インタラクティブデバイスのデザイン検討におけるヒューマンファクター

NII-Electronic Library Service

Human

Factors

in

Design

Validation

of

lnteractive

Devices

E=h

rti-rn=-Monica BORDEGONI

$i1IrvV<#

PolitecnicodiMilano

The

_paper

describes

_practices

forvalidating engonomics as-pect$ofinteractivedevicesusing thestate-of-the-art of virtual

prototyping

methodologies and technologies.The approach

we propose

ls

quite

innovativeand isbased on mixed

virtual-real

_paradigm,

where a virtual reaMy application idaugmented withdynamichaptjcsystems consistjng of ad hocdevicesthat can be _programmed and configured forsimulating the

behav-iorofspecificobjects.

Fig,1.Productdesignand manufacturing _phases

1.

Introduction

The _productdesign_phase issubject to several recycles where designchoices are reviewed several times on the ba-sis of testingresults. When confidence

in

_productdesignis

reached, designisreleased

for

manufacturing

(Figure

1).

Mistakesinthedesign_phasehavethemost impact,and the amount of subsequent rework

downstream

is

stilisubstantial.

Itiswell known thatthe more designchanges occur closer

tothemanufacturing _phase,themore the

impact

on _product

developmentcosts increases,lnorder toreduce mistakes and

designchanges, _{practicesperforming}de$igntestingand

vali-dationefficientlyand effectively

-starting

since the early

con-ceptual design_phase-havedeveloped.

A common and traditional

_practice

makes use of_physical

prototypesfor

_performing

validation tests,Actually,thedesign

phase

issupported byCAS and CAD too]sthat_produceand releasedigitaldescriptionsof

_products.

Therefore,

mathemati-calmodels areimmediatelyavailable forsubsequent re-use. In therecent

_years,

several tooissupporting virtual validation and

simulation of

_products

havebeendeveloped.Thespreading of

these tools_hasbeen also supported byincreasingly_powerful hardwareenabling theexecution ofcomplex simulations. The practicebasedonsimulation isnamed "virtual

prototyping"'.

Compared tophysicalprototypingpractice,virtual

proto-types are in_generallessexpensive, allowfornot one but $ev-eral simulation runs on asinglemodel, areeasilyconfigurable and support variants.

Furthermore,

testsare repeatable, and results of validation are often immediately availablefor

_product

Fig,2.Validatien_{practices:physical}and digital_prototyping

design

review. We can say thatvirtual _prototypesofien _provide insightsthat

_physical

testingwould not reveal,Anyway, if digi-talprototypingdoes not completely substitute_physicaltests,

they helpoptimising and eliminatingredundancy intest

facili-ties,accelerating

life

testing,etc,

Currently,

companies tendtoreduce theuse and numbers

of _{physica[prototypes,}opting forusing them atthevery end of

the

design

_phases,and extensively using

digital

_prototyping

early in_productdesign

(Figure

2),

The_paper

describes

±

he

kinds

of_producttestingand valida-tion_practicesthatcan be_performedusing currentlyavailable

and emerging technologies.

2.

Product Design

Validation

two kindsof validation areusually_performed

in

the_product design:

1

,

Engineeringfunctionalaspects validation; 2.Ergonomicsissuesvalidation.

Thefirstkindofvalidation aims at testing_product

behaviour

and engineering aspects, likeaerodynamics, crash, stress,

de-T-ifif>\ffXkkg speclalissueofjapanesesocietyforthescienceofdesign vol.15-4 ne.60 200B NII-Electronic Mbra

25

y

formatjon,noise, vibration, etc. foday,these testscan

be

done

on thevirtualprototypeusing analysis and simulation tools

(for

example, tools likeANSYS, LMS, etc.).These±oolsallow

buildingsortsofvirtualIabteststhatareoftencomparable to

physical

hardwareones,

Forwhat concerns thevalidationof ergonomics issues,the ISO

9242

ergonomics standard _pointsout threeaspects to

consider:

. Effectiveness_:if_and how _atask

(objective)

is_{achieved ;}

. Efficjency_{: effortrequired} for_{achieving a}

_pre-defined

levelof effectiveness ;

. Satisfaction_: degree_{ofcomfort perceived}bythe_usen Ergonomicscan bedescribedas theapplication of scien-tificinformationconcerning humans tothedesignof objects, systems and environment forhuman use

(Ergonomics,

2001).

Some

type$ofergonomic validation are run usjng virtual pro-totypes:typically,safety tests.The validation of other ergo-nomics aspects of_products,likecomfort, usability,efficiency,

effectiveness, etc. are subject to _personalfeelingsand tastes

of users, related to their_preferences,emotions, skill,cultural

and socialaspects, etc.

Computer

representations cannot fully

address every ergonomic issueofthe_product;therefore

physi-cal _prototypesmay stjll_provenecessary.

Inthefollowing,we analyse the_possibilityof using reallvir-tuaiusers forrunning ergonomics testsof realfvirtual _products

(Figure

3).

ReaiusetiReatproduct: Users_performtraditionaltestsusing

physical

prototypes.

Wrtuatusetiurrtual _product:Virtualhumans are used for test-ingfew aspects of

_products,

throughthesimulation of digital

models, Some toolslikeEDS-Jack, Catia,etc,support the definitionof computational models of humans, the computa-tionofart

_positioning,

and thereforethe

_prediction

of postures and comfort. Theleftsnapshot of Figure4shows avirtual

hu-man drivingacan ltis

_possible

testinguser comfort by mea-suringvjrtuaihuman postureparameters

Iike

bending

degree

oflimbs,forcesexerted on theseat,etc,The snapshot on the right-hand side of

Figure

4

shows a stepof thesimulation of

the

field

of view ofa virtualcar

driver

when turning Ieft.This

study isparticularlyusefulfortestingthelevelof occlusion of acar doorpost

(Cu]ubert

Montanera,2002).

Whileseveralaspects can betested

in

avirtualenvironment, conversely, therearenotmodels thatcan

be

used

for

predict-ingmanipulation forcesapplied byusers.

26T-vrf

),\ensksug speclalissueefjapanesesecietyforthescieficeefdesign vol.15-4 ne.60 200S PRODUCT o=t's e ¢ L /111';11/1'11ill/1/'/1

'u.sets-pe.1mporm

testsusingvttu.al Eptl,'toty'plg,g･bf'.･il･' praduct&･eR"tfotimeht

///tt//t////..t/////t/../

_{/････t//./･///t,･/1} Simulatienef virtuathuman lnvirtuaienvironment Vsersperform traditionaltests usmgphysical protetypes

r

real vlrtual USER

Fig.3.Relationbetween user and _productconcerning the_product

vaiidation _pha$e

Fig.4.Snapshot ofthesimulation ofavirtualhuman drivingavirtual

cac and itsfieldofview when turningleft

ReaiusertWrtuat _product:users _pertormtestsusing virtual prototypingof productand environment. This jssue is

NII-Electronic Library Service

3.

Virtual

Prototyping

The

validationofergonomics aspects ofa_product

includes

thevalidation ofits

"look

and feel".Boththeseaspects require the

involvement

otrealuser$

in

thevalidation_process.

User$

interactwiththerepresentation

-real

orvirtual-of

_products

in

order totest

its

"look"

(attractiveness,

colour, textures,etc.) and

"feel''

{behaviour,

textures,etc.).Whilethefirstinvolves

only aspects relatedtovision,thelatterinvolvesaspects

re-lated

to

_physical

contact with objects. Thisistypicalof several products,especially consumer

_products

_(like

control devices

of cardashboardsand airplane cockpits, washing machines,

ovens, musiclvideo

_players,

etc,),butalso industrial

_products

(like

controldevicesof machine tools).

The use of

_physical

_prototypesinergonomics validation has

allthedisadvantagesand limitations_previouslydescribed.In

particular,

variantsand configurations cannot

be

_quicklyset up

and tested.Therefore,itseems interestingstudying the inter-action of real humans and virtual_products.

Inorder tounderstand which are the capabilities and ef-fectivenessof ergonomics validatlon _performedusing vlrtual prototyping,itisnecessary to have a view of the

state-of-the-artof technologies enabling virtual _prototyping,

Systems

where users interactwith virtual environments make use of VirtualRealitytechniques and technologies

{Earnshaw,

1

995).

Mainly,Virtual

Reality

state-of-the-artconsists ofvisualization techniques :Head Mounted Display

-

HMD

CVirtual

_Research Systems),largescreens, workbench

{Kruger,

1994),

CAVE

(Cruz-Neira,

1993),and others, that_giveusers theimpression of beingimmersed intothevlrtual world, Besides,more real-ismis

_provided

bystereo-viewing technology,AIImentioned technologiesallow the_perceptionof visual aspects

(look}

of

products.

On

theother hand,thosetasksinvolvinginteractionof users

withsystems requiremore appropriate media and modalities.

Some technologieshaveappeared recently and support

in-novative ways ofinteraction:whole handdevices

CVirtual

fech-nologies),3D devices

_{Logitech),

etc.The major [imttationof

these

devices

concerning taskslikevalidating interactive

de-vices

(ergonomics

testing)isthefac±thattheydo not

_provide

information

about thesense oftouch.

Some new devicesappeared inthelast

_years

on themarket, named

haptic

devices.

The

term ['haptic"

is

_{used to}

indiGate

both

tactileand

kinestetics

aspects ofthesense oftouch

(Bor-degoni,

2002).

Some

devicescan beworn on a finger,like

Fig.5.A user interactingwith avirtual pieceotcable using two

foM devices.The twospheres en thevisual scene correspond totheuser'$ fingersintherealworld moving thehapticdevices

thePHANfoM devices

_(SensAble

lechnologies);some others

areexoskeleton ref[ecting forceon theuser's hand,likethe SarcosDextrousArm Master;thereare actuated _gloves,like

the

Vlrtual

fechnologies

Cybergrasp

CVirtual

fechnologies),

and

bi-directionalactive deviceswith threedegreeoffreedom,like theHapticMaster

_(FCS-CS),

and $everal others. Whiiethese

devicesare genera[-purpose,thereare other devicesthatare

oriented tospecific applications, especially forthe medical field,such asforlaparoscopy,endoscopy, etc,

(MMVR

2003). Another interestingtypeofdevice

i$

the so-called

"matrix

dispiay"

thatsupports

both

thevisualand

haptic

simulation ofobjects on aunique

device

(lwata,

200]).

Object

charac-teristicsaresimulated throughhapticmodels, and _perceived throughvision and thesense of touchbyusers wearinglusing hapticdevices.

State-of-the-artof _{general-purpose}hapticdevicessupports

the simulation of activities like"inspect

an object", "push a button",Figure5 shows a user wearing a PHANfoM device, inspectingavirtualcable.

Othermore complex tasks,like"turn a knob"aredifficul±, if not impossible,to_performwith thesedevices,Thisismain[y

due tothecurrent developmentstage of commercia[ devices thatdo not allow thesimulation of complex situationslike6 d,o.f.operations, operations requiring torquecomputation, rendering othighforces,and so on.

Alternativeapproaches

for

thosekindsofvalidationconsist of configurable hardware

_prototypes.

Most of theexamples

have

been

realised

in

theautomotive sector.

Some

research centers of automotive companies use

_physical

_prototypes

thatallow fortheconfiguration of several typesofvehicles, fvrfy#asskmg specialissueotjapanesesecietyferthesc]enceetdeslgn vel.15-4 ne.60 2008 NII-Electronic Mbra

27

y

bymodifying $ome _parameters,like_position,orientation, etc.,

ofsome

_physical

_parts

(Caputo,

2001)

_{Di

Girolamo,

2001).

Programmable VehicleModel

_(PVM)

of Prefix

{Prefix}

isan example of configurable vehicle.

These

solutions are usually

very expensive and require longre-configuration time,and are

mainly developedforaunique _producttype.

An alternative approach to theuse of

haptic

devicesand

configurable hardware_prototypesconsists of dynamichaptic systems whose behaviorcan

be

_programmed.

4.

Dynamic

Haptic

Systems

Dynamic

haptic

systems consist of ad hocdevicesthatcan

be

programmed and configured

for

simulating thebehaviorof

specificobjects.

For

demonstrating

theconcept, we havedesignedand built

some hapticsystems thatcan be dynamicaHy

_programmed

forthevalidationofergonomics and usability designaspects. Figure

6

shows theCAD model of a dynamic knob _prototype

(Gandolfi,

2002), Figure7 shows thedeveloped _physical prototypeof theknob.Itconsists of a mechanical system, a

controllerconnected toa computer running a _programfor set-tingtheknob_parametersjnreal-time. The user interfaceof the configuration_programisshown inFigure8.

The

knob

prototypecan be mounted and _positionedinan

environment. Forexample, itcan be mounted on a car dash-boardor on awhite

_goods

_panel

board,Userscan beasked to tryand use theknob.Iftheydo not likeitsresponse and behavior,and want tomodify it,theknob parameterscan

be

changed

in

real-time.When the

_preferred

configuration is reached, corresponding

_parameters

and datacan bestored

by

the

control system

for

subsequent use

in

design

review

phases.

These

hardware

prototypessimulate _products

behaviour,

butusually do not implementaspects concerning the`[look" _of a _product,

Usually,

they areuglymechanical systems,

simulat-ingthebehaviourof a device,

but

not theappearance ofboth

the deviceand thewhole environment. For example, theknob developedand shown in

Figure

7

can

be

used tosimulate a

knob mounted on acar dashboard.Inorder to _provideusers

a view of how thewhole environment will

be

once thecar

is

bujlt,knob visual aspects are not satisfactory as they are im-plementedinthehardwareprototype:thevisual appearance of

theknobdoes not lookreal and cannot bechanged; besides thattheusers do not see the environment where the knob

28T-wy\assNzae

speclalissueofjapanesesocietytorthesciemceefdesign vol.15-4 ne.60 200S

Fig.6.CADmodel ofadynamic knob _prototype

Fig.7.Knob dynamichardwareprotetype

m -"Z

.dwN

OneHblvop

Ne'b'Utt'op'e'li'mp.Q'ipttu roalavadi.

LptW{11 Etts.exfi. Tpku

"..

{uaan.,:L":al'k'op

'

irr

IS-- rt

op'pa=rmrmj

as" V inpae-/nyak=wi FMun Fig.8. tu Setvtny"tsuua IVtsmoaw in

,t

dmo e3!, 1 ltteec em' IS ee400

,ts/ta

i3 /,/SeeO 2S/M :d e::･tee ]Jbsc

ii

ll:?･

!k,'

_i#1

Ii;

u

S'`i.es,

.mam

i'sc.

osw

1: lllUO Osw Et 11smo npms tfi :7･pm Orm

gl:---un-,-iL::-.ur.-..-r::J

.

"

Userinterfaceofthe_programforsettingtheknob_parameters

shou[d

be

mounted on

(the

car dashboardinour example).

On

theother hand,ifwe

_provide

thevisualrepresentation of

thecar

dashboards

only,theuserwillnot beable tovalidate

the ergonomics aspects ofdevices mounted on the

dash-board,

like

the

knob.

A so]ution forovercoming this

_problem

ismixing realand

NII-Electronic Library Service

5.

Mixed Real-VirtualPrototyping

When only real or only virtual informationisavailabie ina

system, itmight not be enough byitselfso as users can fully interactwith thesystem and accomplish some _giventasks,ln some cases, itmight beuseful conceiving systems thatmix virtual and real information.

Thesesystems are made oftwo _parts

(Figure

9):

. one _partconsists ofrea[ environment and objects;

. theother _partconsists otvirtual representation of both envi-ronment and objects,

An

application can

implement

the

human-computer

interac-tionas a mix ofvirtual and real modalities, where virtual and

real

information

is

_presentatvariable_percentage,

lf

the

infor-mation

fully

comes fromtherealworld we talkabout Reality;

if

the

information

fully

comes

from

thevirtualworld we talk

about

Virtuality.

If

both

themodalities are_present

in

the appli-cation,we talkabout MixedReal-Virtualinteractionmodality,

From a _{practicalpoint}of view, mixed real-virtual interaction

can beimplementedadding informationtothealready existing

one.

Ifa system is

primarily

basedon realinformation,the

quan-tityof informationthatcan be made ava[lable foruse can be improvedadding virtual informationtoit.These systems are

called Augmented Realitysystems

{Figure

1

O).

The

quantity

ofinformationavailableintheaugmented interactioninterface isincreasedof avalue thatdepends on the_quantityof virtual informationthatisadded.

Conversely,ifan applicationismainly based on virtual rep-resentation of environment and objects, realobjects can

be

added tothem.Inthiscase, we talkabout Augmented Vir-tualitysystems

(Figure

1

1).

Also

in

thiscase, thequantityof

informationavailableintheaugmented interactioninterface

is

increased

ofava[ue that

depends

on thequantityofreal

infor-mation thatisadded.

The type ofinteraction interfacethat isimplemented ina

system depends on several _parameters,such as the

applica-tionrequirements, the kindsoftechnology used and its limita-tions,thetypeof

data

and

information

initially

available,etc.

An application might implement various types of interac-tionmodalities thatareavailablesimultaneousiy, and can be

changed continuously

{Milgram,

1994>.Duringtheinteraction with a mixed real-virtual application environment, users can vary continuously the

type

of interaction,adopting the interac-tionmodality thatismore suitable and convenient, _giventhe

Fig.9.Mixingreaeand virtualintormation,according todifferent modalityshare

Fig.10.Augmented Realitysystems

itk'=tibl-8g-Em=c

_Higlt Lrw i

'Yiijtilai･

Quantltytf inionrnatlon Fig.11.AugmentedVirtualitysystems

context and thetaskto_perform

{Figure

]2>,

Based on thisreasoning, theergonomics application using the

dynamic

hapticsystem we havedevelopedisoftype Aug-mented Virtuality.Infact,a realisticvisual representation of the environment and theinteractiondevicesisusually available,

since itcomes fromthedesign_phase

{CAD

models), We have added informationabout thosecharacteristics oftheknobthat cannot be _perceivedthroughthevirtual informationonly,by integratinga real _physicaldeviceintothesystem. The virtual

7'iftyvaneeksce speeialissueotjapanesesocletyferthescieneeofdesign vol.15-4 no.60 20e8

NII-ElectronicMbra

29

Hg,12, Continuous variation of typeof interactioninamixed virtualapplication

information

(digital

model of car dashboardand knob)is

aug-mented withrealinformation

_(physical

dynamic systems)

yield-ing

an

Augmented

Virtualitysystem.

6.

Conclusions

This

_paper

has_presentedtheresults of a research work we havecarriedouton thetopicof ergonomics validation of inteF

activedevices,The approach we haveexperimented using

vir-tualrealityscenarios augmented with dynamichaptic

devices

has _proventobeeffective for_preliminaryevaluation of ergo-nomics aspects, mainly concerning effectiveness, efficiency

and usability inusing interactivedevices.More experiments are

_going

tobeset up,forimprovingthemethod and

propos-ingitforother application areas,

besides

theautomotive one.

Acknowledgments

Theauthors would liketothankFrancoDeAngelisand

Gior-gio

Colombo oftheKAEMaRT Group

-

Dipartimentodi

Mec-canica

-

PolitecnicodiMHano who have contributed to this work,

Engonomics.

_(2001).

InternationalEncyelopaediaof

icsand Human Factors,Vol,

1,

faylor

and

Francis,

don.

Earnshaw R.A.,GiganteM.A.,

Jones

H.J.,

₍₁

995).

Virtual

itySystems,Academic

Press.

FCS-CS, HapticMaster,

URL

: botics2.htm

GandolfiE

_(2002).

Studio

e sviluppo

di

dispositivi

haptic pervalidazioni inambienti virtuali

finalizzati

allosvjluppo

prodotto,Laurea

Thesis,

Universita'diParma

--

Facolta'di

lngegneria,

2002.

IwataH.,YtinoH.,

Nakaizumi

E,Kawamura R.,

_(2001).

Project

FEELEX

:AddingHapticSurfacetoGraphics'',

ingsof SIGGRAPH 2001 _,LosAngeles,CA.

KrugerW.,FrolichB,

₍₁

994).

The responsive workbench, ]EEE

Computer

Graphics

and

Applications.

Log]tech,

URL

:www.logitech.com

MilgramR,lakemura H.,UtsumiA.,KishinoE

₍₁

994). mented

Reality

:aclassofdisplayson therea[ity-virtuaiity continuum,

(SPIE

Vol.

2351,

feIemanipulatorand

resence technologies,l994),

MMVR

2003,

Proceedingsof MedicineMeets VirtualReality

Conference

2003, NewportBeach,CA,January2003.

Prefix,URL :www.prefix.comlpvm.html

SensAble

fechno[ogies,

lnc.,URL :www.sensable.com

VirtualRealityforIndustrialApplications.

_{1

998).Fan Dai

_(Ed

.),

Springer.

Virtual

Research Systems, lnc.,URL :www.virtualresearch.

comVjrtual

lechnologies,lnc.,URL :www.virtex.com

'

References

BordegoniM.,CuginiU.,

_(2002).

Visual-HapticMultimodal

Human Computer Interaction,Yearbookof theAdificial

Methodological

Aspects

and

Cultural

lssues,Vol.I,

ited

by

M.Negrotti,PeterLangEd,,2002,

Caputo

E,Di

Girolamo

G.,MonacelliG.,Sessa E

_(2001),

The

design

of a virtualenvironment

for

ergonomic studies, Proc.ofXllADM lnternational_{Conference,Rimini,ltaly.} Cruz-Neira

C.,

Sandin

D.J.,DeFantirA.

_(1993).

screen projection-basedvirtuaireality:The designand implementationofthe

CAVE,

inComputer Graphjcs, ited

by

J.T.

Kajiya.

Culubert

Montanera,

fovar

Gomis

G.,

(2002).

Use ofaVirtual Human forergonomics analysis of a carcockpit, Laurea ThesisUPC-Barcelona and Dip.Meccanica-Politecnicodi

Milano{Erasmusprogramme).

DiGirolamoG,,MonacelliG.,MartorelliM.,Vaudo

G,

_(2001),

Use of VirtualMock-ups

for

Engonomics

_Design,Proc.of

7thlnternational

Conference

AT/A,

Florence,

ltaly.

30T-if-)7MvereEg

speciali$sueefjapanesesocietyferthescienceofdesign