Design in context of use: An experiment with a multi-view and multi-representation system for collaborative design

Science Arts & Métiers (SAM)

is an open access repository that collects the work of Arts et Métiers Institute of

Technology researchers and makes it freely available over the web where possible.

This is an author-deposited version published in:

https://sam.ensam.eu

Handle ID: .

http://hdl.handle.net/10985/14175

To cite this version :

Bo LI, Frédéric SEGONDS, Céline MATEEV, Ruding LOU, Frédéric MERIENNE - Design in

context of use: An experiment with a multi-view and multi-representation system for collaborative

design - Computers in Industry - Vol. 103, p.28-37 - 2018

Any correspondence concerning this service should be sent to the repository

Design

in

context

of

use:

An

experiment

with

a

multi-view

and

multi-representation

system

for

collaborative

design

Bo

Li

*

,

Frédéric

Segonds

,

Céline

Mateev

,

Ruding

Lou

,

Frédéric

Merienne

a

LISPENEA7515,ArtsetMétiers,UBFC,HESAM,InstitutImage,Chalon-sur-Saône,France

b

LCPI,ArtsetMétiersParisTech,151boulevarddel'Haôpital,75013Paris,France

ARTICLE INFO Articlehistory:

Received28September2017

Receivedinrevisedform11September2018 Accepted11September2018 Availableonlinexxx Keywords: Virtualreality Multi-viewsystem Digitalmock-up Multidisciplinarycollaboration Ergonomics ABSTRACT

Thecurrenttrendofproductdesignleadstoachangeinthecollaborativeworkingstyle.Tofindthemost efficientwaytoexchangeinformationonthedigitalmock-upofaproduct,asynchronousco-located collaborativedesignenvironmentwithrecenttechnologiesisinneeded.Anewgroupwareofmulti-view systemallowsmultipleuserstohaveindividualvisualinformationofadomain-specificrepresentationof digitalmock-up.Inthispaper,weproposeacasestudyforthedevelopmentandtestingofaco-located multi-view system in collaborativevirtual reality, aimingat enhancingthemultidisciplinary early collaborativedesign.AnergonomicmethodofPersonasisintroducedtotheevaluationofthetool, consideringvarioususerperformance.Withamultidisciplinarymugdesignscenario,experimentsare presented,validatingthebenefitsoftheproposedsystem.

1. Introduction

Designingisacomplexactivity,andfailureofsupportcanbe expensiveintermsof time,peopleandmoneyandcanhavea largeeffect onpractice [1]. Foryears, rapidproduct develop-ment (RPD) is relying on the use of new knowledge-based systems such as haptic systems and virtual reality (VR) [2]. Thereisanincreasinglystrongneedforimmersivevisualization andinteractivetoolsthatmaybeusedinindustrialapplications

[3].Recentdevelopmentsintheuseofmulti-viewtechnologies allowmultiplepoint-of-viewsandmultiplerepresentationsof anobject.Moreprecisely,withthedevelopmentofcomputer– human interaction (CHI) andVR technology, the approach of displaying multi-representation of a digital mock-up (DMU) through a multi-view system can be used in product design. Moreover, concurrent engineering has changed design habit from traditionalsequential engineeringto a parallelmode to reducetheoverallproduct developmenttime[4].Intheearly stages of design, stakeholders involved tend to propose a collaborativetoolthatissuitedtothe concurrentdesignstyle

[5].Finally,prospectiveergonomicsisnowadaysakeyfactorin innovation,ofpriorimportanceinproductdesign,lifecycleand especiallyinitsuse[6–8].Thus,extractedmotivationsforthis paperare:

 Tosimulatevirtualproductdesignsituationswithamulti-view systemandprovetheefficiencyofthemethod.

 To mix ergonomics analysis and early product design on an industrialuse-casetovalidatetheproposedmethod.

 Tovalidatethemulti-viewsystemonause-caseandmeasurethe benefitsthedesignteamcouldhavewhenusingthistechnology appliesintheconsumergoodsindustry.

Usingdesignresearchmethodology(DRM)approach[1],the followingdescriptivestudyisaimedatunderstanding collabo-rative designwith multi-point-of-view. We follow a compre-hensive descriptive study, which involves a literaturereview andanempiricalstudy.Afterreviewingtheliteratureof multi-view systems in industry, we determine our research focus (Section 2). Then, we develop our research plan based on a multi-viewsystemforearlycollaborativedesign(Section3)and undertakeanempiricalstudytoevaluateit(Section4).Finally, theresultsarethendiscussed(Section5)beforedrawingoverall conclusions.

2. Reviewoftheliterature

Theliteratureontheuseofmulti-viewistobediscussedhere first withthe definitionof multi-view system, followedby the general applications in the _field of virtual reality and multi-representation display. We focus then on an introduction of Personasmethodandaninnovativefuturecollaborationstylein earlydesignthroughmulti-viewsystem.

* Correspondingauthor.

2.1.Multi-viewsystem

The traditional computer supported collaborative design is basicallybasedonseveralmono-viewsystemsforcommunication indistanceandspecialsoftwareplatformsforproductinformation sharinginclient/servermode[9–11].However,multi-viewsystem isaco-locatedandsimultaneoussolutionofsharinginformation andworkingcollaboratively.

Multi-view system is a visual–perception interface which allowshumanstoseesimultaneouslymultipleimagesthrougha uniquesharedmedium.Themechanismofmulti-viewsystemis: several images are emitted simultaneously from the display medium and then received respectively by human vision, as showninFig.1.In[12,13],multipleviewsaredescribedasframes withview-dependentpixels.Theseviewscanbedisplayedasan output package simultaneously. To create multi-view system, varioustechnologiesexistintheliteratureandtheyareclassified intothreemodes:passive,active,andautomatic.

2.1.1.Passivemode

Theemittedimagesareprojected inthedifferentsubspaces of colorimetryorpolarizationoflightthroughpassivefilters.Anaglyph imagesarecreated byputtingimagesindifferentanaglyphiccolor channels,thenthese imagescan beseenseparatelywithred/cyan filters.Polarizingfiltersrestrictthelightinoppositevibrationdirections beforeprojection.Usingsuitablereceivingfiltersforeach correspond-ingprojectingfilter,imagescanbeencodedwithaseparation. 2.1.2.Activemode

Withinaveryshortperiod,imagesareactivelydisplayedone aftertheotherinsequence.Themechanismistosynchronizethe emitterandthereceiver.E.g.shutterglasses,whichisareceiver, canchangethetransparencyalternately,synchronizingwiththe refreshrateoftheemitter,whichisusuallyfromaprojectorora screen. As knownin humanvisual system (HVS), 60Hz is the lowestfrequency for humantohavea reflection ofcontinuous imageswithoutflickfusion[13].Modernoptoelectronicdisplays withtheirsupportingreceivingdevicescanoperatemuchhigher refreshrates,thusmultipleviewsmorethantwocanbecreated

[13].E.g.a120Hzprojectorandthesupportingshutterglassescan createtwoviewsinsequencethateachhasa1/120speriod. 2.1.3.Automaticmode

Automaticmodedoesnotneedanyequipment(e.g.Glasses) atimagereception.Imagescanbeseenseparatelyfromdifferent positionsin spatialdimensionbeyond thesame screen. E.g.A displayisplacedbehindaparallax-barrier,whichisanopaque sheet with patterned holes stamped out of it, or behind a lenticularsheet,whichiscomposedbyanarray ofmagnifying lenses.Lightfromanindividualpixelinthedisplayisvisibleonly fromanarrowrangeofviewingangles.Thus,theimages seen througheachholeorlenswillchangeaccordingtovisionspots

[12].Withadvancedscreentechnologies,multiplespatialviews more than two can be realized and the combination of the mechanisms above can produce the system with even more views[14].

2.2.Multi-viewapplications

The applications of multi-view can be categorized into two aspects:multiplepoint-of-viewsinVRandmultiple representa-tions.

2.2.1.Multi-viewinVR

Apoint-of-view(POV)ofanobjectisgeneratedfromacertain spatialposition,e.g.inFig.2,aDMUofanairplanehasseveralPOVs seen from different space positions. Multi-view system can be appliedtodisplaythemultiplePOVsofanobject,especiallywhen thesemultiplePOVsareusedtocreatethefeelingofstereoscopyin VR.

Becominganewwayforvisualizationand interaction,VR is widelyusedinproductdesign andmanufacturing,especially in automotiveindustry[15,16].Thisvirtualprototypeoftheproduct can help evaluate its design by simulating theusage (driving), manufacturing(assembly),etc.Comparedtotheactivitiesusing physicalprototype,VRmayhelpinsavingtimeandcost.

Totransfera3Dvirtualworldtocomputergraphics,2Dimages arecalculatedassnapshots of3Dworldaccordingtoapoint of view. Binocular vision is natural for humans. For each eye, visualizingfroma3Dworldisliketakingsnapshots.Thehuman braincanmergethetwosnapshots takenfromthetwoeyesto createstereoscopicrepresentation.

Thus,anyVRdeviceforcreating3Dimagesbelongsto multi-viewsystems.Thethreemechanismsofmulti-viewsystemalso work with VR devices, e.g. passive or active mode multi-view system: 3D glasses, including anaglyph, polarized[17–19], and shutterglasses[20].Inatwo-userVRapplication[15]ora multi-viewtable[18],themulti-viewsystem,acombinationofpassive modeandactivemode,has4POVsforfoureyes.Autostereoscopic displays [12,21] also work as an automatic mode multi-view systemtoreceivetwoPOVsforeacheyefollowingdifferentspatial positionsoftheeyes.

2.2.2.Multiplerepresentations

In addition toshowingmultiple POVs,multi-view system is usedtodisplayingmultiplerepresentations.Arepresentationisa perceptionofanobject.Peoplehavedifferentrepresentationsofan object and they make statements toinfluence the opinions or actionsofothers.Theusersofamulti-representationapplication choose the representations according to their different roles, differentinterests,anddifferentpreferences.Inengineeringdesign showninFig.2,anairplaneDMUisapackageofdataofallthe

Fig.1.Themechanismofmulti-view.

Fig.2.Theapplicationsofmulti-view:multi-POV(spacetransformsofDMU)and multi-representation(domaindataofDMU).

domainsgatheredarounda3Dmodeloftheproduct.Considering his/herowndomain,e.g.airplaneexteriordesign,userwillchoose one representation of the DMU towork with.To display each aspectofamulti-presentationobjectorapplicationtoeveryone involved,amulti-viewsystemcanbeused.

Literatures with the applications of multi-view system in displaying an object's multi-representation are reviewed. In

[19,20],amulti-viewsystemisusedtoshowtworepresentations ofatextparagraph.Sentenceswithsamemeaningarepresentedto twouserswithtwodifferentlanguagerepresentations.

A photo album with two representations is displayed by a multi-viewtablefortwousersin[18].Eachrepresentationfollows theuser's interest, e.g. architecture structure. On a street map between two cities, users can see separately two different representationsonthismulti-viewtable:distancerepresentation andtrafficrepresentation.

Seen from different angles, a skeleton representation of a humanbodyandanorganicrepresentationofthesamebodyare displayedseparatelytotwousersonanautostereoscopic multi-viewsystem[13].

Different from multiple representations, the application of multiplePOVs(discussedinSection2.2.1)maybepresentedwith onlyonerepresentation.E.g.in [15,22],twoormoreusershave theirownPOVsofacarfromdifferentspatialpositioninaCAVE system,buttherepresentationofthiscarisalwaysanassembly view.Multiplerepresentationsmayalsobepresentedinasingle POV.E.g.fromacertainPOV,thedifferentrepresentationsofan objectareoverlapped.

Differentfromsingleuser,multipleusersemphasizeproblems ofcollaborationamongstakeholders.Intheliteratures,veryfew applicationshaveaddressedtheuseofmulti-viewsystemonthe multi-representationofaproductduringitsdesignactivities.This paperfocusesontheproductdesigncollaborationwithdifferent representations of a product. The collaborative product design withmulti-viewsystemwillbediscussedinnextpart.

2.3.Collaborationproductdesign

For years, changes in design teams prone the importance of synchronousphysical co-location to solve design problems. Thus, members are operating in physical proximity which is the most efficientwaytoexchangeinformationontheproduct(inFig.3).When usingtheDMUasameansofcollaboration,designstakeholdersfrom differentbackgroundsmustcollaborateonasinglerepresentationof theproduct.However,itcouldbeusefulforthemtohavetheabilityto display,throughnewtechnologiessuchasmulti-viewsystem,their own“infield”representationtofacilitatedecisions.

Indeed, in the case of collaborative product design, project stakeholdersarerequestedtoworktogetherandinteracttoreach anagreementand makeshared decisions.Thelevel ofdecision couplingassessesthedegreeofcollaborationhere.Designersfrom thewholegroupworktogethertodesignproduct,followingthe customers’requirements.Theprojectleader,aswellastheproject group(a groupof designersfromvarious companieswho have competences and skills in various fields),attempt to solve the design question together.Collaborative activity is synchronized andcoordinatedthroughoutthecollaborativeprocess.

Whendealingwiththefutureuseofaproduct,thevisualization in3Disagoodmeanstosimulatethebehaviorofacustomerfacing thisproduct.Thus,ithelpsdesignerstoformalizethe“designin contextofuse”oftheproduct.Meanwhile,displayingthemultiple representationsofthisproductforbothcustomeranddesignerin real time can improve efficiency of the design activity in a collaborativeway.Bothmulti-POVfor3Dvisualizationand multi-representationforcollaborationcanberealizedbyusinga multi-viewsystemamongthecollaborators.Nextchapterispresenting thePersonamethodthatiscommonlyusedtodefinemodelsof usersinearlyproductdesign.

2.4.ThePersonasmethod

ThePersonamethodwascreatedbyCooper[24]anditwasmainly usedinthefieldofproductdesign.Personasareanarchetypeofusers thattaketheformofacardwithaphotographofthepersona,his/her name,sociological,demographic,andpsychologicalinformationwhich can be embellished with storyboardto promoterealism [25].The Personamethodisatoolfordesignprocess.Asmodelsofusers,with behavior, attitude, personalmotivation, andintentions, they allow designerstounderstandtheneedsoffutureusersastotheuseofa technology[26].

Personasarecommonlycreatedatthebeginningofthedesign processandareusedduringtheentiredesignprocess.Ideally,they arebuiltthroughtheparticipationofateam,becausethecreation processisenrichedbythedebatesofthemembersoftheteam[27]. Theycanbeusedindifferentmanners,dependingthenatureofthe projectandthedesigners’preference[28,29].Theyare material-ized based on ethnographic research, interviews, and user observations[30].Followingthecollectionofdataontheusers, ananalysisisrealized.Alistofbehavioraldataanddemographic variablesisdrawinguptodeterminethemaintrends.Fromthis behavioralpatternemerge,whichwilldeterminetheconstruction ofdifferentPersonas,whichwillbedescribedinnarrativeform,to createrepresentativerepresentationsofusersthatare represen-tativeofawholetargetpopulation[31,32].

The entertaining aspect of this method allows designers to stimulate their creativity and encourage the development of innovative ideas [33]. Indeed, they allow them to distance themselves from their way of apprehending the system since theynolongerreasonfromtheirpointofviewbutfromthatofthe personas[34].Inthedesignprocess,theyimprovethegeneration ofideaswhicharemorenumerous,moreoriginalandmoreflexible thana conditionwithoutpersonas [7]. Personasarecommonly usedwithscenariostohelpdesignersthinkandactlikeendusers. Butthemethodisalsocriticizedbecauseofthedesigners’lackof familiaritywiththis method[35],theabstractivenatureor the impersonality[29].To overcomethesedrawbacks, weused the methodofpersonaasroleplay.

Role-playingisamethodappliedtounderstandthecontextof useandthepointofviewoftheusers[36].Role-playsimulationare interactionsofseveralparticipantswhoadopttheroleof stake-holderswithvaryingpointsofview.Raijmakersetal.[37]have used this method as a co-creation tool to explore different perspectiveandpossiblerolesformultiplestakeholders.Intheir study,theyaskparticipantstoembodyastakeholderroleandto thinkasiftheywereintheirshoesduringworkshopsessions.This methodallowedthedesignerstofeelclosertotheusers andto projectonfutureuses.

2.5.Researchfocus

As we discussed in Section 2.3, VR is widely used for the preliminarytestandpre-evaluation.However,thereislittlework tosystematicallyevaluatetheuserexperiencessuchasusability, natural and intuitive interaction between human and physical objects,andusefulness[38].Multi-viewVRsystemcouldbeused to present multiple POVs of an object, as well as multiple representationsofacontent.In ourstudy,theaimistotestthe interactionbetweenend-usersanddesignersina multi-disciplin-arypoint ofview. Themainadvantageistoinclude asearlyas possibletheend-user in thedesign loop. It alsoprevents from creatingaphysicalmock-upthatiscost-fullandtimeconsuming. Personascouldalsoefficientlybeusedinavirtualenvironment. Buisineetal.[39],haveexperimentedtheuseofpersonathrough avatar invirtualenvironment. Thegoal was tomake engineers embodypersonatomakethemthinkasusers.Theexperiments weremadeusing“SecondLife”asavirtualworldofcollaborative design.Resultsshowthatideaproductionwasclosertotheuser needs.Engineersproducedmoreideastoanticipateuser experi-ence. The embodiment of an avatar in a virtual world was appreciated and helped them to feel closer of the users. The Persona method seems interesting to create a situation of interactionbetweenthedesignersandtheusers.Designersmainly use the method to consider user needs, but users are not necessarilypresentduringthedesignprocessanddecisionmaking. Inourstudy,weshowtheinterestofpersonasduringausercasein avirtualenvironment.Wepushedthismethodtoinvolverealusers playing therole of personas andinteracting with thedesigner, throughamultiviewsystem.

3. Researchplan:amulti-viewsystemforearlycollaborative designproposal

According to the literature review, multi-view system can improve the designing of multi-user interface, e.g. multi-user virtualrealityenvironment.Also,multi-viewsystemmayhelpin thefieldofcollaborativedesignwhereseveralexpertsareworking in multi-disciplinary activities with multi-representation of a DMU.

Toenhancetheearlycollaborativedesignamongexpertsfrom differentdomainstocommunicate,Personasmethodallows

end-users toparticipatein thedesignprocess andprovide practical feedbacktodesigners.Buildingpersonasfromtheearlieststagesof designwasveryimportantfordesignerstofamiliarizewithend users.Wewantedtomakethemmoreawareofusers’pointofview, bygivinglifetothepersonas.Forthatpurpose,weaskrealusersto embody the Personas profiles during a collaborative design session,ininteractionwithamulti-viewsystem.

Inthispaper,amulti-viewsystemtoolisintegratedinanearly design activity.Itaims atdevelopingamulti-viewcollaborative systemprototypeandevaluatingtheusabilityofCHIsystemand performanceofcollaborationthroughexperiments.Tosolvethis scienti_fic problem, the following scheme of the approach is proposedinFig.4:

The proposed approach is a method for evaluating the contribution of a multi-view system and the multiple users’ experience based ona multi-viewcollaborative design module whichiscomposedofamulti-viewhardware/softwaresystemand anearlycollaborativeworkscenario:

 Define anearly collaborativescenario consideringthespecial representationthatisadoptedbyeachPersonas.

 Establishavirtualimmersivemulti-viewsystemwhichconsists ofstereoscopy,headtrackingandhandtracking.

 Evaluate the usability of the system and theperformance of collaborationbyPersonasmethods.

 Analyze the evaluation results of a multi-view system and discussthecontributiononcollaborativeproductdesign.  Definerecommendationstoimprovethesystemregardingusers’

pointofview.

The detailsofthese stepswill bediscussedin thefollowing sectionofexperiment.

4. Empiricalstudy 4.1.Setup

4.1.1.Defineanearlycollaborativedesignscenario

An experiment of early collaborative design is conducted betweentwocollaborators.OneofthemisaPersonawhohasa typical character setting, such as age, profession. Another collaborator is a designer who can control the functional parameters in design. A multi-view system on DMU's multi-representationisusedtoallowtwopeopletocollaborateinreal time.

The interest of the personas is to propose to the users to embody an end user profile, which acts according to a given scenario.Wehavecreated2differentpersonasforthestudy:  A34-years-oldwomanwholikes drinkingteaduringtheday,

mainlyinhercarbetweentwoconsultations.

 A 42-years-old man, head of a company,mug userin public transportationandinhisoffice.

Theproducttodesignisamugthatcanbefilledwithwater,tea, coffee,etc.Theideaistojointwousersinonescenecollaboratively todesignamug.Oneuser,arealend-userwhoplaystheroleofthe persona,istestingthemuginvirtualenvironmentbymovingthe mugtoapreferableposition.Anotheruser,thedesigner,ismore interestedindesignanalysisdata,e.g.size,thermalanalysis,etc.If thesize ofthemugis changed,both userswillseethis tightly relatedelement.Sincebothusershaveotherindividualelements thatonlyeachonehimselfcanunderstand,amixedcouplingstyle ofthismulti-viewsystemwillbeapplied.

ThisapplicationinvolvestworepresentationsofamugDMU.As displayedwith3DVRsystem,eachrepresentationhastwoPOVsto

form stereoscopy, 4 views are needed for this multi-view system. It was chosenwiththefollowingcriterions:wellknownbyendusers, multi-contextsofuse,simplicitytodesign,timetoimplementina3D.

Thespecificityofthisstudyliesinthefactthatweproposetoa designerandausertocollaborateonthedesignofaproduct,inreal timeandwithacommontool.

4.1.2.Establishamulti-viewsystem  Hardwaresetup:

TwoprojectorsofDIGITALPROJECTIONDLP1TITANTM_1080p 3DandtwoshutterglassesofVolfoni1EDGETMareusedinthis multi-viewsystem.Eachprojectorcanrunat120Hz.So,oneuser withshutterglassescanreceivetwoviewswithstereoscopy,aswe describedinSection2.1.Hence,twoprojectorshavefourviews. Twodifferent representations for different users are displayed throughpolarizedfilters.Twolinearfiltersinasamepolarization directionareplacedbeforeapairofprojectorandshutterglasses separately.Twoexperimentalscenesaredevelopedbyusing3D gameengineandaredisplayedbytwoprojectorsseparately,as showninFig.5.

 Softwarefunction:

To realize an early collaborative design application, each collaborator hasan experimental scene with special functions.

AswedescribedinSection4.1.1,oneofthecollaboratorsisaUSER (takingtheroleofapersona),theotherisaDESIGNER(Table1). Different actions realized in this multi-view system for two collaboratorsarelistedinTable1.

The userof themugexperimentsits usageunder a specific context(insideacar)invirtualenvironment(Fig.6a):manipulate themug;putthemugintothecarmugholder;putthetea-baginto themug;takethedocumentwithoutoverthrowingthemug.

Meanwhile the designer of the mug optimizes the design parametersandanalysisdata(Fig.6b),e.g.size,thermalanalysis, etc.Ifthesizeofthemugischanged,bothuseranddesignerofthe mugwillsee.Bothhaveotherindividualelementsthatonlyeach one himself can understand. The user has the accessories accompanying the use of the mug. The designer has the tool panelofdesignthatonlyhe/shecansee.Ifuserputateabaginthe mug,he/shecandesignahandleforthemugtotietheteabag.User canaskdesignertoshowthetransparentmugtoverifytheteabag (Fig.6c).Arealtimedisplayingofthesizeofthemugisshownin

Fig.6d.Designercanuseavirtualwandinthescenariotoselect andcontrolobjects.Thereflectionsofobjectsonthewindshield canbesodisturbingforusersthathe/shecanaskthedesignerto followhis/heradvices(Fig.6a).

4.1.3.Evaluateusabilityandcollaboration

Theevaluationofusabilityandcollaborationofthesystem consiststwoparts.Firstly,ergonomicevaluationmethods,such

Fig.5.Multi-viewsystemof23Dviews(4views)usedintheexperiment. Fig.4. Proposedscientificapproach.

asobservation,interview,andopenquestions,areadoptedto participantsbyaprofessionalergonomist.Secondly, quantita-tiveanalysiswithaLikertscalequestionnaireisusedduringthe experiment. The variables adapted to our research question based onliterature [40,41] canbe measuredfrompoint 0 to point4:

 Involvement: the participants’ feeling of how well they participateintheiractivities.

 Learnability:thequalitythatasystem offersdealingwiththe levelofeaseoflearning.

 Satisfaction:user'sacceptanceoftheperformanceofthetoolsor system.

 Awareness:theabilitytocommunicatedirectly,usersofshared systemsshouldbeawareofeachother'spresenceandactivity.  Collaborative effort: how much work can two collaborators providetoaccomplishaspecifictaskinacollaborativecontext. Amongwhich,learnabilityandsatisfactionarecriteriaforthe usabilityoftheCHIsystem,whiletheperformanceofcollaboration concerns about the involvement, awareness and the effort providedduringacollaborativetask.

4.2.Collectingdata

The experiment requires simultaneously2 participants(one willplaytheroleoftheuserandtheotherwillplaytheroleof designer).Theyareaskedtoberecordedduringtheexperimentby camera.

10 participants are recruited for the experiment to form 5 partnerships.Participantsarestudentsfromaschoolofengineer. Theyareallmen,between19and23yearsold.Theyarenovicein productdesign.

(1) Introduction(5–10min):Theyarefirstlyintroducedwiththe object andprocedure of theexperiment. Thenin a training scenario,theyare askedtomanipulatevirtualobjects.They mustbetaughthowtodothefollowingmovements:catching anobject,movingit,droppingit,returningit,etc.Atraining taskistobetested severaltimes tomake thecollaborators familiarwithourequipment.Assoonastheparticipantsare ready,theexperimentstarts.

(2) Testphase(20min):Thetasksaskedthegroupofparticipants toachieveduringthis“productuser/designer”experimentare

Table1

Definitionoftheactionsofusers(personas)anddesignersinthescenarios.

USER'sactions DESIGNER'sactions

Reachableobjects

-Highlightobjectswhichcanbemoved(mug,teabag,folder,cupholder) -Highlighteffects(coffee,cupholder)

Synchronousobservation: -Analysisresult -Thermalanalysis

-Performanceanalysis(Teabag) Personasmotionsimulation

-Moveandrotatethemug -Pourwatertocoffeemug -Grabteabaganddropintomug -Grabfolderandopenit

Exteriordesign -Geometrydimension -Color

-Part/assemblydesign(muglid,handle) -Shapedesign(changemugform)

listed in Table 2. These 4 tasks are intended to represent situationsoftheuseofaproduct.

(3)Feedbackofexperiment(20min):Participantsareaskedtofill aquestionnaireandarecordedinterviewtogathertheirpoint ofviewsontheexperience.

4.3.Processing,analyzingandinterpretingdata

The main results of this case-study are highlighted in this chapter.From background investigation questions in the ques-tionnaire, we can _find out that 50% of the participants have experienceinprojectswithdesignactivities.Butonly20%ofthem havebeeninvolvedincollaborativedesign activities withother people.Thus,theopinionsandperformancesoftheseparticipants mayhavepracticalimplicationsinmulti-usersystemandtheway ofcollaborativedesign.

Fromtheresult,wecanexplorethatthisexperimentallowedus to identify elements about the co-activity between users and designersrelatedtotheuseofthemulti-viewsystemontheone hand,andelementsofimprovementsofthissystemontheother hand.

4.3.1.Ergonomicsanalysisofcollaborationactivities

Themainresultoftheanalyzingtheinterviewsrecordedare listedasbelow:

 Collaborationbetweenusersanddesigners:

From the familiarization phase, spontaneous exchanges be-tween users and designers quickly become noticeable. These exchangesmake possibletoagreementsonthemodificationsof themodelandcontinuethroughouttheexperiment.Throughthe observationoftheparticipantsduringtheexperiment,asFig.6,the usergivesdirectionstothedesigner“alittlebit”“it'sok”sothathe/ shecanmodifytheDMUtogettheexpectedresults.Userguidance allowsdesignerstomodifythesettingsoftheDMUasrequired,but designerscanalsoproposespontaneouschangestoarouseuser's interestinthingstheymightnothavethoughtof,suchascolor changesforexample“Ididn’tknowthatyoucouldchangethecolor, it'sgoodtobeabletodoit”.

Usersanddesignersraisetheinterestofthisdevicefordesign. Regardingtheirexperience,theyexpressdifferentpointofviews. Fortheusers,itisveryinterestingtobeabletocommunicate with the designer. Being involved in a design project is not common,anditisinterestingforthemtobeabletoexpresstheir opinion“speakingwiththedesignermakehimcloser”.Ontheother hand,seeinglivemodificationshelpsuserstoexpresstheirneeds, somethinggenerallydifficultforanon-expertinthedomain“it's veryinterestingtoseethe changesinlive,ithelpstoverbalizethe thoughts_”.Usersappreciatethe_fluidityoftheexperimentandthe

scopeofpossiblemodifications“it'sgreattoseetheobjectchanging inlive”.Ontheotherhand,thedeviceisalsoappreciatedforputting itselfinasituation,userscanprojectthemselvesinacontextofuse withtheproduct“ithelpstoproject”(Fig.7).

On thedesigners’ side,the contributionof the presenceof users is undeniable. This allows them to think differently and considerelementstheywouldnothavethoughtofalone,suchas theembarrassmentofthereflectionofthemugonthewindshield, dependingon thecolor “we seethings we didn’t necessarily see withouttheuser_”.Finally,designersappreciatethetimesavingthis tool allowstoreduceiterations.The presenceof usersand live modificationsisdeemedrelevanttosavetimeindesign“thistool gives the opportunity to reduce iterations and to collect user impressionsinlive”.

Finally,theplayfulaspectofmulti-viewsystemisappreciated and encourages both parties toparticipate, compared toa test situationwithoutdigitalsupport“it'snicetomanipulateanobject withvirtualreality!”

Usersanddesignersalsomentioneddifficultiesand improve-mentpointsforthedevice.

 Difficultiesandareasforimprovement

Ontheuserside,themaindifficultylayininteractionswiththe virtualmodel.Indeed,someactionswerehardlyfeasibleandnot veryfluid,likeputtingthemugonthecupholder“itfallswhenitis placedonit”.Trackinglatencywasalsomentioned.An embarrass-mentalsoresidedinthefieldofviewbecausetheobjectstobe caughtwerelocatedbehindtheuser'shand,whichdidnotallow himtoseethemcorrectly.

Fordesigners,themajordifficultywasthelackofinformation ontheviewofusers.Indeed,usershadaviewofthecarwithallits attributes,whilethedesignersonlysawthemug.Therefore,they couldnotassistusersinadjustingthemodificationstotheobject. Inaddition,designershaveofferedustoaddmoreparametersto vary,suchasmorechoiceofcolororshapeofthemug.

Table2

Experimentaltasks.

Task Productuser Designer

Task1:Putthemuginthecupholder Theuserispositionedinthevehicle.His/hermugis onthedashboard.He/shemustputitonthecup holder

Openthecupholder;manipulatethemugandtriestoputitonthe cupholder,butthesizeofthemugisnotsuitable

Thedesignerchangesthesizeofthemug immediatelywhentheuseraskshim.

Task2:Putteabag

Theusermustgrabtheteabagandputitinthemug

Leavethemugonthecupholderandmovetheteabag.Theusermay requirethemugtobetransparenttoseetheconditionoftheteabag.

Thedesignercanmakethemugtransparent tobetterseetheteabag.

Task3:Catchthefoldercupholder Theuserisaskedtograbthefolder

Theusermusttakethefolderfromthedashboard,buttheheight positionofthemugonthecupholdermayblockthemovement.

Thedesignerchangesthesizeofthemugso thatthefoldercanpassoverthecupholder. Task4:Colorchange Thecolorofthemugisreflectedonthewindshield Changethecolorofthemugsothatthe

reflectionstops

4.3.2.Quantitativeanalysisofcollaborationactivities

Themainresultsofrepresentativethemesofthequantitative questionnaireareshowninFig.8andarelistedbelow:

 Involvement

The questions about the impression of the experiment are asked. The results show positive feedbacks of the general impressions. An average score of “Excellent” and “Interesting” impressionsisupto3.25/4.FromthefrequencyanalysisinFig.9, over85%oftheparticipantsisbeyonda“fair”scoreininvolvement. Thus.Duringthisexperiment,participantsarehighlyinvolvedin thecollaborativework.

 Learnability

Intermsoftheeaseoflearningofthemulti-usersystem,aswell asthecollaborativescenario,ameanscoreof1.8/4(4meansvery hard learnability) is present. In the questionnaire, 20% of the participantsthinktheexperimenttaskisveryeasytocomplete. Half of the participants have some difficulties about the collaborativetaskortroubleswithhardware.

Fromananalysiscomparingusersanddesigners,usersappear toconsiderthecollaborationtaskeasytocomplete.Theyachieve anaveragescoreof1.4/4.However,designersusuallyhavetrouble withthelearnabilityofthescenariothattheygot2.2/4point.  Satisfaction

Thequestionsabouttheconceptofcollaborativeworkingstyles intheexperimentscenarioreceivehighmarks.Participantsseems

likethewaythemulti-user systemworks.A scoreof 3.75/4(4 meansverysatisfied) hasbeengivenfor thesatisfactionof the usabilityofthesystem.InFig.9,75%oftheparticipantsarevery satisfiedwiththecollaborationstyleduringthis kindofdesign activity, inwhich collaborators canexchange multi-disciplinary informationinrealtime.

 Awareness

For each participant, the awareness of the collaborator is questioned,aswellastheconfidenceofthecollaborator.During the experiment, participants show great awareness of their collaborators.Throughfrequentcommunications,theawareness ofcollaboratorreachesahighmarkof3.5/4(4meansknowingvery clearlyaboutone'scollaborator)andthescoreoftheconfidenceof one'scollaboratoris3.5/4(4meansveryconfident).

 Collaborativeeffort

The effort that collaborators provide to accomplish the experimenttaskduringthecommunicationbetweenthetwoof themisquestioned.Theresultshowsanaverageof3.25/4(0means having tough difficulties)for thecollaborationdifficulty. Partic-ipants do not meet terriblecommunication problemsand they havehighevaluationofthecontributionsoftheircollaborators. 5. Discussion

 Usabilityofmulti-userCHIsystem

Wewillfirstdiscusstheevaluationoftheusabilityof multi-view system. In both ergonomic interview (Section 4.3.1) and quantitativequestionnaire(Section4.3.2),participantsexpressed satisfaction on the usability of CHI system for the arranged collaborativetask.Theyfeltinterestedinthemulti-viewsystem whichpresentsmulti-representationofaproduct.Thescorefrom the questionnairealso has a very positive feedbackof theCHI system.

Fromthequantitativequestionnaire,wefindthecriterionon theusageofthesystem,learnability,isslightlybelowthemedian, whichmeansthatparticipantshaveproblemswiththeusageofthe system.ParticipantsdidnotthinkthisCHIsystemwaseasytolearn duetothetechnicalproblem.Fromtheinterviewresponses,there aremanyrecommendationsabouttechnicalimprovementsofthe trackingsystemandinteractionsystem.Thecomplexoperationof theinteractiondevicescouldbeatechnicalproblemwhichaffect theuserexperience.That'salsowhyalong-timetrainingprogram

Fig.8.Score meansof thevariables inthe quantitativequestionnaire ofthe experiment.

wassetuptokeeptheparticipantstobefamiliartoourdevices beforetheexperiments.

 Performanceofcollaboration

Thenwewilldiscusstheperformanceofcollaborationandits evaluationoftheexperimentaltask.Sinceourparticipantshave littleexperienceincollaborativedesignactivitieswithapartner, theyseemcuriousandinterestedinthemulti-viewsystemandthe multi-representationof aproduct.Fromtheinterviews, wecan conclude that participants had positive comments on the synchronization of virtual world changes in both view during design activities. One's modification on the product can be synchronizedandvisualizedinrealtimetotheotherparticipant. Thisreducesthetimeofcheckingmodificationsinanotheruser's view.Thetimesavingindesignprocesscanberealizedbetween theuseranddesigner. Comparedtotraditionaldesign activities, transferring,checking,andre-designingwillcauseawasteoftime. From quantitative questionnaires, the criteria of the perfor-manceofcollaboration:involvement,awareness,andcollaborative effort,allhavea scorethatis over3points(Fig.8).Duringthe collaborativetask, participantsfelt theirindependent work got involved in a collaboration. Synchronization on the multiple representationsofthesameobject,e.g.themodificationofthecup anditsparameters,allowsparticipantstohaveanawarenessofthe existenceandcontributionoftheircollaborators.Participantsalso feelthat the synchronization of multiple representationsof an objectleadstoareductionintheeffortthateachpersonneedsto provide.Theseresultsofevaluationofcollaborationrevealthatour conceptofintroducingmulti-viewsystemintoearlycollaborative designisafeasiblesolutionofpresentingmulti-representationofa product.Bothquantitative questionnaires and ergonomic inter-viewsdemonstratedthattheadvantageofco-locatedface-to-face communicationisobviousduringacollaborativetask.

 Personasmethodincollaboration

Personasmethodletsparticipantsactassomespecialusersofa product.Inthecollaborationwithadesigner,apersonateststhe productfromhis/herpointofview.Atthesametimeapersona allows the designer to express opinions according to both professionalknowledgeanduserexperience.Theroleplayingof twocharactersneedstopracticetherequirementof therolein trainingprogram.However,inthefuturedevelopment,awayto improvethesetupcouldbe:includingexpertdesignersintothe experimenttomakemorepersuasiveandauthoritativeevaluation. 6. Overallconclusionsandfuturework

Communication can be realized among designers through computer_–humaninteractiontechnology.Sometimeswhenexperts with different backgrounds want to work collaboratively, the computerhumaninterfacesusedbyexpertsbecomea barrierof transferring,checking,andmodifyingdesigndataamongexperts.

Multi-viewsystemisakindofinterfacethatcansupportseveral userstoworktogetherbutwithdifferentvisualizationsaswellas interactions.Theaimofthispaperistointegrateamulti-viewVR system into an early collaborative design case (Section 2). An evaluation is conducted, and the advantageof this multi-view systemisconcluded.

AresearchplanhasbeenproposedinSection3.Themechanisms andtechniquesto builda multi-viewsystem arediscussed.The relationshipamongmulti-view,multi-POVand multi-representa-tionisexplained.Theanalysisoftheneedofmulti-userinteractionin earlycollaborative design has beendiscovered, followed by the evaluationmethodofthiskindofcollaboration.

An experiment using Personas method has been conducted (Section 4). A multi-view system has been developed and integrated into a multidisciplinary mug design scenario. From the results of the experiments that we have conducted, the advantageofintroducingmulti-viewsystemintoearly collabora-tive design is positive.The interviews ofparticipantsalso have positive comments on the multi-viewcollaborative design and givesomeadvicesonthefurtherimprovementoftheprototyped multi-viewsystem.

In the future,more improvements onthecomputer–human interaction techniques in the multi-view system should be considered.Thenexpertdesignerswillbeinvitedtotestin the multi-viewexperimentsformoreauthoritativeevaluation.

Acknowledgements

Thank China Scholarship Council for its financial support (http://en.csc.edu.cn/).

References

[1]L.T.Blessing,A.Chakrabarti,DRM:ADesignResearchMethodology,Springer, 2009.

[2]A.Bernard,A.Fischer,Newtrendsinrapidproductdevelopment,CIRPAnn. Manuf.Technol.51(2)(2002)635–652.

[3]P.Fillatreau,J.-Y.Fourquet,R.LeBolloc’H,S.Cailhol,A.Datas,B.Puel,Using virtualrealityand3Dindustrialnumericalmodelsforimmersiveinteractive checklists,Comput.Ind.64(9)(2013)1253–1262.

[4]R.P.Smith,S.D.Eppinger,Decidingbetweensequentialandconcurrenttasksin engineeringdesign,Concurr.Eng.6(1)(1998)15–25.

[5]F.Segonds,F.Mantelet,J.Nelson,S.Gaillard,PropositionofaPLMtoolto supporttextiledesign:acasestudyappliedtothedefinitionoftheearlystages ofdesignrequirements,Comput.Ind.66(2015)21–30.

[6]G.Acosta,K.Morales,D. Lagos,M.Ortiz, Addressinghumanfactorsand ergonomicsindesignprocess,productlifecycle,andinnovation:trendsin consumerproductdesign,HumanFactorsandErgonomicsinConsumer ProductDesign:MethodsandTechniques,(2011),pp.133–154.

[7]A.Liem,E.Brangier,Innovationanddesignapproacheswithinprospective ergonomics,Work41(Suppl.1)(2012)5243–5250.

[8]J.-M.Robert,É.Brangier,Prospectiveergonomics:origin,goal,andprospects, Work41(Suppl.1)(2012)5235–5242.

[9]J.Chung,K.Lee,Aframeworkofcollaborativedesignenvironmentforinjection molding,Comput.Ind.47(3)(2002)319–337.

[10]H. Kan, V.G. Duffy, C.-J. Su, An Internet virtual reality collaborative environmentforeffectiveproductdesign,Comput.Ind.45(2)(2001)197–213. [11]S. Zhang, W. Shen, H. Ghenniwa, A review of Internet-based product

informationsharingandvisualization,Comput.Ind.54(1)(2004)1–15. [12]W. Matusik, C. Forlines, H. Pfister, Multiview user interfaces with an

automultiscopicdisplay,ProceedingsoftheWorkingConferenceon AdvancedVisualInterfaces,ACM,2008,pp.363–366.

[13]X. Wu, G. Zhai, Temporal psychovisualmodulation: a newparadigm of informationdisplay[exploratoryDSP],IEEESignalProcess.Mag.30(1)(2013) 136–141.

[14]T.Balogh,T.Forgács,T.Agács,O.Balet,E.Bouvier,F.Bettio,E.Gobbetti,G. Zanetti,Ascalablehardwareandsoftwaresystemfortheholographicdisplay ofinteractivegraphicsapplications,Eurographics(ShortPresentations)(2005) 109–112.

[15]P.Martin,P.Bourdot,D.Touraine,Areconfigurablearchitectureformultimodal andcollaborativeinteractionsinvirtualenvironments,2011IEEESymposium on3DUserInterfaces(3DUI),IEEE,2011,pp.11–14.

[16]P.Zimmermann,Virtualrealityaideddesign:asurveyoftheuseofVRin automotiveindustry,Prod.Eng.(2008)277–296.

[17]W.Fujimura,Y.Koide,R.Songer,T.Hayakawa,A.Shirai,K.Yanaka,2x3D:real timeshaderforsimultaneous2D/3Dhybridtheater,SIGGRAPHAsia2012 EmergingTechnologies,ACM,2012,pp.1.

[18]R.Lissermann,J. Huber,M.Schmitz,J.Steimle,M.Mühlhäuser,Permulin: mixed-focuscollaborationonmulti-viewtabletops,Proceedingsofthe32nd AnnualACMconferenceonHumanFactorsinComputingSystems,ACM,2014, pp.3191–3200.

[19]K.Nagano,T.Utsugi,K.Yanaka,A.Shirai,M.Nakajima,ScritterHDR: multiplex-hiddenimagingonhighdynamicrangeprojection,SIGGRAPHAsia2011 Posters,ACM,2011,pp.52.

[20]P.Mistry,Thirdeye:atechniquethatenablesmultipleviewerstoseedifferent contentonasingledisplayscreen,ACMSIGGRAPHASIA2009Posters,ACM, 2009,pp.29.

[21]S.Kim,X.Cao,H.Zhang,D.Tan,Enablingconcurrentdualviewsoncommon LCDscreens,ProceedingsoftheSIGCHIConferenceonHumanFactorsin ComputingSystems,ACM,2012,pp.2175–2184.

[22]A.Kulik,A.Kunert,S.Beck,R.Reichel,R.Blach,A.Zink,B.Froehlich,C1x6:a stereoscopicsix-userdisplayforco-locatedcollaborationinsharedvirtual environments,ACMTrans.Graph.30(6)(2011)188.

[23]S. Sharifi, K. Pawar, Product development strategies for agility, Agile Manufacturing:The21stCenturyCompetitiveStrategy,(2001),pp.175–192. [24]A.Cooper, etal.,TheInmatesare RunningTheAsylum:WhyHigh-Tech

ProductsDriveUsCrazyandHowtoRestoretheSanity,vol.261,Sams, Indianapolis,2004.

[25]F.Long,Realorimaginary:theeffectivenessofusingpersonasinproduct design,ProceedingsoftheIrishErgonomicsSocietyAnnualConference,Vol. 14,IrishErgonomicsSocietyDublin,2009,pp.1–10.

[26]E.Brangier,C. Bornet, Amethodto producerepresentationsfocused on consumers’needs,Persona:HumanFactorsandErgonomicsinConsumer ProductDesign:MethodsandTechniques,TaylorandFrancis,2011,pp.37–61. [27] R.Barlow-Busch,Case11:acasestudyinpersonas,User-centeredDesign Stories:Real-WorldUCDCaseStudies,ElsevierScience,2010,pp.209–240. [28]Y.-N.Chang,Y.-K.Lim,E.Stolterman,Personas:from theorytopractices,

Proceedingsofthe5thNordicConferenceonHuman–ComputerInteraction: BuildingBridges,ACM,2008,pp.439–442.

[29]T.Matthews,T.Judge,S.Whittaker,Howdodesignersanduserexperience professionalsactuallyperceiveandusepersonas?ProceedingsoftheSIGCHI ConferenceonHumanFactorsinComputingSystems,ACM,2012,pp.1219– 1228.

[30]J.Pruitt,J.Grudin,Personas:practiceandtheory,Proceedingsofthe2003 ConferenceonDesigningforUserExperiences,ACM,2003,pp.1–15. [31] K.Goodwin,Perfectingyourpersonas,CooperInteractionDesignNewsletter

(2001)295–313.

[32]K.Goodwin,Gettingfromresearchtopersonas:harnessingthepowerofdata, CooperNewslett.(2002).

[33]J.Pruitt,T.Adlin,ThePersonaLifecycle:KeepingPeopleinMindThroughout ProductDesign,MorganKaufmann,2010.

[34]J.Spool,Threeimportantbenefitsofpersonas,UserInterfaceEng.(2007). [35]C.N.Chapman,R.P.Milham,Thepersonas’newclothes:methodologicaland

practicalargumentsagainstapopularmethod,ProceedingsoftheHuman FactorsandErgonomicsSocietyAnnualMeeting,Vol.50,SAGEPublications Sage,CA/LosAngeles,CA,2006,pp.634–636.

[36]G.Seland,Systemdesignerassessmentsofroleplayasadesignmethod:a qualitativestudy,Proceedingsofthe4thNordicConferenceonHuman– ComputerInteraction:ChangingRoles,ACM,2006,pp.222–231.

[37]B.Raijmakers,M.Thompson,E.vandeGarde-Perik,Newgoalsfordesign,new rolesfordesigners,ProceedingsofCumulusHelsinki2012Conference(2012) 1–11.

[38]D.W.Seo,H.Kim,J.S.Kim,J.Y.Lee,Hybridreality-baseduserexperienceand evaluationofacontext-awaresmarthome,Comput.Ind.76(2016)11–23. [39]S.Buisine,J.Guegan,J.Barré,F.Segonds,A.Aoussat,Usingavatarstotailor

ideationprocesstoinnovationstrategy,Cognit.Technol.Work18(3)(2016) 583–594.

[40]M.Gerhard,D.J.Moore,D.J.Hobbs, Continuouspresencein collaborative virtualenvironments:towardsahybridavatar-agentmodelforuser representation,InternationalWorkshoponIntelligentVirtualAgents, Springer,2001,pp.137–155.

[41]H.Hrimech,F.Merienne,Interactionandevaluationtoolsforcollaborative virtualenvironment,Int.J.Interact.Des.Manuf.4(3)(2010)149–156.