触摸屏论文

inImmersiveVirtualEnvironments

Ji-SunKim∗DenisGraˇcanin†VirginiaTech,USA

VirginiaTech,USA

ABSTRACT

iPhoneandiPodToucharemulti-touchhandhelddevicesthatpro-videnewpossibilitiesforinteractiontechniques.WedescribeiPhone/iPodTouchimplementationofanavigationinteractiontech-niqueoriginallydevelopedforalargermulti-touchdevice(i.e.Lemur).TheinteractiontechniqueimplementedonaniPhone/iPodTouchwasusedfornavigationtasksinaCAVEvirtualenviron-ment.Weperformedapilotstudytomeasurethecontrolaccuracyandtoobservehowhumansubjectsrespondtotheinteractiontech-niqueontheiPhoneandiPodTouchdevices.Weusedtheprelimi-naryresultstoimprovethedesignoftheinteractiontechnique.IndexTerms:I.3.6[MethodologyandTechniques]:Interac-tiontechniques—;H.5.2[UserInterfaces]:Inputdevicesandstrategies—1

INTRODUCTION

Theacceptanceandsuccessofimmersivevirtualenvironments(IVEs)fornavigationtasksstronglydependsonasuccessfulinter-actiontechnique.TherearemanyapproachestonavigationtasksinIVEs.Someapproachesusephysicallocomotionoftheuser,suchastreadmillwalking,andsomeuseadevice,suchasjoystick.Gen-eralacceptanceofnavigationinteractiontechniquesthatusesomekindsofphysicallocomotionisassumedtobehigher[2].

RecentintroductionofiPhone[3]andiPodTouchopensnewpossibilitiesforinteractiontechniquesondevicesthathaveamulti-touchscreen.Multi-touchdevicescansensemultipletouchessi-multaneously[4].Althoughthetechniquewasavailableatleastsincemid80’s[8],ithasnotbeenusedextensivelyuntilrecentlyduetolimitedavailabilityofdevicesthatsupportit.Single-touchscreensfoundtheirapplicationatpubliclyaccessibleinstallations(suchasmuseums,ATMs,touristinformationkiosks,etc.)butdidnotfindtheirwaytotheeverydayuse.Problemswithprecision,armfatigueandresponsetimearethelargesthurdles.IntroductionofiPhoneorspecialdevicessuchasThinSight[5]madealargecommunityawareofmultitouchdevicesandtheirpotential.Thesemultitouchdevicesareusedmoreandmoreininteractiondesign[1,9].However,thetaskofdesigningamultitouchinteractionisfarfromtrivial.Effectiveinteractiontechniquesarecrucialforthesuccessofnewinteractionparadigm.2

PROPOSEDINTERACTIONTECHNIQUE

Inourpreviouswork[7],wedescribedanavigationinteractiontechnique,namedFingerWalkinginPlace(FWIP).Ourgoalwastotakeadvantagesofourinherentwalkingknowledgeanduseitforsteeringcontrolwithfingers.FWIPisdividedintotwoparts,oneforthetranslationofaviewpointandonefortherotationof

∗E-mail:hideaway@vt.edu†E-mail:

gracanin@vt.edu

‡E-mail:matkovic@vrvis.at§E-mail:quek@vt.edu

IEEE Virtual Reality 2009

14-18 March, Lafayette, Louisiana, USA978-1-4244-3943-0/09/$25.00 ©2009 IEEE

KreˇsimirMatkovi´c‡FrancisQuek§VRVis,Austria

VirginiaTech,USA

aviewpoint,named“walking-in-place”and“rotation-in-place”,re-spectively.For“walking-in-place”,weusefinger(s)thatmoveasifhumanlegsmoveonatreadmill.For“rotation-in-place”,onefingeristouchingonespotwithoutmovingandanotherfingerisdraggingtotheleftortheright.Ausercanmoveinanydirection,andchangeorientationinavirtualworld.Virtuallocomotionspeediscontrolledusingthesizeandthefrequencyoffingermovements.InordertoevaluateFWIPfornavigationtasksinIVEs,weusedaLemurdevice[6]1withthemultitouchsurface/screensizeof12-inchdiagonallyandresolution800x600.

(a)Controlmode(b)Walkingmode

Figure1:UserinterfacesforiPhone/iPodTouch

Theoriginalone-handedFWIPdevelopedfortheLemurdeviceismodifiedbecausethetouchinformationdetectedbyiPhone/iPodTouchdeviceisdifferentfromtheonedetectedbytheLemurdevice(stationarytouchevent).

Themodificationismadeonlyfortherotationtechniquebecausetherotationtechniquerequiresmultipletouches.WhenusingtheLemurdevice,ausersimplykeepstouchingwiththethumbwhiletheindexfinger(oranotherfingertheuserprefers)isusedfordrag-ging.WhenusingiPhone/iPodTouch,theusertouchesthescreenwithtwofingersatthesametime,dragsboth,andtakesbothofthescreenatthesametime.

Fortheexperiment,wedesignedtheuserinterfaceswithtwomodes,thecontrolmodeforevaluatorsandthewalkingmodeforhumansubjects.AsshowninFigure1(a),evaluatorscancontrolthesystemsetupspecifictotheexperiment.Ahumansubjectcancontrolaviewpointusingthemultitouchscreen,asshowninFig-ure1(b).Inthepilotstudy,weobservedthatmostofthesubjectsaccidentlytouchedthe‘back’buttonortouchthenon-detectableareawhiletheyaremovingtheirfingerswithoutlookingatthescreen.WeattachedtherubberbandstolimitthewalkingareasontheiPhonescreen(Figure1(b)).TwocomponentsofFWIPforiPhone/iPodTouchareillustratedinFigure2(a)“walking-in-place”,andFigure2(b)for“rotation-in-place”.

3PILOTSTUDY

InordertoevaluateFWIPoniPhone/iPodTouch,weconductedapilotstudyusingaCAVEinstallation.Inthisstudy,iPhone/iPod

1In

July2007,thecompanyhasadoptedthenameofStantum

(http://www.stantum.com),whileJazzMutantremainsthebrandnameforLemurandotherproductsunderJazzMutant.

261

(a)Forward“walking-in-place”

(b)Clockwise“rotation-in-place”

Figure2:FWIPusingiPhone/iPodTouch

Touchisonlyusedasasurfaceforfinger–walking.Navigationdi-rectionisonlychangedbyfingermovement.InordertoconducttheexperimentwiththeconstraintthatuserswouldnotphysicallymoveintheCAVEimmersivespace,weplacedaniPhone/iPodTouchonthetabletoprovideapersistentspatialreference(Figure3).

Figure3:FWIPwithiPhone/iPodTouch

ThispilotstudywasconductedwiththeobjectivetodeterminehowFWIPcanbeappliedforamultitouchhandhelddevice,suchasiPhone,fornavigationtasksinIVEs.WeaskedsubjectstoholdaniPhone/iPodTouchwiththenon-dominanthand,alignitwiththeverticallineofthefrontwallintheCAVEspace,andmovetheirfingerswiththedominant-handonthescreensurface.OurobservationfocusesonhowsubjectsrespondedtoFWIPusinganiPhone/iPodTouchfornavigationtaskinIVEs.

Figure4:ControlAccuracy

Weusedtwovirtualmazes.Thesimpleandcomplexmazessizesareapproximately245by195unitsand230by230units,respectively.Thecumulativeangletoturnaviewpointfromonemarkertoanotherislargerinthecomplexmaze(699vs.479de-grees).Theroutethroughthemazeispresentedwithmarkers(i.e.cross-rectangularobjects,Figure3).Theroutelengthpresentedwithmarkerobjectsisalittlelongerinthecomplexmaze(377vs.310unitsfromthefirsttothelastmarker).

262

Wedidnotprovideanyvisualfeedbackexceptfortheshadowsofvirtualobjects.Wemeasuredhowmanymarkerobjectssubjectsdirectlyhit.Thisdependentvariablewasmeasuredtoinvestigatehowaccuratelytheycancontroltheinterfaceforthenavigationtask(Figure4).Theaveragehitratesinthesimpleandthecom-plexmazesare80.8%and83.5%,respectively.Theresultssug-gestthatusersmaynavigatewithsufficientprecision.EventhoughFWIPisalreadyevaluatedwithanotebooksizemultitouchdevice(i.e.Lemur),weobservedthatsomedifferencesbetweenLemurandiPhone/iPodTouchmayinfluencethetaskperformance;(1)thewalkingsurfaceareaofiPhone/iPodTouchissmallerthanLemur;(2)therotationtechniqueisdifferentfromLemur.Interestingly,twosubjectstriedtokeeptouchingwiththumbandonlydragindexfingerontheiPhonesurface(i.e.“rotation-in-place”forLemur),eventhoughtheyneverusedFWIPbefore;and(3)eventhoughthisisaone-handedtechnique,asubjectshouldusethenon-dominanthandtoholdtheiPhone,whilethedominanthandisusedforFWIP.Weaskedsubjectstoalignthedevicewiththefrontverticalline,butitisnotstrictlyaligned,comparedtotheLemurdevicethatwasfixedonthetable.

Inaddition,twosubjectscommentedthatthetouchsensitivityofiPhone/iPodTouchistoohigh,andtheysometimesfeltjumpydur-ingnavigation.Wefoundithappenedwhentheyfrequentlymadeveryshortfingermovement.FWIPimplementationshouldbead-justedtofingermovementsoniPhone/iPodTouch.

4CONCLUSION

Inthispaper,wedescribedtheimplementationandpresentedsomepreliminaryresultsofFWIPappliedforiPhone/iPodTouchdevices.FWIPwaspreviouslyimplementedforLemurdevicebutduetodifferentsensingtechniquesandsmallertouchsurfacewehadtoslightlymodifytheinteractiontechnique.

Futureworkwillproceedintwodirections.First,wewillusetheresultsofourongoingformativestudytofinetunetheproposedin-teractiontechnique.Second,wewillperformacomparativestudyofseveralnavigationinteractiontechniqueswithdifferentinputde-vices,includingjoystick,Lemur,andiPhone/iPodTouch.There-sultsofthestudyshouldprovidesomeguidelinesabouttheuseofmultitouchdevicesfornavigationandrelatedtasks.REFERENCES

[1]H.Benko,A.D.Wilson,andP.Baudisch.Preciseselectiontechniques

formulti-touchscreens.InProceedingsoftheSIGCHIconferenceonHumanFactorsincomputingsystems(CHI’06),pages1263–1272,NewYork,2006.ACM.

[2]D.A.Bowman,E.Kruijff,J.J.LaViola,Jr.,andI.Poupyrev.3DUser

Interfaces:TheoryandPractice.Addison-Wesley,Boston,2004.

[3]S.Grissom.iPhoneapplicationdevelopmentacrossthecurriculum.

JournalofComputingSciencesinColleges,24(1):40–46,Oct.2008.[4]J.Han.Themultitouchscreen.http://multi-touchscreen.com/.Last

accessed27Oct.2008.

[5]S.Hodges,S.Izadi,A.Butler,A.Rrustemi,andB.Buxton.ThinSight:

Versatilemulti-touchsensingforthinform-factordisplays.InProceed-ingsofthe20thannualACMsymposiumonUserinterfacesoftwareandtechnology(UIST’07),pages259–268,NewYork,2007.ACM.[6]JazzMutant.Lemur.http://www.jazzmutant.com/lemuroverview.php.

Lastaccessed30Dec.2008.[7]J.Kim,D.Graˇcanin,K.Matkovi´c,andF.Quek:FingerWalkingin

Place(FWIP):ATravelingTechniqueinVirtualEnvironments.InPro-ceedingsof9thInternationalSymposium,SmartGraphics2008,pages58–69,2008

[8]S.Lee,W.Buxton,andK.C.Smith.Amulti-touchthreedimensional

touch-sensitivetablet.ACMSIGCHIBulletin,16(4):21–25,Apr.1985.[9]T.MoscovichandJ.F.Hughes.Multi-fingercursortechniques.In

ProceedingsofGraphicsInterface2006(GI’06),pages1–7,Toronto,Ont.,Canada,2006.CanadianInformationProcessingSociety.