Indrum Qualica en PDF

STELIAN BRAD
EMILIA BRAD BOGDAN MOCAN

MIRCEA FULEA
TOOLS AND METHODS OF

COMPETITIVE DESIGN IN ROBOTICS
CONTENTS
Preface.......................................................................................................5
Workshop 1 Ranking Requirements in Competitive Robot Design Using the AHP Method.....7
1.1. Objective..............................................................................................9
1.2. Background..........................................................................................9
1.3. Task Description and Methodology..........................................................13
1.4. Example 1...........................................................................................17
1.5. Example 2...........................................................................................21
1.6. Workfow.............................................................................................26
1.7. Homework...........................................................................................26
Workshop 2 Computing the Consistency Ratio within the AHP Method.............................27
2.1. Objective............................................................................................29
2.3. Workfow.............................................................................................37
2.4. Homework...........................................................................................37
Workshop 3 Planning Engineering Specifcations in Competitive Robot Design Using the
QFD Method..............................................................................................................39
3.1. Objective............................................................................................41
3.2. Background.........................................................................................41
3.4. Example 1...........................................................................................56
3.5. Example 2...........................................................................................59
3.6. Workfow.............................................................................................62
3.7. Homework...........................................................................................63
Workshop 4 Functional Analysis in Competitive Robot Design using the FAST and FBD
Methods....................................................................................................................65
4.1. Objective............................................................................................67
4.2. Background.........................................................................................67
4.4. Example 1...........................................................................................76
4.5. Example 2...........................................................................................80
4.6. Workfow.............................................................................................83
4.7. Homework...........................................................................................84
Workshop 5 Idea Generation in Competitive Robot Design using the 6-3-5 and Mind-Map
Methods....................................................................................................................85
5.1. Objective............................................................................................87
5.2. Background.........................................................................................87
5.4. Example 1...........................................................................................98
5.5. Example 2...........................................................................................98
5.6. Workfow.............................................................................................98
5.7. Homework...........................................................................................98
Workshop 6 Variant Selection in Competitive Robot Design using the Pugh and Combinex
Methods....................................................................................................................99
6.1. Objective...........................................................................................101
6.2. Background.......................................................................................101
6.3. Task Description and Methodology........................................................109
6.4. Workfow...........................................................................................117
6.5. Homework.........................................................................................117
Workshop 7 Causality Analysis in Competitive Robot Design using the AIDA Method.......119
7.1. Objective...........................................................................................121
7.2. Background.......................................................................................121
7.4. Example 1..........................................................................................128
7.5. Example 2..........................................................................................133
7.6. Workfow...........................................................................................138
7.7. Homework.........................................................................................138
Workshop 8 Innovative Problem Solving in Competitive Robot Design using the TRIZ and
ARIZ Methods..........................................................................................................139
8.1. Objective...........................................................................................141
8.2. Background.......................................................................................141
8.4. Example 1..........................................................................................174
8.5. Example 2..........................................................................................176
8.6. Example 3..........................................................................................178
8.7. Workfow...........................................................................................182
8.8. Homework.........................................................................................182
REFERENCES...........................................................................................................183
Preface
Robotsarecomplexindustrialproducts,beingalsosituatedinthecategoryof
themostsophisticatedmechatronicequipment.Sincetheirimplementationin
practice in the 1960s, robots encountered several stages in developing their
subsystemsandcomponents.Theglobalcompetitivemarkets,aswellasthe
relativenewandextensibleareasofrobotapplicationrequiresignificantvalue
addedinthedesignofrobotsolutions.Thecomplexityoftechnicalaspectsthat
shouldbeconsideredduringrobotdevelopmentprocessisharderandharder
manageablewiththeclassicalengineeringapproaches.Necessitytobringthe
sectorofroboticsatacomparablelevel,bothtechnicallyandeconomically,as
some of the world key industries (i.e. car industry) is already revealed by
severalindicatorsontothemarket.Thisrequiresnewwaysofapproachingthe
robot design, considering tools and methods that some other strong
competitiveindustrieshavealreadytakenintoaccountseveralyearsago.
For a long period of time, robot design wasbased on classical approaches,
characterizedbyintuitioningeneratingsolutionsandorientedonfollowing
relatively rigid algorithms in modeling and calculating the robot structure.
ModernCAD/CAEtoolssupportedengineersinincreasingtheefficiencyand
inoptimizingsomeoftheconstructiveelements(seeoptimizationoftheshape,
material, etc.) during the design process, but a significant weakness was
revealed in the area of competitive concept formulation, planning and
development.
Thisworkbookincludesasetofsevenworkshopspreparedforstudentswhich
attendcoursesinthefieldofroboticsandwhichareinterestedtoenrichtheir
knowledge with some basic tools of competitive engineering design. These
toolscoveraspectsofrobotplanning,robotanalysis,robotselectionandrobot
innovation.
Theworkbookisthoughtasapracticalguideforstudentsinordertosupport
them in developing skills and competencies for competitive product
engineering.Thetasksrequiredintheworkshopshavebeencarefullyselected
suchthattoformabilitiesthatcanbeimmediatelyapplicableinrobotdesign
projects.

Theauthors
Workshop 1 Ranking Requirements in Competitive Robot Design Using the AHP Method
Ranking Requirements in Competitive

Robot Design Using the AHP Method
8
ToolsandMethodsofCompetitiveDesigninRobotics
9
1.1. Objective
In the initial stages of competitive product design a key aspect is to
identify, structure and rank the set of requirements related to the
applicationwhere the productis goingtobe integrated. For example, a
robot gripper is attached to the robot arm for grasping and handling
variouspartsundercertainworkingconditions.
Thepurposeofthisworkshopistodevelopbasicskillsandcompetenciesin
applyingawellknowncompetitiveengineeringtool,calledAHP(Analytic
HierarchyProcessTechnique),whichisoftenusedbyengineerstorank
requirementswithinthedesignprocess.
1.2. Background
Duringthedesignprocess,engineerswillhavetomakemanyjudgments
and to take decisions. They will have to determine key requirements
relatedtoagivenapplication.Oncerequirementsareidentified,thenext
importantstepthroughcompetitiveproductdesignistherankingprocess
oftheserequirements.Onthisway,engineerswillbeabletosetpriorities
whentheywillfacewithlimitedresources.
Some of the ranking techniques that engineers apply in practice are:
Prioritization Matrix Method (PMM), Analytic Hierarchy Process
Technique (AHP), Agreed Criteria Approach (ACA), Individual Criteria
Approach(ICA),aswellassomehybrids ofthem. AnalyticalHierarchy
ProcessTechniqueisahighlydevelopedmathematicalsystemforpriority
setting.ThemainoperationalstepsoftheAHPmethodare:
Definetheproblem;
Structureahierarchyrepresentingtheproblem;
Perform pairwise comparison judgments on requirements with
respecttotheproductgoal.
Whencomparingrequirements,Saatyseigenvectormethodisused;itis
presentedintable1.1.
10
Table1.1.Saatysintensityofimportancescale.
Inten- Defnition Explanations
sity of
impor-
tance
1 Equal The two activities or requirements or customer

importance groups, etc. contribute equally to the objective
3 Somewhat more Experience and judgment slightly favor one item

important over another
5 More important Experience and judgment strongly favor one

item over another
7 Much more An item is strongly favored and its dominance is

important demonstrated in practice
9 Absolutely more The evidences favoring one item relative to

important another is over the highest possible order of
afrmation
0.33 Somewhat less Experience and judgment slightly un-favor one

important item over another
0.20 Less important Experience and judgment strongly un-favor one

item over another
0.14 Much less An item is strongly un-favored and its

important dominance is demonstrated in practice
0.11 Absolutely less The evidences un-favoring one item relative to

important another is over the highest possible order of
afrmation
2, 4, 6, Intermediate When compromise is needed

8 values
11
0.5, Intermediate When compromise is needed

0.25, values
0.16,
0.12
IntheAHPmethod,eachpairwisecomparisonrepresentsanestimateof
the ratio of the priorities of the compared elements. Estimates of the
priorities(orweights)arecalculatedforeachpairwisecomparisonmatrix
foreachlevelofthehierarchy.Tosynthesizetheresultsoveralllevels,the
priorities at each level are weighted by the priority of the higherlevel
criterionwithrespecttowhichthecomparisonwasmade.
TounderstandtheworkingprincipleoftheAHPmethod,anillustrative
example is given in table 1.2. There are considered five requirements
denotedwith:A,B,C,DandE.Theproductdevelopmentteamcreatesa
listlikeintable1.2andestimatestheimportanceofcustomerpairwise.
Theresultsareshownintable1.2.
Thenextstepistonormalizethevaluesineachcolumn.Fornormalizinga
valueinacolumnitisnecessarytomakethesumofthevaluesinthat
columnandafterthattoreplaceeachvaluewiththeratiobetweenthat
valueandthesumofthatcolumn.
12
Table1.2.ExampleofapplyingtheAHPmethodinranking
requirements.
Requirement D
Requirement C
Requirement B
Requirement A
Requirement E
Requirement A 1 3 5 9 7
Requirement B 0.33 1 1 0.33 0.14
Requirement C 0.20 1 1 3 5
Requirement D 0.11 3 0.33 1 9
Requirement E 0.14 7 0.20 0.11 1
Total on column 1.78 15 7.53 13.44 22.14
Normalization:
Requirement D
Requirement C
Requirement B
Requirement A
Requirement E
% importance
Raw weight
Requirement A 0.56 0.20 0.67 0.67 0.32 0.48 48
Requirement B 0.19 0.07 0.13 0.03 0.01 0.09 9
Requirement C 0.11 0.07 0.13 0.22 0.22 0.15 15
Requirement D 0.06 0.20 0.04 0.07 0.41 0.16 16
Requirement E 0.08 0.46 0.03 0.01 0.04 0.12 12
13
Afternormalization,thesumineachcolumnwillbeequalto1.Theraw
weightintable1.2iscomputedinthefollowingway:
thesumoneachrawforeachsegmentismade(example:forsegment
Athesumis2.42);
thevaluesofthesumforeachsegmentareadded(intheexample
fromtable1.2thesumis5);
each value of the sum for each segment is divided by the sum
obtained at point b), resulting the values from the Raw Weight
column.
TheAHPmethodisapowerfultooltoprioritizeresourceswhendesign
productsorproductcomponents.
1.3. Task Description and Methodology

The task within this workshop is to rank requirements formulated in a
projectrelatedwiththedesignofarobotgripper.Withinthegivenproject,
ateamofexpertshasgeneratedthefollowingsetofrequirements:
1.Lownoise;
2.Aslowaspossibleproductioncost;
3.Easytoassemble;
4.Easytoattachontherobotflange;
5.Highaccuracy;
6.Designedforeasyrecycling;
7.Easytomaintain;
8.Aslowaspossibledrivingforces;
9.Lowweight;
10.Assmallaspossiblegauge;
11.Highreliability(MTBF);
12.Tobeapplicableforhandlingmorepartfamilies.
For applying the AHP method in this case study, the software package
QualicaQFDwillbeused.Thenextstepshavetobefollowed:
14
Step1:OpentheapplicationQualicaQFDandselectWorkbookWizard
(see Fig. 1.1). Click OK. A list of possible templates will be opened.
Followthelist,selectTreeDiagramWithAHP(seeFig.1.2)andNext.
Fig.1.1.ThefirstscreenwhenusingtheQualicaQFDapplication.
15
Fig.1.2.ThewindowwiththelistofpredefinedtemplatesinQualicaQFD.
Fig.1.3.ThewindowwithAHPgraphicalinterface.
16
Step2:Openanewdatabasefollowingstepbysteptherecommendations
givenbytheuserinterface.SelectTopLevelComparisonMatrix.You
willcomeuptoawindowlikeinfigure1.3.
Step3:StarttointroducethesetofrequirementsintherowscalledInput.
AutomaticallytheywillappearinthecolumnscalledOutput,too.You
canintroducedatadirectlybyclickingtheroworyoucanselectthegroup
ItemsfromthetopleftsideoftheframecalledDatabaseandintroduce
thesedatawithin.Whenyouwillclicktherightbuttonofthemouseinside
anyoftherowsofthemainframe,awindowwillappear.SelectReplace
treeandafterwardsselectITEMS.Datawillbecomevisible.
Step4:Selectstepbystepeachboxontherightsideofthemaindiagonal
anddoubleclicktheleftbuttonofthemouse.AlistwiththeSaatysscale
willbecomevisible(seeFig.1.4).Selectthevalueyouconsiderthemost
appropriate. Do this action until all boxes above the main diagonal are
filled.
Step 5: Go in the menu Tools and select Recalculate Now. In the
columnentitledImportancewillbeshowntherankofeachrequirement
inpercentages.
Step6:YoucanuseseveraloptionsgivenbyQualicaQFDtovisualizethe
finalresults.Also,youcansavethevariousscreensaspictures(.jpg,.bmp)
forfurtheruse.
AHP requires good expertise in the field of application for which
requirementsareanalyzed.Ifthepersonorpersonsinvolvedwithinthe
comparison process is orare not qualified in that field, the final results
mightbelessreliable.PracticerecommendsthatanAHPprocesstoberun
byateamgrouping34experts.
17
Fig.1.4.ThewayoffillingboxesusingSaatysscaleofcomparison.
For further details on the facilities of Qualica QFD during task

developmentthestudentisencouragedtousetheHelpwindow,where
exhaustive explanations are provided. Before starting the work, please
consultthefollowingtwoexamples.
1.4. Example 1
The first example concerns with ranking a set of requirements
relatedtocompetitivedesignofaroboticinstallationfortelescopemirror
grinding. Typically, the manufacturing process of smallsize telescope
mirrors consists of a roughcut of the mirror blank by conventional
machiningandthenbroughttoanopticalshoptogetthefinalsurfaceby
lapping,usingabrasivesofsuccessivelyfinergrit.Thisprocessisdivided
intothreedistinctphases:grinding,polishingandfiguring.
Thegrindingprocessstartswiththeroughgrindingoperationusingalarge
toolandcoarsegrittorapidlybringthesurfaceoftheblanktothebestfit
18
sphere.Thebestfitsphereisthespherefromwhichtheleastamountof
materialhastoberemovedtoobtainthedesiredasphericsurface.
Inthesecondstageofthegrindingprocess,asmallertooloratoolofthe
same size as the mirror is used to obtain a surface close to the desired
asphericsurface.Thisdependsonthefocallengthandthediameterofthe
mainmirror.Duringthegrindingprocess,thesurfaceerrorwilldropfrom
aprox.0.5mm(asitcomesfromthefirstoperation)toafewm.Grinding
of glass surfaces consists of conchoidal (shelllike) fracturing under
pressure.Fractureisthetensilefailureofabrittlesolidunderloadswhich
exceedtheelasticlimitsofthematerial.Duringthefinegrindingoperation,
theglassisfracturedbytheabrasivegrainsunderthehighpressurethat
occursatthecontactpoints.
Usingsuccessivefinergrit(200mdownto5m),boththeroughnessof
the surface layer and the depth of the damaged layer are reduced. The
thicknessofthedamagedlayerisontheorderof5to15mattheendof
thefinegrindingoperation.Surfaceroughnessdropsfrom5mto0.2m.
Thisisstillnotsmoothedenoughcomparedtothewavelengthofvisible
lightforthesurfacetobespecular.Inordertoimprovethesurfacequality
thenextphasesarepolishingandfiguring.
Because of the motions required during the grinding, polishing and
figuring operations, the best surface quality in the case of smallsize
telescopemirrorsisgottenmanually.
Roboticinstallationsareusuallymetonlyfortheproductionoflargeand
verylargesizetelescopemirrors,usedbyprofessionalobservatories.These
technologies are very expensive, are not customizable and are not
affordablefortheamateurastronomers.Inthiscontext,thechallengeisto
design a reliable concept of an affordable robotic installation that can
manufacturecustomizedsmallsizetelescopemirrors.
Inamanualproductionprocess,8workerscanproduce100mirrors/year.
For an overall labor cost/ worker/month of 500 $ (in the case of
Romania)andapaybackperiodof2years(consideringanumberof2robot
19
operatorsandallotherannuallyexpendituresi.e.maintenancecosts)the
economic justified investment in the robotic installation is maximum
$31.000.
The robotic installation must also satisfy a set of specifications that are
directly linked to the set of performance characteristics of a telescope
mirror.Thekeyperformancecharacteristicsare:
Mirrordiameter;
Mirrorthickness;
Mirrormass;
Mirrormaterial;
Surfaceaccuracy;
Chamferededges.
To plan properly the product specifications, a first step in competitive
designoftheroboticinstallationfortelescopemirrorgrindingistorankthe
set of performance characteristics. In this respect, the AHP method is a
reliable solution. Figure 1.5 shows the results of this analysis, as it was
captured from the Qualica QFD software application. The column
Importanceingroupfromfigure1.5presentstherelativeimportanceof
eachrequirementforthegivenindustrialapplication.
20
Fig.1.5.RankingthekeyperformancecharacteristicswiththeAHPmethod.
21
Fig.1.6.Theroboticinstallationforgrindingcustomizedsmallsizetelescope
mirrors(Brad&Srb,2004).
Forexemplification,figure1.6showstheroboticinstallationfortelescope
mirrorgrinding.
1.5. Example 2
The second example concerns with the ranking process of requirements
belongingtomoreobjectivefunctions.Inthisrespect,therankingprocess
willconsistoftwostages.InthefirststagetheAHPmethodisappliedfor
ranking the objectivefunctions. In the second stage the AHP method is
22
usedtorankthesetofrequirementsbelongingtoeachobjectivefunction.
The application considered in this example refers to an outdoor wall
cleaningroboticinstallation.
Anareainserviceindustrythatseemstobecomeofinterestforautomation
is the one of cleaning outdoor walls of tall buildings. Several robotic
solutionsarealreadyontothemarket.Theirpricesvaryaround30000USD.
Forincreasingthecompetitivenessofaservicerobot,akeyobjectivefor
designersistogenerateacustomisablesolutionthatmaximizestheratio
performance/costforthebeneficiary.Servicerobotsarebuiltinalow
number of units or very often, in a single unit. This requires low
developmentcostsinordertoobtaineconomicjustifiablesolutions.
In many western countries the use of robots for cleaning the walls of
modernbuildingsdesignedwithlargeglassareashasbecomesomething
usual.InRomania,humanworkersmainlydotheseoperations;butseveral
modern banks and hotels started to show interest for automatic wall
cleaninginstallations.Thiscomesupfromsomereasonslike:difficultiesin
determininghumanoperatorstoworkintheverycoldperiodsoftheyear;
humanoperatorsforsuchjobsstartedtobecomeexpensiveandthereare
fewer and fewer disposed to apply for this type of job, which is not
permanently; occasional workers are harder and harder found exactly
when they are needed; the frequency of the cleaning operations is
increasing(1time/month);inthecaseofhotels,therehavebeenreported
complaintsoftheclientsaboutthecleaningqualityofthewindows,etc.
The objectivefunctions considered for design optimization are the
following: a) minimum cost possible (below 25000 USD); b) easy to
maintain;c)easytouse;d)highreliability;e)highquality.Theapplication
oftheAHPmethodtorankthesetofobjectivefunctionsisshowninfigure
1.7.
23
Fig.1.7.RankingobjectivefunctionswiththeAHPmethod.
Foreachobjectivefunction,asetofrequirementshasbeengenerated.Two
oftheobjectivefunctionsarefurtherdetailed:
1)Fortheobjectivefunctionminimumcostpossible(below25000USD),
thesetofrequirementsis:a)usecommercialavailableparts;b)reduced
numberofcomponents;c)usesmallsizepartswheneverpossible;d)use
partsthatdonotrequirecomplexmanufacturingoperations;e)useparts
thatreducelossesinthesystem;f)followtherequiredcyclogramofthe
cleaningoperations;g)modularsystemdesign;h)lowcomplexityofthe
roboticsystem.
2)Fortheobjectivefunctioneasytomaintain,thesetofrequirementsis:
a)useofstandardcomponents;b)usepartseasytomanufacture;c)easy
24
access to all components; d) easy to maintained even by less qualified

persons;e)easytoreset;f)rapidchangingsystemofthetool.
TheapplicationoftheAHPmethodtoranktherequirementsbelongingto
thetwoobjectivefunctions:minimumcostpossible(below25000USD)
andeasytomaintainisshowninfigure1.8andfigure1.9.
Fig.1.8.Rankingrequirementsassociatedwithminimumcostpossible(below
25000USD)objectivefunction.
25
Fig.1.9.Rankingrequirementsassociatedwitheasytomaintainobjective
function.
26
Fig.1.10.Thegeneralviewoftherobotstructure(Brad&Mocan,2003).
Forexemplification,theroboticinstallationforwallcleaningisillustrated
infigure1.10.
1.6. Workflow
Thetaskrequiredatsection1.3istobeperformedin2hoursof
laboratorywork.Inthefirstthirdofthetime,studentsmustrunthetask
individually.Theyhavetosavetheresultsaspictureformat(.jpg).Picture
mustbeprintedout.Inthemidpartofthelaboratorywork,teamsof23
studentswillbeorganized.EachteamwillruntheAHPonceagainforthe
sametask.Theresultswillbeafterwardsavedandprintedout.Inthethird
partofthelaboratorywork,eachteamwillcomparetheresultitgotwith
theindividualresultsofeachteammember.Conclusionsconcerningtothe
effectivenessandefficiencyofapplyingtheAHPmethodindividuallyand
inaworkteammustbeformulated.Thenextstepconsistsincomparing
theresultsworkedoutbytheteams.Conclusionsmustbeformulated.
1.7. Homework
Analyzevariousgrippersfromcataloguesagainstthesetofrequirements
andsetupconclusions.
Workshop 2 Computing the Consistency Ratio within the AHP Method
Computing the Consistency Ratio

within the AHP Method
28
29
2.1. Objective
ThefinalstageoftheAHPmethodistocalculateaConsistencyRatio(CR)
inordertomeasurehowconsistentthejudgementshavebeenrelativeto
largesamplesofpurelyrandomjudgements.IfthevalueobtainedforCRis
greater than a threshold, the judgements are considered inconsistent
(actuallyuntrustworthy),astheyaretooclosetorandomness,andtheAHP
exercisehastoberepeated.
Thepurposeofthisworkshopistolearnhowtocomputeandinterpretthe
ConsistencyRatiowithintheAHPmethod.

ThetaskwithinthisworkshopistocomputetheConsistencyRatioforan
AHPexerciseandtointerpretitsvalue.Theproductunderconsiderationis
anissuenotificationsystemforrobotizedassemblylines.
Notification systems regarding issues on robotized assembly lines are
aimed towards alerting the maintenance staff about the occurrence of a
defectorunconformityintheassembly(orproduction)process.Theroleof
suchsystems istosignalsuchevents,asquickas possible, andtooffer
support in solving the issues in the shortest time possible, in order to
minimizetheeffectsonproduction.
Therequirementlistwhichwasconsideredfortheissuenotificationsystem
underconsiderationis:
1. lowcosts
2. easyimplementationandmaintenance
3. shortandeasysystemdevelopment
4. quickproblemsignalling
5. highefficiencyandeffectiveness
6. easyproblemsignalling
7. supplementarycontextinformation
8. smartproblemsignallingassistance
30
9. supplementaryadditionalrelevantinformation
10. increasedinformationrelevance
11. manageableknowledgebase
12. knowledgebaseefficientmanagement
First,thesoftwarepackageQualicaQFDwasusedtocompletetheAHP
rankingexercise(seeFig.2.1).
Fig.2.1.RequirementsrankingcompletedwithQualicaQFD.
Theanalysisshowedthatthemostimportantrequirementsrefertoquick
problemnoticingandtheeffectivenessofthesystem.It'salsoworthnoting
thatcostsandmaintenanceeffortsarenotconsideredtobecritical.
Inwhatfollows,thebasicelementsandstepsof the AHPmathematical
theorywillbeexplained,andtheCRwillbecomputed.Theexplanations
31
are based on G. Coyle's book Practical Strategy: Structured tools and

techniques(PrenticeHall,2004).
Let'sconsidernelementstobecompared(inourcase,the12requirements
above),denotedC1...Cn,andlet'sdenotetherelativeweight(orpriority
orsignificance)of Ci withrespecttoCj byaij andformasquare matrix
A=(aij)ofordernwiththeconstraintsthata ij=1/aji,forij,andaii=1,alli.
Suchamatrixissaidtobeareciprocalmatrix.Inourcase,ifwewould
representthematrixinaspreadsheet(e.g.LibreOffice'sCalcorMicrosoft's
Excel),itwouldlooklikeinFig.2.2.
Fig.2.2.Therequirementsmatrix(inaspreadsheet).
32
Qualica uses symbols to represent the figures, with each symbol

representing a numericalvalue. If we copypaste the comparisonvalues
fromQualicatoourspreadsheet,wegetthenumericalvaluesasshownin
Fig.2.3.AswecompleteonlytheupperpartofthematrixduringanAHP
exercise,nosymbolsareshowninthelowerpartofthematrix,butthe
correspondingvaluesareautomaticallycomputed.
Fig.2.3.Thenumericalvaluesintherequirementsmatrix.
The matrix in Figure 2.3 and the one shown by Qualica are actually
identical. The next step is the calculation of the list of relative weights,
importance,orvalue,oftherequirements(technically,thislistiscalledan
33
eigenvector). There are several methods for calculating the eigenvector;

however,multiplyingtogethertheentriesineachrowofthematrixand
thentakingthenthrootofthatproductgivesaverygoodapproximationto
thecorrectanswer.Thenthrootsarethensummedandthatsumisusedto
normalisetheeigenvectorelementstoaddto1.00or100%.Thisisshownin
Figure2.4.
Fig.2.4.Theeigenvectorcalcullation.
Inthespreadsheet,allthecomparisonvaluesforoneline(correspondingto
each requirement) are multiplied on column O. In column P the
eigenvectoriscalculatedusingthefollowingformula:[=O i^(1/12)],where
i takes values from 5 to 16 (the spreadsheet lines containing the
34
requirements).ColumnQcontainsthenormalisedvaluesoftheeigenvector
(their sum being 100). If you compare them to the values obtained in
Qualica, the results are similar (minor differences might exist due to
differentroundingoperations).
Theweightsareconsistentiftheyaretransitive,thatisa ik=aijajkforalli,j,
and k. Such a matrix might exist if the aij are calculated from exactly
measureddata.Thenfindavector oforder n suchthatA = .For
such a matrix, is said to be an eigenvector (of order n) and is an
eigenvalue.Foraconsistentmatrix(i.e.anidealhumanjudgement),=n.
However,formatricesinvolvinghumanjudgement(inourcasethismeans
rankingrequirementsrelatedtoaproductconcepts),theconditiona ik=aijajk
doesnothold,ashumanjudgementsareinconsistenttoagreaterorlesser
degree.InsuchacasethevectorsatisfiestheequationA =maxandmax
n. The difference, if any, between max and n is an indication of the
inconsistency of the judgements. If max = n then the judgements have
turnedouttobeconsistent.
Tocalculate max wehavetofirstmultiplyourmatrixwithcomparison
valuesbythecalculatedeigenvector.Figure1.5showshowthiswasdone
inthespreadsheet.Thehighlightedcolumnrepresentstheresultingvector.
35
Fig.2.5.Multiplicationofthematrixbytheeigenvector.
Next,theobtainedvectorisdividedbytheeigenvectorandweobtainthe
vectorhighlightedinFigure2.6.
36
Fig.2.6.Divisionbytheeigenvector.
Basedonthisvector,maxcanbecalculated.maxrepresentstheaverageof
thevector'svalues.Inourcase,theobtainedvalueformaxis13.30.
Finally, the Consistency Index is calculated from (max n) / (n 1).
According to Coyle, that needs to be assessed against judgments made
completelyatrandomandSaatyhascalculatedlargesamplesofrandom
matricesofincreasingorderandtheConsistencyIndicesofthosematrices.
AtrueConsistencyRatioiscalculatedbydividingtheConsistencyIndex
forthesetofjudgmentsbytheIndexforthecorrespondingrandommatrix
(see Figure 1.6). Saaty suggests that if that ratio exceeds 0.1 the set of
37
judgmentsmaybetooinconsistenttobereliable.Inpractice,CRsofmore
than 0.1 sometimes have to be accepted. A CR of 0 means that the
judgementsareperfectlyconsistent(butmostprobablythereisnorealcase
inwhichavalueof0willbeobtainedfortheCR).
Inourcase,theConsistencyIndexis(13.3012)/(121)whichmeansCI
=0.118.
Fig.2.7.IndicesforcomputingtheConsistencyRatio.
The(final)ConsistencyRatiowillbecomputedbydividingtheobtained
valueforCIbytheindexhighlightedinFigure2.7.Thus,theCRforour
AHPexerciseis0.118/1.48=0.079whichissignificantlylowerthan0.1.
Thus, the judgements we made for ranking the requirements are
consideredtobeconsistent.
2.3. Workflow
Thetaskdescribedinsection2.2istobeperformedin2hoursoflaboratory
work.Inthefirsthalfofthetime,studentsmustrunthetaskindividually
andcalculatetheConsistencyRatioforoneoftheAHPexercisesalready
completed.Allthecalculationsshouldbemadeusingaspreadsheeteditor
(like OpenOffice Calc or MS Excel) and results should be saved. In the
secondpartofthelaboratorywork,obtainedCRsshouldbeinterpreted
(discussed)andtheinitialAHPexercisesshouldberemadeiftheyproveto
beinconsistent.Conclusionsmustbeformulated.
2.4. Homework
Identifyrequirementsforpurchasingarobotforpalletizingapplications(at
least78requirements).RankthemfirstusingtheQualicaQFDsoftware
andthencomputetheCRusingaspreadsheet.Formulatetheconclusions.
38
Workshop 3 Planning Engineering Specifications in Competitive Robot Design Using the QFD
Method
Planning Engineering Specifications in

Competitive Robot Design Using the
QFD Method
40
41
3.1. Objective
Duringrobotdesignprocess,engineersconfrontwithmanytechnical,time,
financialandhumanresourcerelatedconstrains.Fromthisperspective,the
efforttoperformacertaintaskmustnotexceedthelevelthatisjustifiedby
thevaluewhichtherespectivetaskbringswithinthesystem.Todecide
properlyonsuchaspects,engineersshouldusevariousplanningtoolsat
differentstagesofthedesignprocess.Thefirstimportantplanningphase
concerns with the translation of requirements defined for the given
application into the socalled product engineering characteristics (or
product specifications).Product specifications are measurable,
controllable and quantifiable technical attributes which describe the
productperformances.
applying the QFD (Quality Function Deployment) method for planning
productspecifications.
3.2. Background
QualityFunctionDeploymentisastructuredplanningandcommunication
methodologythatassistsengineersduringproductdevelopmentprocess.
QFD is a complex and systematic approach, consisting of a reunion of
methodsandmeanslinkedbyspecialalgorithmsthattogetherprovidea
robustwaybywhichamultifunctionalteamidentifiesandtransfersthe
needsandexpectationsofthestakeholdersthrougheachstageofproduct
developmentandimplementation.
QFDisapowerfuldesignmanagementtoolcommonlyusedforsupporting
thedecisionmakingprocess.QFDutilizesaseriesofspecializedmatrices
forinterpretingthestakeholdersrequirementsintotherelevantproduct
features, engineering characteristics, manufacturing processes and
operations,implementationanddisposalaspects,etc.
42
Itworksproperlywithinanorganizationwheretheconcurrentengineering
concept is implemented, because QFD requires teamwork and
interdependency between different specialists (marketers, designers,
productionengineers,productmanagementengineers,etc.).Itrequiresthe
horizontalintegrationofthoseorganizationalfunctionsthatmustplando
checkact,suchastobringtheexpressedandunexpressedrequirements
intotheproduct.
The first stage in running a QFD project is to define and rank all
stakeholder requirements or in order words, to set up the socalled
requirement matrix. In this respect, market research approaches and
toolshavetobeapplied.Thereadercouldreviewthefirstworkshopin
order to remember one of the tools that is used for prioritizing
requirements(AHP).
The second stage of the QFD process consists in establishing the key
engineeringcharacteristics.Thisisacriticalpointinachievingasuperior
balance between development costs, product performances and timeto
market.Inthisrespect,fromthewholesetofproductrelatedengineering
characteristicsonlythesignificantonesshouldbeconsideredfurtherfor
evaluation. For supporting the process of generating engineering
characteristics, tools like MindMap and 635 (that are described in the
next workshops) could be used. For deciding which are the key
engineering characteristics, an evaluation process using a matrix that
shows the relationships between stakeholder requirements andtechnical
attributes, as well as a matrix that shows the correlations between the
considered engineering characteristics should be worked out. Figure 3.1
illustratestheseaspects.
The relationship coefficient aij between the ith stakeholder requirement
andthe jthengineeringcharacteristic, i=1,,n; j=1,,m,expressesthe
strengthofthelinkage(alsocalleddegreeofrelationshiporimpact)
betweenthetwoentities.
43
Fig.3.1.Evaluationofengineeringcharacteristicsforbalancingresources.
Forcalculatingthevalueweightofeachtechnicalattribute,thefollowing
formulaisused:
n
w j Ri aij ; j 1, m (3.1)
i 1
where:
nisthenumberofstakeholderrequirements;
misthenumberofengineeringcharacteristics;
wjisthevalueweightofthejthengineeringcharacteristic,j=1,,m;
Riistherankoftheithstakeholderrequirement,i=1,,n;
aij is the relationship coefficient between the ith stakeholder
requirement and the jth engineering characteristic, i=1,,n; j=1,
,m.
Inpracticethenormalizedformofthevalueweightisused,wjr,j=1,,m.
Thenormalizedvalueweightiscalculatedwiththeformula:
44
r wj
wj 100; j 1, m
m
(3.2)
w
k 1
k
Thevalueweight wjr ofthe jthengineeringcharacteristic, j=1,,m,isa

measure of the maximum possible contribution that the respective
engineering characteristic can bring in satisfying the set of input
requirements,relativetotheotherconsideredengineeringcharacteristics.
To describe the degree of relationships between requirements and
engineeringcharacteristics,thereareusedsymbolsornumericalvalues,as
itisshownintable3.1.
Table3.1.Symbolsandnumericalvaluesusedtodescribetheimpact.
45
Table3.2.Thethreescenariosindefiningtheimpactnumericalvalues.
Becauseofdifferentopinionsindefiningthescalethatexpressesthelevels
ofrelationships,thisguideproposesanalternative,whichhasbeentested
withvery good results in many projects. The alternative is touse three
scalesinthesametime,eachscalebelongingtoasocalledmembership
scenario,resultingthreemagnitudesforeachvalueweight.Theresultsare
analyzedinterms of evolution(increasing/decreasingthe valueweight)
andintermsofamplitude(thedifferencebetweentheextrememagnitudes
ofeachvalueweight).
Thenumericalvaluesfornorelationship,weakrelationshipandmedium
relationshiparekeptthesameineachscenario:0,1and3.Fortheother
levels of relationship: strong, very strong and extremely strong, the
proposedrulesaredescribedintable3.2.
The value weights in (3.1) should not be exclusively used to determine
priorityitems.Theyrepresentartificialnumbersandinthisrespectthey
couldsupporttheselectionofkeyengineeringcharacteristicsbutonlyin
strongrelationwithotherfactorssuchas:customercompetitiveevaluation,
complaints,salespoints,companygoals,correlationsbetweenengineering
characteristics,etc.
46
The identification and analysis of correlations between engineering

characteristics(seethecorrelationmatrixinfigure3.1)isanotherimportant
aspect indefining the key items on which the developers should focus.
Whentryingtoimproveacertainengineeringcharacteristic,ifthereisa
correlationwithanotheroneoranotherones,engineersshouldseehow,
themoveoftheengineeringcharacteristicinthedirectionofgoodnesswill
influencetheimprovementdirectionsoftheothers.Inthisrespect,there
could occur both positive and negative impacts, on different scales of
influence.Inpractice,itisusedthescalepresentedintable3.3.
Table3.3.Scalefordefiningthedegreeoftechnicalimpact.
An important aspect that should be determined and introduced as

informationwithinthecorrelationmatrixisthedirectionofimpact.Ifthere
areconsideredtwoengineeringcharacteristicsTRiandTRj,ij,i<j,then
the following three situations are possible when there is a correlation
betweenthetwoengineeringcharacteristics:
a) TRionlyinfluencestheTRj,whenresultaonedirectionalimpact
fromlefttoright();
b) TRjonlyinfluencestheTRi,whenresultaonedirectionalimpact
fromrighttoleft();
47
c) TRi influences the TRj and also TRj influences the TRi, when
resultabidirectionalimpact().
Thesymbols,andcouldbevisualizedwithinthecorrelationmatrix
above the symbols that express the levels of correlations between
engineeringcharacteristics.
When considering the direction of goodness of each engineering
characteristic, three situations could appear: more is better (symbolized
with );lessisbetter(symbolizedwith )andtargetisbest(symbolized
withO).
Thistypeofinformationcouldbeplacedinarowwithintheengineering
characteristicmatrixorinarownearthevalueweightrow.
An alternative to the correlation matrix, which is used by some
practitioners,isthesocalledRelationshipNetworkDiagram.Itconsists
ofasetofcirclesforrepresentingthetechnicalattributes(acircleforeach
technicalattribute)andanetworkofarrowsthatconnectthecircles.Along
thearrowsarerepresentedthelevelsofimpact(seetable3.3).Thearrows
enterintothosetechnicalattributesthatareinfluencedbytheothers(see
onedirectionalimpactfromlefttoright,onedirectionalimpactfromright
toleftandbidirectionalimpact).Anexampleisgiveninfigure3.2.
48
Fig.3.2.Anexampleofrelationshipnetworkdiagram.
Anotherimportantaspectforoptimizingtheuseofexistentresourcesisthe
identification of the minimum set of engineering characteristics that are
relatedwiththewholesetofstakeholderrequirementsandwiththemost
significantcorrelations.Definingthiskeysetofengineeringcharacteristics
does not imply neglecting the analysis of the other engineering
characteristics. It couldbe used whennot enoughtime orresources are
availabletosolveacertainproblemandtheteamshouldcometoafeasible
solution.Insuchsituations,onlythekeysetofengineeringcharacteristics
willbecomeofinterest.
Engineeringcharacteristicsfullysatisfythestakeholderrequirementsonly
wheneachofthemattainsasocalledtargetvalue.Settingthesetargetsis
of very great importance. Traditional approaches for setting the target
valuesoftheengineeringcharacteristicsrelyheavilyontheexperienceand
intuition of the design engineers. This leads most probably to feasible
solutions rather than optimal ones. The target values will drive all
subsequent development activities. In this respect, the approach for
defining the right target value for each engineering characteristic is of
49
majorconcern.Whenestablishingthetargetvalues,severalaspectsshould
be considered, like for example: correlations between engineering
characteristics, current performance of the product (if an old version
already exists), relationships between engineering characteristics and
stakeholderrequirements, currentperformance of competitors products,
availablebudget,etc.Inordertodefineoptimumtargetsfortheconsidered
key engineering characteristics, it is necessary to run a quantitative
optimization process. However, it is not the purpose of this guide to
describeaspectsrelatedtosuchoptimizationprocesses.

Thetaskwithinthisworkshopistoplantheengineeringcharacteristicsofa
robot gripper against the set of requirements formulated in the first
workshop.Importanceofeachrequirementcorrespondswiththeresults
obtainedfromtheapplicationoftheAHPmethodinthefirstworkshop.
Thesetofengineeringcharacteristicsthatareconsideredinthisworkshop
isthefollowing:
Stroke/finger[mm];
Graspingforceat6bar[N];
Payloadcapacity[kg];
Compressedairconsumptionperdoublestroke[cm3];
Fingerclosingtime[s];
Fingeropeningtime[s];
Grippermass[kg];
Inertialmoment[kgcm2];
Repeatability[mm];
Maximumlengthofthefinger[mm];
Graspingforceensuredonlybythespring[N];
Openingbetweenjaws[mm];
Gauge[cm3];
Efficiency(drivingforce/graspingforce)[%];
Abilitytoexchangefingers&flanges[Yes/No];
50
Timetoexchangefingers&flanges[s];
Materialhomogeneity[%];
Mobileparts(relativetofixedparts)[%];
Numberofparts[#];
Power[W];
Partcomplexity[$].
TheQualicaQFDsoftwarepackagewillbeconsideredforperformingthe
task.Thefollowingstepshavetobedone:
Step1:OpentheapplicationQualicaQFDandselectWorkbookWizard.
ClickOK.Alistofpossibletemplateswillbeopened.Followthelist,
select1stHouseofQuality(seeFig.3.3)andNext.
Step2:Openanewdatabasefollowingstepbysteptherecommendations
givenbytheuserinterface.SelectHouseofQuality.Youwillcomeupto
awindowlikeinfigure3.4.
Step3:IntroducethesetofrequirementsinthecolumncalledNeedsand
theirlevelofimportanceinthecolumnImportancetocustomer.
Step4: Introducethesetofengineeringcharacteristicsintherowcalled
CTQs. Introduce the optimization trend of each engineering
characteristicintherowcalledOptimization.Introducethemeasurement
unit of each engineering characteristic in the row Measurement Unit
fromthebottomsideofthegraph.
51
Fig.3.3.ThewindowwiththelistofpredefinedtemplatesinQualicaQFD.
52
Fig.3.4.ThewindowwithQFDgraphicalinterface.
53
Fig.3.5.TherelationshipmatrixwithintheHoQ.
Step5:CompletetherelationshipmatrixintheHouseofQuality(seeFig.
2.5).Thisisperformedbydoubleclickingeachboxfromtheintersectionof
acertainrequirementwithacertainengineeringcharacteristic.Awindow
withlevelsofrelationshipisdisplayed.Youneedfirstlytoaddtwomore
levels of relationship within the system, by clicking on in that
window.Thetwosupplementaryrelationshiplevelsare:27verystrong
relationship; 81 extremely strong relationship. See in figure 2.5 these
results. When no relationship exist between a given input and a given
output, the box will be kept blank. The process is done until all
relationshipsareanalyzed.
Step6: Becausestudentscouldhavesomelackofexperienceinusingthe
QFDmethodathighstandards,theyneedtocheckouttheaccuracyofthe
54
datatheyintroducedwithintherelationshipmatrix.Someguidelinesfor
verifyingthequalityofanalysisarefurthergiven:
a)Emptyrows:ifcertainrequirementshavenotrelatedengineering
characteristicsmeansthereisnowaytoassurethoserequirements.
Gobackandcheckonceagain.
b)Emptycolumns:engineeringcharacteristicsthatdonotrelatewith
somerequirementsmeansthosecharacteristicsarenotrelevantfor
productoptimization.Gobackandcheckonceagain.
c) Rows with no strong relationship (or no very strong or no
extremely strong relationship): requirements are difficult to be
achieved without at least one strongly related engineering
characteristic.Ifinyourmatrixoccurssuchcases,gobackandcheck
onceagain.
d) Rows that repeat identical relationships: if two or more
requirements have 100% identical relationships with the set of
engineering characteristics means that those requirements are not
properlystructured.Insuchsituations,gobackandcheckonceagain.
e)Clustersofrelationships:ifsomeclustersoccurintherelationship
matrix,meansthateitherrequirementsorengineeringcharacteristics
orbotharepoorlydefined.Thisrequiresareanalysisofyourresults.
f)Rowswithtoomanyrelationships:ifarequirementitemisrelated
tothemostoftheengineeringcharacteristics,itmayactuallybea
cost,reliabilityorsafetyitem.Suchitemsmustberemovedinother
planning matrices (cost deployment, reliability deployment, safety
deployment).Incasethissituationappearsinyourmatrix,youhave
totakebacktheanalysis.
g) Columns with too many relationships: if an engineering
characteristic is related to the most of the requirements, it may
actually be a cost, reliability or safety item. Such items must be
removed in other planning matrices (cost deployment, reliability
55
deployment, safety deployment). In case this situation appears in

yourmatrix,youhavetotakebacktheanalysis.
h)Diagonallinesacrossmatrixwithfewotherrelationships:inthis
case, the set of requirements may in fact be engineering
characteristicsworked(formulated)differently.Requirementsarethe
voiceofthecustomer,notthevoiceofengineers.Thismeansyour
analysisisnotproperlyperformed.
i)Toomanyweakrelationships:insuchsituations,theconclusionis
thattheengineeringcharacteristicsarenotproperdefinedandsome
othermustbeformulated.Ifinyouranalysisexistssuchcases,you
need to perform the task once again because each engineering
characteristicshouldrelatestronglywithatleastonerequirement.
Step 7: Complete the roof of the QFD. The roof shows the correlation
between engineering characteristics. Two engineering characteristics are
positivelycorrelatediftheimprovementofonecharacteristicsupportsthe
improvementeffortoftheotherengineeringcharacteristic.Inthecontrary
case, they are negative correlated. There might be cases where no
correlationexistsbetweentwocharacteristics.Inordertoperformthistask,
click in the boxes at the intersection between two engineering
characteristicsandselecttheappropriatelevelofcorrelation.Dothisjob
systematically, starting with the first engineering characteristic. Note all
negative correlations. They generate technical conflicts and from this
perspectivetheyrequireinnovation.Figure2.6showshowthisprocessis
performedbymeansofQualicaQFD.
Step8:Calculatethevalueweightoftheengineeringcharacteristics.Thisis
performedbygoingatToolsandselectingRecalculateNow.
56
Fig.3.6.CorrelationmatrixwithintheHoQ.
InthecolumnentitledImportance,thevalueweightofeachengineering
characteristicwillbeshowninpercentages.Someotherinformationcould
beprovidedbyQualicaQFD,too.ConsulttheUserGuideofthesoftware
inthisrespect.
3.4. Example 1
ThefollowingexampleshowshowQFDmethodhasbeenappliedtoplan
engineering characteristics of a robotic installation for telescope mirror
grinding(seeFig.1.6).Inthisexample,theprocessbeginsfromthestage
wherethecustomerrequirementsareidentifiedandranked(seetheresults
oftheAHPmethodwithintheworkshop1,figure1.5).
A set of 10 engineering characteristics were considered important for
designingtheroboticinstallation:
57
1. TranslationOxaxis[mm]&[m/min];
2. TranslationOyaxis[mm]&[m/min];
3. TranslationOzaxis[mm]
4. MirrorrotationonOzaxis[rot/min];
5. ToolrotationonOzaxis[rot/min];
6. ActiveforceonOzaxis[N];
7. AccuracyonOxaxis[mm];
8. AccuracyonOyaxis[mm];
9. PassivemirrorrotationonOxaxis[deg];
10. PassivemirrorrotationonOyaxis[deg].
TheapplicationofQFDmethodmakesthe customerrequirements askey
driversinthedesignprocess,aswellascatalystsformodifications.The
scaleusedtorelatetherequirementstotheengineeringcharacteristicsis:1
weakrelationship,3mediumrelationship,9strongrelationship.Figure
3.7showstheresultsoftheHouseofQualityforexample1.Itcanbeseen
that the active force on OZ axis is the most critical characteristic so, a
specialattentionmustbeconsideredondesigningthesensingsystemfor
theOZrobotaxis.
58
Fig.3.7.Houseofqualityforexample1.
Anotherfourengineeringcharacteristicsarecriticalforsuccessfuldesignof
the robotic installation. In this respect, configuration design is
recommendedforOXandOYaxes,aswellasformirrorandtoolrotation.
Configuration design is the term used when engineers integrate
59
commercialavailablemodules(thataretestedandguaranteed)withinthe
technicalsolution.
3.5. Example 2
The second example concerns with QFD application for redesigning an
existent product in order to increase its competitiveness. The product
under consideration is a pressureflow measuring head, a complex
mechatronicproductwhichincorporatesbothmechanics,hydraulicsand
electronics. The planning process here presented begins from the stage
wherethecustomerrequirementsareidentified.Table3.4revealsthesetof
requirements that has been collected from potential customers. The
requirements were ranked using the AHP method (see the column
Importanceintable3.4).
For the pressureflow measuring head, a number of 13 engineering
characteristics have been considered important for product redesign, as
follows:1)workingpressure[bar];2)flowrate[l/min];3)workingfluid
[gr/cm3]; 4)static characteristic [%]; 5) energy consumption [W]; 6)
dynamiccharacteristic[Hz];7)resolution[%];8)sensibilitytotemperature
variations[%];9)responsetime[ms];10)sensibilitytoflowpulsations[%];
11)modularitylevel[15];12)dimensionsmechanicalsystem[mm 3];13)
dimensionselectronicsystem[mm3].
Table3.4.Thesetofrequirementsforredesigningthepressureflowmeasuring
head
No. Requirements Importance

1 Nominal pressure of 320 [bar] 12.8
2 The maximum admissible pressure 400 [bar] 2.9
3 The fow rate: 0-60 [l/ min] 12.1
o o
4 The temperature of the working fuid: -25 +80 4.8
5 Maximum pressure drop on the sensor: 10 [bar] 6.1
6 Minimum working pressure: 30 [bar] 6.1
60
7 Precision of pressure regulation: 1% of the

14.6
maximum pressure of the system
8 Precision of fow regulation: 1.5 % of the maximum
14.6
fow rate of the system
9 Response time of the sensors for a step signal: 15 [Hz] 13.4
10 Use of diferent types of pumps with variable
7.3
displacement
11 Small gauge 5.2
61
Fig.3.8.HouseofQualityforexample2.
To deploy properly the critical customer requirements throughout the

designstages,QFDmethodhasbeenused.Thescaleusedfordescribing
relationshipsbetweenrequirementsandengineeringcharacteristicswas:1
62
weak,3medium,9strong,27verystrong,and81extremelystrong.
Figure3.8showstheresultsofthisanalysis.
Weseethatresolutionanddynamiccharacteristicaretwoofthekey
engineeringperformancecharacteristicsofthesensor.Inordertodesigna
sensor with high performances in terms of these two characteristics, a
specialattentionmustbepaidonhowtheelectronicsystemisredesigned.
Fig.3.9.Theredesignedversionofthepressureflowmeasuringhead(Brad&
Vaida,2005).
For exemplification, figure 3.9 shows the redesigned solution of the

pressureflow measuring head. An easyvisible improvement is that the
electronicmoduleintheredesignedversionisseveraltimessmallerthanin
theinitialversionanditislocatedinsidethemechanicalsystem(10times
smaller:anintegratedcircuitreplacesBGAelectroniccomponents).
3.6. Workflow
Thetaskdescribedatsection3.3isplannedtobeperformedin2hoursof
laboratorywork.Studentsmustrunthetaskindividually.Iftheworkisnot
carriedoutinthescheduledtime,studentsmustgoonuntilthetaskis
completed.Conclusionsaboutdifferenttimedurationsrequiredtoperform
thesametaskbydifferentpeoplehavetobeformulated.Oncethetaskis
completed, students have to save the results as picture format (.jpg).
63
Picturesmustbeprintedout.Inthenextstage,individualresultswillbe
analyzed and compared each other. Conclusions will be afterward
formulated(qualityofresults,differencesbetweendifferentworks,etc.).
3.7. Homework
Analyzevariousgrippersfromcataloguesagainstthesetofengineering
characteristics and set up conclusions. Run a QFD project related to a
positioningsensor,workinginteamsof23people.Formulateconclusions
abouttimerequiredtoperformthetask, aswellasabout the tradeoffs
occurredduringdiscussionsbetweentheteammembers.
64
Workshop 4 Functional Analysis in Competitive Robot Design using the FAST and FBD
Methods
Functional Analysis in Competitive

Robot Design using the FAST and FBD
Methods
66
67
4.1. Objective
Onceaconceptorasolutionforagivenproductisdefined,anextremely
important aspect for engineers is to have a full understanding of the
functionforeachelementofthesystem,aswellasoftheinterrelationships
anddependenciesofallthefunctionsthatformthesystem.
applying two interrelated methods, FAST (Function Analysis System
Technique)andFBD(FunctionBlockDiagram),toanalyzethesolutionofa
givensystemintermsoffunctionalinterrelationships.
4.2. Background
FASTisausefulapproachforanalyzingthedesignofagivensystemwith
respecttothefunctionalinterrelationshipsbetweendifferentcomponents
and subsystems belonging to that system. The generated diagram, by
applyingFAST,representsapossiblesupportforevaluatingthecostsand
reliabilityoftheconsideredsystem,too.
ThefirststageinderivingaFASTdiagramconsistsingeneratingalistwith
all known functions of the system. These functions are described by a
nounverbphrase.Inlistingthenounverbphrases,animportantaspectis
theconsiderationofthelevelofdetailforeachsubsystem.Forexample,if
themotorsthatwilldrivetheaxesofarobotarepurchased,themotorswill
beconsideredasblackboxes(compactsubsystems)andtheirfunction
willbe:toprovidetorque.
Ifthemotorshavetobedesignedwithinthecompany,theywillhavetobe
split into their constitutive components and for each component to be
definedtherelatedfunction(s).Thefunctiontoprovidetorqueisinthis
casetheultimatefunctionofthesubsystemmotor.
ThenextstepindevelopingaFASTdiagramconsistsinlinkingtogether
eachoftheindividualfunctions,inordertoformanetwork.
68
Fig.4.1.AnexampleofFASTdiagramforagenericsystem.
Oneofthefunctionsischosenasthestartingpoint.Thenetworkisbuiltby
askingtwotypesofquestionsrelativetoeachindividualfunction:Why
doesthefunctionexist?andHowdoesthefunctionoccur?Theanswers
toWhy?areplacedtotherightandtheanswerstoHow?areplacedto
theleft.Attheend,thenetworkwillhavetheultimatefunctionatthefar
right of the FAST diagram and the set of parametric functions for the
systemoperationatthefarleft.Duringnetworkdevelopmentitispossible
tobeidentifiedsupplementaryfunctionstothoseincludedintheinitiallist.
Figure4.1illustratesanexampleofFASTdiagram.Itcanbeseenthatthere
are m functions of level 1, where the functions 1.2, 2.3 and n.2 are
parametricfunctionsandtheparametricfunctionsaregettingdownstream
totheleveln.
InsomecasesofFASTdiagram,itisnotnecessarilytogetdowntothepart
levelorviceversa,uptothecustomerrequirementslevel.Itdependson
theaspectthatshouldbeanalyzedbyengineers.Also,itshouldbenoted
69
that the process of generating the FAST diagram is very dynamic,

especiallywhenlessexperienceexistsinthisrespect.Thismeansthatthe
final variant of the diagram could result after several refinements. For
ensuringtheflexibilityinrunningthisprocess,itisrecommendedtouse
specialized software packages or postit cards. The functions are firstly
writtenonthesecardsandthecardscanbetheneasilyplaced,replaced,
addedorrejectedinthediagram.
Whentheanalysisisgottendowntothelevelofparts,FASTdiagramsare
efficient to apply on relatively small systems. In the case of largescale
systems, their constitutive subsystems are mainly subjected to such
analysis.Inthecaseoflargescalesystems,FASTcouldbealsoelaborated
for the complete system, but from practical point of view it is
recommendedtodefinetheboundariesoftheanalysisdowntothelevelof
themainsubsystems.Oncetheinterdependenciesbetweenthemainsub
systemsareunderstood,eachsubsystemcanbeextractedfromthemain
FASTdiagramandadditionalFASTdiagramscouldbedevelopedtothe
levelofeachsubsystem.
Inconclusion, for the same system, a FASTdiagramcould take various
forms depending on what the engineers are interested to study. The
generationprocessoftheFASTdiagramisthekeysuccesscharacteristic.
Pleasenotethattherearetoolsquitesimpleintheorybutquitedifficultto
applyproperlyinpractice.FASTisoneofthesetools.Theresultdepends
onengineersskillsandknowledgeinrunningthisprocess.Inthisrespect,
anyalgorithmthatmakesastepforwardinsustainingtheconstructionof
the FAST diagram is an admirable initiative. A possible framework for
deriving in practice a functional analysis process is given in figure 4.2.
Having prepared a FAST diagram, engineers have a clear view of the
overallsystem.
70
Fig.4.2.Aframeworkforderivingproperlyafunctionalanalysisprocess.
Fig.4.3.ThegraphicrepresentationofaFBD.
71
Toolsmentionedinfigure4.2aredescribedwithinthespaceallocatedfor
theotherworkshopsinthisbook.OncetheFASTdiagramisderived,the
constitutive functions of the system could be now ranked using, for
example,theAHPmethod.Theresultedkeyfunctionalareascanbefurther
analyzed.Inthisrespect,asocalledfunctionblockdiagram(FBD)will
beconfiguredforeachkeyfunctionofthesystem.TheFBDisthecorefor
criticalparameteranalysis.AFBDcomprisesaboxrepresentingthenoun
verbfunction,alistofinputcontrolledparametersthatenablethefunction
to operate, responses that may be used to measure the function
performance,noisefactorsthatconstrainthefunctiontooperateatitsfull
potentialandthefailuremodesassociatedwiththefunction(seeFig.4.3).
BuildingaFBDrequiressomestepstobefollowedinawelldefinedorder.
Theyare:
1. Theverbnounfunctionisdefinedandintroducedintoabox;
2. The expected response of the function is identified (it should be a
measurablefactor);
3. The response is considered from the customer acceptance point of
viewandupper,intermediaryandlowerfailurelimitsaredefined;
4. Thecontrolparametersareidentified,analyzedandthecriticalones
areintroducedinthediagram;
5. Thenoisefactorsareidentifiedandincludedinthediagram.
Inordertodeterminewhichoftheparametersarecriticaltothefunction,
anevaluationprocessshouldbecarriedout.Thiscouldbeperformedby
settingupanAHPevaluation.
Aftertheevaluationstage,engineerswilldecidethesetofparametersthat
areconsideredcritical,basedonexperience.Selectionofcriticalparameters
isanarbitrarydecisionprocessandingeneraltherearetakenintoaccount
5% to 10% of the parameters from the considered initial set. When
discussingaboutrobustdesign,eachofthecriticalparametersandnoise
factors will have a nominal value and a range of variation (tolerance).
Knowingthepossiblefailuremodesandthenoisefactors,engineerscan
72
focusondesigningtherightvaluesofthecriticalparameterssuchasto
achievethedesiredresponse.Forcingengineerstothinkatallaspectsthat
are required to fill a FBD represents a positive attitude for preventing
drawbacksinthenextphasesofproductdevelopment.
ToolslikeFASTandFBDareextremelyusefultosupportotherphasesof
competitive product design and development like, for example, the
reliabilityanalysis.

ThetaskwithinthisworkshopistoworkoutaFASTdiagramfortherobot
wristshowninfigure4.4,aswellastheFBDsfortwofunctionsoftherobot
wrist.Thetwofunctionsaretothestudentschoice.Therobotwristhas3
DOF andthe kinematic chains are gearingbased. The kinematics of the
robotwristissuggestivelyillustratedinfigure4.4.
73
Fig.4.4.A3DOFrobotwristwithgearedtransmission.
74
Fig.4.5.TheentrancewindowinMSVisio.
Tosupportthetaskexecution,thetechnicalplatformMSVisiowillbeused.
ForperformingaFASTprocess,whenenterintheMSVisioenvironment,
selectthecategoryFlowchartandthetemplateBasicflowchart,asitis
showninfigure4.5.ClickonBasicflowchartshapesandtheenvironment
fromfigure4.6willberevealed.
75
Fig.4.6.TheworkingenvironmentforderivingaFASTprocessinMSVisio.
Ontheleftsideofthescreen,alistwithvariousshapesispresented.By
clickingonthedesiredshapeanddraggingitintotheworkingsheet,you
canbuilttheFASTdiagram.Todothis,thefollowingstepsareimportant:
Step1:ClickontheshapeParallelmodeanddragittothetopleftsideof
the working sheet. Once you position the shape in the working sheet,
doubleclickontheshapeandintroducethetextHow?.
Step2:Repeatthesameactionsasforstep1andbuildtheboxforWhy?
onthetoprightsideoftheworkingsheet.Youcanmodifythesizeandthe
type of the font using the menu Formatand afterward the submenu
Text.
Step3:BuildtheFASTdiagramusingtheshapesProcessandDynamic
connectorfromthelistofpredefinedshapes.Ifyoudoubleclickonthe
shapeyoucanintroducetextwithin.Thetypeofthedynamicconnector
canbemodifiedbyleftclickingonitandselectingFormatandLine.A
windowwithoptionswillbeopened.
BecausetheMSVisioenvironmentisveryflexible,youcandynamically
modifythestructureoftheFASTdiagramaccordinglywiththeneeds.
76
To draw FBD diagram, select Borders and titles in the left side
frameworkoftheMSVisio.UsetheshapeNoteboxneontobuildthe
boxesrequiredtorepresentaFBD.Itgivesyouapredefinedshapethat
supportstheintroductionoftitlesandcomments,asintheboxesofaFBD.
To represent arrows, select Basic flowchart shapes and use Dynamic
connectorpredefinedshape.Figure4.7exemplifiesthesesteps.
Fig.4.7.BuildingaFBDframeworkusingMSVisio.
Drawings that are worked out with MS Visio can be saved in picture
format(.jpg).Thisisusefulwhentransfertheresultsoftheworkinthe
technicalreport.ItisdonebyselectingFileandSaveasoption.
4.4. Example 1
In this example, the FAST diagram is applied to generate a possible
solutionforarobot mechanicalhand. Afirststepdeals withtheuseof
FAST as a tool for developing and refining ideas about the considered
functionalsystem.Figure4.8putsintoevidencetheFASTdiagramforthe
incipientconceptualstageoftherobothand.
77
Infigure4.8,threeboxesarehighlightedindarkgrayandtwoboxesare
highlightedinlightgray.Thoseboxeshighlightedindarkgrayshowthat
engineers should generate ideas for the three subsystems. Those boxes
highlighted in light gray show that the two systems (the robot and the
connectionelement)donotbelongtotherobothand.Thelinksfromthe
FASTdiagramshowtheconnectionsbetweendifferentboxes.
Fig.4.8.FASTdiagramintheincipientconceptualphaseofarobothand.
Forexample,thedrivingsystemandthetransmissionmechanismcannot
be defined separately because of their interdependence. To generate
various concepts of the subsystems, engineers could use creativity
methodslikeMindMap,635,TRIZ/ARIZ,etc.Theyaredescribedinthe
otherworkshopsfromthisbook.Oncetheconceptsaredefined,toolslike
78
Pugh method are recommended for selecting the best of the proposed
variants.Pughmethodisalsopresentedinthisbook.
Figure4.9illustratesapossibleresultfortheexampleunderconsideration.
Thesystemusesanaircylinderasdrivingsystemandapairofslidercrank
mechanisms as transmission system. The grasping device includes the
extensions of the crankmechanisms and a pair of chucks with special
shapes to ensure autocentering of the object and to diminish the
drawbacksinpositioning.
When the piston is pushes to the right by the pneumatic pressure, the
elements of the two crankmechanisms rotate around the two fulcrums.
Theserotationsmakethegraspingactionattheextendedendofthecrank
elements,usingspecialshapedchucks.Movingthepistontotheleftdoes
thereleasingaction.
Thenextstepofanalysisusestheconceptualschemeoftherobothand.At
thisstage,engineersareinterestedindefiningtherightactionsthatshould
bedoneduringtheconceptualphaseofthesystemtoavoidfailuresinthe
next phase of system development, that is, in the constructive phase. It
shouldbenotedthatthestatedobjectiveoftheanalysisprocessplaysa
significant role on how the FAST structure will look like. The analysis
process goes down to more and more details until the system is fully
understood.
ThisexampleisquitedifferentinapproachingtheFASTmethodthanthe
taskwhichisrequiredinthisworkshop.ThisispossiblebecauseFASTisa
universaltechniquethatcanbeusedforvariouspurposesinengineering
design.
79
Fig.4.9.Arobotmechanicalhandbasedonapairofslidercrankmechanisms.
Fig.4.10.AFBDforthefunctiondrivethegraspingdevice.
For the function drive the grasping device, a FBD could look like in
figure4.10.
80
4.5. Example 2
Thesecondexampleconcernswiththefunctionalanalysisoftherobotic
installationfortelescopemirrorgrindingillustratedinworkshop1,figure
1.6.TheresultsofFASTdiagramforthiscaseare showninfigure3.11.
According to the FAST diagram, the following modules are generated
withinthesystemandrequiretobedesignedorselected:
Graspingmechanismofthemirrorsuchthattoensuretherequired
graspingforce;
Fingertips;
Transmissionsystemofthetool;
Systemtorotatethetoolholder;
Forcesensingsystemforthetool;
TransmissionsystemOx;
Graspingmechanismofthemirror;
TransmissionsystemOy;
Transmissionsystemtoensurethenecessarymovementforrotating
themirror;
Toolholdertograspthetoolthatperformsactionsonthepart(the
telescopemirror);
Transmissionsystemforpressingdownthemirror;
MechanismtoensureOxmirrorpassiverotation;
MechanismtoensureOymirrorpassiverotation.
81
Fig.4.11.TheFASTdiagramfortheroboticinstallationshowninfigure1.6.
82
Thefunctionsensetheforce(seeFig.4.11),alongtheOZaxis,isvery
criticalbecauseoftherequesttoavoidthebreakofthetelescopemirror
duringthegrindingprocess.Forthisfunction,aFBDisshowninfigure
4.12.
Fig.4.12.TheFBDforthefunctionsensetheforce.
83
Fig.4.13.Technicalsolutiontosustainthefunctionsensetheforce:aforcesensor
andasafetyforcefeedbackspringontheOZtranslationrotationaxis(Brad&Srb,
2004).
Figure 4.13 presents the technical solution which is provided to solve

properlythefunctionsensetheforcewithinthetechnologicalprocessof
telescopemirrorgrinding.
4.6. Workflow
Thetaskwithinthisworkshophastobeperformedin2hours.Students
willworkindividually.Oncetheworkiscarriedout,resultsaresavedin
pictureformat(.jpg)andafterwardareprintedout.Individualresultswill
bepostedandanalyzedinteam.Conclusionsaroundtheresultswillbe
formulated.
84
4.7. Homework
DoaFASTandFBDanalysisforamodulethatismetwithinthestructure
of industrial robots. This module might be one of the followings: an
encoder,anACmotor,aplanetaryspeedreducer.
Workshop 5 Idea Generation in Competitive Robot Design using the 6-3-5 and Mind-Map
Methods
Idea Generation in Competitive Robot

Design using the 6-3-5 and Mind-Map
Methods
87
5.1. Objective
Throughoutthedesignprocess,engineersarefacedwithtasksthatrequire
imagination to generate ideas related to various engineering problems.
Different methods are used in practice to support this kind of process.
Some of them can be applied by individual designers, some others are
moreeffectiveiftheyareusedwithinateam.
applyingtwomethods,635andMindMap,usedbyengineerstosupport
the idea generation process. The first method, 635, is a teamoriented
method.Thesecondonecanbeappliedbothbyindividualsandteams.
Withinthepresentworkshop,studentswillexperiencethesemethodsin
teamoriented structures. They will have the opportunity to understand
factors predisposing creative effort and will benefit from the working
proceduresthesemethodspropose.
5.2. Background
The 635 method is one of the most spread and accepted methods for
generatingideasinproblemsolvingorientedprocesses.Itisaworkteam
approachanditspopularitycomesfromthereasonitdoesnotimplyany
verbalcommunicationbetweentheteammembers.
Theworkingprocedurerecommendsorganizingateamof6peoplethatare
askedtogenerateasetof3ideas/sessionwithrespecttoaspecifiedsubject,
withatimingof5minutes/sessionforgeneratingtheseideas.Thenumber
ofsessionsisequalwiththenumberofteammembers.Evenifthemethod
recommendsthecombinationteamsizeideas/sessionminutes/session
= 635, other combinations like: 536, 437, 6310 have been
experienced,too.Infact,thecombinationsshouldbedesignedsuchastofit
bettertotheparticularsituation.
88
The graphic support of this method looks like the one in table 5.1. It
consists of a matrix with several rows and columns where ideas are
introducedwithintheboxesofthematrixaccordingtothe635algorithm.
Table5.1.Ideagenerationworksheet.
Idea 1 Idea k
Team member 1
Team member i
Team member n
The 635 method follows a welldefined working procedure, as it is

describedbelow:
Step 1: The moderator prepares the group and informs them about the
problem;
Step2:Themoderatordistributestoeachgroupmemberaworksheetlike
theoneintable5.1;
Step 3: Each groupmember will generate k ideas into his/her
correspondingrow(forexample,thegroupmembernumber2willalways
fillintherownumber2);
Step4:Whenthesessioniscompleted(afterawelldefinedintervaloftime
T), each groupmember will transfer the worksheet to his/her leftside
neighborhood and will get the worksheet from his/her rightside
neighborhood;
Step 5: The activities from the step 3 are resumed, each groupmember
being asked to generate new ideas into his/her corresponding row, but
89
having as support the ideas already generated by his/her rightside

neighborhood;
Step6:Theprocesscontinuesuntilallsessionsarecompleted.
Thewholeprocessisruninsilence.Astheworksheetscirculate,eachgroup
memberseesthecontributionsoftheothergroupmembersastheyreceive
anotherworksheet.Ateachchangeover,thenumberofideasgrowsand
thesemaystimulatethegenerationofnewideas.
Attheendoftheprocess,alotofideasshouldbegenerated.Considering
thegrouphas n membersandateachsessioneachmemberproposed k
ideas,attheendwillresultkn2ideas.Theyareanalyzedandthosethatare
consideredreliablearekeptforfurtherdevelopments.Becausenoverbal
communication between participants is required, the 635 method also
supportstheuseofInternetenvironmentfortransferringtheworksheets
betweentheteammembersiftheyarelocatedatlargedistances.Insuch
cases,thetimerequiredtocompleteonesessioncouldberedefined(tensof
minutes,hoursordays).Thegroupwillmeettogetheronlyattheendof
theprocess(atapredefineddate)foranalyzingtheresults.
Besidesthetechnicalbenefitsofthe635method,significantreductionof
timeisgeneratedincomparisonwithsomeothersimilartechniques,like
forexamplethebrainstormingapproach(23timeslessduration).
Inroboticstherearealotofsituationswherethe635methodmakesface
successfully.The635methodisveryefficientintermsofapproachingthe
processofideageneration,inthewaytheideasaresharedbetweenthe
teammembers.Incaseofproblemsthatrequirehigherlevelsofinnovation
(wherestrongtechnicalcontradictionsshouldbesolved),the635method
could be combined with ARIZ or TRIZ algorithms. These methods are
presentedinanotherchapterofthisbook.Duringtheselectionprocessof
thereliableideasfromthesetresultedattheendphaseofthe635method,
variousspecifictoolscanbeused.OneofthesetoolsisthePughmethod,
whichisalsotreatedinthisbook.
90
MindMapmethodisatoolforrapidexternalisationofideasgeneratedin
the brain. Individuals or groups canproduce aMindMap sheet, but in
bothcasesspeedisimportant.
The process starts by writing in the central area of the sheet the core
problem.Thenextstepconsistsingeneratingthefirstordersubsystemsor
areasthataredirectlyrelatedtotheinvestigatedtopic.Thenextstepisthe
identificationofthelowerorderareas. Inthisrespect,eachof the areas
fromthefirstlevelaretakeninturnandissuesrelatedtoitareidentified
andrepresentedonthemapbywords,briefphrases,pictographs,etc.
Theinvestigationcangosimultaneouslyonmorelevels.Thismeansthatit
isnotnecessarilytocarryouttheinvestigationonthefirstlevelareasand
afterthattostartitonthesecondlevelsandsoon.Ifideasaboutthetopics
onthesecondlevelsorontheotherlevelscomeduringthedevelopment
processofthefirstlevelareas,theywillbeimmediatelywritten.Oncethe
map is generated, connections betweenvarious areas will be defined. If
contradictions exist between different areas, a distinctive sign will be
placed to highlight that innovation should be brought there in order to
solvetheproblem.ToolslikeARIZcouldsupporttheinnovativeprocess.A
pictureofaMindMapsheetisshowninfigure5.2.
MindMapcouldsupporttheQFDprocessduringthegenerationofvarious
Hows (the items subjected to evaluation) within the matrices. This
shouldbedoneproperlyinconjunctionwiththesetofWhats(theinputs
into the analysis process). MindMap method is recommended for
applicationwhentheteamortheindividualintendstoclarifyandhavea
viewonthesystemthatisgoingtobedeveloped.
91
. c=contradiction
Sub-area 1.1
(Level 2) Area k
(Level 1)
Area 1
(Level 1)
Sub-area k.1
. Main (Level 2)
Sub-area 1.h .
(Level 2) topic
.
.
Area n .
(Level 1) c
.
.
Fig.5.2.Thelayoutofamindmappingprocess.
Ifthemapdevelopmentgoeswell,studentswillfindthattheyarerapidly
addingideasinaveryeasyway.Whenthishappens,itmeansthatstudents
haveestablishedagoodcommunicationwiththeirsubconscious.

The task within this workshop is divided into two phases. In the first
phase,the635methodwillbeappliedtogenerateideasfordesigningthe
conceptofaservicerobotwhichhastobeusedforcarwashing/cleaning.
In the second phase, the MindMap method will be experimented for
defining the technical subsystems and the key functions of the robot
conceptelaboratedinthefirstphase.
For the application of the 635 method the following steps will be
followed:
Step1:Studentswillbeorganizedinteamsof4people.Beforestartingthe
work,eachteammemberwillprepareaworksheetasintable5.2.
Step2: Eachteamwillstartthe635algorithm,whereforeachstep5
minutes will be allocated. At each step, each student will generate a
numberofideas,rangingfrom0to3(dependingonhis/herinspirationin
92
thatmoment).Considering4peopleinateam,theprocesswilltake20min.
Attheend,eachteamwillgenerateanumberofideas(4worksheets/team:
maximum48ideas).
Step3:Themostrelevantideasfromeachteamwillbefurtherrepresented
onaseparateworksheetandexposedonaboard.
Step4:Theresultsofeachteamwillbeanalyzedwiththewholegroupof
students.
Table5.2.435worksheetofthe635method.
Team X Idea 1 Idea 2 Idea 3
Student A
Student B
Student C
Student D
Step5:Thebestconceptwillbeselected.Inthisrespect,studentswillhave
toconsiderasetofcriteria.Someofthesecriteriacouldbe:a)robotcost;b)
robotdexterity/versatility;c)possibilitytowash/cleanawiderangeofcar
models. At least another two new criteria must be elaborated and
consideredbythestudentsintheiranalysis.
93
FortheapplicationofMindMapmethod,studentsareencouragedtouse
theMSVisionenvironment.Inthisrespect,thefollowingstepshavetobe
followed:
Step1: OpentheMSVisionenvironment.FromthewindowCategory
select Brainstorming, as it is shown in figure 5.3. Click on the icon
Brainstorming diagram and a framework with specific shapes for
performing the MindMap process is displayed. On the left side of the
screen select the icon Main topic and drag the drawing inside the
worksheet.Doubleclickonthedraggedshapeandintroducethenameof
thecoretopic.
Step2:UseDynamicconnectortodrawlinksbetweenthemaintopicand
othertopicsrelatedtoit.Tointroducenewtopics,usetheiconTopicor
Multipletopic.Thetextcanbechangedbyclickingontheselectedtopic.
The process continues until the team considers the mindmapping is
completed. Figure 5.4 shows a MS Visios screencapture of a mind
mappingprocess.
94
Fig.5.3.TheMSVisioenvironmentforMindMap.
95
Fig.5.4.AresultofamindmappingprocessinMSVisio.
96
Fig.5.5.Theresultofamindmappingprocessforacleaningrobot(selections).
97
Table5.3.Exampleofasingle435worksheetingeneratingconceptideasfora
firefighterservicerobot.
Idea 1 Idea 2 Idea 3
Capable to
investigate a
High level of hazardous area and
universality (easy relay vital Autonomous robots/
to adapt for information - details vehicle which can
Student:
domestic, industrial of heat carry out tasks using
Marius
and chemical fre) accumulation, wheel-based
see the modular ventilation points mechanisms
concept and structural
features see
camera
Autonomous robots Robust and reliable Capable of spraying
Student: with night-vision (can safely go into large amounts of
Paul cameras and heat dangerous areas) water and foam
sensors see protection case see special gun-tool
Adaptable for large-
scale petroleum
Easy to charge with
fres, hazardous
Student: water (at least 50 m
Easy to transport chemical leaks and
Mihai of water hose
fres where theres ' a
resistant to fre)
risk of building
collapse
Be able to identify
some remains and
Be able to function in could identify
Suitable for fghting the most difcult of whether certain
Student:
fres in subway circumstances such areas were close to
George
tunnels as areas of intense collapse and
heat and dust therefore too
dangerous for
human search teams
98
Fig.5.6.Sketchoftheconceptforafirefighterservicerobotfrom635method.
5.4. Example 1
Example1presentsamindmapprocessdonetodefinethecoretechnical
subsystems/characteristics for a service robot used in outdoor wall
cleaning.Selectionsoftheresultsareillustratedinfigure5.5.
5.5. Example 2
The second example reveals the results of a 635 process for concept
definitionofafirefighterservicerobot(seetable5.3andfigure5.6).
5.6. Workflow
willworkinteamsoffourpeople.Forthemindmappingprocess,oncethe
workiscarriedout,resultsaresavedinpictureformat(.jpg)andafterward
areprintedout.Individualresultswillbepostedandanalyzedinteam.
Conclusionsaroundtheresultswillbeformulated.
5.7. Homework
Do individually aMindMap analysis for defining the core aspects of a
servicerobotforsurgicaloperations.
Workshop 6 Variant Selection in Competitive Robot Design using the Pugh and Combinex
Methods
Variant Selection in Competitive

Robot Design using the Pugh and
Combinex Methods
100
101
6.1. Objective
Invariousphasesofthedesignprocess,engineersneedtoselectthebest
variantofasolutionfromseveralavailableorpossible.Becauseinmany
cases systems under consideration are complex, the selection process is
quitedifficultwithoutthesupportofpropertools.
applyingtwomethods,PughandCombinex,usedbyengineerstosupport
thesolutionselectionprocess.
6.2. Background
At different stages of the design process there are situations where
decisionsshouldbemade.Inmostofthecases,thesedecisionsareinthe
formofchoices,thatis,toestablishwhichoftheavailablealternativesfits
the best to the specified criteria. One of the most known methods for
selectingalternativesisthatdevelopedbyStuartPugh.Thebasicideaisto
compare the proposed alternatives relative to a reference variant with
respecttoasetofpredefinedcriteria.
Thesimplestsituationiswhennodegreesofimportancearegiventothe
consideredcriteria.Insuchcases,oneofthealternativesisconsideredthe
referenceandtheothersarecomparedonebyonerelativetothereference
variantallocatingasymbolwithrespecttoeachcriterionasfollow:[S]if
theanalyzedalternativesatisfiesatalmostthesamelevelthecriterionlike
the reference variant; [+] if the analyzed alternative satisfies somehow
betterthecriterionthanthereferencevariant;[]iftheanalyzedalternative
satisfies somehow worse the criterion than the reference variant. At the
end,therearecountedhowmany[S],[+]and[]areforeachalternative.
Theonehavingthehighestvaluefor [+][],with [+][]>0,will
beselected.Ifnoneoftheconsideredalternativesfromthesethasthevalue
[+][]>0,thenthereferencevariantwillbetheoneconsideredfor
furtherdevelopments.
102
Table6.1.AnexampleforthesimplestvariantofthePughselectionmethod.
No. Criterion Idea 1 Idea 2 Idea 3 Idea 4 Idea 5
1 Criterion 1 S S S S +
2 Criterion 2 Reference + + + + S
3 Criterion 3 (idea 0) S + S +
4 Criterion 4 S
5 Criterion 5 S
[S] 5 2 1 3 2 1
[+] 0 1 2 1 2 1
[] 0 2 2 1 1 3
[+] [] 0 1 0 0 1 2
Priority 2 X 4 3 1 X
Iftherearesituationswheresomealternativeshave[+][]=0andall
theothershave[+][]<0,thaneithertheonehavingthehighestvalue
of[S]orthereferencevariantwillbefurtherselected.Thissimplevariant
ofthePughmethodismainlyusedforselectingideasfromaconsidered
set.Anexampleisgivenintable6.1.
Forselectingsolutionsattheconceptualstageofproductdevelopment(in
thecaseofrobots,atthepointinwhichthekinematicchainisdefined),the
Pughmethodtakesintoaccountthedegreeofimportanceforeachcriterion
(calculated, for example, with the AHP method) and the following
comparisonscale:0if the alternative satisfies the criterionat almostthe
same level with the reference variant; 1 if the alternative satisfies the
criterionlessthanthereferencevariant;2alternativesatisfiesthecriterion
much less than the reference variant; +1 if the alternative satisfies the
criterion better than the reference variant; +2 alternative satisfies the
103
criterion much betterthan the reference variant. Other evaluation scales

canbedefined,too.Inthisrespect,theSaatysscalecouldbeoneofthem
(seetable1.1).
Foreachalternativeitiscalculatedamark,asthesumofalllocalproducts
between the degree of importance of the considered criterion and the
numbergiventotheconsideredalternativewithrespecttotherespective
criterion.Themarkcouldbe>0,<0or=0.Thebestalternativeistheone
havingthehighestpositivemark.Ifallalternativeshavethemark<0,then
thereferencevariantwillbeselectedforfurtherdevelopments.Iftwoor
morealternativeshavethesamebestmark,supplementarycriteriashould
beconsideredforrefiningtheselectionprocess.Anexamplethatusesthe
socalledroughscaleofevaluation(2,1,0,+1,+2)isgivenintable6.2.
Thesameexample,butusingtheSaatysscaleofevaluation,ispresentedin
table6.3.Itcanbeseenthat,inthecaseofSaatysscale,thevalue+1could
berefinedwithfourvalues:2,3,4or5dependingofthesituation,+2with
6,7,8or9,etc.Thiscouldbereflectedinthefinalresultsbutnotintheway
majorinappropriatedecisionstobedone.
Table6.2.AnexampleofthePughmethodforconceptselection.
No. Criterion R Concept Concept Concept Concept Concept
(Concept 0)Reference
1 2 3 4 5
1 Criterion 1 2 0 2 +2 0 +1
2 Criterion 2 3 +1 1 +1 0 +1
3 Criterion 3 5 +2 0 +1 +1 1
4 Criterion 4 4 2 0 1 1 1
5 Criterion 5 6 0 0 1 0 0
Mark 0 5 7 2 1 4
Priority 4 1 X 2 3 X
Table6.3.AnexampleofthePughmethodforconceptselectionusingSaatys
scale.
104
(Concept 0)Reference
No. Criterion R Concept Concept Concept Concept Concept
1 2 3 4 5
1 Criterion 1 2 1 0.11 7 1 3
2 Criterion 2 3 5 0.33 3 1 5
3 Criterion 3 5 9 1 3 3 0.33
4 Criterion 4 4 0.14 1 0.2 0.2 0.33
5 Criterion 5 6 1 1 0.33 1 1
Mark 20 68.56 16.21 40.78 26.80 29.97
Priority X 1 X 2 3 4
For example, in table 6.2 only at the fourth level of priority there is a
modificationincomparisonwithtable6.3.Thiswillnotaffectthedecision
becauseonlythebestalternativeistakenintoaccount,thatis,theconcept1
willbeselectedinbothsituations.
The reader should note that the Pugh method is not used to rank the
alternativesortoorderthem.ThePughmethodissimplyusedtoselectone
singlevariant,thatis,thebestpossibleonefromaconsideredset.Thereis
nocriteriontodefinewhichformatoftheevaluationmatrixiscorrector
incorrect. The important thing is that design team members to be
familiarized with the format chosen and to accept the cost and effort
involved.TheessentialroleofthePughmethodistodetermineengineers
thinkaboutasmanyaspossibledesignalternativesandtoevaluateeach
alternativeagainstasetofdesirablecriteria.Onthisway,peopleareforced
todefine criteria from the early stages of product development process,
diminishingonthiswaytheriskfordevelopingimpropersolutions.
AnalternativetothePughmethodisthe5pointVDI2221guidelinefor
concept evaluation. It works with the following point scores:
0=unsatisfactory;1=justtolerable;2=adequate;3=good;4=verygood
(ideal).Anexampleusingthe5pointVDI2221approachisgivenintable
6.4.Inthisexample,theconcept4isrejectedbecauseisdoesnotsatisfyat
allaveryimportantcriterion(thecriterion5thathastheimportanceR=7).
105
Table6.4.Anexampleofevaluationmatrixusingthe5pointVDI2221
guideline.
No. Criterion R Concept 1 Concept 2 Concept 3 Concept 4
1 Criterion 1 1 0 1 3 4
2 Criterion 2 3 2 3 3 2
3 Criterion 3 5 4 2 3 2
4 Criterion 4 4 3 2 2 1
5 Criterion 5 7 3 1 2 0
Mark 59 35 49 24
Priority 1 3 2 X
Theconcept1willbefinallyselectedbecauseithasthehighestmark,even
ifitdoesnotsatisfyonecriterion(criterion1thathasaverylowimportance
R=1).
Pugh method can be slightly modified to select the best variant on the
marketforacertainsystem.Inthisrespect,asetofmeasurablecriteriaand
theirrankswillbefirstlydefined.TheAHPmethodisusedtorankcriteria.
Thesecriteriacouldbecataloguedata,forexample.
Acriticallimitforeachcriterionwillbeestablished.Itisveryimportantto
see what does minimum and maximum really mean when define the
criticallimit?Forexample,maximizationinthecaseofrobotrepeatability
meanstohaveasloweraspossiblevaluesofthischaracteristic.
Table6.5.AnexampleofthemodifiedPughmethodinassemblyrobotselection.
(Model 1)Reference
Robot
No. UM R Limit
characteristic
Model 1
Model 3
Model 2
Model 2
Model 3
1 Number of axes 1 4 4 4 4 100 100 100

2 Max speed in joints m/s 1 4.5 5.28 5.80 5.32 100 109.84 100.75
106
3
3 Working space cm 2 300000 907460 1406563 333679 100 155 36.77
4 Accuracy mm 4 0.25 0.12 0.10 0.20 100 120 60
5 Repeatability mm 4 0.030 0.020 0.025 0.030 100 80 66.66
6 Maximum payload kg 2 10 10 15 10 100 150 100
7 MTBF h 2 9500 10000 12000 11000 100 120 110
8 Cost $ 2 70000 65000 66000 55000 100 98.48 118.18
Weight 1800 1817 1437
Best ? ?
Remark: data are taken from real SCARA type robot models (their names are
not given from confdentiality reasons).
Afterthedeterminationoftheboundaryvalueforeachcriterion,oneofthe
alternativeswillbeconsideredasreference.Forthereferencevariantwill
be given a 100% achievement with respect to each criterion. The other
variantswillgetpointsapplyingthesimple3rulerelativetothereference
variant.Attheend,thescoreisgivenbyaddingthelocalscores,calculated
as the product of the degree of importance and the calculated local
percentages.Anexamplethatclarifiestheapproachisgivenintable5.5.In
this example, the model1and the model2have veryclose weights. In
order to decide which of the two solutions will be selected, some
supplementary aspects should be considered. The team members may
select one of the important criteria as a supplementary subject of
discussion.If,intheexamplefromtable6.5,repeatabilityisconsideredabit
morecriticalfortheassemblyprocess,thanthemodel1willbechosen.
Inordertosolvebettersuchdelicatesituations,thesocalledCombinex
methodcouldbeused,too.Thismethodismuchmorecomplexthanthe
Pughmethod.Themethodisappliedin7steps.Theyarefurtherpresented:
Step 1: Identification of selection criteria: they should be quantifiable and
measurableparameters.
Step2: Rankingtheselectioncriteria: variousmethodscouldbeappliedto
carryoutthisstep.TheAHPmethodisoneofthem.
107
Step3: Developmentofutilitycurves: thisisthestepthatdifferentiatesthe

CombinexmethodfromthemodifiedPughmethod.Eachcriterionistaken
and the engineers should determine the value that would make the
customer100%satisfiedandthevaluethatwouldmakethecustomer0%
satisfied. After that, pairs of values and their related satisfaction
percentages,inbetweenthosecorrespondingto0%and100%,shouldbe
determined. Using the resulted data, the socalled utility curve (or
satisfaction curve) will be built. The utility curves are tools for
quantifyingthecustomerlevelofsatisfaction.
Step4: Assignmentofscorestoeachvariant: eachalternativeisassessedin
turnagainsteachofthecriteriaandscoresaregivenbasedontheutility
curves.
Step5:Alternativeevaluation:foreachalternative,therankofeachcriterion
ismultipliedwiththecorrespondingscore,resultingthetruevalueofthe
alternativeagainsttheconsideredcriterion.Thesumofalllocalweightsis
performed,obtainingtheglobalweightofeachalternative.
Step 6: Bestoftheworld selection: looking to the alternative with the
highestweight,thebestvariantwillbeselected.
Step7:Applyingatestofreasonableness:asintheexamplespresentedbefore,
apurejudgmentbasedonlyonfiguresisnotrecommended.Inthisrespect,
a supplementary analysis based on experience should be performed in
ordertodecideonthefinalvariant.
Forexemplification,therobotmodelsandthecriteriafromtable6.5are
further considered. For each of the 8 criteria the utility curve has to be
defined.Theresultsarepresentedinfigure6.1andtable6.6.
108
No. axes Speed joints Working space Accuracy

100% 100% 100% 100%
90%
50% 50% 50% 50%

3 4 6 4.5 5.5 6 8 300000 1000000 0.05 0.25
Repeatability Payload MTBF Cost

100% 100% 100% 100%
85% 90%
80%
50% 50% 50% 50%
0.01 0.03 10 15 21 25 9500 10000 20000 40000 65000 70000
Fig.6.1.Utilitycurvesfortheselectioncriteriaintheexamplefromtable6.6.
Inthe example from table 6.6, the model 2 willbe selectedbecause the
weightinthecaseofCombinexmethodisequivalentwiththetotalvalueof
theproductinthecustomerseyes.
Table6.6.AnexampleoftheCombinexmethodinassemblyrobotselection.
(normalized to 1)R
No. Robot characteristic UM R

Model 1
Model 2
Model 3
Model 1
Model 2
1 Number of axes 1 0.06 4 4 4 50.0 50.0 Model 3

50.0
2 Max speed in joints m/s 1 0.06 5.28 5.80 5.32 81.2 96.0 82.8
3
3 Working space cm 2 0.11 907460 1406563 333679 93.4 100.0 52.4
4 Accuracy mm 4 0.22 0.12 0.10 0.20 95.0 99.0 85.0
5 Repeatability mm 4 0.22 0.020 0.025 0.030 92.0 85.0 50.0
6 Maximum payload kg 2 0.11 10 15 10 50.0 85.0 50.0
7 MTBF h 2 0.11 10000 12000 11000 90.0 92.0 91.0
109
8 Cost $ 2 0.11 65000 66000 55000 80.0 74.0 88.0

Weight 83.49 87.85 68.62
Best *
Remark: data are taken from real SCARA type robot models (their names are not
given from confdentiality reasons).

The task within this workshop is divided into two phases. In the first
phase,thePughmethodwillbeappliedtoselectthebestconceptofarobot
gripperfromthe3availableconceptspresentedinfigure6.2,figure6.3and
figure 6.4. The following selection criteria will be considered (see, in
brackets,theirdegreeofimportance):
Lowdrivingforces[R=4];
Lownumberofparts[R=5];
Easytorecycle[R=2];
Highreliability[R=4];
Easytomaintain[R=3];
Easytomanufacture[R=3];
Easytoassembly/disassembly[R=1].
Fig.6.2.TherobotgrippermodelA.
110
Fig.6.3.TherobotgrippermodelB.
Fig.6.4.TherobotgrippermodelC.
Inordertoperformthistask,thetechnicalplatformfromQualicaQFDcouldbe
considered.Inthisrespect,whenopentheQualicaQFDenvironment,selectthe
templatePughNewConceptSelection,asitisshowninfigure6.5.Awindow
liketheoneinfigure6.6willbeopened.
111
Fig.6.5.ThewindowinQualicaQFDforselectingthetemplatewhichsupportsthe
applicationofthePughmethod.
112
Fig.6.6.TheworksheetforPughmethodinQualicaQFD.
113
Fig.6.7.ScreencaptureshowinghowconceptsareevaluatedinQualicaQFD.
114
Fig.6.8.ScreencaptureshowingtheresultsofaPughprocessinQualicaQFD.
Figure6.7showshowthePughmethodisapproachedinQualicaQFD.For
thefirstvariantitisselectedthesymbolOneutralwithrespecttoeach
criterion.Forthenextvariants,itwillbeselectedasymbolthatcomesout
from the comparison of that variant with the first one (considered as
reference), withrespect toeachcriterion. At the end of the process, the
resultsaredisplayedlikeinfigure6.8.
In the second phase of the workshop, the Combinex method will be
experimentedforselectingthebestrobotmodelfromthethreeavailablein
figure 6.9, figure 6.10 and figure 6.11. The selection criteria are those
presentedbelow:
1. Highpayloadcapacity[kg];
2. Attractiveprice[$];
3. Maximumreachofhighvalue[mm];
4. Highspeedinrobotarmjoint[/s];
5. Continuouspathcontrol[#];
6. Guarantiesconditions[years];
115
7. Goodrepeatability[mm];
8. Highworkingforcesforassemblyoperations[N];
9. Highreliability[hours];
10. Shortdeliverytime[days];
11. Freecostinstallationservices[hours];
12. Trainingcosts[$];
13. Aftersaleservices[$/hour];
14. Offlineprogrammingfacilities[#].
Fig.6.9.RobotmodelA.
Fig.6.10.RobotmodelB.
116
Fig.6.11.RobotmodelC.
Studentswillchoosethedegreeofimportanceforeachcriterion.Also,the
utility curve of each criterion stays in the students responsibility. Both
degreesofimportanceandutilitycurvesmustbedefinedsuchthattobe
appropriateforthepurposethesetypeofrobotsaregoingtobeused.For
thisexercise,itisconsideredthatthemainapplicationoftheselectedrobot
willbespotwelding.
Table6.7.Datafortherobotmodels.
No. Criterion UM Model A Model B Model C
1 Payload capacity kg 25 30 26
2 Price $ 100 000 99 000 102 000
3 Maximum reach mm 1 570 1 750 1 800
4 Average speed in robot arms /s 120 130 125
nd rd
5 CP control # circular 2 spline 3 spline
6 Guaranty years 2 3 3
7 Repeatability mm 0.15 0.20 0.10
8 Working force N 2 500 2 400 2 100
9 Reliability h 9 600 9 800 11 000
10 Delivery time days 30 25 20
11 Free-cost installation services h 15 30 35
12 Training costs $ 2 500 3 200 2 600
13 After-sale services $/h 500 450 520
14 Of-line programming # no yes yes
117
Table6.7presentsthedataofthethreerobotmodels,astheyareprovided
bytherobotmanufacturers.
6.4. Workflow
willworkindividually.Oncetheworkiscarriedout,resultsareprinted
out.Individualresultswillbepostedandanalyzedinteam.Conclusions
aroundtheresultswillbeformulated.
6.5. Homework
DoindividuallyaPughanalysisforselectingaspraypaintingrobot.Define
yourownselectioncriteria.RankthemusingtheAHPmethod.
118
Workshop 7 Causality Analysis in Competitive Robot Design using the AIDA Method
Causality Analysis in Competitive

Robot Design using the AIDA Method
120
121
7.1. Objective
Many times during the design process, engineers should take
interdependentdecisions.Twoormoredecisionsareinterdependentwhen,
consideringacertaindecisiontheboundariesoftheotherdecisionorofthe
otherdecisionsaredisturbed.
applyingtheAIDAmethodforcausalityanalysisduringthecompetitive
designprocess.
7.2. Background
TheAIDAmethod(AnalysisofInterconnectedDecisionArea)isawell
knownapproachformappinginasystematicwayallcompatiblesolutions.
Inotherwords,AIDAmethodleadstothedefinitionofallcombinationsof
optionsthatarefeasiblewithrespecttotheconsideredinterdependencies.
Thefirststepinsettinguptheanalysisconsistsinidentifyingthedecision
factors, called decision areas. Such decision factors could be: the
functionstheproductshouldfulfill,theworkingprinciples,theshapeof
the constitutive parts of the product, the materials, the manufacturing
technologies,etc.Theselectionofthedecisionareasisstronglyconnected
withtheconceptualconstructivedesignphaseoftheproduct.
Duringtheconceptualstageoftheproduct,decisionsaremoregeneraland
comprehensivethaninthenextstagesofproductdevelopment.Thesecond
step in deriving the AIDA method consists in identifying the options
withineachdecisionarea.Oncetheseoptionsareestablished,asocalled
option graph is built. Within this graph, two options that are not
compatible are linked with a line. The option graph is the mean for
identifying compatible solutions to a certain design problem, that is, to
establishallcombinationsofoptions(oneoption/decisionarea)thatdonot
containanypairofincompatibleoptions.Anexampleofoptiongraphis
122
given in figure 7.1. For handling complex applications, the use of

specializedsoftwarepackagesisrecommended.
Fig.7.1.AnexampleofoptiongraphinthecaseofAIDAmethod.
Whengeneratingoptions,engineersshouldworkoutacausalityanalysis
worksheetwhere thesetofproposedoptionswithinadecisionareaare
analyzedagainstasetofglobalconstrainingfactors.Inordertoidentify
options that are less sensible tovarious constrains, inpractice there are
usedspecialtablesforcausalityanalysis.Suchaframeispresentedintable
6.1.Itconsistsofthreeareas.Inthefirstarea,theinfluenceofeachconstrain
oneachoptionisinvestigated.Inthisrespect,threelevelsofinfluenceare
considered:1,2and3,with1thehighestlevel.Intable6.1,thecoefficient
aij,i=1,,n;j=1,,m,showsthelevelofinfluenceoftheithconstrainon
the jthoption.Inthesecondarea,theinfluenceofeachoptiononeach
constrain is investigated. The same levels of influence are used and the
sameruleisappliedasinthefirstarea.Intable6.1,thecoefficientbij,i=1,
123
,m; j=1,,n,showsthelevelofinfluenceofthe ithoptiononthe jth

constrain.Inthethirdarea,theproductsaijbji,i=1,,n;j=1,,m,areput
intoevidence.Thesumsalongeachrowandalongeachcolumnarefurther
performedandafterthat,dividedwiththenumberofoptions(m)andthe
numberofconstrains(n)respectively.
TheconstrainhavingthelowestvalueofC/mhasthestrongestcausality
onthesetofoptions.TheoptionhavingthelowestvalueofO/nhasthe
strongestcausalityonthesetofconstrains.Aslongasthesetofconstrains
isreferringtothewholesystemthatshouldbedesigned,theoptionhaving
thehighestvalueofO/nfitsbettertotheneeds,becauseithasthelowest
causalityonthesetofconstrains.
124
Table7.1.Causalityanalysisworksheet.
Worksheetsliketheoneintable7.1canbepreparedinthewellknown
softwareenvironmentMicrosoftExcel.
125

ThetaskwithinthisworkshopistoapplytheAIDAmethodduringthe
designprocessofanspraypaintingindustrialrobot.Beforestartingthis
process, students must consult various commercial available spray
painting robots, as well as technical requirements of a spraypainting
operation. Students can use the Internet facilities to collect data in this
respect.

Fig.7.2.Examplesofspraypaintingindustrialrobots.
Examplesofspraypaintingindustrialrobotsareshowninfigure7.2.Their
CAD models are realized with the WorkspaceTM environment. In this
workshop, the following decision areas are important for the spray
paintingrobotunderconsideration:(1)positionofthemotorsfortherobot
wrist; (2) type of transmission system forthe robot wrist; (3) robot arm
configuration;(4)robotwristconfiguration;(5)typeofdrivingsystem.
To performthe taskinthis workshop, students have to follow the next
steps:
Step1:Alistwithperformancecriteriaoftherobotmustbeworkedout.It
mustresultfromadeepanalysisandunderstandingoftheperformances
requiredforaspraypaintingoperation.
126
Step2:Foreachdecisionarea,alistofoptionsmustbedefined.Atleast3
options/decisionareahavetobeworkedout.
Step3: Foreachdecisionarea,acausalityanalysismustbeperformed.In
thisrespect,aworksheetofcausalityanalysisforeachdecisionareamust
bepreparedinMicrosoftExcel.
Step 4: The option graph and the simplified option graph have to be
completed afterward. The MS Visio environment can be used in this
respect.SelectthecategoryBusinessprocessandafterwardthetemplate
TQMDiagram.Theresultshouldlookasfigure7.3reveals.
Using the shapes called Operation and Line connector, the option
graphcanbeeasilybuilt.Inordertodrawconnectorswithnoarrowsat
theirends,afterthelineisdraggedandproperlylocated,rightclickonthe
line, select Format and afterward Line. A window with various
optionsisopened.InthatwindowatthegroupLineendsselectBegin
andEndwiththeoption00:None.Tointroducetextwithinthegraph,
selecttheshapecalledTextblock.Figure7.4showshowanoptiongraph
isdoneinMSVisio.
127
Fig.6.3.ThetemplateselectedinMSVisiofordrawingoptiongraphs.
128
Fig.6.4.AnexampleofoptiongraphrealizedinMSVisio.
Step5:ThecompatiblevariantshavetobeselectedattheendoftheAIDA
process.
7.4. Example 1
ForunderstandingbetterhowtheAIDAmethodworks,afirstexampleis
presented in continuation. In a certain stage of the development of an
articulatedrobot,thefollowingdecisionareashavebeendistinguished:(1)
position of the motor for the upper arm; (2) type of speed reducer; (3)
constructionofthebody;(4)typeoftransmissionsystemfortheupperarm.
In the previous phases of the project, a number of decisions have been
alreadytaken:
thedrivingsystemwillusea.c.typemotors;
theupperarmoftherobothastobestaticallybalanced;
thecenterofgravityofthefrontsideoftheupperarmtobeascloseas
possibletotherotatingjoint;
129
thelowerarmshouldbeaslightaspossible;
themotoractingtheupperarmshouldhavealowmass.
The options for each decision area are presented in table 7.2. For each
optionthe causality analysis is carriedout. The results are presentedin
tables7.3,7.4,7.5and7.6.Inthesetables,the5decisionsthathavebeen
previouslytakenareconsideredasconstrains.
Table7.2.Optionsfortheconsidereddecisionareas.
Decision area Option
A1. Near the joint of the upper arm
A. Position of the motor for the
A2. At the base of the lower arm
upper arm
A3. Direct drive system
B1. Harmonic gearing

B. Type of speed reducer B2. Planetary gearing
B3. Cycloid gearing
C1. Classical structure

C2. Modular structure
C. Construction of the body
C3. Modular structure with extensible
arm
D1. No transmission (excepting the
D. Type of transmission system for speed reducer)
the upper arm D2. Parallelogram mechanism
D3. Screw-nut and bar mechanism
Table7.3.ThecausalityanalysisfortheoptionsinthedecisionareaA.
A1 A2 A3
a 3 3 3
b 3 3 3
c 3 3 3
d 2 3 1
e 3 3 3
a b c d e
130
A1 1 3 3 1 3
A2 3 3 3 1 3
A3 1 3 3 1 1
A1 A2 A3 C/3
a 3 9 3 5.0
b 9 9 9 9.0
c 9 9 9 9.0
d 2 3 1 1.6
e 9 9 3 7.0
O/5 6.4 7.8 5.0
Table7.4.ThecausalityanalysisfortheoptionsinthedecisionareaB.
B1 B2 B3
a 1 1 1
b 3 3 3
c 3 3 3
d 3 1 2
e 3 3 3
a b c d e
B1 2 3 3 3 3
B2 2 3 3 1 3
B3 2 3 3 2 3
B1 B2 B3 C/3
a 2 2 2 2.0
b 9 9 9 9.0
c 9 9 9 9.0
d 9 1 4 4.6
e 9 9 9 9.0
O/5 7.6 6.0 6.6
According to the results in table 7.3, the best option is A2, that is, the
positionofthemotorfortheupperarmtobeatthebaseofthelowerarm.
Itcanbealsoseenthattwoconstrainsarelesscriticalwithrespecttothe
131
optionsinthedecisionareaA.Theyare:(b)theupperarmoftherobothas
tobestaticallybalanced;and(c)thecenterofgravityofthefrontsideofthe
upperarmtobeascloseaspossibletotherotatingjoint.Thismeansthat
thefulfillmentofthesetwopreviousdecisionswillnotbedependentonthe
decisionsthatwillbetakenatthisstageofdevelopmentwithintheareaA.
Table7.5.ThecausalityanalysisfortheoptionsinthedecisionareaC.
C1 C2 C3
a 3 3 3
b 3 3 2
c 3 3 1
d 3 3 2
e 3 3 1
a b c d e
C1 3 3 3 3 3
C2 3 3 3 3 3
C3 2 1 1 2 1
C1 C2 C3 C/3
a 9 9 6 8.0
b 9 9 2 6.6
c 9 9 1 6.3
d 9 9 4 7.3
e 9 9 1 6.3
O/5 9.0 9.0 2.8
Table7.6.ThecausalityanalysisfortheoptionsinthedecisionareaD.
D1 D2 D3
a 3 3 3
b 3 3 3
c 3 3 3
d 1 3 3
e 3 3 3
a b c d e
132
D1 3 3 3 1 1
D2 3 3 3 3 3
D3 3 3 3 3 3
D1 D2 D3 C/3
a 9 9 9 9.0
b 9 9 9 9.0
c 9 9 9 9.0
d 1 9 9 6.3
e 3 9 9 7.0
O/5 6.2 9.0 9.0
Theresultsintable7.4revealthattheoptionB1(harmonicgearing)inthe
areaBisthebestchoice.Theresultsintable7.5showthattwooptionsC1
(classicalstructure)andC2(modularstructure)areavailable.Theresultsin
table7.6defineD2(parallelogrammechanism)andD3(screwnutandbar
mechanism) as two possible alternatives. Using information from tables
7.37.6,theoptiongraphcanbesimplified,thatis,toworkintheareaA
onlywithA2,intheareaBonlywithB1,intheareaCwithC1andC2and
in the areaD withD2 and D3. The simplifiedoption graph is given in
figure7.5.
133
Fig.7.5.Thesimplifiedoptiongraphfortheexampleintable7.2.
According to the results in figure 7.5, the following combinations are

compatible: A2B1C1D2 and A2B1C1D3. The engineers could choose
anyofthetwovariantsforfurtherdevelopments.
7.5. Example 2
ThesecondexampleconcernswithaprojectwhereAIDAmethodhasbeen
usedtodefineconsistentsolutionsofthelayoutforanassemblyrobotic
cell.Thefollowingdecisionareashavetobetakenintoconsideration:a)
robotpositionwithinthecell;b)typeofrobot;c)typeoffeedingsystem;d)
type of robot gripper to handle parts. The parts which have to be
assembledareillustratedinfigure7.6.
134
Fig.7.6.Thepartassembleunderconsiderationinexample2.
Table7.7.Theoptiontableforexample2.
Decision area Option
A1. On the foor
A. Robot position within the
A2. Suspended
cell
A3. On a column
B1. SCARA type
B. Robot type B2. TRICEPT type
B3. STEWART type
C1. Index table & belt feeder
C. Type of feeding system C2. Belt feeder
C3. Modular kits in magazines
D1. Two fxed grippers with tool exchanger
D. Type of gripper D2. Multi-gripper
D3. Multifunctional gripper
AmajorobjectiveindefiningtheroboticcellistoachieveI.anaffordable
solution. To this, the II.working force and the III. accuracy of the
robotareanothertwocriticalfactors.Thesekeyfactorsmustbeguiding
lineswhenconsidertheoptions.
135
Accordingtothis,theoptionspresentedintable7.7havebeenselected.The
causalityanalysisfortheoptionsbelongingtothedecisionareaAisshown
intable7.8.
Accordingtotheresultsintable6.8,thebestoptionisA1.Onthefloor.It
can be also seenthat two constrains are less criticalwithrespect tothe
optionsinthedecisionareaA.Robotpositionwithinthecell.Thetwo
less critical constrains are the following: affordable solution; and
working force. This means that the fulfillment of these two previous
decisionswillnotbedependentonthedecisionsthatwillbetakenatthis
stageofdevelopmentwithintheareaA.
Table7.8.ThecausalityanalysisfortheoptionsinthedecisionareaA
A1 A2 A3
I 3 2 2
II 3 3 3
III 3 3 2
I II III
A1 3 3 3
A2 2 3 2
A3 3 2 2
A1 A2 A3 C/3
I 9 6 6 7.0
II 6 9 6 7.0
III 9 6 4 6.3
O/3 8.0 7.0 5.3
Tables7.97.11showthecausalityanalysisfortheareasB,CandD.
Table7.9.ThecausalityanalysisfortheoptionsinthedecisionareaB.
B1 B2 B3
I 3 2 1
II 3 2 1
III 2 2 1
136
I II III
B1 3 3 2
B2 2 2 2
B3 3 1 1
B1 B2 B3 C/3
I 9 6 2 5.6
II 6 4 2 4.0
III 6 2 1 3.0
O/3 7.0 4.0 1.6
Table6.10.ThecausalityanalysisfortheoptionsinthedecisionareaC.
C1 C2 C3
I 2 2 3
II 3 2 2
III 3 1 2
I II III
C1 2 3 1
C2 2 2 3
C3 3 2 1
C1 C2 C3 C/3
I 4 6 3 4.3
II 6 4 6 5.3
III 9 2 2 4.3
O/3 6.3 4.0 3.6
Table6.11.ThecausalityanalysisfortheoptionsinthedecisionareaD.
D1 D2 D3
I 3 2 1
II 2 3 3
III 3 1 1
I II III
D1 3 2 2
D2 2 1 1
137
D3 1 2 2
D1 D2 D3 C/3
I 9 4 2 5.0
II 4 3 3 3.3
III 3 2 2 2.3
O/3 5.3 3.0 2.3
Fig.7.7.Theoptiongraphforexample2.
Theresultsintable7.9revealthattheoptionB1(SCARAtype)intheareaB
isthebestchoice.Theresultsintable6.10showthatoneoptionC1(index
table&beltfeeder)isthemostsuitablesolution.Theresultsintable7.11
defineD1(twofixedgripperwithtoolexchanger)asthebestalternative.
Using information from tables 7.8 7.11, the option graph can be
simplified,thatis,toworkintheareaAonlywithA1,intheareaBonly
withB1,intheareaCwithC1andintheareaDwithD1.Thesimplified
optiongraphisillustratedinfigure7.7.
Accordingtotheresultsinfigure7.7,thebestcompatiblecombinationis:
A1B1C1D1.
138
7.6. Workflow
Thetaskistobeperformedin2hoursoflaboratorywork.Studentswillrun
thetaskindividually.Attheendofthework,resultswillbeprintedand
analyzedingroup.Conclusionsaroundtheresults,aswellasaroundthe
potentialofAIDAmethodinengineeringdesignwillbeformulated.
7.7. Homework
ApplytheAIDAmethodtodevelopacustomizedtrainingprograminthe
field of robotics. Apply the same tool to optimize the curricula for a
specializationinindustrialrobots.
Workshop 8 Innovative Problem Solving in Competitive Robot Design using the TRIZ and
ARIZ Methods
Innovative Problem Solving in

Competitive Robot Design using the
TRIZ and ARIZ Methods
140
141
8.1. Objective
There are many situations in practice when engineers have to solve
conflictingproblems.Usually,engineersmakecompromisesandtradeoffs
between conflicting features that describe the technical system and thus
theydonotachieveanidealsolution.
According to many countless opinions, over 90% of the problems that
engineersfacehavebeensomewheresolvedbefore.Inthisrespect,mostof
the solutions could be derived from knowledge and information that
already exist. In solving difficult problems, engineers should find a
guiding path from their own knowledge and experience to the
knowledgeandexperiencesthatalreadyexistintheirfieldofactivityorin
otherindustries.
TRIZ is one of the methods that help engineers in solving difficult
technologicalproblemsthroughidentificationandeliminationofconflicts
that are presented within systems (see, for example, the negative
correlationsintheroofmatrixofQFD).
applyingtheTRIZmethodandtheARIZmethod(aderivativeofTRIZ)in
solvinginnovativelyconflictingproblemsinengineeringdesign.
8.2. Background
TRIZisasystematicapproachforgeneratingcreativesolutions,ifthereare
possible, using the present day scientific progress. The father of TRIZ
method is the Russian scientist Genrich S. Altshuller and TRIZ is the
RussianacronymofwhatinEnglishiscalledTheoryofInventiveProblem
Solving(TIPS).Altshullerstartedworkingonthismethodatthemidof
the 20th century and generated the first theories in solving problems
inventivelyafterstudyingover200000patents,whenheconcludedthat
thereareabout1500technicalcontradictionsthatcanbesolvedrelative
easybyapplyingasetoffundamentalprinciples.Uptotheendofthe20 th
142
century, Altshullers team has screened over two million patents. From
these,only40000incorporatedinventivesolutions.
Fig.8.1.Ageneralpatterninsolvingrelativesimpleengineeringproblems.
Aninventivesolutionisonethatsolves,withnocompromise,theconflicts
thatexistbetweentwoproblems.
Inpractice,therearetwocategoriesofproblemspeoplehavetomakeface.
The first category includes problems with already known solutions that
people usually find out from books, journals, reports or from experts.
Figure8.1showsthegeneralpatternthatisusedtosolvesuchcases.
Thesecondcategoryincludesproblemswithunknownsolutionsthatmay
containcontradictoryaspects.Altshullercalledtheminventiveproblems.
Attempting to solve such problems, people try to apply trialanderror
approaches or some improved tools like brainstorming, morphological
analysis, etc. These methods are simple and easy to learn and use, but
when ideas are generated without rules the problemsolving process
remainsstochasticandwhenappliedtochallengingproblemstheyremain
too intuitive to stimulate significantly the creativity. The use of
psychologicaltoolsinsolvinginventiveproblemsmakesverydifficultor
impossiblethetransferofexperienceandintuitiontootherpeople.When
applyingpsychologicaltools,mostofthestepsintheattempttosolvethe
problem follow a vector of psychological inertia. This vector is
determined by a complex set of constraining factors depending of
recognizedperceptions,ownexperience andknowledge,commonsense,
culturalbackground,etc.thatcausessearchingforsolvingtheproblemin
traditional directions, while the ideal solution may lie far from these
directions.
143
TRIZmakesastepforwardinapproachingthesolvingproblemprocess.It
uses a structured and scientific formulated algorithm in generating
powerfulsolutions.Thealgorithmisfocusedonsupportingthegeneration
ofsuperiorengineeringsolutionsthatareabletoeliminateconflictswithin
technicalsystemswithnocompromises.
TheTRIZmethodisbuiltontwoimportantlaws.Thefirstlawisnamedthe
Law of Increasing Ideality. According to this law, technical systems
should evolve toward increasing their degrees of ideality. Ideality is
definedasbeingtheratiobetweenthesumofusefulfunctionsandeffects
thatgovernthesystemandthesumofitsharmfulfunctionsandeffects,as
itisshownbelow:
n
UF
i 1
i
ID m .(8.1)
HF
j 1
j
In(8.1)thefollowingnotationsareused:IDideality,UFitheithuseful
function/effect, UFj the jthharmfulfunction/effect, n thenumberof
useful functions/effects, mthe number of harmful functions/effects.
Useful functions/effects are described by the benefits in system
functioning. Harmful functions/effects are described by all undesired
inputsinsystemdesignsuchascosts,consumedtimeandenergy,danger,
footprints,etc.Accordingto(7.1),theidealstateofasystemiswhereonly
benefitsandnoharmfuleffectsexist.TheTRIZmethodintendstomove
thesystemtoitsidealstate,whichcouldbeastatewherethemechanismis
absentbutitsfunctionsarepresent.Thelawofincreasingidealityforces
engineerstoeliminateorsolveanytradeoffordesigncontradiction.
ThesecondlawofTRIZsaysthatobjectiverulesgoverntheevolutionof
technical systems. Altshuller called it the Law of Engineering System
Evolution.Thereareeightmainrulesofevolution,asfollows:
144
Rule1:Emergenceofanewsystem:Anewtechnicalsystememergewhenthe
following situations appear: (a) there is a need; (b) there is at least a
functionalabilitytomeetthatneed.
Rule2:Increasingideality: Afterthesystememerges,ithastobeimproved
for providing more useful functions for less harmful functions. This
requiresboththeincorporationofresourceswithinthesystem(oraround
it)andtheeliminationofsome(orall)ofthecontradictionsinthesystem.
Rule3:Unevendevelopmentofsubsystems:Intime,thesolutionsforsomeof
theconstitutivesubsystemsadvancesfasterfromtechnicalpointofview
thanthesolutionsrelatedtotheothersubsystems,creatingdissonances.
Engineerswillhavetoworkatimprovingthelaggingsubsystemsinorder
toreducediscrepanciesinthesystem.
Rule4:Systemdynamics:Asengineeringsystemsevolve,theybecomemore
dynamic,beingabletoperformmorefunctionsorincreasingtheirdegree
offreedom.
Rule5:Transitiontobiandpolysystems:Combiningatechnicalsystemwith
anotherwillleadtoahybridsystemthatincorporatesbenefitsfromthetwo
initialsystems.
Rule6:Harmonizationofrhythms: Asasystembecomesclosertoitsideal
state,thedissonancesbetweenitsconstitutivesubsystemsarereducedtoa
pointthat furtherevolutionof the system is only incremental. Alonger
periodofstabilityinsystemdesignandfunctionsisreachedatthislevel.
Rule7:Transitiontomicrolevelsandincreaseduseofenergyfields:Atacertain
levelofevolution,systemsbecomemoreefficientintransferringtheenergy
into the useful functions. The use of molecular and energy fields is
incorporatedintothesystemandveryprobablythesystemwilldecreasein
size.
Rule 8: Increasing levels of automation: In its ultimate state, the system
becomestotallyautoadaptable(autonomousandselfregulating)withno
needofhumanintervention.
145
Fig.8.2.StepsinapplyingTRIZ.
WhenapplyTRIZ,thefollowingstepshavetobeconsidered(seealsoFig.
8.2):
Step1:Definetheparticularproblem: Thesystemhastobedescribedbythe
followingitems:(a)systemcomponents;(b)operatingenvironmentofthe
146
system;(c)resourcerequirements;(d)primaryusefulfunctionsandeffects;
(e)primaryharmfulfunctionsandeffects;(f)idealresult.
Step 2: Reformulate the problem in terms of TRIZ philosophy: The problem
beingstudiedhastobereformulatedintermsofphysicalcontradictions
and all possible drawbacks that could appear should be identified and
analyzed in terms of tradeoffs, influences on other characteristics that
define the system, potential for occurrence of other problems trying to
solvetheactualproblem,etc.ThisstepisknowninliteratureasthePrism
ofTRIZ.
Step3:Searchinasystematicwayforpreviouslywellsolvedsimilarproblemsand
identify the closest one to your needs: Searchingforpreviously wellsolved
problems requires the use of large knowledge databases (patents, etc.).
Altshullers studies on over 2 million patents have conducted to the
identification of 39 standard technical characteristics that may cause
conflicts within a system. Altshuller called these characteristics
Engineering Parameters. They are presented in table 8.1. From these
engineering parameters, there should be found those that are in
contradictioninthesystem.Inthisrespect,afirststepconsistsinfinding
themainparameterthatneedstobechangedwithinthesystem.Afterthat,
the parameter being in conflict with the main parameter has to be
identified.Basedonthetwoparameters,thestandardtechnicalconflictwill
befurtherstated.
Table8.1.EngineeringparameterscausingconflictsaccordingtoAltshuller.
No. Characteristic No. Characteristic No. Characteristic
1 Weight of moving 14 Strength 27 Reliability

object
2 Weight of nonmoving 15 Durability of moving 28 Accuracy of
object object measurement
3 Length of moving 16 Durability of 29 Accuracy of
object nonmoving object manufacturing
147
4 Length of nonmoving 17 Temperature 30 Harmful factors

object acting on object
5 Area of moving 18 Brightness 31 Harmful side efects
object
6 Area of nonmoving 19 Energy spent by 32 Manufacturability
object moving object
7 Volume of moving 20 Energy spent by 33 Convenience of use
object nonmoving object
8 Volume of 21 Power 34 Repair-ability
nonmoving object
9 Speed 22 Waste of energy 35 Adaptability
10 Force 23 Waste of substance 36 Complexity of device
11 Tension/ Pressure 24 Loss of information 37 Complexity of control
12 Shape 25 Waste of time 38 Level of automation
13 Stability of object 26 Amount of substance 39 Capacity/
Productivity
Step 4: Look for analogous solutions: Having defined the physical

contradiction in the system, the next step consists in trying to find a
solution to solve it properly. The typical approaches in solving
contradictionsare:(a)ignorethem;(b)acceptacompromise.Thisisnot
desirablebecausethesystemcannotbeimprovedinameaningfulway.
The best approach is that of trying to find an elegant solution that
eliminatesthecontradiction.Inhisresearchesonpatents,Altshullerfound
that a relatively small number of principles, which he called Inventive
Principles,areabletosolvemostoftheengineeringcontradictions.There
are40inventiveprinciplesthatcouldsupportengineersindefininghighly
inventivesolutionstoaconflictingproblem.Theyarepresentedintable8.2.
Theseprincipleshavealargeapplicabilityineliminatingcontradictionsin
variousengineeringfieldsanddisciplines.
148
Table8.2.Theinventiveprinciples.
No. Inventive principle Description
1 Segmentation a) Divide an object into independent parts
b) Make an object sectional (for easy assembly and
disassembly)
c) Increase the degree of an objects segmentation
2 Extraction a) Extract (remove or separate) a disturbing part
(Extracting, or property from an object
Retrieving, b) Extract only the necessary part or property from
Removing) an object
3 Local Quality 1. Transition from a homogeneous structure of
an object or outside environment (action) to
a heterogeneous structure
2. Diferent parts of the object have to carry out
diferent functions
3. Each part of the object to be placed under
the most favorable conditions for its
operation
4 Asymmetry a) Replace a symmetrical form with an
asymmetrical form or with several asymmetrical
forms
b) If an object is already asymmetrical, increase
the degree of asymmetry
5 Combine/ a) Combine/ consolidate in space homogeneous
Consolidate objects or objects destined for contiguous
operations
b) Combine/ consolidate in time homogeneous or
contiguous operations
6 Universality a) An object can perform multiple diferent
functions; therefore there is no need for other
elements (or other elements can be removed)
7 Nesting a) An object is placed inside another one, which in
(see the Russian toy turn, is placed inside a third object, etc.
Matrioshka) b) Through a cavity, an object passes in another
object
149
8 Counterweight a) Compensate an objects weight by means of

another object that is able of providing a
corresponding lifting force
b) Compensate for the weight of an object by
interaction with forces (aerodynamic or
hydrodynamic types) coming from an outside
environment
9 Prior counteraction a) Perform a counteraction in advance
b) If the object is or will be under an undesirable
tension (stress), provide a corresponding
counter-tension in advance
10 Prior action a) Carry out in advance the required actions or
changes (completely or partially) to an object
b) Arrange/place objects in advance in a way they
can go immediately into action when required
and they do this from the most convenient
position
11 Cushion in advance a) Compensate for the relatively low reliability of an
object taking countermeasures in advance
12 Equipotentiality a) Change the working conditions in such a way to
be no need the object to be lifted or lowered
13 Inversion/ Reversion 1. Instead taking an action that is dictated by
the specifcations of the problem, implement
an opposite action (i.e. heating instead of
cooling)
2. Make a mobile (movable) part of an object
(or the outside environment) immobile
(immovable) and vice versa
3. Turn the object upside-down
14 Spheroidality a) Replace linear parts with curved ones; replace
fat surfaces with spherical ones; replace cubical
shapes with ball ones
b) Use rollers, balls and spirals
c) Replace a linear motion with rotating motion;
use a centrifugal force
150
15 Dynamicity 1. Some characteristics of an object or its

environment must be automatically adjusted
or altered to ensure optimal performance at
each stage of the considered operation
2. Divide an object into elements that are able
of changing their position relative to each
other
3. If an object is immobile, make it mobile or
interchangeable
16 Partial or excessive a) If it is difcult to obtain 100% of a desired efect,
action try to achieve somehow more or less of the
desired efect/result
17 Translation into a new Remove problems by translating the object from
dimension one-dimensional movement or placement into a
two-dimensional movement or placement; from
two-dimensional movement or placement into a
three-dimensional movement or placement
Use a multi-level assembly of objects instead of a
single level (layer)
Incline the object or turn it on its side
Use the opposite side of a given surface
Project optical lines of an object onto neighboring
areas or onto the reverse side
18 Mechanical vibration a) Set an object into oscillation
b) If oscillation exists, increase its frequency, even
to the ultrasonic level
c) Use the resonant frequency
d) Use piezo-vibrations instead of mechanical
vibrations
e) Use ultrasonic vibrations in conjunction with an
electromagnetic feld
151
19 Periodic action a) Replace a continuous action with a periodic

action (an impulse)
b) If the action is already periodic, try changing its
frequency
c) Use pauses between impulses in order to provide
additional actions
20 Continuity of a useful a) Carry out an action without pauses (in a
action continuous mode) and all parts of the object to
operate constantly at their full capacity
b) Remove idle and intermediate motions
c) Replace back-and-forth motions with rotary
ones
21 Rushing through a) Perform harmful or hazardous operations at very
high speeds
22 Convert harm into a) Utilize harmful factors (especially provided from
beneft environment) to obtain a positive efect
b) Remove a harmful factor by combining it with
another harmful factor
c) Increase the level of the harmful actions until
they cease to be harmful
23 Feedback a) Introduce feedback
b) If feedback already exists, try to reverse it
24 Mediator a) Use an intermediary object to transfer or carry
out an action
b) Temporarily connect an object to another one
that can be easily removed
25 Self-service a) Make the object capable of servicing itself and
carrying out supplementary and repair
operations
b) Make use of waste material and energy
152
26 Copying Use simple and inexpensive copy instead of an

object that is complex, expensive, fragile or
inconvenient to operate
Replace an object by its optical copy or image. A
system can be used to reduce or enlarge the
image
If visible optical copies are used, replace them
with infrared or ultraviolet copies
27 Dispose a) Replace an expensive object with several
inexpensive objects, comprising some of the
properties (i.e. longevity)
28 Replacement of a 1. Replacement of a mechanical system by
mechanical system an optical, acoustical or olfactory (odor)
system
2. Use an electrical, magnetic or
electromagnetic feld for interaction with
the object
3. Replace:
1.Stationary felds with moving felds
2.Fixed felds with those which change in time
3.Random felds with structured felds
4. Use a feld in conjunction with
ferromagnetic particles
29 Pneumatic or a) Replace solid parts of an object with gas or liquid
hydraulic systems. These parts can now use air or water
construction for infation, or use air or hydrostatic cushions
30 Flexible membranes a) Replace classical constructions with
or thin flms systems made from fexible membranes
or thin flms
b) Isolate an object from its outside
environment using fexible membranes
or thin flms
153
31 Use of porous a) Make an object porous or add supplementary

material porous elements (inserts, covers, etc.)
b) If an object is already porous, fll the pores in
advance with some substances
32 Changing the color a) Change the color of an object or its surroundings
b) Change the degree of translucency of an object
or surrounding processes that are difcult to see
c) Use colored additives to see objects or processes
that are difcult to see
d) If such additives are already used, employ
luminescent traces or trace elements (atoms)
33 Homogeneity a) Those objects that interact with the main object
have to be made out of the same material or
materials having close properties as the main
objects material
34 Rejecting and a) After an element of an object has completed its
regenerating parts function or become useless, it should be rejected
or modifed during work process (i.e. discarded,
dissolved, evaporated, etc.)
b) Restore any part of the object that is used-up,
exhausted or depleted
35 Transformation of a) Change the physical state of the object
object properties b) Change the density or concentration of the state
c) Change the degree of fexibility
d) Change the temperature
e) Change the volume
36 Phase transformation d) Use the efects that are generated
during the phase transition (change) of
a substance. For instance, during the
change of volume, liberation or
absorption of heat and so on
37 Thermal expansion a) Use the thermal expansion or contraction
property of the material
b) Use diferent materials with various thermal
expansion coefcients
154
38 Use strong oxidizers a) Do the transition from a certain level of oxidation

to the next higher level by:
1.Replacing normal air with enriched air in
oxygen
2.Replacing enriched air with oxygen
3.Replacing oxygen with ionized oxygen
4.Replacing ionized oxygen with ozoned
oxygen
5.Replacing ozoned oxygen with ozone
6.Replacing ozone with singlet oxygen
39 Inert environment a) Replace the normal environment with an inert
one
b) Carry out the process in a vacuum
c) Introduce a neutral substance or additives
40 Composite materials a) Replace a homogeneous material with a
composite one
The set of 39 engineering contradictions presented in table 8.1 can be

creativelysolvedusingthesetof40inventiveprinciplespresentedintable
8.2.Inthisrespect,Altshullerdevisedanapproachofmappingtheconflicts
to the inventive principles. The approach is called the Contradiction
Matrix (some other names used for this approach are: Table of
Contradictions;EngineeringParameterTable).
Thecontradictionmatrixispresentedintable8.3andconsistsofanetof39
rowsand39columns.
155
Table8.3.Thecontradictionmatrix.
156
157
158
159
160
The use of the contradiction matrix is simple but requires following a

sequenceofsteps.Thefirststepistoselect(intherowsofthecontradiction
matrix)theparameterthatneedstobeimproved.Thenextsteprequires
161
matchingtheselectedparametertooneormoreoftheparametersthatare
in conflict with it (in the columns of the contradiction matrix). In the
intersectingboxesthereareplacedaseriesofnumbersthatcorrespondto
theappropriateinventiveprinciples(astheyareorderedintable8.2)tobe
considered for eliminating the contradiction. The placement of the
inventiveprinciplesintheboxesofthecontradictionmatrixistheresultof
Altshullersstudyworkonthepatentdatabase.
Forexample,itisconsideredanupperrobotarmwheretheparameterthat
hastobeimprovedis1.Weightofmovingobject(seetable8.1),inthe
senseithastobereduced.Inthesametime,therobotarmhastobeenough
resistanttoexternalforcessuchastherobottohaveanashighaspossible
payload capacity. The length of the arm is imposed by kinematic
considerations.Inthisrespect,theconflictingparameteris11.Stress(see
table8.1).
Thestandardtechnicalcontradictionisformulatinginthefollowingway:
themoretryingtoimprovethestandardengineeringparameterWeightof
moving object, the more the standard engineering parameter Stress
becomes worse. Looking in the contradiction matrix (table 8.3) at the
intersection of row 1 with column 11 there are obtained the inventive
principles10,36,37and40(seetable8.2),thatis:10.Prioraction;36.
Phasetransition;37.Thermalexpansion;40.Compositematerials.Itis
obvious that making the upper robot arm according to the inventive
principle40,thecontradictionissolved,thatis,thearmisbothlighterand
enoughrigidtohandlehighpayloads.
The TRIZ methodology can be used for solving various problems. The
matrixofcontradictionandtheinventiveprinciplesaretoolsthatcanbe
successfully applied in solving problems that require only minor
improvementstoanexistingsystem.
For problems of higher difficulty, Altshuller proposed more precise
approaches. One of these instruments is called the Substance Field
Analysis (SuField analysis). It was thought for expressing function
162
statements and identifying functional failures when a system acts upon

another system. According to this approach, the systems are called
substances andthe actionis called field. SuField analysis describes
howvarioussubstances(elementsofthesystems)andfields(mechanical
pressure, inertia forces, vibrations; electrical; magnetic; thermal;
informational; electromagnetic; etc.) interact each other to produce a
certaineffect.
AtransitionfromasystemW1toasystemW2canbeconsideredaschanges
inthesubstancefieldstructureofsystemW1.SuFieldanalysisgeneratesa
modelthatreflectsthecriticalfeatureofsystemW1includingthephysical
contradiction.Themodelusuallycontainsthreecomponents:
SubstanceS1:itdescribesthearticle,thematerial,theprocess,thepart
orthesubsystemthatshouldbecontrolled,processedorgenerally
speakingthatitemuponwhichitshouldbeacted;
SubstanceS2: itdescribesthetool,theelement,etc.thatwillcontrol,
processor,generallyspeaking,actonthesubstanceS1;
FieldF:itistheenergyusedtocontrol/tointeract,etc.
Themodelcanberepresentedinadiagrammaticway.Forexample,figure
7.3showshowtheenergyFactsonanelementS2tomodifythepartS1.In
otherwords,S2willtransformthefieldFinakindofenergythatwillacton
S1suchastosolvetheproblem.Theproblemappearsduetotheabsenceof
some components from the pair substancefield and from here, the
conclusion is that something should complete the considered substance
field.ThekeyinSuFieldanalysisconsistsinidentifyingtherightfield F
andtherightsubstanceS2,oncethesubstanceS1hasbeendefined.Looking
at actions like being fields, undesirable or insufficient actions can be
counteredbyapplyingintensifiedoroppositefields.
163
Fig.8.3.AnexampleofdiagrammaticrepresentationforaSuFieldmodel.
Fig.8.4.AnexampleofaSuFieldmodelwithaninterveningagent.
In SuField analysis there could appear situations where the useful

interactionbetween S1 and S2 isaccompaniedbysomeharmfuleffects.In
suchcases,theharmfuleffectscouldbeeliminatedbyintroducinganagent
S3 between S1 and S2,where S3 wouldbe S3{S1, S2, S1, S2, S1+S2}.The
most effective solution is when engineers succeed having S3=S2. The
diagramforsuchacaseisillustratedinfigure8.4.Frompatentanalysis,
Altshullerprepared79standardSuFieldsolutionsforelegantelimination
of engineering contradictions. At present, there are several software
packages that include these standards in their knowledge bases (i.e.
Improver,Ideator,etc.).
164
For a better understanding of the SuField approach, a simple example

from robotics is further given. It is considered a positioning axis of a
Cartesian robot that uses a ball screw mechanism for ensuring accurate
positioning.Inordertoachievehighprecisionpositioning,themechanism,
consistingoftwonutsandaballscrew,hastobepreloaded.Thepreloadis
done with a space plate that is introduced between the two nuts. The
backlashiseliminatedbutthepreloadproducessignificantfrictionbetween
theballscrewandthetwonuts.Frictionischaracterizedbytwophases:the
microdynamic phase and the macrodynamic phase. The first phase
describes the ball screwnut behavior before the sliding occurs. At this
phase,theelasticdeformationattheboundarylayer(thecontactsurfacesof
thenutswiththeballs)isdominantandthisgeneratesapositioningerror
up to severalhundrednanometers. Therefore, it is almost impossible to
obtain nanometer resolution using a ball screw mechanism. This task
becomes even harder when the application requires working at high
speeds.
Keepingtheballscrewmechanism,improvementsindesignarerequiredto
achieve a nanometer resolution. The physical contradiction can be
formulated as follows: to ensure a high resolution, the ball screw
mechanismshouldbepreloaded,butontheotherhanditshouldnotbe
preloaded since this action destroys the boundary layer and makes
impossibletheachievementofananometerresolution.
Inthiscase,thesubstance S1 isthepreloadforcethatmustbecontrolled,
becauseitgivesthemagnitudeoftheelasticdeformationattheboundary
layer.ModelingtheproblemusingtheSuFiledanalysis,apair(S2,F)must
beintroducedintothesystemsuchthatS2willtransformFintoaforcefor
exertingasupplementarylightpreloadofnanometerresolution.
Thepreloadforceiscontrolledbythethicknessofthespaceplatebetween
thetwonuts.IfS2isapiezoelectricactuator,interposingitbetweenoneof
the nuts and the space plate, the thickness of the space plate can be
controlled because piezoelectric actuators have the characteristics of
165
nanometerresolutionandfastresponse.Integratingaloadcell,usingstrain
gages,intothepiezoelectricactuator,thenonlinearity,hysteresisandDC
driftsproblemsareovercame,too.Inthiscase,Fisthepiezoelectricforce,
generatedbythepiezoelectricactuator,whichwillactonthespaceplate.
Aconventionalservomotorwillperformthegrossmotion.
Fig.8.5.TheSuFieldmodelforthecasestudy.
Afterthat,piezoelectricactuatorswillgenerateahighpreloadtoeliminate
thebackslash,toincreasethestiffnessoftheballscrewmechanismandto
adjustthepositioningatthedesiredtarget.TheSuFielddiagramofthis
problemisillustratedinfigure8.5.
Another powerful tool proposed by Altshuller and his collaborators for
properlysolvingdifficultproblemsisARIZ(orASIPinEnglish:Algorithm
for Solving Inventive Problems). ARIZ is a systematic approach for
identifying solutions that elegantly overpass the contradictions. ARIZ is
consideredthemainanalyticaltoolofTRIZ.TheobjectiveinARIZisto
systematicallybreakdownaproblemintosmallersetsofminiproblems
andtofindconflictfreesolutions.ARIZassumesthatthelevelofdifficulty
insolvingacertainproblemlargelydependsonthewaytheproblemis
formulated.Theclearertheformulationwillbe,thesimplerthesolution
willbeobtained.ARIZisanadaptablealgorithm,inthesensethesame
problemcanbesolvedinvariouswaysdependingonwhowillsolvethe
problemandhowtheproblemwillbeapproached.Thepathsomeonewill
166
follow using ARIZ will be in accordance with his/her knowledge,

experienceandcreativecapabilities.Thealgorithmonlysavestheperson
from performing wrong steps and nothing more. Using ARIZ, different
peoplewillveryprobablygeneratedifferentsolutionstothesameproblem.
ARIZ produces results that significantly depend on the applicants
inventive potential, knowledge and practice. In other words, not every
problemcanbesolvedbyanybodybysimplyusingARIZ.Morethanthis,
ARIZislikeacomplexsport,whereenoughpracticalskillsandtrainingare
required in order to apply it at high standards. Some practitioners
mentionedaboutmorethan6monthsspecialtraininguntiltheysucceeded
tounderstandbetterhowARIZreallyworks.
ARIZdrivestheprocessofproblemsolvingthroughachainofsequential
reformulations and reinterpretations of the problem such that it is
transformedfromvagueideasintoalucidformulationofthecoreconflict.
Fig.8.6.TheflowchartofproblemreformulationusingARIZ.
167
In this process, the original formulated problem will often migrate to

unsuspected areas that define the system where easier solutions can be
provided.ThismeansthatusingARIZ,acompletelydifferentproblemthan
theoneinitiallyformulatedcouldbeidentified.Thebasicflowofproblem
reformulationusingARIZisshowninfigure7.6.Itisbasedontwomajor
ideasthathavebeenalreadymentionedinthissection:thesystemconflict
andtheidealityprinciple.
In the first phase, ARIZ requires to translate a vague (or even wrong)
defined initial problem into a miniproblem, using the following rule:
everythinginthesystemremainsunchanged,buttherequiredfunctionis
realized.Thenextphaseleadstotheformulationofsystemconflictand
thegenerationofasimplifiedschemeoftheconflict,whichiscalledthe
modeloftheproblem.
Specifyingtheconflictdomain,theareaofanalysisisnarrowed,ensuringa
better focus on conflicting aspects. The next phase of ARIZ consists in
assessingtheavailableresources(energy,materials,information,etc.)by
selecting a critical system resource in the conflict domain and then
formulatingthesocalledIdealFinalResult(IFR).Inmostofthecases,
achieving the ideal result means inducing within the available resource
conflicting physical properties (i.e. both electroconduction and electro
insulation).
Forovercomingvariousphysicalcontradictions,ARIZusesfourrules,as
follows:
RuleI(separationofoppositepropertiesintime):Duringoneintervalthe
systemShasthepropertyP;duringanotherintervalthesamesystem
ShastheantipropertyofP,denotedwithaP.
RuleII(separationofoppositepropertiesinspace):ForacertainpartZof
thesystem S isgiventheproperty P,whiletotheotherpart SZ is
giventheantipropertyofP(aP).
168
Rule III (separation of opposite properties between the system and its
components): The whole system S has the property P, while its
componentsS1,S2,havetheoppositepropertyaP.
RuleIV(coexistenceofoppositepropertiesinthesamesubstance):Oneand
the same substance exhibits mutually exclusive properties under
differentconditions.
Foreliminatingthephysicalcontradictions,theuseofvariousmethodsis
recommended. The use of knowledge bases with all sorts of previous
solvedproblemscouldaidthisprocess(i.e.knowledgebaseofphysical,
chemicalandgeometriceffects).
Ifnosatisfactorysolutionisfoundattheendofthisprocess,theruleis
tryingtoreformulatetheinitialproblembecauseitcouldbewrongdefined.
Ifthisisstillunsatisfactory,amoregeneralproblemhastobestatedand
thentheprocesstoberestarted.
Altshuller considers that the scientific structure of the creative work
includesseveralsteps,asfollows:
Continuous study of the fields and technologies of the leading
industries;
Continuous study of the fields and technologies of the follower
industries;
Collection of information about physical effects, new materials,
methodsofsolvingtechnicalproblems,etc.;
Continuousstudyofpatentinformation;
Continuous study of new developments in the theory of inventive
problemsolving;
Accumulationofexperiencebysolvingexerciseproblems.
ThefirstversionofARIZwaspublishedin1959.NewversionsofARIZ
have been further developed and released (1961, 1964, 1965, 1968, 1971,
etc.).Eveninourdays,ARIZisstillinacontinuousimprovementprocess.
169

ThetaskwithinthisworkshopistoapplytheTRIZandARIZmethodsin
solvinginnovativelysomeengineeringproblemsfromthefieldofrobotics.
In the first phase the TRIZ method will be experienced to generate
innovative solutions in designing a robot gripper. The innovation is
required to solve various conflicting problems whichoccur betweenthe
quality engineering characteristics of a robot gripper. The engineering
characteristicsconsideredinthisworkshoparethefollowing:
1. Stroke/finger[mm];
2. Graspingforceat6bar[N];
3. Payloadcapacity[kg];
4. Compressedairconsumptionperdoublestroke[cm3];
5. Fingerclosingtime[s];
6. Fingeropeningtime[s];
7. Grippermass[kg];
8. Inertialmoment[kgcm2];
9. Repeatability[mm];
10. Maximumlengthofthefinger[mm];
11. Openingbetweenjaws[mm];
12. Gauge[cm3].
In the first stage, students must determine the whole set of conflicts
betweenthequalityengineeringcharacteristics.Inthisrespect,theymay
usetheQualicaQFD,theroofmatrix.Afterward,alistwiththeconflicts
hastobecompleted.Foreachconflict,thepairofengineeringparameters
fromtable8.1hastobeidentifiedandaddedinthelist.
Eachpairofengineeringparameterswillbetakenintoaccountturnby
turnandTRIZmethodwillbeappliedtoidentifyinnovativesolutions.To
run faster this process, the software package called Innovation
Workbenchwillbeused.Thefollowingstepshavetobefollowed:
170
Step1:OpentheInnovationWorkbenchandclickonTools(Fig.8.7).
Fig.8.7.ThemainwindowoftheInnovationWorkbench.
171
Fig.8.8.TheContradictionTableasitappearsinInnovationWorkbench.
172
Step2:SelectContradictiontable.Awindowliketheoneinfigure8.8
willbeopened.Youmustfirstlyselecttheattributeyoudesiretoimprove
(seeFig.8.8.a.).Afterward,selectthedesignattributethatdegradesasa
result (see Fig. 8.8. b.). The software will display a set of innovation
principlesyoumayusetoeliminatethetechnicalcontradiction(seeFig.8.8.
c.).
Step3:Selecteachinnovationprinciple.Ontherightsideofthewindow,
the software will display explanations and examples related to that
principle (see Fig. 8.8. d.). Use the information to generate innovative
solutions.
ThesecondpartoftheworkshopconsistsinapplyingtheARIZmethodto
solveinnovativelythefollowingproblem:
Itisconsideredaroboticcellforassemblingcomputercomponents.The
robot gripper performs complex motions with delicate workpieces
(grasping,handling,insertion).TherobotisaSCARAtyperobot,astheone
infigure8.9.Therobotgripperissuppliedbytwovacuumlinesandfour
pneumaticlinesandhasseveralsensorswhichtransmitsignalsthroughthe
robot controller by means of electric cables. Both the electric supplying
cablesandcompressedairtubesarelinkedtotherobotcontrollerbymeans
ofanumbilicalcordacorrugatedplastichosewhichcoveralltubesand
wires(seeFig.8.9.).Thecorrugatedplastichosehastheroletokeepthefine
particlesresultedfromthefrictionbetweenthetubesandwires.Duringthe
workingprocess,theplastichosebroke.
173
Fig.8.9.TherobotunderconsiderationfortheARIZproblem.
Thebrakeofthehosewascausedbytheplasticmaterialfatigueassociated
with high speeds and amplitudes of motion in the robot joints which
generated excessive bending of the hose. Because of this failure, the
workingenvironmentwascontaminatedwithmicroparticlesresultedfrom
thefrictionofcablesandtubes.
Itisarequesttodefineasolutionforpreventingthebreakofthehoseorto
think some alternative solution that is able to solve the problem
innovatively.Theuseofaplasticmaterialwithahigherdurabilityisnota
goodsolutionbecauseitismoreexpensiveandatacertainmomentitwill
however break (even if later than the first hose). This problem can be
174
elegantlysolvedbymeansofARIZmethod.Inthisrespect,studentscould
use either a manual approach or the approach given by Innovation
Workbench.
8.4. Example 1
ThefirstexampleconcernswiththeapplicationofTRIZmethodinsolving
aconflictingproblemspecifictorobotdesign.Theproblemistodesigna
lessexpensiverobotcontroller(i.e.aPIDcontroller,independentlyoneach
robotaxis)withoutaffectingrobotaccuracy.Itisknownthatrobotshave
variable geometry, this generating coupled and nonlinear effects in the
robotjoints.Inotherwords,themotioninacertainrobotarmwillgenerate
negativeeffectstotheaxesrelatedtotheotherrobotarms.So,thegoalis
toreducethecomplexityofdevice(parameterno.36intable8.1)without
reducing the accuracy of the process (parameter no. 29 in table 8.1).
ApplyingtheInnovationWorkbench,thefollowinginnovativeprinciples
arerevealed:
1.Copying:
Use a simple and inexpensive copy instead of an object which is
complex,expensive,fragileorinconvenienttooperate;
Replaceanobjectbyitsopticalcopyorimage.Ascalecanbeusedto
reduceorenlargetheimage;
If visible optical copies are used, replace them with infrared or
ultravioletcopies.
2.Mediator:
Useanintermediaryobjecttotransferorcarryoutanaction;
Temporarilyconnectanobjecttoanotheronethatiseasytoremove.
3.Changethecolor:
Changethecolorofanobjectoritssurroundings;
Changethedegreeoftranslucencyofanobjectorprocesseswhichare
difficulttosee;
175
Use colored additives to observe objects or processes which are

difficulttosee;
Ifsuchadditivesarealreadyused,employluminescenttracesortracer
elements.
Becausetheavailableresourceistherobot,byanalyzingtheresultsafter
the application of TRIZ, the following ideas could be extracted and
aggregatedfordefiningareliablesolution:
1. Use a simple and inexpensive copy instead of an object which is

complex;
2. Use an intermediary object to transfer or carry out an action;
3. Change the degree of translucency of an object or processes which
are difficult to see.
Fromthefirstidea,theconclusionistodesignarobotmechanicalstructure
thatissimplerintermsofdynamicbehavior.Fromthesecondidea,the
conclusionistousesomeextraentitiesforimprovingtherobotdynamic
behaviorinordertoknowinadvancetheleveloftorque/forcetotherobot
axesforagiventask.Fromthethirdidea,theconclusionistoelaboratethe
dynamicsmodeloftherobot(seetheEulerLagrangeformalism)suchthat
toseetheeffectofeachkeyrobotscomponentonrobotdynamics.
Toachievethegoal,thefollowingsolutionmaybeformulated:inthefirst
stage, elaborate the model of the robot chain using the EulerLagrange
formalism.Identifytherelationshipsbetweenthevariousparametersofthe
robotstructure(masses,lengths,momentsofinertia)inconjunctionwith
theequationsthatdescribetheresistanttorques/forcestotherobotaxes.
Try to see which of the terms in these equations can be reduced or
eliminated by defining proper relationships between some of the
parametersthatdescribetherobotstructure.Oncetheserelationshipsare
established,trytodesigntherobotstructureaccordingly.Attheend,the
result is a robot structure characterizes by less coupled and nonlinear
effects to the axes, so, a simpler robot controller can be implemented
176
without affecting the robot accuracy when continuous paths have to be

followedduringtaskperformance.
UsingtheinnovativeideaUseasimpleandinexpensivecopyinsteadof
anobjectwhichiscomplex,anothersolutionmightbetheuseofrobust
controlalgorithmsinsteadofcomplexcontrolalgorithms.
8.5. Example 2
TheuseofARIZinroboticsisfurtherexemplifiedinasimplecasestudy.It
isconsideredtheconveyorfromfigure8.10thatreceiveslongbarsfroma
tubeanddeliversthemtoarobotstation.Becausethebarsareorientedina
chaotic way, the machineloading robot, placed at the other end of the
conveyor,hastobeequippedwithvisionsystems.Thisthingincreasesthe
investmentintheroboticcell.Theproblemstatementinthisexampleisthe
following:Findasimple,reliableandcosteffectivesolutiontoalignthe
bars.
Fig.8.10.CasestudyforARIZ.
The application of ARIZ to solve this problem is presented in the next

paragraphs:
177
Problem statement: Find a simple, reliable and costeffective solution to

alignthebars.
Miniproblem:Partshavetobealignedwithoutmajordisturbancesofthe
initialsystem.
Systemconflict:Foraligningthepartsitisnecessarytouseaspecialdevice
butasupplementarydevicecomplicatesthesystemandincreasestheprice
significantly.
Modeloftheproblem:Itisnecessarythatsomeoftheexistingelementsof
thesystemtoberesponsibleforalignment.
Analysisoftheconflictdomain: Theconflictdomainisthesurfaceofthe
conveyorbelt.
Analysis of available resources: The single available resource is the
conveyor.
Idealfinalresult(IFR):Theconveyoritselfhastoaligntheparts.
Physicalcontradiction:Inordertoaligntheparts,conveyorhastoactwith
aspecifiedspeedatoneendofthepartandwithadifferentspeedatthe
otherendofthepart(producingarotationofthepart).Thisisnotpossible
becausethebeltmoveswithonesinglespeedinonedirection.
Eliminationofthephysicalcontradiction:ItisusedtheruleIII:separation
ofoppositepropertiesbetweenthesystemanditscomponents:ThewholesystemS
hasthepropertyP,whileitscomponentsS 1,S2,havetheoppositepropertyaP.
Accordingtothisrule,theresultisthatthewholeconveyorwillmoveata
certainspeedbutsomeofitspartswillmoveatdifferentspeeds.
178
Fig.8.11.Innovativesolutionforexample2.
Solution: The width of the conveyor belt is narrowed at a value that is

muchlowerthanthelengthofthepartsandtothesidesoftheconveyor
beltwillbeplacedanothertwobeltsmovingintheoppositedirectionthan
the conveyor belt. This will determine a longitudinal alignment of the
parts.Thesketchoftheengineeringsolutionispresentedinfigure8.11.
8.6. Example 3
The third example shows how TRIZ and ARIZ are together applied to
defineinnovativesolutionsforthetooloftheroboticinstallationusedto
grindtelescopemirrors(seeFig.1.6).Itisrequiredalowcostandreliable
conceptforthemirrorgripper,bothtosolvethegraspingrequirementsof
the mirror and to assure the passive motions for imitating the human
handsmovements.
Inthefirststage,TRIZ/ARIZhavebeenappliedtodefineaninnovative
conceptforgraspingthemirror.Thefollowingstepshavetobedone:
A1.Problemstatement:Thegripperhastograspthemirrorinanumberof
6points,withequalforcesandinthesametime.
B1.Miniproblem: Thegraspingprocesshastobeperformedwithalow
costsimpledeviceinordertokeepthesolutionattractivefromeconomic
pointofview.
179
C1.Systemconflict:Fixingthemirrorin6pointsaccurately,simultaneously
and with the same force, usually requires complicated mechanical and
sensorialsolutionswhichleadtotheincreasingofcostsbothfromdesign,
manufacturing, assembly, quality assurance and maintenance points of
view.
D1.Modeloftheproblem:Someelementsofthegrippermustbelinkedin
suchawaythattheycanachievetherequiredtask.
E1.Analysisoftheconflictdomainandresources: Theconflictdomainis
representedbythetradeoffbetweenthecomplexityandthecostofthe
mirrorgraspingmechanism.Thesingleavailableresourceisthegrasping
device.
F1.The ideal final result: The mirrorgripper must be both cheap and
reliableintermsofapplicationrequirements.
G1.Physicalcontradiction: IntermsoftheTRIZengineeringparameters,
the conflict is between the characteristics: complexity of device (the
gripper)anddurabilityofmovingobject(themirrorduringthegrinding
operation).
H1.Elimination of the physical contradiction: The application of TRIZ
contradictionmatrixleadstothefollowingpossibleinnovativeprinciples:
arrange/placetheobjectsinadvanceinawaytheycangoimmediately
into action when required and they do this from the most convenient
position;replaceasymmetricalformwithanasymmetricalformorwith
severalasymmetricalforms;divideanobjectintoelementsthatareable
ofchangingtheirpositionrelativetoeachother;maketheimmobileparts
mobileorinterchangeable.
I1.Engineeringsolution:Combiningtheinnovativeprinciplesproposedat
pointH1andapplyingtheruleIIofARIZ(forsomepartsofthesystemis
givenacertainpropertyPseethesameshape,thesamemobility,etc.,
while for the rest of the system is given the antiproperty aP see for
180
examplefixedcomponents,oppositeshapethanthoseofthemobileparts,
etc.)thesolutionfromfigure8.12hasbeenproposed.
Fig.8.12.Innovativeconcepttograspthemirror(Brad&Srb,2004).
181
Fig.8.13.InnovativeconcepttoensureOXandOYpassiverotationsandOZactive
rotation(Brad&Srb,2004).
ThesecondapplicationofTRIZ/ARIZmethodswasnecessaryfordefining
acheapandreliablesolutionofthemirrorgripperthatimitatesthepassive
movements of the human hands during the manual grinding operation.
Thefollowingresultshavebeenobtained:
A2.Problem statement: The gripper must have two passive rotations
aroundtheOXandOYaxes,whileallowingthetransmissionofarotary
movementaroundtheOZaxis.
B2.Miniproblem:Toachievethepassivemotionsusingcheapandsimple
tomanufacturemechanisms.
C2.Systemconflict:Inordertoachievesuchtasks,complexandexpensive
mechanismscomposedfrompartswithcomplicatedsurfacesareused.This
increasesthecosts.
182
D2.Modeloftheproblem: Someelementsofthegrippermustensurethe
generationofthetwopassivemotions,whilethewholesystemcanrotate
aroundaverticalaxis.
E2.Analysis of the conflict domain and resources: The conflict domain
standsinthetradeoffbetweenthecomplexityandthecostofthegripper.
Thesingleavailableresourceisthegripperconcept.
F2.Theidealfinalresult:Thegrippermustbecheap,easytomanufacture,
andinthesametimetocopyascloseaspossiblethehandmovements.
G2.Physicalcontradiction:ConsideringtheTRIZparameters,theconflictis
generatedbetweenthecharacteristics:complexityofdevice(thegripper)
andaccuracyofmirrormanufacturing.
H2.Eliminationofthephysicalcontradiction:FromtheTRIZcontradiction
matrixthefollowinginnovativeprincipleshavebeenextracted:usesimple
and inexpensive copies instead of the complex object; use an
intermediaryobjecttocarryouttheaction;temporarilyconnectanobject
toanotheronethatcanbeeasilyremovedifnecessary.
I2.Engineeringsolution:Combiningtheinnovativeprinciplesproposedat
point H2 and applying the rule IV of ARIZ (one and the same system
exhibits mutually exclusive properties under different conditions) the
solutionfromfigure7.13hasbeengenerated.
8.7. Workflow
Thetaskistobeperformedin2hoursoflaboratorywork.Studentswillrun
thetaskindividually.Attheendofthework,resultswillbeprintedand
analyzedingroup.Conclusionsaroundtheresults,aswellasaroundthe
potential of TRIZ and ARIZ methods in engineering design will be
formulated.
8.8. Homework
In case the laboratory work is not carried out in 2 hours, students will
continuethetaskathome.
183
REFERENCES
[ALT00] Altshuller, G., The Innovation Algorithm. TRIZ, Technical

Innovation Center, Worcester, ISBN 0-9640740-4-4, 312
pp., 2000.
[BRA03] Brad, S., Mocan, B., About the Competitive Design of an
Outdoor Wall-Cleaning Robot, Proceedings Int. Conf.
MTeM03, Cluj-Napoca, Ed. Roprint, ISBN 973-656-490-8,
pp. 81-84, 2003.
[BRA04a] Brad, S., Fundamentals of Competitive Design in Robotics.
Principles, Methods and Applications, Ed. Academiei
Romne, Bucureti, ISBN 973-27-1065-9, 410 pp., 2004.
[BRA04b] Brad, S., Srb, M., Innovative Design of a Robotic
Installation for Ttelescope Mirror Grinding, in: Robotica &
Management, Ed. Timpul, Vol. 9, No. 2, ISSN 1453-2069,
pp. 13-18, 2004.
[BRA05] Brad, S., Vaida, C., Competitive Redesign of a Measuring
Head for Pressure and Flow, microCAD International
Conference, Miskolc, Hungary, ISBN 963 661 646 9, pp. 19-
24, 2005.
[COY04] Coyle, G., Practical Strategy: Structured tools and
techniques, ISBN 9780273682202,
Prentice Hall (2004)
[***02] ***, Qualica QFD 2.5.0.1 User Manual, 2002.

[***03] ***, MS Visio 2003 User Manual, 2003.
[***04] ***, Innovation Workbench 2.9.4 User Manual, 2004.

Indrum Qualica en PDF

Uploaded by

Copyright:

Available Formats

Indrum Qualica en PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Indrum Qualica en PDF

Uploaded by

Copyright:

Available Formats

STELIAN BRAD

EMILIA BRAD BOGDAN MOCAN

TOOLS AND METHODS OF

Ranking Requirements in Competitive

1 Equal The two activities or requirements or customer

3 Somewhat more Experience and judgment slightly favor one item

5 More important Experience and judgment strongly favor one

7 Much more An item is strongly favored and its dominance is

9 Absolutely more The evidences favoring one item relative to

0.33 Somewhat less Experience and judgment slightly un-favor one

0.20 Less important Experience and judgment strongly un-favor one

0.14 Much less An item is strongly un-favored and its

0.11 Absolutely less The evidences un-favoring one item relative to

2, 4, 6, Intermediate When compromise is needed

0.5, Intermediate When compromise is needed

Requirement B 0.33 1 1 0.33 0.14

Requirement D 0.11 3 0.33 1 9

Requirement E 0.14 7 0.20 0.11 1

Total on column 1.78 15 7.53 13.44 22.14

Requirement B 0.19 0.07 0.13 0.03 0.01 0.09 9

Requirement C 0.11 0.07 0.13 0.22 0.22 0.15 15

Requirement D 0.06 0.20 0.04 0.07 0.41 0.16 16

Requirement E 0.08 0.46 0.03 0.01 0.04 0.12 12

1.3. Task Description and Methodology

For further details on the facilities of Qualica QFD during task

access to all components; d) easy to maintained even by less qualified

Computing the Consistency Ratio

2.2. Task Description and Methodology

are based on G. Coyle's book Practical Strategy: Structured tools and

Qualica uses symbols to represent the figures, with each symbol

eigenvector). There are several methods for calculating the eigenvector;

Planning Engineering Specifications in

Thevalueweight wjr ofthe jthengineeringcharacteristic, j=1,,m,isa

The identification and analysis of correlations between engineering

An important aspect that should be determined and introduced as

3.3. Task Description and Methodology

deployment, safety deployment). In case this situation appears in

No. Requirements Importance

7 Precision of pressure regulation: 1% of the

To deploy properly the critical customer requirements throughout the

For exemplification, figure 3.9 shows the redesigned solution of the

Functional Analysis in Competitive

that the process of generating the FAST diagram is very dynamic,

4.3. Task Description and Methodology

Figure 4.13 presents the technical solution which is provided to solve

Idea Generation in Competitive Robot

The 635 method follows a welldefined working procedure, as it is

having as support the ideas already generated by his/her rightside

5.3. Task Description and Methodology

Team X Idea 1 Idea 2 Idea 3

Variant Selection in Competitive

criterion much betterthan the reference variant. Other evaluation scales

1 Number of axes 1 4 4 4 4 100 100 100

Step3: Developmentofutilitycurves: thisisthestepthatdifferentiatesthe

No. axes Speed joints Working space Accuracy

50% 50% 50% 50%

Repeatability Payload MTBF Cost

50% 50% 50% 50%

0.01 0.03 10 15 21 25 9500 10000 20000 40000 65000 70000

[02] , Qualica QFD 2.5.0.1 User Manual, 2002.