VRCollide!ComparingCollision-AvoidanceMethodsBetweenCo-locatedVirtualRealityUsersANTHONYSCAVARELL I

VirtualRealityVirtualReality(VR)isadigitalmediumthatfocusesonprovidinggreatimmersion andpresence.◦ Immersion:“adescriptionofthetechnologythat...deliversaninclusive,extensive,surrounding,andvividillusionofreality”[1][9]

◦ Presence:“astateofconsciousness”[1][9]

◦ Head-MountedDisplays(HMDs)focusonprovidingthisexperiencethroughaheadsetthatcompletelycoversonesvisionandhearing.MostcosteffectiveversionofVR.

Multi-UserVirtualRealityExperiencesthatinvolvemultiplesimultaneous(concurrent)users

Oftenrelyuponlocal/remotenetworking[2][7]

Powerfulapplicationsineducation[7],virtualheritage[6],and3Dcontentcreation[10].

RecRoom PoolNationVR TiltBrush

TheProblemWhataboutmulti-userexperiencesthatinvolveco-locatedusers– sharingthesamevirtualandrealspace?

whyisco-locationimportant?◦ Physicalproximityhelpsencourageresearchandcreativity[4]◦ WhathappenswhenVRbecomespopularenoughthatmultiplepeopleinonehouseholdwanttouseVRsimultaneously?EithersharingVirtualEnvironmentsorwithinunrelatedVE’s.

PriorWorkRedirectedWalking(asubtleformofmotioncompression)[4]

Avatars◦ Streuber,S.,&Chatziastros :Multi-UserVirtualReality[9]◦ TiltBrush,RecRoom,PoolNationVRetc.

PriorWorkcont.EnvironmentalCollisions◦ Thoughnotaboutavoidinguserswecanlooktohowotheravoidcollisionswithenvironmentalboundaries.

◦ SteamVR’s “Chaperone”implementationisthemostfamousexample◦ ButHololens real-timeenvironmentalrecreationsinvirtualspacecouldalsobeuseful.

Chaperone(grid) Hololens 3Dscanning

Methodology– ThreeMethods

MovetoyellowcolumnBoundingBoxAvatar CameraOverlay

Methodology- Apparatus

MovetoyellowcolumnPhysical Environment Virtual Environment

ComputerResearcher

Participant

Potential Simulated User Positions

Participant

Vive HMD

Play Area 5m x 5m

4m Diameter

targets

activetarget

Methodology- Apparatus

Movetoyellowcolumn

PhysicalViewVirtualView MobileControlApp

Methodology- IntroductionTest3differentmethodsofcollisionavoidance(simulatingan“other”user)inthehumanusers’VE.3Methods

1. Avatar - primitivehumanshape2. BoundingBox - Distancedependentvisibleboundinggrid3. CameraOverlay - ”AR”usingfrontcamera

DataCollected◦ Movementtime- Trialendtime– trialstarttime◦ In-appsurveydata- completedaftereverytrial.◦ Numberofcollisions- betweenuserand”simulateduser”◦ PreandPostexperimentsurveys- (effectiveness,pleasantness,safeness,suitability,open-endedquestion).◦ AnxietyLevels- bytrackinghowhardtheanalogtriggerontheVivecontrollerwaspressed(unsuccessful).

Methodology- DesignOverview◦ Awithin-subjectsdesign◦ 1IV,3levels◦ 3DVs

Numbers◦ 12Participants(4female,8male)betweentheagesof18and35◦ Priorexperiencenotnecessary(thoughallhadnonetolittleexperienceinVR).

Technology◦ 1ViveVRHMDandWin10PC(4-coreXeon,16GBRAM,GTX970)◦ Unitydeveloped(efficiencyofdevelopmentandValvesupportedVRplug-ins)

Methodology- Design◦ 18trialsofeachmethodperparticipant.Randombutconsistentorderforeachparticipant.◦ Eachof6participantsfollowsadifferentorderofmethodtestingviaabalancedLatinsquare.◦ 18*3=54trialsperparticipant◦ 54*12=648experimentaltrialstotal

Hypothesis

AftertheCameraOverlay controlmethodwefeelthattheAvatarmethodwillbethemostpreferred,providefastermovementthroughaspacewhileavoidingtheleastnumberofcollisionswithotherusers.

Results– MovementTimeSignificantdifferencesbetweenallmethodsusingANOVAandTukeyHSDpost-hoctesting

Results– NumberofCollisionsSignificantdifferencesbetweenallmethodsusingANOVAandTukeyHSDpost-hoctesting.

NosignificancedetectedbetweenAvatar andCameraOverlay though.

Results– In-AppSurveyResponses

Post-Trialcomfortlevelswerecomparedwiththenon-parametricFriedmantest;andtheywerefoundtobenon-significant.

Results– Post-ExperimentSurveyPost-ExperimentsurveyresponseswerefoundtobesignificantusingaFriedmantest.Furtherpost-hocanalysisshowsnorealdifferentbetweenCameraOverlay andAvatarbutsignificantdifferencesbetweeneachofthoseandtheBoundingBoxmethod.

Discussiono Overviewo BoundingBox thesafestwithlesscollisionso CameraOverlay allowedforthequickestmovementacrossthespaceo BothCameraOverlay andAvatarwerejudgedtobecomfortabletonavigate

oWith6participantstherewasnotenoughdatatoquantitativelyanalyzepost-surveydatabutsomeinterestingnoteso The“surprise”elementoftheBoundingBox deemedunpleasantbymostparticipantso MostwerecomfortablewithCameraOverlayMethod.Comfortincreasedwhenseeingphysicalenvironment.o BoundingBoxmovementtimesweresignificantlylongerasparticipantswereverycautiousafterbeing“surprised”o Avatarwaspersonifiedbysomeparticipants.Wasdescribedas“cute”and“annoying”.o MostuserswereveryawareofHMDwiresbutstillfoundituncomfortabletohavetoavoid.CameraOverlay wasdeemedusefulto

seethesewires.o TheBoundingBox ismuchclearerindefiningpersonalspaceboundaries.

ConclusionPreventingusercollisionsinvirtualspaceisinteresting!

ForuserswithinthesameVEanAvatarbasedmethodshouldbebetterrecognizableasanotheruserwithinthespacebutless“safe”asusershaveahardertimejudgingpersonalspaceboundariesinVR.

ForusersindifferentVEsanalwaysvisibleAvatar basedmethodmaybreakimmersion/presence.

PerhapsthereisroomforahybridmethodthatcombinestheBoundingBox andAvatar(perhapsthemed)?

FutureWorkExplorehybridmethods(Avatar/BoundingBox)

Increasenumberofparticipants

ExploreBoundingBox visibilitythresholdsthatcouldincreaseusercomfortwhilekeepingpresencehigh.

Usemultiplehumanparticipantstoallowformorenaturalmovementwithinthespace.

ExploremultiplesimultaneousVEsforco-locatedVRusers.

References1. Bowman,D.A.,&McMahan,R.P.(2007).Virtualreality:howmuchimmersionisenough?.Computer,40(7),36-43.

2. Carlsson,C.,&Hagsand,O.(1993,September).DIVEAmulti-uservirtualrealitysystem.InVirtualRealityAnnualInternationalSymposium,1993.,1993IEEE(pp.394-400).IEEE.

3. Lepouras,G.,&Vassilakis,C.(2004).Virtualmuseumsforall:employinggametechnologyforedutainment.Virtualreality,8(2),96-106.

4. Gloor,P.A.,Grippa,F.,Putzke,J.,Lassenius,C.,Fuehres,H.,Fischbach,K.,&Schoder,D.(2012).Measuringsocialcapitalincreativeteamsthroughsociometric sensors. InternationalJournalofOrganisational DesignandEngineering, 2(4),380-401.

5. Holm,J.E.(2012).CollisionPredictionandPreventioninaSimultaneousMulti-UserImmersiveVirtualEnvironment(Doctoraldissertation,MiamiUniversity).

6. Lepouras,G.,&Vassilakis,C.(2004).Virtualmuseumsforall:employinggametechnologyforedutainment.Virtualreality,8(2),96-106.

7. Monahan,T.,McArdle,G.,&Bertolotto,M.(2008).Virtualrealityforcollaborativee-learning.Computers&Education,50(4),1339-1353.

8. Slater,M.,&Wilbur,S.(1997).Aframeworkforimmersivevirtualenvironments(FIVE):Speculationsontheroleofpresenceinvirtualenvironments.Presence:Teleoperators andvirtualenvironments,6(6),603-616.

9. Streuber,S.,&Chatziastros,A.(2007,December).HumanInteractioninMulti-UserVirtualReality.InProceedingsoftheInternationalConferenceonHumansandComputers(pp.1-7).

10. http://www.theverge.com/2016/9/30/13122502/tilt- brush-getting-multiplayer-mode-google-virtual-reality