A METHODOLOGY USING COMPUTERTOOLS IN THE PRODUCT DEVELOPMENT

PROCESS

Ulf Lor entzon�

and Gunnar Bolmsjö�

�LundUniversity, Dept.MechanicalEngineering,

Divisionof Robotics,P.O.Box 118,SE-22100 Lund,Swedenhttp://www.robotics.lu.se,email: ulorentzon@robotics.lu.se

�LundUniversity, Dept.MechanicalEngineering,

Divisionof Robotics,P.O.Box 118,SE-22100 Lund,Swedenhttp://www.robotics.lu.se,email: gbolmsjo@robotics.lu.se

AbstractThispaperpresentsa descriptionof a methodological framework for de-signandmanufacturingof prototypesusingcomputertoolsasanaid dur-ing the entire designprocess.An Importantaspectin this methodologyis informationintegration, not only betweencomputertoolsbut primar-ily betweenhumanresources,their differentexpertiseareas,permanentandtemporary experts,etc. In this context a holistic approach is of vitalimportancethatput thehumanresourcesin centerof thedevelopingpro-cesswhile maintainingintegratedtools for specificanalysisduring thedesignprocess.Thedescribeddevelopmentmethodology, HumanCenteredVirtual Pro-totyping, hasprovedto workwell in a designgroupwhereengineersandotherpeople, with differentfieldsof expertise, areworkingtogether.

1 Intr oduction

Oneof themostimportantfeaturesfor adesignmethodologyis thatit providesawork-ing environmentthat supportsinnovative designsolutionsand is flexible enoughtoprovide thenecessarymeansto positively enhancethegroupdynamicsduringthede-velopmentanddesignprocess.To quoteMichaelSchrage(Schrage,2000)“The mostimportantraw materialof innovationhasalwaysbeentheinterplaybetweenindividualsandtheexpressionof their ideas”,which is a valid point. Themethodologydescribedin thispaper, HumanCenteredVirtual PrototypingHCVP, is basedontheextensiveuseof computertoolsduringthedesignprocessandhasbeenusedin severalcases,whicharedescribedin thepaper, with goodresults.

Traditionally it hasbeencommonpracticethat the designprocessis carriedoutby engineerswith a high degreeof specialization,but with computertools andnewmethodologiesit is possibleto put thehumanresourcesin thecenterof theprocessandbettertakeadvantageof thepositive interactionbetweenthepeopleinvolved.

2 Designmethodology

Thedesignmethodologydescribedin this paperhasbeenevolvedfrom methodssuchasthosedescribedby (Dai, 1995)(Dai andGöbel,1994). By putting the humanre-sourcesat the centerof the project, this methodologybetter take advantageof thehumanresources.The methodologyis focusedon creatinga working environmentthatsupportthegenerationof innovativedesignsolutionsandprovideapositivegroupdynamics.

CAD 3D models

Simulation& AnalysisCAM

NC prog. Simulationresults

Human Resources

Robot sim.

Visualisation

FEM analysis

Controller test

Kinematics test

Dynamics

User feedback

Figure1: A depictionof the“HumanCenteredVirtual PrototypingHCVP” methodol-ogy.

Normally a designprojectstartswith definition of demandsandrequirementsfortheproduct,thesearethentransformedto technicallymeasurableitemsandform aba-sisfor thegenerationof differentsolutions.Thereareseveralwell documentedmethod-ologiesfor waysto performthedesignprocessandgeneratesolutions,astheaxiomaticapproachto design(Suh,1990)andotherapproaches,mostof which decomposetheproblemsto smallseparateunits,but thesemethodsarenotdealtwith in thispaper. TheHCVP methodologytakesa moreholistic view of the designprocessbut someareasmaystill needto bedecomposedandstudiedin detail to obtainthedesiredresult.

As seenin figure 1 the humanresourcesarecentrallyplaced,surroundedby thecomputertools, which shouldreflect the actualsituation. The humanresourcesrep-resenta group of peoplethat are involved during sometime of the designprocess.This group is not a fixed unit but may changeover time as the projectdevelop andthe needfor certainform of expertiseis required. Typically thereis a fixed coreof

peopleworking throughoutthe whole project, to supplycontinuity andmakingsurethattheprojectnot losemomentum,likethemechanicalandelectricaldesigners,whileothermayenterandexit thegroupduringsometime in theprojectandmake valuablecontributions.Thehumanresourcesmaybespecialistsin solidmodeling1, simulation,NC-processing,andspecialistsin areassuchas: marketing; ergonomics;aesteticde-sign; thatall cancontributewith valuableinput during sometime of the project. Allinvolved are working together, at somepoint of the project, and by using the solidmodelsin the simulationsoftware2, performingiterationsof refining the modelsandperformingthesimulations.

Oncethegrouphasgenerateda reasonablenumberof differentconceptualdesignsthe “HCVP” maybegin. Evenat thestageof generatingconceptualdesignsolutions,theuseof computertools canbeof greatvalue,sincesimulationandvisualizationofsimplified conceptsmay identify someproblemareasandgive new ideas. The firststepis to create3-D modelsof themajorcomponents.In anearlystagethesemaybedesignedof verysimplegeometriesin orderto quickly getstartedwith thesimulationsto verify themechanisms.Themodelingandsimulationshave a circulardependencysincethe modelsare simulatedand, using the datafrom the simulations,graduallyimproved.

The simulationscanhave several differentpurposes:simplegeometrycheck;vi-sualizationsthat form a basefor userevaluation;virtual testingof kinematicsmodelsandcontrollerparameters;generationof off-line robot programs;FEM-analysis;justto mentiona few possibilities.Thegoodvisual interfacethatcanbeobtainedby visu-alizationsmakesit easyto includenew peopleto thedesigngroupduring the projectandquickly let the new input be madevisible in the simulations.Theuseof simula-tions/visualizationswill make it easierto “cultivateknowledgeablenetworks, to seeksolutionsfrom disciplinesnot representedin yourcompany” (Davis andMeyer, 1999).

Oncethedesignprocesshasreachedasatisfactorylevel of completionit is possibleto usethemodelsto generateCAM-datathatcanbedown loadedto aNC-machineandmanufacturethecomponent.

This designcycle is representedin figure 2, wherethe differentgenerationsof acomponentaredisplayed.Thecomponentusedto describethiscycle is a“finger” usedin the packageproject. This finger is a part of the baseplateof the unit andhasaspurposeto centerandgraspthe bottomendof a consumerpackagewith screw top.Oncethepackageis graspeda “Top unit” movesdown to thetop of thepackageand,with a principlesimilar to theoneusedin thebaseplate,graspsandunscrewsthecap.

In frame“A” the first simplefinger geometriesaredisplayed.The fingersrotatesarounda shaftandthe shapeof the gripping surfaceis only a circular surfaceoff-setrelative the shaft. Thesefingershave beentestedusingthe simpletestbeddepictedin frame“B” andthe main concernin this casewasthe geometricalinterferenceandpossiblegrip width. Somesimulationiterationslater it becameobviousthattheshapeof the gripping surfacewas important in order to obtain a good gripping force andwide gripping range.Onesolutionto this problemwasto usea specialcurve for thegripping surface. In frame“C” the finger is equippedwith a segmentof a “tractrix”curve which provedto be a goodapproximationfor the surface. The finalizedfingerdesignis displayedin frame“D”, in which friction materialis included.Thefinalized3-D modelof thefingerformeda basefor thegenerationof NC-milling programming,

1PRO/Engineer(PTC)wasusedasthemainsolidmodelingtool andto generateNC-datafor themachin-ing of thecomponents.

2IGRIP/Envision/UltraArc(Delmia)wasusedto performall simulations.

A B C

D E F

Figure2: Thegenerationsof a component.A: earlymodels,B: simplesimulationtestbed,C: a moredevelopedcomponent,D: the 3-D CAD model,E: generationof NCmachiningsequence,F: photoof thefinishedcomponent.

which is showed in frame“E”. Finally the completefinger is displayedin a photoinframe“F”.

Thesimulationsmakesit possibleto, at anearlystagein thedesignprocess,virtu-ally testandvalidatesomeof theimportanttechnicalissues.Theuseof NC processingandmachiningmakesit possibleto comparevirtual andphysicalmodelsandto rapidlybuild componentswhich canbeusedto validatethesimulationresultsandmake moreexactfuturesimulationspossible.

TheHCVPmayalsobeputin awidercontext asadetailedpartof theVSOPmodel.TheVSOPmodelvisible in figure4 is anapproachthatis morefocusedonproductandprocessdevelopmentin themanufacturingindustry.

3 Casestudies

Themethodologydescribedin thelastsectionwasusedin severalprojectsandin thissectioneachcasewill representandexemplify someaspectsof thesimulations.In thefirst casedetailedanalysisof aroboticdeviceandsomegeneralrobotdesignissuesaredescribed.Thesecondcaseexemplifiesthemethodologyusedin amoreindustrialper-spective, theVSOPmodel,with focuson productionplanningandtestingof differentsimulationtypesandtheir industrialapplicability. Thethird andfourth casewill showhow the methodologycanbe usedin the designprocessof a specificcomponentandalsoput focuson theuseof realisticvisualizations.

3.1 The ASIMOV Wheelchair manipulator

Theprojectwasinitiatedby “AMS3” andLundUniversityto find amoreflexible alter-native to the existing wheelchairbasedmanipulators.The manipulator(U.Lorentzon

3TheSwedishNationalLaborMarket Board

andM.Olsson,1997)wasdesignedwith amodularapproach,whereeachjoint is asep-arateunit, includingmotor, transmissionandcontroller. Themodularapproach,with aflexible configuration,wasimportantsinceit wouldmakeusercustomizationpossible.

Simulationswere usedto: test several different robot configurationsand studythe work space,functionality andinterferencewith the wheelchair;analyzethe jointtorquesandinterlink forcesin orderto betterchoosemotorsandtransmission;workasa motionplannerfor thetestingof theprototype(Olsson,HedenbornandBolmsjö,1997).

Figure3: A dynamicanalysisof a roboticdevice usedto selectmotorsandreducers.To theright a torquehistogramfor onejoint is displayed.

Sinceit is of greatimportancethat themanipulatorhasa low weight,not to causeunbalancein the wheelchair, a solution had to be found with low massthat wouldhave sufficient performance.After having madethe modelingof the parts,the masspropertiesof thesewereknown andusedasin-datafor thesimulations.Thesimulationperformedanumberof differentmovementsthatshouldrepresenta3 minutesworkingcycle with the manipulatorcarrying the load of 1 kg. The result of this simulationis seenin figure 3 including a torquehistogramfor oneof the joints. Furthermore,thesimulatorwasusedto, in real-time,generateandperformthemotioncontrolof thephysicalrobotduringtechnicalvalidationandtesting.Thesimulationsmadeit possibleto designandbuild themanipulator, usinga sevenaxisconfigurationwith a lengthof1,2m, andobtainapayload/weightratioof 1/12.

3.2 The VSOP project

Themaingoalsfor theprojectwasto find new techniquesto speedup thedesignpro-cessandshortenthe time for productionintroductionof a product,develop workingmethodsandfinding waysto introducethesetheindustrialapplications.Themain in-dustrialfocusgroupin thisprojectwerethemediumsizecompanieswhichnotalreadyare using the simulationtechnologyto any significantextent but with the recoursesto managethesetypeof projects.Theprojectwasplannedto demonstrateandintro-ducesimulationsmethods,thatmake it possibleto usethesimulationtechnologyin aproductive way to the industrialpartners.In short, it shoulddemonstratehow to re-ducethedevelopmentprocesstimeof new productwhile maintainingor improving thequality of theresultingproduct.During this work, a methodologywasdevelopedovertime,adaptedfor theinvolvedcompany, basedon theVSOP-modelasshown in figure

4. This projecthasbeenrunningover four years(Bolmsjö, LorentzonandRandell,1999).

MarketNeed

RequirementSpecification

ConceptualModels

PrototypeModels

ProductModels

Manufacturing(Do it)

ManufacturingProcesses

PreliminaryProcess Plan

Pro

cess

Inve

nto

ryP

relim

inar

yP

rep

arat

ion

sF

inal

Pre

par

atio

ns

Man

ufa

ctu

rin

g D

evel

op

men

t P

roce

ss (

Ho

w?

)

Pro

du

ct D

evel

op

men

t P

roce

ss (

Wh

at?

)

Co

nce

ptu

alD

esig

nP

relim

inar

yD

esig

nD

etai

led

Des

ign

➯

➯

➯

Feed−back

➯

➯

Process Plans,Operation Programs,Jigs, Toolings, etc.

Market

➯

Figure4: TheVSOPapproachto development.

Themainindustrialpartnerin thisprojectwasBT-industries,manufacturerof fork-lift trucksandothersimilar productsfor handlinggoods.BT wasat thetime whentheprojectstartedfocusingon two basicissues:how to increasethequalityof theproductdevelopmentprocesswhile at thesametime reducethetime to market to half or less,andhow to reducecostsduringthefirst ramp-upphaseof manufacturingnew products.As indicatedin figure4, themodelindicatea concurrentworking schemethroughouttheproductdevelopmentprocessintegratingproductandmanufacturingdevelopmentprocessinto oneconcurrentoperationvia information links. The project focusedontheselinks thatessentiallyprovidedmeansto connectdifferenthumanresourcesduringaprojectasneeded,with theright informationat theright timethatcouldbepresentedandunderstood.

As mentionedabove,oneimportanttool in theprojectwassimulationtools to fa-cilitate muchfasteriterationloopsthannormallyusedin similar projects.This provedto be extremely importantas the developmentof a productinvolvescomplex issuesrelatedto aesteticdesign,manufacturingprocesses,ergonomicsrelatedto manualas-

sembly, userissues,productivity andquality. Thus,humanresourcesduringtheprojectwasnot thesameindividualsthewhole time. Instead,differentpersonsrangingfromprocessexpertsto executivemanagerswasinvolvedin theprojectashumanresourcesasneeded,andthe needfor providing the right kind of informationat the right timebecameanimportantissuesto bring resultsforwardin theproject.

The simulationtools usedin the projectaddressedtwo issues:the dynamicallo-cation of resourcesin the shopfloor by using a discreteevent simulationtool, andthe specificprocessessuchasusability of the productor robot arc-weldingby usinga time continuousrobotsimulationtool. For discreteeventsimulation,developmentswasmadeto integrateinformationbetweenmaterialresourcesplanningtools to thesimulationandconcurrentdevelopmentof factoryandsimulationmodelsfor reductionof thedevelopmenttime. Usabilitysimulationswasmadefor earlyvirtual experimentsconcerningvariousideasof theproductat theconceptualstage.Whenthedesign,stillin adigital form, becamemorespecific,simulationswasintroducedfor manualassem-bly of critical partsof the productaswell asrobotic arc welding. During this phase,aspectsrelatedto ergonomicscouldbe dealtwith beforethe productexistedand,forroboticwelding,designsolutionsthat introducedmajorproblemsin theweldingpro-cessand/orweldjig designcouldbehandleswithin thevirtual environmentbeforetheyoccurredin therealworld.

Experiencefrom theprojectconfirmedour assumptionthat realworld projectsin-volve humanresourceswith individualsthat take part in differentwaysin theproject,have different level of expertiseand responsibilitiesand so on. Tools that providesolutionson bridging the informationgapis in this context extremelyimportantandfacilitatedreal time experimentsat work shopswith expertsfrom differentfieldsandproducingresultsthat otherwisewould be impossibleto obtaingiven the sametimeandresources.

3.3 The PACKAGE project

ThePackage4 projectis anongoingprojectandtheaim of theprojectis to make con-sumerpackagingmoreaccessibleto disabledandolderusers.Theproductin focusforthis paperis called“Magic-Hand”andis aproductthatwill openpackageswith screwtops. The productis a true mechatronicdevice consistingof mechanicalparts,servocontrolledmotors,micro controllerandsoftware. The first setof prototypesarede-signedto demonstratethemechanicalfunctionalityandbesubjectedto usertests.The“Magic-Hand” will be equippedwith someadditionalfeaturesto make it possibletoscanthebarcodesof thepackagesandmatchthiscodeagainstadatabaseof products’global tradeitem numbersandprovide theuserwith informationaboutthecontentofthepackage.

HCVP wasusedduringthedesignof themechanicalcomponents,astheexamplewith the finger in figure 2. During this productdevelopmentprocesssimulationsisa valuabletool to createuserscenariossuchas the onedisplayedin figure 5. Suchscenarioswasutilizedbothwithin thedesigngroupandausergroupto study, visualizeandanalyzeissuesandsolutionsbeforethey actuallyexist asphysicalentities.

Thecurrentstatusof theprojectis thatasetof prototypehavebeenbuilt andwill besubjectedto usertestsconcerningfunctionality, physicaldesign,accessibilityin homeenvironment,hygienics,aesteticsandsoon.

After thefirst userteststhesetof prototypeswill beconvertedinto barcodereading

4fundedby theEuropeanCommission’s “Information SocietyTechnologies(IST) Programme”

Figure5: A pictureshowing the“Magic-Hand” in asimulationscenario.

units with enhancedfunctionality andsubjectedto a new seriesof usertestsforminga basefor future improvements.Thedesignis currentlyin a prototypephaseandwillhave to bemodifiedto beableto beproducedin largeseriesat a reasonableprice.

3.4 The MATS project

The MATS project5 is alsoan ongoingprojectbut in an earlierstagethanthe Pack-ageproject. Accordingto theprojectcontract“the mainobjective of theprojectis todesign,develop,manufactureandevaluateprototypeversionsof aninnovativemobilemechatronicassistitive technologysystemthatwill providedramaticimprovementsinthe prospectsfor employment,the quality of life andthe potentialfor integrationofdisabledandelderlypeopleinto aninclusive society”. A flexible robotarm(Topping,2001)thatmovesbetweendockingstations,canbeattachedto a wheelchair, andwillwork in a SMART homeenvironmentto assistthe disableduserin the activities ofdaily living.

At this stage,in theproject,theHCVP methodhasbeenusedto studyendeffectorsolutionsandrobotconfiguration.A processasdescribedin figure2 will alsobemadeon somecritical componentsin this project.

Therehasalsobeenmadeseveral simulationscenariosin domesticenvironment,suchastheonesin figure6. Thesesimulationsareusedin collaborationwith theotherpartnersof thisprojectandbedisplayedfor a groupof peoplethatmaybefutureusersof theproductandbeevaluatedto give theprojectpartnersfeedback.

4 Results

Experienceof thecaseshaveclearlyshown thatresultsfrom thesimulations,likeveri-ficationsof particularmechanicalproblemsandvisualizations,will beavaluableinput

5fundedby theEuropeanCommission’s “Information SocietyTechnologies(IST) Programme”

Figure6: Two virtual scenarioswhich will form abasefor userevaluations.

to thedesigngroupandgenerateapositiveworkingenvironment.It is hereworthwhileto notethatthe“designgroup” includesindividualswith avarietyof expertise,suchasthoseinvolvedin issuesrelatedto userrequirements,market, electronics,mechanics,manufacturingandso on. The shorttime betweensolid modelingandsimulationre-sultswhichmakesit possibleto verify andtestdifferentsolutionsalsohelpsthedesigngroupto reducetheoveralldevelopmenttime,usingaminimumof physicalprototypes.

Therearesomeclearadvantagesof a simulationbaseddesignprocess:shorterde-velopmenttime; betterquality of design;lessneedfor physicalprototypes;possibilityto adaptthedesignfor manufacturingeasierandto includetheuserperspective to theproject. The HCVP methodologyin combinationwith the right typesof simulationscangivegoodresultsasthecasesshow.

5 Discussion

Oneissuethat is a matterfor discussionis thequestionof how muchthemodelscanbesimplifiedwithout the lossof importantinformation. Theexperiencegainedin thedescribedcaseshaveshown thatat theearlystages,in thedesignprocess,ratherdrasticsimplificationscanbemadeto themodelsaslong asthevital parametersareincluded.But whentheprojectgoestowardthefinal designandpreparationfor manufacturingasmuchdetailsaspossibleshouldbeincludedin orderto avoid mechanicalinterferenceandsimilar problems.Thetime it takesto do a moredetailedmodelingusuallyis wellspent.

6 Conclusion

Theextensiveuseof computertoolscannotreplacetheold fashionwayof brainstorm-ing andpapersketchesto generateideas,but will beveryusefulin orderto validateandvirtually testtheideas.TheHCVPmethodologyhasprovedto work well in actualde-signprojects.

References

Bolmsjö, G., Lorentzon,U. andRandell,L. (1999). Integrationof developmentofproductandproductionsystemusingadvancedsimulationtools,in J.-P. K. H. vanBrusselandB. Lauwers(eds),Proceedingsof the32ndCIRPInternationalSem-inar on ManufacturingSystems(New SupportingTools for DesigningProductsandProductionSystems), Leuven,Belgium,pp.389–398.

Dai, F. (1995). On the role of simulationand visualizationin virtual prototyping,Prodceedingsof theInt. WorkshoponVirtual RealityandScientificVisualization,Hangzhou,China,pp.23–26.

Dai, F. andGöbel,M. (1994). Virual prototyping- an approachusingvr-techniques,Proc.of the14thASMEInt. Computers in EngineeringConference, Minneapolis,Minnesota.

Davis, S.andMeyer, C. (1999).Blur, WarnerBooks,chapter5.

Olsson, M., Hedenborn,P. and Bolmsjö, G. (1997). Eight axis lightweight ser-vice manipulator:Simulationbasedvalidationof controllersoftwarein robotics,Proceedingsof World Congresson ManufacturingTechnology Towards 2000,pp.541–550.

Schrage,M. (2000).SeriousPlay, HarvardbusinessSchoolPress,chapter1.

Suh,N. P. (1990).ThePrinciplesof Design, Oxford UniversityPress.

Topping,M. (2001). Flexibot - a multi-functionalgeneralpurposeservicerobot, In-dustrialRobot:An InternationalJournal28(5): 395–400.

U.LorentzonandM.Olsson(1997).Eightaxislightweightservicemanipulator:Virtualprototypingof mechanicalstructureandcontrollersoftware,in J.G.LoughranandS.A.Domanti(eds),Proceedingsof theWorld CongressManufacturingTechnol-ogyTowards2000, Cairns,North Queensland,Australia.