEFlCIENTOFDROWSINESSJUDGMENTINHISSYSTEM
"ROWN MENTIONED THAT THE DETERIORATION OF STEERING SKILLS IS THE MOST VALID AND ACCESSIBLE MEASURE OF DROWSINESS WHICH ACCOMPANIES DRIVERSFATIGUE4HIFFAULTETAL STATESTHAT THEEFFECTOFFATIGUECANBEOBSERVEDBYCHANGES OFSTEERINGWHEELMOVEMENTAMPLITUDE/VERALL IT CAN BE CONCLUDED THAT STEERING WHEEL IS A SUITABLEDATASOURCEFORDROWSINESSANDFATIGUE DETECTIONSYSTEM
34%%2).'2%6%23!,2!4%
"ELZ HYPOTHESIZED THAT STEERING WHEEL REVERSAL RATE DECREASES AS A FUNCTION OF TIME ON THE ROAD OR DRIVERS FATIGUE 3TEERING WHEEL REVERSAL GENERALLY INCREASED OVER TIME THUS SUGGESTINGFATIGUEINDUCEDREDUCTIONINVIGILANCE ANDDECREASEDDRIVERPERFORMANCE4HETRENDWAS FREQUENTLYOBSERVED(IGHINDIVIDUALVARIABILITY HOWEVER WAS THOUGHT BY THE RESEARCHERS TO BE THE FACTOR THAT LIMITED THE UTILITY OF THIS MEASURE
7IERWILLEAND-UTO FOUNDTHATDURING ALONGDURATIONSIMULATORBASEDDRIVINGTASKTHE NUMBER OF STEERING REVERSALS GREATER THAN TWO
&)'52%-ODULESINDRIVERSlTNESSMONITORINGSYSTEM
Training On-board system
Sensor Sensor
Pre-Processing
Processing
Feature Extraction
Sensor Sensor
Pre-Processing
Processing
Feature Extraction
Supervised Training
Aromatherapies treatment
Experts Evaluation
GPS
DMC
Communication via SMS / LAN Drowsiness
Rate
Internal Warning (Audible) ANN Structure
ESKB
Online Identification (With ES and ANN)
ESKB Expert System Knowledge Based ANN Artificial Neural Network
DEGREES INCREASED WHILE THE NUMBERS OF SMALL STEERINGREVERSALSONEHALFTOTWODEGREES WERE FOUND TO DECREASE !VERAGE STEERING REVERSAL AMPLITUDEANDSTANDARDDEVIATIONALSOINCREASED OVER TIME "ELZ INDICATED THAT THE IMPLICATION OFTHISTRENDISNONFATIGUEDDRIVERSDETECTAND RESPOND TO ENVIRONMENTAL CHANGES QUICKLY WITHTIGHTPRECISECORRECTIONSWHEREASFATIGUED DRIVERS APPEAR TO HAVE AN INCREASED DETECTION THRESHOLDFORWHATMAYCONSTITUTEANECESSARY CHANGEANDARENOTLIKELYTORESPONDASQUICKLYAS NONFATIGUEDDRIVERS&ATIGUEDDRIVERSTHEREFORE ARE MORE LIKELY TO MAKE FEWER MORE COARSE CORRECTIONS)NTHISSTUDYTHESTEERINGREVERSALRATE WASCHOSENASTHEMAINFEATURESEXTRACTEDFROM
THE STEERING WHEEL FOR CLASSIlCATION OF DROWSY DRIVER
4O CAPTURE STEERING WHEEL POSITION A QUADRATURE OPTICAL INCREMENTAL ENCODER IS ATTACHEDTOSTEERINGWHEELINTHECARSIMULATOR 4HEARRANGEMENTALLOWSTHETRANSITIONSCOUNTING ANDSTATEVIEWINGOFTHEOPPOSITECHANNELDURING THIS TRANSITION7ITH THIS INFORMATION IT CAN BE DETERMINED IF ! LEADS " AND SUBSEQUENTLY THE DIRECTION4HE REVERSAL RATE IS THEN CALCULATED FROM THE POSITION OF THE STEERING WHEEL USING THERELATIONSHIPBELOW
WITH
WHERE
322K3TEERING2EVERSAL2ATEATTIMETKΔT 3K 3TEERINGPOSITIONATTIMETKΔT 3K 3TEERINGPOSITIONATTIMETK ΔT ΔT SAMPLINGTIMEINTERVAL
6)$%/)-!'%"!3%$&%!452%3
6IDEOIMAGEBASEDFEATURESWASUSEDEXTENSIVELY BY RESEARCHERS THE WORLD OVER IN DETERMINING SUITABLEPARAMETERSFORDROWSINESSANDFATIGUE DETECTION%XAMPLESOFFEATURESEXTRACTEDFROM LIVE VIDEO IMAGES OF DRIVERS ARE SUCH AS HEAD MOVEMENTHEADNODDINGANDEYECLOSURE !SURVEYDONEBY.4(3! SHOWSTHAT OF THE DRIVING POPULATION HAVE NODDED OFFFORATLEASTAMOMENTORFALLENASLEEPWHILE DRIVING AT SOME TIME IN LIFE OF DRIVER ADMITHAVENODDEDOFFWHILEDRIVINGTHATTHEY STARTLEDAWAKE4HEPERCENTAGESHOWTHATHEAD MOVEMENT DATA OR HEAD NODDING IS A VALID INDICATOROFDROWSINESS
(EADMOVEMENTISATIMESENSITIVEINDICATOR )TCANBEINTERPRETEDASADROWSINESSINDICATOR 0OPIEULETAL )NFACTTHEINCREASEOFTHE VARIABILITYOFTHEHEADPOSITIONISLOGICALANDCAN BE EXPLAINED BY THE CONJUNCTION OF TWO MAIN FACTORS &IRST LOTS OF RESEARCHES IN THE lELD OF PSYCHOPHYSIOLOGYOFVIGILANCEANDTIREDNESSHAVE SHOWNTHATASASUBJECTBECOMESDROWSYTHEREIS
AREDUCEDMUSCLETONETHROUGHOUTHISBODY,AL ETAL )NCARDRIVINGGLOBALRELAXATIONOFTHE BODYLEADSTHEDRIVERTOhSHRIVELvINHISSEAT4HIS GLOBALTRENDOFTHEPOSTUREHASANINmUENCEON THE BEAD POSITION MEAN HEAD POSITION TENDS TO LOWER WITH TIME &URTHERMORE FEELING MORE AND MORE UNCOMFORTABLE WITH TIME THE DRIVER TRIESREGULARLYTORESTOREAGOODDRIVINGPOSITION WHICH LEADS TO AN INCREASE IN THE VARIABILITY OF POSITIONSANDSPEEDSOFTHEHEAD0OPIEULETAL (EADMOVEMENTCANALSOBEINTERPRETED AS AN INCREASE IN THE SUBSIDIARY OR COLLATERAL ACTIVITIES AS INDICATIONS OF THE INDIVIDUAL LEVEL OFAROUSAL,ALETAL "EHAVIORALVARIATIONS RUBBING YAWNING NODDING SINGING INDUCE HEADMOVEMENTSVARIATIONSANDCOULDBEAMONG THE CAUSES OF THE INCREASE OF THE VARIABILITY IN HEAD MOVEMENTS %XPERIMENTS MADE ON LONG TIME PERIODS REVEALED THAT EVOLUTIONS OF HEAD MOVEMENTS INDICATORS WERE CONSISTENT WITH THOSEOFDRIVERSPERFORMANCEINDICATORS0OPIEUL ETAL
)-!'%02/#%33).'
&IGURESHOWSTHEmOWOFIMAGEPROCESSINGSTEPS IN$&-43!LLIMAGEACQUISITIONANDPROCESSING TASKS WERE DEVELOPED AS SOFTWARE OBJECTS CLASSES USING#THROUGHTHEUSEDOFOBJECT ORIENTEDPROGRAMMINGAPPROACH!N!CTIVE8™
CONTROL FOR FACE IDENTIlCATION WAS DEVELOPED UTILIZINGTHEABOVECLASSES4HIS!CTIVE8™CONTROL WASUSEDIN$&-3TOPROVIDE$&-3WITHONLINE IMAGEBASEDFEATURES
&)'52%%STIMATINGTHEHEADLOCATION Image Acquisition
Skin Segmentation
Adaptively Find Estimated Centre
Location of Face
Determine Face Area
(EAD-OVEMENT$ETECTION
(EADMOVEMENTINTHEXANDYAXIS(-8AND (-9 RESPECTIVELY WERE CALCULATED FROM THE ESTIMATEDCENTEROFTHEFACEPOSITION4HECENTER POSITIONWASESTIMATEDUSINGSTATISTICALMETHODS APPLIEDONTHESEGMENTEDAREAOFTHESKIN%ACH AREA WITH HIGH PROBABILITY OF BEING A SKIN IS MARKED AS AND AREA WITH LOW PROBABILITY OF BEINGASKINISMARKEDAS!CENTERPOSITIONOF DOMINANTFACEFROMANINPUTVIDEOIMAGEISTHEN ESTIMATEDUSINGTHEFOLLOWINGRELATIONSHIP
WHERE
AND
322K3TEERING2EVERSAL2ATEATTIMETKΔT 3K 3TEERINGPOSITIONATTIMETKΔ 3K 3TEERINGPOSITIONATTIMETK ΔT ΔT SAMPLINGTIMEINTERVAL
2'"XY ISTHE2ED'REENAND"LUECOMPONENT OFAPIXELATLOCATIONXY
)NTHEPRESENCEOFMULTIPLEFACESWITHINONE SCENEONLYTHEDOMINANTFACEWHICHISCENTERED AND IS SIGNIlCANTLY LARGER THAN THE REST WILL BE IDENTIlEDASTHEFACE&IGURESHOWSTHEEXAMPLE OFTHEHEADDETECTIONPROCESS-OVEMENTSINTHE XANDYAXISAREDElNEDASTHEDIFFERENTBETWEEN THE CURRENT ABSOLUTE POSITION OF THE CENTER LOCATIONXKYK ANDTHEPREVIOUSABSOLUTEPOSITION OFTHECENTERLOCATIONXKnYKn
2EACTION4IME42
2EACTION4IMEWASUSEDINTHEEXPERIMENTSASA NUMERICALVALUEOFDRIVERSlTNESS4HEORETICALLY AN ALERT AND lT DRIVER WILL RESPONSE FASTER TO A STIMULUSEGSUDDENEXISTENCEOFOBSTACLEON
THEROAD ASCOMPAREDTOALESSALERTANDlTDRIVER 4HE2EACTION4IME42WASTHEREFOREDElNEDAS THETIMEDIFFERENCEBETWEENASTIMULUSWASGIVEN TOATESTDRIVERANDTHETIMETAKENBYTHETESTED DRIVERTORESPONSETOTHESTIMULUS)TISINVERSELY PROPORTIONALTOTHEDRIVERSALERTNESSANDlTNESS )N THE EXPERIMENTS THAT WERE CARRIED OUT THE STIMULUSWASGIVENINTHEFORMOFAREDLIGHTAND THE DRIVER IS REQUIRED TO PUSH A BUTTON ON THE STEERINGTOINDICATETHATHEISALERTANDlT
-%4(/$
%XPERIMENT3ETUP
4HE EXPERIMENT TOOK PLACE IN A FIXED BASED SIMULATORTHE!UTOMOTIVE3IMULATORFOR$RIVERS
"EHAVIORAL!NTHROPOTECHNIC3TUDY!3)3 !3ONY 03BASEDRACINGGAMEWASUSEDTOPROVIDEIN CITYDRIVINGSCENARIO%ACHSUBJECTISALLOWEDTO DRIVEUNTILTHEYFEELTIRED4HEREISNOlXDISTANCE ANDTIMELIMIT4HEEXPERIMENTWASCARRIEDOUTAT AROUNDPM4HEDROWSINESSLEVELESTIMATED IS A LITTLE BIT HIGHER DURING THAT PERIOD DUE TO BODYCIRCADIANCYCLEANDAFTERMEALFACTOR 4HEVARIABLESMEASUREDFROMTHEEXPERIMENTS ANDDISCUSSEDINTHISPAPERWERE
(EAD MOVEMENT IN X AND Y AXIS (-8 AND (-9
3TEERING2EVERSAL2ATE322 2EACTION4IME$ATA42
$RIVERS &ITNESS -ONITORING AND 4RAINING 3YSTEM $&-43 A LOCALLY DEVELOPED SYSTEM FOR DATA ACQUISITION FEATURE EXTRACTION AND IDENTIFICATION OF DRIVERS FITNESS WAS USED EXTENSIVELY THROUGH OUT THE EXPERIMENTS4HE 'RAPHICAL5SER)NTERFACE'5) OF$&-43ISSHOWN IN&IGUREBELOW
2%35,4
/NLINE(EAD$ETECTIONAND)SOLATION
ONLINEHEADDETECTIONANDISOLATIONFROM THELIVEVIDEOSTREAMOBTAINEDFROMTHECAMERA WASACHIEVEDTHROUGHOUTTHEEXPERIMENT&IGURE BELOWSHOWSTWOEXAMPLESOFSUCCESSFULHEAD DETECTIONANDISOLATION
/6%2!,,2%35,43
4HE RESULTS OBTAINED FROM THE EXPERIMENTS WERE SHOWN IN4ABLE 0EARSON CORRELATION COEFlCIENTρWASUSEDTOCALCULATETHESTRENGTH OF RELATIONSHIP BETWEENσ322 (-8 AND (-9
/RIGINAL&RAME 0ROCESSED&RAME
3AMPLE
3AMPLE
&)'52%3AMPLE/FAREASFROMLIVEVIDEOFRAMEIDENTIlEDASFACEAREA 4!",%3UMMARYOFρVALUES
3AMPLE
!VG
&)'52%'5)OF$&-43
WITH424HEAVERAGESOFTHEABSOLUTEVALUEOFTHE CORRELATIONCOEFlCIENTFOREACHRELATIONSHIPWERE BETWEEN AND 4HESE VALUES INDICATESTHATTHEREARERELATIONSHIPBETWEENTHE INVESTIGATED VARIABLES THOUGH THE RELATIONSHIP WASNOTVERYSTRONGONTHEAVERAGE(OWEVERAT INDIVIDUALSAMPLELEVELTHEREARESAMPLESTHAT INDICATESTHATWHILETHERELATIONSHIPMIGHTNOT BESTRONGBETWEEN42ANDσ322THERELATIONSHIP
&)'52%2ELATIONSHIPBETWEEN42WITHSTANDARDDEVIATIONOF322FORB SUBJECTANDA SUBJECT A
Reaction Time vs Standard Deviation SRR
Standard deviation SRR
R2 = 0.7828
Reaction Time
1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0
200 220 240 260 280 300 320 340
B
Reaction Time vs Standard Deviation SRR
Standard deviation SRR
R2 = 0.5774
Reaction Time
2.5
2
1.5
1
0.5
0
0 50 100 150 200 250 300 350
BETWEEN42AND(-8ANDBETWEEN42AND(-9 ISQUITESTRONGSAMPLEANDSAMPLE ANDVICE VERSASAMPLE )NSHORTTHEREARECASESINWHICH THERELATIONSHIPBETWEEN42ANDANINDEPENDENT VARIABLEISWEAKBUTTHERELATIONSHIPBETWEEN42 ANDOTHERINDEPENDENTVARIABLESARESTRONG4HIS BEHAVIOURCANBEATTRIBUTEDASTHEINDIVIDUALTRAIT OF EACH SAMPLEρMAX SHOWS THE LARGEST VALUE OFρFOREVERYSAMPLEANDTHEAVERAGEVALUEOF
A
Reaction Time vs HMX
Standard deviation HMX
R2 = 0.7803
Reaction Time
3
2.5
2
1.5
1
0.5
0
0 0.5 1 1.5 2 2.5 3 3.5 4
B
&)'52%!NALYSISOFHEADMOVEMENTFORSUBJECT Reaction Time vs Standard Deviation of HMY
Standard Deviation HMY
R2 = 0.8131
Reaction Time
3
2.5
2
1.5
1
0.5
0
0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016
ρMAX IS RELATIVELY HIGHER THAN THE AVERAGE CORRELATION COEFlCIENT VALUES4HE ABSOLUTEVALUEOFρWASUSEDINTHECALCULATION SINCE THE INTEREST WAS ON THE STRENGTH OF THE RELATIONSHIPRATHERTHATTHEWAYTHERELATIONSHIP GOES
2ELATIONSHIPBETWEEN42 AND σ322
OF THE ANALYZED DATA SHOWS POSITIVE
CORRELATIONBETWEEN2EACTION4IMEAND3TANDARD
$EVIATIONOF3224HEBESTLINEARRELATIONSHIPlT SHOWSASTRONGPOSITIVECORRELATIONBETWEENTHE INVESTIGATEDVARIABLESWITHA0EARSONCORRELATION COEFlCIENTVALUEOFρ&IGUREA AND B SHOWEXAMPLEOFTHERELATIONSHIPBETWEEN FOR42 ANDσ322 TWO RANDOM SUBJECT AND SUBJECT"OTHGIVEPOSITIVERELATIONTHATSHOWS ASREACTIONTIMESLOWERTHEDRIVERTENDTOMAKE
LARGERSTEERINGDEVIATIONASSIGNOFTIREDNESSOR EARLYDROWSINESS
2ELATIONSHIP"ETWEEN(-8AND(-9WITH42 4HEBESTLINEARRELATIONSHIPBETWEEN42AND(-8 HASA0EARSONCORRELATIONCOEFlCIENTOFρ
&IGURESHOWSTHEPLOTOF42VS(-8AND(-9 FOR3UBJECT)TCANSEENTHATHEADMOVEMENTS SHOWGOODCORRELATIONTOREACTIONTIMEWHICHIS INDICATOROFDROWSINESS
#/.#,53)/.
4HE EXPERIMENTAL RESULTS PROVED THAT THERE IS EXISTENCEOFASIGNIlCANTRELATIONSHIPBETWEEN THE SELECTED PARAMETERS AND DRIVERS FITNESS (OWEVERINDIVIDUALVARIABILITYWASALSOSIGNIlCANT AND IT WAS OBSERVED THAT IT IS NOT POSSIBLE TO DERIVE A SINGLE MATHEMATICAL RELATIONSHIP THAT CAN GENERALIZED THE RELATIONSHIP BETWEEN THE SELECTED PARAMETERS AND DRIVERS lTNESS AS THE CORRELATION BETWEEN INVESTIGATED VARIABLES
VARIES )N OTHER WORDS THE EXACT NATURE OF THE RELATIONSHIP VARIES BETWEEN INDIVIDUAL AND THEREFORE A MODEL WHICH IS PERSONALIZED TOWARDS AN INDIVIDUAL WOULD BE ABLE TO BETTER DESCRIBE THE RELATIONSHIP &URTHERMORE IN THIS PAPERONLYLINEARRELATIONSHIPWASASSUMEDAND INVESTIGATED BETWEEN THE SELECTED PARAMETERS AND DRIVERS FITNESS )T IS HIGHLY LIKELY THAT THE RELATIONSHIP THAT OCCUR IS NONLINEAR IN NATURE )N THE FUTURE NONLINEAR RELATIONSHIP BETWEEN THESELECTEDFEATURESANDDRIVERSlTNESSWILLBE INVESTIGATEDUSING!RTIlCIAL.EURAL.ETWORKAND
%XPERT3YSTEM.EVERTHELESSTHEMAINOBJECTIVE OF DETERMINING THE EXISTENCE OF RELATIONSHIP BETWEEN THE SELECTED PARAMETERS AND DRIVERS lTNESSWASFULlLLEDANDSHOWED
!#+./7,%$'%-%.43
4HEAUTHORSWOULDLIKETOTHANKTHE-ALAYSIAN -INISTRYOF3CIENCE4ECHNOLOGY%NVIRONMENT FORSPONSORINGTHISWORKUNDERPROJECT)20!
32'RANT
2%&%2%.#%3
-ADNI!-AND7ELLS2&!N!DVANCED3TEERING WHEEL 3ENSOR 4HE *OURNAL OF !PPLIED 3ENSING 4ECHNOLOGY
(ARTLEY , (ORBERRY4 -ABBOT . +RUEGER '0 2EVIEW OF &ATIGUE $ETECTION AND 0REDICTION4ECHNOLOGIES.ATIONAL2OAD 4RANSPORT #OMISSION ONLINE WWWNRTCGOV AUPUBLICATIONREPORT ASPLOPUBLIC -ARCH
*OHN (OPKIN 5NIVERSITY 3URVEY OF /N"OARD 4ECHNOLOGIES!PPLICABLETO#OMMERCIAL6EHICLE /PERATIONS ONLINE HTTPCVISNFMCSADOT GOVDOWNDOCSCVISNDOCS?GENERALONBOARD?
TECH?SURVEYPDF-AY
-ALLIS--3LOW%YE-OVEMENTSASA0OTENTIAL -EASUREOF&ATIGUEAND!LERTNESSONLINE HTTP WWWMOMRPORG#OGNITIVE$AY$!9 3LOW%YE-OVEMENTSASA0 OTENTIAL-EASUREOF&ATIGUEAND
!LERTNESS?-ALLISPDF-AY 9OSHIYUKI 5 $RIVER"EHAVIORAND!CTIVE3AFETY
2$2EVIEWOF4OYOTA#2$,
7YLIE #$ 3HULTZ4 -ILLER *# #OMMERCIAL -OTOR6EHICLE$RIVE&ATIGUEAND3TUDY!LERTNESS ONLINE HTTPWWWTCGCCATDCPROJECTS HFACTORSBHTML*ULY
.ATIONAL4RAFlC (IGHWAY AND 3AFETY !DMINISTRATION .4(3! .ATIONAL 3URVEY OF $ISTRACTED AND $ROWSY $RIVING !TTITUDES AND "EHAVIOURS ONLINE WWWNHTSADOTGOVPEOPLEINJURY
DROWSY?DRIVINGSURVEYDISTRACTIVEINDEX HTM-AC
7IERWILLE 77 AND -UTO 7( 3IGNIFICANT CHANGES IN DRIVERVEHICLE RESPONSE MEASURES FOREXTENDEDDURATIONSIMULATEDDRIVINGTASKS 0ROC&IRST%UROPEAN!NNUAL#ONFERENCEON(UMAN
$ECISION -AKING AND -ANUAL #ONTROL $ELFT .ETHERLANDS$ELFT5NIVERSITYOF4ECHNOLOGY
&UKUDA * !KUTSU % !OKI + !N %STIMATION OF DRIVERS DROWSINESS LEVEL USING INTERVAL OF STEERING ADJUSTMENT FOR LANE KEEPING *3!%
REVIEW
+NIPLING22AND7IERWILLE776EHICLEBASED DROWSY DRIVER DETECTION #URRENT STATUS AND FUTUREPROSPECTS0ROC)6(3!MERICAn 6ERWEY 7 " AND :AIDEL $ - 0REVENTING
DROWSINESS ACCIDENTS BY AN ALER TNESS MAINTENANCE DEVICE !CCIDENT !NALYSIS 0REVENTION6OLUME
*AILANI - - . (ANIF - - 3 2UZELITA .
$EVELOPMENT OF GENERAL PURPOSE OBJECT ORIENTED EXPERT SYSTEM SHELL FOR EMBEDDED SYSTEMAPPLICATION%33%! 4HERD)NTERNATIONAL
#ONFERENCEONADVANCESIN3TRATEGIC4ECHNOLOGIES
+IRCHER ! 5DDMAN - 3ANDIN * 6IRTUAL 4ECHNOLOGY6EHICLE #ONTROL AND $ROWSINESS 64)RAPPORT!!UTOLIV2ESEARCH
"ELZ 3- !N /N2OAD )NVESTIGATION /F
#OMMERCIAL-OTOR6EHICLE/PERATORS!ND3ELF
2ATING/F!LERTNESS!ND4EMPORAL3EPARATION!S )NDICATORS/F$RIVER&ATIGUE0H$$ISS)NDUSTRIAL AND3YSTEM%NGINEERING&ACULTYOFTHE6IRGINIA 0OLYTECHNIC)NSTITUTEAND3TATE5NIVERSITY 3HERMAN 0* %LLING - AND "REKKE - 4HE
0OTENTIAL OF 3TEERING7HEEL )NFORMATION TO
$ETECT $ROWSINESS AND !SSOCIATED ,ANE
$EPARTURE
4HIFFAULT 0 AND "ERGERON * &ATIGUE AND INDIVIDUALDIFFERENCESINMONOTONOUSSIMULATED DRIVING 0ERSONALITY AND INDIVIDUAL DIFFERENCES
"ITTNER 2 (ANA + 0OUSEK , 3MREKA 0 3CHREIB 0 AND6YSOKY0$ETECTINGOF&ATIGUE3TATES OFA#AR$RIVER&IRST)NTERNATIONAL3YMPOSIUMON -EDICAL$ATA!NALYSIS)3-$!
"ROWN ) $ 0ROSPECTS FOR 4ECHNOLOGICAL COUNTERMEASURESAGAINSTDRIVERFATIGUE!CCIDENT
!NALYSISAND0REVENTION
7IERWILLE77 AND -UTO7( 3IGNIFICANT CHANGES IN DRIVERVEHICLE RESPONSE MEASURES FOREXTENDEDDURATIONSIMULATEDDRIVINGTASKS)N 0ROC&IRST%UROPEAN!NNUAL#ONFERENCEON(UMAN
$ECISION-AKINGAND-ANUAL#ONTROL
0OPIEUL *# 3IMON 0 AND ,OSLEVER 0 5SING
$RIVERS (EAD -OVEMENTS %VOLUTION AS A
$ROWSINESS )NDICATOR 0ROC )%%% )NTELLIGENT 6EHICLES3YMPOSIUM)63
,AL 3+, #RAIG ! ! CRITICAL REVIEW OF THE PSYCHOPHYSIOLOGY OF DRIVER FATIGUE "IOLOGICAL 0SYCHOLOGY
