to

uzalm, Un

Keywords:AssemblyVirtual

opeust

e manueir req

models of the same product share the capacity of one assembly line

execute action sequences to solve problems (Rittle-Johnson et al.,2001). This means that an assembly operator knows how andwhen certain procedures should be performed in order to accom-plish a given task. By having procedural skill related to a specic

commonly performed on pre-series (prototype) vehicles (Krammertions such as: highariants are built toeated exposure to

the use of trainingith several advan-g and exibility inining. Kraus andl training is costavel expenses for

trainers/trainees as training can be delivered on-site; minimisingdown-time as training can be exibly undertaken around trainees'work schedule; and less demanding on personnel resources astrainees can train independently. Boud et al. (1999) found thatoperators who have used a virtual training system learned newprocedures effectively and performed better on real assembly tasksthan those using solely written instructions. Similar trends havealso been shown in the application of virtual training in other areassuch as aircraft maintenance (Barnett et al., 2000), machine

* Corresponding author. ITRC Building (B03), University Park, Nottingham NG72RD, UK. Tel.: 44(0)115 951 3808.

Contents lists availab

Applied Erg

els

Applied Ergonomics 46 (2015) 144e157E-mail address: setia.hermawati@nottingham.ac.uk (S. Hermawati).to support the competitiveness of automotive manufacturing. As aconsequence, operators on the nal assembly lines are required toswitch effortlessly between assembly operations for one model tothe next.

Nof et al. (1997) dened assembly as the aggregation of allprocesses by which various parts and subassemblies are built togetherto form a complete, geometrically designed assembly or product (suchas a machine or an electronic circuit) either by an individual, batch orcontinuous process. Assembly operations can be considered asskill-based operations that require procedural skill i.e. an ability to

et al., 2011). This approach has substantial limitacost; only a low number of vehicles and product vkeep the cost down; and parts wear from repassembly and disassembly operations.

In contrast to training on pre-series vehicles,in virtual environments has been associated wtages such as a standardised approach to traininconducting, progressing and evaluating traGramopadhye (2001) also argued that virtuaeffective for several reasons: elimination of trand robustness (Hajarnavis, 2012). It is also common that several new product and its variants are introduced. Operator training isTraining

1. Introduction

Final assembly lines in automotiva low degree of automation due to thhttp://dx.doi.org/10.1016/j.apergo.2014.07.0140003-6870/ 2014 Elsevier Ltd and The Ergonomicsused to analyse the results of interviews with various stakeholders (17 and 28 participants at OPEL andVOLVO, respectively). The results show that there is a strong case for the implementation of virtualtraining for assembly tasks. However, it was also revealed that stakeholders would prefer to use a virtualtraining to complement, rather than replace, training on pre-series vehicles.

2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.

facturers typically haveuirement for exibility

assembly task, an operator will have a mental representation of theassembly task details (e.g. the number and order of steps involved,and detail of what needs to be done in each step). Therefore,training is crucial in developing operators' procedural skills when aAvailable online 15 August 2014analysis, timeline analysis, link analysis, Hierarchical Task Analysis and thematic content analysis wereAccepted 26 July 2014 conducted at OPEL and VOLVO Group to identify assembly training needs and a subset of requirements;and to explore potential features of a hypothetical game-based virtual training system. StakeholderUnderstanding the complex needs of auassembly lines

Setia Hermawati a, *, Glyn Lawson a, Mirabelle D'CrLina Andersson c, Maria Gink Lovgren c, Lennart Ma Human Factors Research Group, Faculty of Engineering, The University of Nottinghamb Adam Opel AG e General Motors Company, Rsselsheim, Germanyc VOLVO Group, Gothenburg, Sweden

a r t i c l e i n f o

Article history:Received 1 June 2012

a b s t r a c t

Automobile nal assemblywhere multiple variants m

journal homepage: www.Society. All rights reserved.motive training at nal

a, Frank Arlt b, Judith Apold b,skold c

ited Kingdom

rators must be highly skilled to succeed in a low automation environmentbe assembled in quick succession. This paper presents formal user studies

le at ScienceDirect

onomics

evier .com/locate/apergo

tor tperconasseifcrdedns w

wor

be tsks?en d

ughhe trroaca

form)

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e157 145Table 1List of questions in the semi-structured interview.

Roles andresponsibilities

1.1. Summary of your rolea

1.2. What is your current involvement with opera1.3. Please provide a task level description of the o1.4. Does the task that the operators perform vary1.5. What is your production rate? i.e. how many1.6. Do any of the assembly tasks pose particular d1.7. How are issues with assembly operations reco1.8. Do you receive feedback on assembly operatio

Workplace andworkenvironment

2. What are the conditions in which the operatorssocial environmentb

Training 3.1. What are the tasks that the operators need to3.2. How are they currently trained to do these ta3.3. What is the timeframe of the training (i.e. wh3.4. What is your opinion of the training?a

3.5. What are the key skills or knowledge being ta3.6. What is good about the current approach to t3.7. What are the difculties with the current app3.8. Can you suggest how to improve the training?3.9. What do you consider the most important perHuman: Of the operators' job (e.g. time, errors, operations (Lin et al., 2002), surgical operations (Seymour et al.,2002; Larsen et al., 2009), and the military (Gerbaud et al., 2008).

There have been many virtual training systems which weredeveloped to aid the acquisition of procedural skills related to as-sembly tasks. However, most systems are aimed at supportingtraining of maintenance tasks in which knowledge of both assem-bly and disassembly are part of what is acquired during the trainingof the tasks (Webel et al., 2013; Xia et al., 2012; Peniche et al., 2011;Gutierrez et al., 2010; Abate et al., 2009; Oliveira et al., 2007; Wangand Li, 2004; Bluemel et al., 2003; Vora et al., 2002); only a few arededicated solely to support training of assembly tasks (Lili et al.,2009; Brough et al., 2007; Abe et al., 1996). On studying thesepublications further we also found an indication that formal userrequirements elicitation was rarely conducted prior to the devel-opment of the systems. Failure to perform user requirementsgathering means that the system is at risk of being unable toaddress users' real needs and its usability is reduced (Nielsen, 1993;Maguire and Bevan, 2002). To the extent of the knowledge of thispaper's authors, there are limited studies (e.g. Anastassova andBurkhardt, 2009) that focus specically on user needs and re-quirements for training of assembly tasks in the context of

Operational: Of the training (e.g. operator must complBusiness: (e.g. a Timeframe reduction for the training)3.10. What are the requirements for authoring training3.11. Does training improve productivity? If so, how?e

Process and workow 4.1. What software tools do you currently use as part4.2. What is the communication/information ow (i.e.4.3. What information is required by planners/trainers4.4. Within the vehicle development lifecycle, what inrequired to specify assembly instructionsa

Game-basedvirtual training

5.1. What are your initial thoughts about training usin5.2. Do you think this approach could improve trainin5.3. What problems would you anticipate?a

5.4. What are your initial thoughts about a game-like5.5. What are your thought on the capture and feedbasystems engineers?a

5.6. Do you own, use or have been tried an X360, PS3game using a wireless controller?a

5.7. Would you like to be involved in the VISTRA projedeveloped technologies?a

a All stakeholders.b Operators.c Operators, engineers, supervisors.d Engineers.e All stakeholders except operators.raininga

ators job, i.e. the detail of what they actually dob

siderably, i.e. do they do many different operations, or are they mostly similar?c

mbly operations are required per hour?b

ulties to the operators? Do any tasks cause frequent problems?a

?a

hich are difcult or time-consuming? If so, how?d

k? E.g. shift work, noise, lighting, please describe the physical workplace and

rained for (e.g. tting components on vehicle assembly lines)?bb

oes it take place in relation to vehicle launch?)c

t?a

aining?a

h of the training, or what problems exist?a

ance measures and goals?aautomotive manufacturers. This paper aims to ll this gap in thehope that the information could be used to promote the develop-ment of a virtual training system that matches the requirements ofassembly task training. This paper's contribution lies on its wealthof ndings which were gathered from two different automotivemanufacturers to reect the complex needs of training in auto-motive manufacturers.

Furthermore, game-based training has received increased in-terest over the last decade and is applied in a variety of eldsincluding business (Leger, 2006), education (Jong et al., 2008) andmilitary (Beal, 2009). This popularity is attributed to the hypothesisthat it can lead to skill acquisition and retention due to its ability toengage learners (Colquitt et al., 2000; Prensky, 2001) and is sup-ported by empirical research evidence (Corbeil, 1999; Engel et al.,2009; Garris et al., 2002). There have also been indications thatskills learned in game-based training environments transfer toreal-life situations (Gopher et al., 1994; Topolski et al., 2010).Despite the latest evidence of the effectiveness of game-basedtraining there has not been any study that investigates thepossible application of game-based training within amanufacturing setting in the automotive sector. This paper aims to

ete task correctly XX% of the time)?sessions (time available, man-power etc.)?e

of the training process? (e.g. CAD systems or production planning software?)a

who provides what information to who)?a

?a

formation related to training is required and when? i.e. when are engineers

g virtual systems?a

g?a

system for training?a

ck of assembly issues to the product designers and manufacturing

or WII based game console? What is your impression or experience controlling a

ct? i.e. can we contact you again to obtain your feedback on the

ll this gap by investigating end users' opinions towards game-based virtual training. While this was necessarily hypothetical, asno game-based system was available yet for them to experience,the feedback was gathered in an attempt to identify potentiallybenecial features for such a system and highlight areas that couldbe worthy of future investigation.

In summary, the objectives of this paper are as follows: 1) toinvestigate user needs and requirements for training assemblytasks in the context of automotive manufacturers; and 2) toinvestigate users' opinions towards game-based virtual training.

2. Methods

2.1. Participating companies

This study involved two automotive manufacturers (OPEL andVOLVO Group), which were chosen to exemplify the complexity ofboth organisation and assembly tasks in such organisations.

concept of game-based virtual training, the following denition ofgame-based virtual training was given: a virtual training systemwhich can be used to train assembly tasks without physical partsand could be in the form of training software that runs on a desktopcomputer or uses technology similar to Nintendo Wii; also thetraining will be game-like and perhaps involve a competition orcontest.

A total of 45 participants were involved in the interviews, ofwhich 17 participants were from OPEL and 28 participants werefrom VOLVO (see Table 2). The semi structured interviews lastedapproximately 30mine1 h, with between one and four participantsbeing interviewed together in each session. During the interviews,participants were encouraged to talk freely. Involvement from theinterviewer was limited to prompting participants to provide moredetails or expand on key issues. Due to concerns over commercialsensitivity and the resources available for analysis, responses fromparticipants were mainly recorded through handwriting.

Observations at the nal assembly lines lasted between 45 minand 1 h. Notes regarding nal assembly tasks and the environment

the effectiveness of product/system design and development invarious elds (Amiri et al., 2012; Atkinson et al., 2001; Neuman,

osit

er o

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e1571462.2. Interviewees

The stakeholders were identied through discussion and liaisonbetween human factors researchers and representatives of endusers of both automotive manufacturers. During the discussions,the end user representatives were encouraged to adopt broaddenitions of stakeholders i.e. anyone in their organisation who islikely to use or be affected (directly or indirectly) by nal assemblytraining. The end user representatives were requested to provide ageneral overview of the demographics and backgrounds for eachidentied group of stakeholders (see Appendix A).

2.3. The semi-structured interview and observation

A semi-structured interview was developed following a brain-storming session that involved end user representatives, humanfactors researchers and system developers. The interview questionswere organised in ve main sections: 1) roles and responsibilities;2) workplace and work environment; 3) training; 4) process andworkow; and 5) game-based virtual training. During the devel-opment of the interview, feedback from end user representativeswas obtained and used to improve the questions of the semi-structured interview. Table 1 shows the list of the questions ineach category and indicates to whom these questions weredirected. In order to help interviewees understand and visualise the

Table 2Stakeholders, number of participants and their average number of years in current p

Stakeholders Numb

OPEL Manufacturing System Engineer Manager 1Manufacturing System Engineers 2Virtual Engineers 2System Owner 2Launch Team Manager 1Core Team Manager 1Core Team 2Launch Team 1Supervisors 2Team Leaders 2Operators 2

VOLVO Design Engineers 2Production/Introduction Engineer 3Virtual Manufacturing Engineers 2Technical Preparation Engineer 1Team Leaders 3Pilot Plant Team 3

Operators/Key Operators 142004; Brooks, 1998; Neary and Sinclair, 1998). In this study, thestakeholder analysis helped to identify and examine users thatare likely to be somewhat affected by implementation of atraining approach. The stakeholder analysis was based on theresponses from interview questions 1.1 and 1.2.

ion.

f participants Average number of working years in current position

2012Information not provided1232Information not provided22241820.524.5

3.53.38187.59.8of assembly lines were taken. Short informal discussions with op-erators were also performed when required and/or possible. Theobservation complemented the results of the semi-structured in-terviews regarding the details of workplace, work environment andassembly tasks (interview question 1.3 and 2).

2.4. Data analysis

The interview and observation results were analysed with vequalitative methods. The choice of these methods followed theguidance provided by Leonard et al. (2006) which considered fac-tors such as the purpose of the study (explanation driven orimplementation driven) and availability of resources (time, expe-rience, funding and previous research). A detailed description ofthe methods used is given below:

1. Stakeholder analysis has been recognised as important to ensure2.5

1.2

Involvement with current training

) Being trained at pilot line (dedicated training centre/location)by launch team (OPEL) or pilot plant team (VOLVO).Being trained at assembly line by Team Leaders (OPEL) orKey Operator-Teachers (VOLVO).

Train experienced operators at pilot line.) Train operators at assembly line.

inator (VOLVO) Train experienced operators at pilot line.Managing training for operators at assembly line.

Rectifying incorrect assembly operations at main assembly line.

Maintain the system.

ry o

tooeredrs atnoly linraine

Ergonomics 46 (2015) 144e157 1472. Timeline analysis combines functions or tasks with time-relatedinformation. It is useful to create plots that show the temporalrelationship among tasks, their length and timing (Nemeth,2005) as well as to assess task allocation and communicationrequirements identication (Kirwan and Ainsworth, 1992).Timeline analysis was based on the responses from question 3.3and 3.10.

3. Link analysis is useful to provide descriptions of the in-terrelationships between elements/components and their fre-quency (Chapanis, 1959). Although it is traditionally used to

Table 3Categories of stakeholders that were identied from responses of questions 1.1 and

Stakeholder category Stakeholders

Trainee Team leaders (OPEL), Key Operators-Teachers (VOLVO

Operators

Trainer Launch team (OPEL), Pilot plant team (VOLVO)Team leaders (OPEL), Key OperatorseTeachers (VOLVO

Coordinator Launch team manager (OPEL), Pilot Plant Team CoordSupervisors (OPEL), Team Leaders (VOLVO)

Indirect benefactor Core Team, Manufacturing System Engineers (OPEL),Production Engineers (VOLVO)

System champion System Owner (OPEL), (VOLVO)

Table 4Summary of responses to questions related to timeline analysis.

Interview question Summa

3.3. What is the timeframe of the training (i.e. when does it take placein relation to vehicle launch?)

TrainingA staggoperato

3.10. What are the requirements for authoring training sessions(time available, man-power etc.)?

There isassembof the t

S. Hermawati et al. / Appliedarrange equipment and individuals in a workplace and arrangecontrols and displays on a console, it has also been used forother purposes such as monitoring and analysing communica-tions in rescue operations (Thorstensson et al., 2001) andnding information needs (Albinsson et al., 2003). Link analysiswas based on the responses from questions 1.7, 1.8, 4.1, 4.2, 4.3and 4.4.

4. Hierarchical Task Analysis (HTA) is a form of task analysis to studyactions and cognitive processes that are required by an operatorin order to complete the system goals (Kirwan and Ainsworth,1992). It can be applied to a variety of applications (Stanton,2004). The HTA in this study was based on the responses fromquestions 1.3, 1.4, 1.5, 2 and observation results.

5. Thematic analysis provides a descriptive presentation of quali-tative data by identifying common topics and categorising thequalitative data under suitable themes (Franzosi, 2004). It hasbeen shown to successfully support development of techno-logical products (Minocha and Reeves, 2010; Coleman et al.,2010; Artman and Zallh, 2005; Sawasdchai and Poggenpohl,2002). Because it is guided by a specic research question orsub questions by making inferences of an intended population(Lapan et al., 2012) it can be used to provide answers for designquestions that could not be directly addressed by the analysesdescribed above. For this study, thematic analysis helped toselect, focus and transform qualitative data into manageableinformation segments to show their patterns. Thematic analysiswas mainly based on the responses of questions that were notused in other analyses.2.5. Identication of training needs, training requirements andpotential features of game-based virtual training

The results of data analysis were used to identify training needsOn completing the data analysis, the results of the data analysiswere reported back to, and veried by, end-user representatives.Any found inconsistencies, especially related to the timeline anal-ysis and link analysis, were then corrected.

f responses

k place towards the end of vehicle launch at both automotive manufactures.approach was adopted i.e. train the trainers at pilot line who then trainassembly line.specic duration of a training session, especially for a training session ate. A training session at an assembly line largely depends on the availabilityr. The pre-series vehicle is required for training at both pilot and assembly lines.and drafted as a subset of the user requirements. Features of game-based virtual training that could potentially be benecial were alsodrafted. These drafts were presented to the end user representa-tives and the system development experts. The representative endusers judged the importance (high, medium, low) of each identieduser requirement/potential feature while the system developmentexperts judged the technical feasibility (high, medium and low) ofpotential features. During the renement process, answers toquestions such as how important are the different requirements/

Fig. 1. Timeline analysis results of OPEL (top) and VOLVO (bottom).

SIssues related to assembly operations found during training at the assembly andp(VEIslaCDPaNhBdtoD

Mthey can make adjustments on product design and assembly operations.TwD

Ergfeatures for the users? and what is the consequences for users ifwe do not implement a specic requirement/features? werecontinually assessed. In this process, human factors researchersacted as independent auditors to provide more objectiveperspective of stakeholder positions and interests (Crosby, 1992).

3. Results

For each type of data analysis a summary of the relevant semi-structured interview results will be presented and is followed byan account of the analysis results.

3.1. Stakeholder analysis

Table 5Summary of responses to questions related to link analysis.

Interview question

1.7. How are issues with assembly operations recorded?

1.8. Do you receive feedback on assembly operations which are difcultor time-consuming? If so, how?

4.1. What software tools do you currently use as part of the training process?(e.g. CAD systems or production planning software?)

4.2. What is the communication or information ow (i.e. who provideswhat information to who)?

4.3. What information is required by plannersa/trainers?

4.5. Within the vehicle development lifecycle, what information relatedto training is required and when?

a Engineers.

S. Hermawati et al. / Applied148The responses from questions 1.1 and 1.2 are shown in AppendixB, which lists all stakeholders, their roles and responsibilities andtheir current involvement in training. We observe that both auto-motive manufacturers follow the commonly adopted on-the-jobtraining approach which requires a trainee to work side-by-sidewith a trainer on the assembly line. Five categories of stake-holders were identied based on their involvement with trainingi.e. trainee, trainer, coordinator, indirect benefactor and systemchampion. Table 3 shows the detail for each category. The stake-holder analysis also revealed that stakeholders' involvement wasaffected by whether the training takes place at the pilot line orassembly line.

3.2. Timeline analysis

Responses from questions 3.3 and 3.10 for both automotivemanufactures are summarised in Table 4. Detailed informationfrom the responses was used to create timelines (see Fig. 1) whichshow the succession of events prior to the training of new productsat both automotive manufacturers.

The timelines show that the succession of events prior to thecommencement of training of new products at OPEL and VOLVO aresimilar. Virtual builds, which precede all other events, bring prod-uct and manufacturing engineers together to identify and resolveany issues related to the design (at an early stage of product design)and assembly processes of products (at a later stage of productdesign). Virtual builds are aided by software that accommodates 3Dsimulation of parts and assembly processes. This is followed byphysical builds, where a mock up is created by the Launch Team(OPEL) or Pilot Plant Team (VOLVO) and used to further identify andresolve any issues that are missed in the virtual builds. Virtualbuilds and physical builds are parts of the product design processand are aimed at rening the design and assembly planning of anew product. They are part of the current training.

Training on pre-series vehicles only commences once the virtualbuild and physical build are completed. This takes place after acertain time gap and is performed at two different locations: pilotand main assembly lines. This time gap is required to provideadequate hardware (training vehicles) and to organise the trainingsessions. Although the timeline shows that there is a time period

rainers require information on relevant assembly sequences and the reasonshy assembly sequences or steps have to be done in certain ways.etail information is shown in Fig. 2.ilot lines are recorded and collected by the Launch Team (OPEL)/Pilot Plant TeamOLVO)/who will then pass them to Manufacturing Engineers (OPEL)/Productionngineers (VOLVO).sues related to assembly operations that are found once a product has beenunched are recorded by Team Leaders (OPEL/VOLVO) and handled byore Team (OPEL)/Technical Preparation Engineers (VOLVO).uring training at assembly and pilot lines, Manufacturing Engineers (OPEL)/roduction Engineers (VOLVO) receive feedback on assembly operations whichre difcult/time consuming.o further feedback is received by the two stakeholders above once a productas been launched.oth OPEL and VOLVO use commercial CAD and production planning softwareuring virtual builds and production planning. However, no software is usedsupport training. Assembly sequences are printed and distributed on paper.etailed information is shown in Fig. 2.

anufacturing engineers require information regarding assembly issues so thatummary of responses

onomics 46 (2015) 144e157dedicated to train operators for a new product, the stakeholdersmentioned that due to various factors (e.g. limited hardware/pre-series vehicle to train, availability of trainers, and productivitypressure on the lines) assembly line training is frequentlyshortened.

3.3. Link analysis

Responses from questions 1.7, 1.8, 4.1, 4.2, 4.3 and 4.4 are sum-marised in Table 5. Detailed information from the responses wasused to create link diagrams (Fig. 2) which show the informationow and relationships among stakeholders.

Fig. 2 shows that, for a new product, there is already a system inplace to capture assembly issues that indirectly connect operatorsand manufacturing engineers (OPEL)/production engineers(VOLVO). Once a product is launched, a different mechanism tocapture assembly issues is applied. This mechanism keeps themanufacturing engineers (OPEL)/production engineers (VOLVO)out of the loop.

3.4. Hierarchical Task Analysis (HTA)

Table 6 shows a summary of the responses from 1.3,1.4,1.5 and 2while the following summarises relevant ndings from theobservation:

or O

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e157 1491. Operators at OPEL are required to perform assembly tasks in fastmoving assembly lines while operators at VOLVO are required toperform assembly tasks with an appropriate speed (neither toofast nor too slow), especially for assembly tasks which involvemore than one operator.

2. The majority of assembly operations at VOLVO are collaborativeassembly operations due to the size and weight of parts thatwere involved in assembly tasks.

3. There is a higher need for assistance tools at VOLVO such aslifting devices due to the size and weight of assembly parts.

Fig. 2. Link analysis results fCombining the ndings from the observation results and re-sponses from questions 1.3 and 2, HTAs were created for bothautomotive manufacturers (see Fig. 3a and b).

Table 6Summary of responses to questions related to Hierarchical Task Analysis.

Interview question Summ

1.3. Please provide a task level description of the operators job,i.e. the detail of what they actually do2

Detail

1.4. Does the task that the operators perform vary considerably,i.e. do they do many different operations, or are they mostly similar?3

Operaare limat VOL(up toThe ntheir cat theto undcontaiAn unassemencou

1.5. What is your production rate? i.e. how many assemblyoperations are required per hour?2

Cycle

2. What are the conditions in which the operators work?E.g. shift work, noise, lighting, please describe the physicalworkplace and social environment2

PlantsresponAt VOof fouup to(16 op3.5. Thematic analysis

Table 7 provides a summary of responses from interviewquestions that were not used in other analysis. The main themesderived from all shown responses are described below. Bold font isused to highlight the theme.

The interview revealed that one of the main drawbacks of thecurrent approach to training is the fact that actual training timecan be shorter and less extensive than the initial training plan,resulting in a lack of training for operators. An example of a

PEL (left) and VOLVO (right).comment that reected this circumstance was given by a teamleader: One day for training is good, but we don't have time. Difcultfor new people to deal with time stress. Quality is most important. It isright people on right jobs, because some people nd some operation

ary of responses

information is shown in Fig. 3a and b.

tors are required to work at different stations or work cells. Rotations at OPELited among work stations (up to 16 assembly operations), whereas rotationsVO could either be among assembly cells (up to 24 operations) or stations96 assembly operations).umber of product variants at VOLVO is larger than OPEL because VOLVO allowsustomers to custom make their orders; some variants might appear only onceassembly line. Therefore, operators at VOLVO are required to have an abilityerstand SPRINT i.e. a paper-based instruction of assembly procedures whichned technical terminologies and codes that described assembly procedures.derstanding of technical terminologies and codes allows Operators to performbly tasks for a new variant at the main line, even if they had neverntered it before.time at each work station is 1e2 min at OPEL and up to 14 min at VOLVO.

at OPEL and VOLVO are both bright and noisy. At OPEL, each team leader issible for up to four work stations, with 1e2 operators at each work station.LVO, each team leader is responsible for two to four work stations that consistedr assembly cells with 1e2 people at each cells. A work station at VOLVO contains16 Operators (4e6 Operators at OPEL) and up to 24 different operationserations at OPEL).

ErgS. Hermawati et al. / Applied150very difcult, so they need to change when it is possible. Trainingprocedure in theory is good, but sometimes it is not possible to giveenough time for training e explaining documentation to workers, whythings are important, safety issues .

Another drawback of the current training approach, mainlyidentied by engineers (OPEL-manufacturing system engineer,VOLVO-production engineer, design engineer) is the lack of con-trol of what is being delivered in training. This was reected bythe following comment: Operator is taught by an experienced

Fig. 3. a. HTA diagram of an assembly operation at OPEL.onomics 46 (2015) 144e157operator.....The experienced operator will teach his ownway. You don'treally know if the operator really understands how to do the job.Different educators teach in different ways and maybe the wrong way.There are different ways to do a job from its technical description butno one does it the same way which can lead to quality issues. Youshould learn the important stuff. Base knowledge is not being taughtleading to quality issues.....

Most of the stakeholders (OPEL-manufacturing system engineer,system owner, virtual engineer, operators, team leader, core team,

b. HTA diagram of an assembly operation at VOLVO.

tated

ErgTable 7Summary of responses to questions related to thematic analysis. Unless otherwise s

Interview questions

1.6. Do any of the assembly tasks pose particular difculties to the operators? Doany tasks cause frequent problems?

3.1. What are the tasks that the operators need to be trained for (e.g. ttingcomponents on vehicle assembly lines)?

3.2. How are they currently trained to do these tasks?

3.4. What is your opinion of the training?

3.5. What are the key skills or knowledge being taught?

3.6. What is good about the current approach to the training?

3.7. What are the difculties with the current approach of the training, or whatproblems exist?

3.8. Can you suggest how to improve the training?3.9. What do you consider the most important performance measures and

goals? i.e. i) Human: of the operators' job e.g. time, errors; ii) Operational: of

S. Hermawati et al. / Appliedlaunch team; VOLVO-virtual manufacturing engineer, technicalpreparation engineers, team leaders, pilot plant Team, operators),irrespective whether or not they have had experienced with gameconsoles, expressed positive responses towards game-basedvirtual training by stating that it would be fun, interactive,engaging and might be a good way to get acceptance. They alsothought that introducing it in a competitivemanner would possiblydrive operators' efciency although they emphasised that thetraining solution should make clear trainees are not being evalu-ated individually and that the system should be agreed by workcouncils i.e. the organisations (trade unions) that represent oper-ators at the assembly lines.

Interestingly, the general initial thought on game-based virtualtraining was that this approach to training should not replacetraining on pre-series vehicle as operators need to get a touchand feel for objects' mass and size. This concernwas mostly basedon stakeholders' experiences that, while everything looks perfectand easy on paper and computer, difculties in assembling partsare likely to occur in real-life situations because of various factorssuch as restrictions imposed by the parts' weight, lack of t be-tween parts, lack of exibility of parts (i.e. wiring harnesses). Thus,the know-how on techniques or ways to remediate the difcultiescould only be obtained by handling real-parts. Nonetheless, all ofstakeholders viewed game-based virtual training as a good

the training e.g. operator must complete task correctly XX% of the time; iii)Business e.g. a timeframe reduction for the training.

3.11. Does training improve productivity? If so, how?5.1. What are your initial thoughts about training using virtual systems?

5.2. Do you think this approach could improve training?

5.3. What problems would you anticipate?5.4. What are your initial thoughts about a game-like system for training?

5.5. What are your thought on the capture and feedback of assembly issues tothe product designers and manufacturing systems engineers?

5.6. Do you own, use or have been tried an X360, PS3 or WII based gameconsole? What is your impression or experience controlling a game using awireless controller?

5.7. Would you like to be involved in the VISTRA project? I.e. can we contact youagain to obtain your feedback on the developed technologies?, the summary applies to both automotive manufacturers.

Summary

Assembly operations that involve awkward and non-ergonomic postures werefrequently cited. Another issue is harness wiring as the assembly instruction isdifcult to understand, especially related to the cable behaviour.Assembly operations, assembly sequences and how to understand paper-basedassembly instruction.Operators at assembly lines are trained by experienced operators who had beentrained at pilot line. A trainer gives a demonstration of the assembly operations/sequences, followed by a trainee performing the operations/sequences.Theoretical background is only given when there is enough time.There is no standard method of training. There is not enough time and resources(hardware and trainers) to deliver the training.How to read paper-based assembly instruction, parts recognition (partappearance and number), where and when the part has to go, assembly time,quality check.The training is done by experienced operators; there is freedom to askquestions.Lack of training due to time and resources. A lack of standard on what is beingtrained. The number of variations that could be learned through training wasreally limited. Only a few operators mastered assembly operations other thanthose within their stations/cell.Provide structured training; allow dedicated time and resources for training.Human: quality of assembly, errors, resolving problems independentlyOperational: completing the assembly sequence within the required time

onomics 46 (2015) 144e157 151additional tool to have for better preparation in knowing the tools,parts, sequences, etc. prior to hardware training and real produc-tion. Some of the benets which were identied by stakeholdersinclude: i) helping operators to learn assembly operations byenabling parts visualisation in 3D rather than 2D drawings ahead oftheir training on pre-series vehicle; ii) providing opportunities totrain on more product variants that otherwise would not bepossible due to limited hardware provision; iii) reducing cost andtime for training; and iv) providing an opportunity to involve op-erators in the development of the car at an earlier stage.

Stakeholders who have experienced using software as part oftheir roles (OPEL-System Owner, Manufacturing System EngineerManager, Virtual Engineers, Launch Team, Core Team; VOLVO-Virtual Manufacturing Engineer, Technical Preparation Engineer,Pilot Plant Team, Design Engineers) highlighted the need to link avirtual training with existing software systems. Stakeholderswho have had experience in using virtual systems (e.g. OPEL-manufacturing system engineers, virtual engineers, launch team;VOLVO-design engineers, production engineers, pilot plant team)also stressed the importance of a realistic virtual training whichrefers to the ability of a virtual training to represent conditions thatwere encountered in real assembly line. Some examples that weregiven included: parts' behaviour, time pressure, force feedbackduring parts' tting, tolerance issues, etc.

Business: better training prior to launching the product, quality.

Yes. It reduces the rework due to poor quality.Most stakeholders had positive views on training using a virtual system.Stakeholders that have experienced software and virtual system suggestedfactors that had to be considered in creating a virtual training system.All of the stakeholders could foresee the benet of a virtual training system incomplimenting the current hardware-based approach to training.Acceptance, ease of use, realism, system maintenance and updates.Nearly all stakeholders were positive about a game-like system for training.However, the level and type of competition in the game should be consideredcarefully.Most stakeholders agreed that capturing the feedback and assembly issueswould be useful. There was also a suggestion to extend this to capturingknowledge and the experience of operators.More than half of the stakeholders had tried game consoles.

All of stakeholders, except one, gave their consent to be contacted for furtherinvolvement in VISTRA project.

ErgWith regards to anticipated problems of game-based virtualtraining, most of stakeholders (VOLVO-Virtual Manufacturing En-gineer, Technical Preparation Engineer, Team Leaders, Operators,Pilot Plant Team; OPEL-Manufacturing System Engineer Manager,Virtual Engineers, System Owners, Virtual Engineers, Operators)cited acceptance of the technology as a potential issue, with ageand technology afnity as the given causes. Two examples ofcomments from a team leader and an operator provided a goodsummary for this view: need to relate training to product. Mostpeople here don't know what an axle is at rst. Might have use fortraining of variants, if seen in real life rst. Can't teachweight of part sohave to learn this in real life. Expect big individual differences inacceptance. People who worked for many years are doing things theirway and don't like change (Team Leader); Old ones would notaccept it. Theyre negative. Theyve been here long. Don't want tochange. Young people would see it as fun (Operator). Therefore, theyemphasised that the game-based virtual training should be easy touse and not required training in order to use it. If separate trainingwas required to interact with the virtual training, they wereadamant that virtual training might be too much work to add on totraining on pre-series vehicle. This was reected by a comment of ateam leader: May be too much work for us, may be too much workfor workers too, may be need some time to understand it.

4. Discussion

The results of the data analyses were used in generatingtraining needs, requirements and potential features of game-basedvirtual training. It was revealed there were two needs related toassembly training. The rst was the need for early deployment oftraining. This need (identied from the time line and thematicanalyses) emerged from the fact that the current training systemreliance on the provision of pre-series vehicles resulted in a lack oftraining for operators. This was because the training was oftendelivered too close to the start of real production at the assemblyline. The second need (identied from the link analysis and the-matic analysis) was standardised content of assembly training.This need emerged from the fact that the current training systemrelied heavily on the autonomy of experienced operators tomanage and deliver training with limited or no possibilities forstakeholders who have an overall overview of assembly operations(e.g. manufacturing engineers (OPEL)/production engineers(VOLVO)) to contribute to or control the training content. Both ofthese needs could potentially open an opportunity and present astrong case to support implementation of virtual training forautomotive manufactures. Provided that digital informationrelated to part information and assembly procedure is available,virtual training can be delivered earlier and easily fulls the needfor a standardised training content and approach. Based on theinterview results, a sub set of requirements related to what as-sembly training should contain and how it should be measuredwas established (see Appendix C). The interview results were alsoused to identify potentially benecial features of game-basedvirtual training that could be worthy of investigations in furtherstudies (see Appendix D). As a result of OPEL and VOLVO differ-ences in their products' customisation and assembly arrange-ments, there were observable variations towards trainingrequirements and potential features of game-based virtualtraining.

Most stakeholders had positive views towards game-basedvirtual training. This was likely due to game-based virtualtraining being viewed as more motivating than traditional form oftraining (Gee, 2003; Prensky, 2001). Research has shown thatgame-based virtual training can provide players with a continuous

S. Hermawati et al. / Applied152variety of emotional conditions or psychological stimuli which inthe end inuences motivation (Dondi et al., 2004). However,game-based virtual training is also known to be ineffective due tovarious factors such as inappropriate methods/means to assesslearning outcomes and poor contextualisation of games into ameaningful learning context (Squire, 2004; Egeneldt-Nielsen,2005). Another aspect which needs to be considered when intro-ducing game-based virtual training is the negative effect ofcompetitiveness on the learning process (e.g. competitivenessresulting in certain individuals to focus on performance/score atthe expense of learning (Harviainen et al., 2012), demotivationwhen losing (Moseley et al., 2009)) and potential ethical/legal is-sues. Prior to the application of game-based training withinautomotive manufacturers, it is important that work councils areinvolved to resolve various issues such as: are scores of trainingmade explicit among operators?, and is competition amongoperators allowed? As shown in this study, competitiveness,which is one of attributes of game-based virtual training, wouldlikely have to be minimised if the work council objected to theabove questions.

However, the interview results also showed stakeholders' res-ervations towards replacing existing training with game-basedvirtual training. Stakeholders asserted that such training shouldact as a complement to training sessions on pre-series vehicles.This nding supports other studies on the ability of virtual envi-ronment to replace real life encounters (Zhang et al., 2004; Chi-Wai et al., 2011). The most common reason given was the need fortactile feedback while handling parts and the associated assemblysteps. Thus, it is likely that, until virtual training could providesimilar tactile feedback to those in real life, the role of virtualtraining in automotive manufacturers will remain complementaryto training sessions on pre-series vehicles. While this somewhatlimits the current applicability of game-based virtual training, inthe meantime, game-based virtual training could be introduced asa preliminary on-site independent training session that allowstrainees to recognise parts and learn assembly steps interactivelywhile pre-series vehicles for training purposes are built. Further-more, as resources for pre-series vehicles training are likely to belimited due to the time availability of the trainers and smallnumber of pre-series vehicles, game-based virtual training couldalso be used simultaneously during the pre-series vehiclestraining. Thus, pre-series vehicles training will provide traineeswith the touch and feel and know-how to remediate possibledifculties while performing assembly in real-life whereas thegame-based virtual training will teach trainees the assembly stepsand part recognition. The Technology Acceptance Model (TAM),which was proposed by Davis (1989) postulated that the perceivedlevel of usefulness signicantly affects user acceptance of infor-mation technology design. Therefore, to increase the level of po-tential users' acceptance of a virtual training within the context ofautomotive manufacturers, it is essential that potential users areclearly informed that the role of virtual training is to complementexisting training on pre-series vehicles and will only serve as asole means of training when training on pre-series vehicles is notavailable.

The interview results also revealed stakeholders' opinionstowards potential features that could likely improve game-basedvirtual training system's delity i.e. the ability to capture andtransfer real life situations to a virtual environment. Their mostcommonly suggested features were compiled under the categoryof high realism (Appendix D). A further observation of thepotential features in Appendix D shows that most of the sug-gested approaches related to high realism category belong topsychological delity (the degree to which simulated tasksreproduce behaviours that are required for the actual, real-world

onomics 46 (2015) 144e157task) with only a few under the category of physical delity (how

Erga virtual environment and its component objects mimic theappearance of their real-world counterparts) i.e. simulation ofappearance of parts, simulation of parts behaviour for exibleparts. A correct blend between the two is crucial for successfulvirtual training, bearing in mind that psychological delity isassociated more with positive transfer of training than physicaldelity and that the two are not necessarily positively correlated(Stone, 2008). While the compilation of features that supportreal-life encounters were obtained under the context of game-based virtual training; some of them would likely be applicableon other virtual training related to assembly tasks that aim totransfer a trainee's recognition of product variants' and parts'appearances and their behaviour from a virtual environmentsetting to a real-world setting.

This study also identied that, in addition to training assem-bly operations, stakeholders viewed game-based virtual trainingas offering the potential to be used as a platform for knowledgesharing between stakeholders. Wang and Noe (2010) denedknowledge sharing as: the provision of task information andknow-how to help others and collaborate with others to solveproblems, develop new ideas or implement policies or pro-cedures. It has been shown to contribute to reduced productioncosts, faster completion of new product development projects,improvements in team performance, and innovation capabilities(Wang and Noe, 2010). Studies have shown that a functionallysegmented organisational structure likely hinders knowledgesharing across stakeholders within the various functions (Lam,1996; Tagliaventi and Mattarelli, 2006). Virtual training canaddress this issue by allowing experienced operators to sharetheir best practices not only with other operators but also withengineers who can then decide whether to exploit their practicesor modify them. A study has shown that workers with moreexperience are more likely to share their expertise and havepositive attitudes towards knowledge sharing (Constant et al.,1994). Possible exploitation of shared best practices can rangefrom inclusion of best practices as part of work standardisationto utilisation of best practices to support future product andprocess design.

As a consequence of the level of operators' computer self-efcacy, stakeholders desired game-based virtual training forassembly tasks that can be used without extensive training.While this view was obtained under the context of game-basedvirtual training, this attitude would likely be transferable forother virtual training which involved low-level computer self-efcacy. Computer self-efcacy (or perceived computer self-efcacy), which refers to an individual's judgement of his herability to use a computer (Compeau and Higgins, 1995), andcomputer anxiety are important determinants in inuencing anindividual perception of a technology's ease of use and accep-tance (Agarwal et al., 2000). Existing virtual environment us-ability criteria (Stanney et al., 2003; Gavish et al., 2011) could beused to assist decisions regarding user interface design of virtualtraining. It is also important to avoid the temptation to integratea novel device simply because it would become an attractive andattention-grabbing gimmick (Stone, 2008). Difculties withtechnology or interface in which the training content is deliveredhave been cited as a key frustration source and a reason for lowcompletion rates in e-learning programs (Frankola, 2001).Stedmon and Stone (2001) argued that human-centred knowl-edge and expertise is required to ensure full utilisation andexploitation of technology-based training. Therefore, the choiceof interaction devices should take into account end-users capa-bilities/limitations and the nature of the task to be trained. Forassembly operations, alternative input devices such as a visual

S. Hermawati et al. / Appliedgesture interface (O'Hagan et al., 2002) which allows a user toperform natural movement to select, move and grab componentsof virtual objects could be one possible solution. While this paperhas successfully identied potential benecial features of virtualtraining, it should be noted that the stakeholders involved in thisstudy had either little or no experience in actually using virtualtraining and may not be accurately predicting their actual us-ability with such a system.

Involvement of the end user representatives to identifystakeholders was clearly benecial as they were able to performthis task with ease and accuracy due to familiarity to their or-ganisations. The use of semi structured interviews was useful as itensured that important topics were covered while simulta-neously enabling stakeholders to have the freedom to explain andexpand their views. In addition, performing interviews with morethan one stakeholder of the same job function proved to bebenecial as this facilitated discussion among them during theinterview. A drawback of this study is the lack of input frompotential end users during the renement of the trainingrequirement subset and potential features of game-based virtualtraining, relying instead on end-user representatives. There is apossibility that misjudgement occurred while weighing theimportance of each requirement/potential feature. This could becaused by conicts of interest, especially in circumstances whererepresentative end users have stakes towards an issue, and/or therepresentative end users leaning more towards a managerial thanend user role. The combination of representative end users andresearchers within this study provides the opportunity for re-searchers to draw on the contextual insight of representative endusers. It also enabled representative end users to gain insight intohow their assumptions or conicts of interest may cause bias.This approach is mostly suitable when end-users' involvementduring the renement process is impractical, as is the case in thisstudy.

This paper found the usefulness of employing various dataanalysis methods to identify training needs, generate a subset oftraining requirements and compile potential features for game-base virtual training environments. The analyses complementedeach other by providing different insights into the work organisa-tion and training approaches at both automotive industries. Thiseventually allowed the identication and extraction of trainingneeds and requirements from various perspectives. The graphicalrepresentation that was adopted in link analysis, timeline analysis,and HTA was found to be advantageous in supporting vericationby end-user representatives. Thematic analysis, although highlyresource-demanding in comparison to other analysis methods,provided more detailed information regarding views of users onexisting training approaches or a hypothetical game-based virtualtraining system.

5. Conclusions

This paper has performed a comprehensive user study to informthe training needs, a subset of training requirements, as well aspotential benecial features of game-based virtual training for as-sembly tasks. This study presented a strong case towards the needfor an implementation of virtual training for assembly tasks.However, the preference of stakeholders to adopt virtual training asa complementary training to the existing training on pre-seriesvehicles should also be considered. Different approaches in theorganisation of assembly work and business approaches in meetingcustomer demands have also been shown to have direct impact ontraining requirements and how game-based virtual training shouldbe designed. The study beneted from a team that consisted ofhuman factors researchers, representatives of end users and system

onomics 46 (2015) 144e157 153developers.

Acknowledgements

The authors would like to acknowledge the contribution to thiswork from all members of the VISTRA project e ICT-2011-285176,which is funded by the European Commission's 7th FrameworkProgramme.

Appendix A. Level of education, technology awareness, language skills and the likelihood in embracing new technology foridentied stakeholders.

Stakeholders Level of education Technological levels of skills and awareness Languageskills

OPEL Operator Trained to skilledworkers

Less or low technical afnity, used to work with hardware German

Team Leader Skilled workersto technician

As above German

Supervisor As above As above GermanLaunch team As above Middle to high technical afnity, uses software and hardware for his work German,

EnglishCore team As above As above As aboveCore team/launch team manager As above As above As aboveManufacturing systems engineer Engineer High technical afnity, used to work with computers naturally As aboveManufacturing Systems EngineerManager

Engineer As above As above

System Owner As above As above As above

VOLVO Operators Primary school/high school/College

Differs depending on age, education and personal interest; no virtual simulationsexperience

Swedish,English

Key Operator e Teacher As above As above As aboveKey Operator e Hands On As above As above As aboveKey Operator e Safety As above As above As aboveGroup Manager As above As above As abovePilot Plant Team As above As above As aboveTeam Leader As above As above As aboveDesign engineer College Very good knowledge in computer and virtual simulations As aboveIntroduction Engineers As above Differs depending on age, education and personal interest; various virtual

simulation skill levelsAs above

Production Engineers As above Good computer knowledge; different skill levels of virtual simulation depending oneducation

As above

Technical Preparation Engineer As above As aboveVirtual Manufacturing Engineer As above Very good knowledge in computer virtual simulation As above

Appendix B. Stakeholders' roles, responsibilities and current involvement to training.

Position Roles and responsibilities Current involvement in training

OPEL Manufacturing SystemsEngineer Manager

To manage the general assembly of a plant i.e. process planning and virtual engineering. Coordinating manufacturing systemengineers

Manufacturing SystemsEngineers

To design and produce assembly procedures. Plan and decide assembly sequences

Virtual Engineers To prepare and alter virtual build (3D CAD data) when required. Support planning and decision ofassembly sequences

System Owner Responsible for the overall IT system and maintenance of database Support and maintain IT system anddatabase

Launch Team Manager To lead and coordinate launch team Coordinating launch teamCore Team Manager To lead and coordinate core team Coordinating core teamCore Team To optimise the assembly process in terms of cost, quality, time and efciency at each plant. Improving assembly sequenceLaunch Team The Launch Team or Pilot Team is an ad-hoc team that is formed during a launch of a new

product and is responsible to: i) provide feedback during virtual builds; ii) to assist physicalii) train Team Leaders and support them in

Train Team Leaders at the early stage ofproduct launching

nd their operators (up to 6 team leaders and Oversee training in his team

rators for up to 4 stations and is responsiblein assembly line occur/repairs are required;to operator's illness, etc); iii) providea Supervisor

Trainee during the early stage ofproduct launching and trainer duringthe later stage of product launching

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e157154build at the pilot line and provided feedback; itraining Operators when required.

Supervisors A Supervisor manages a group of team leaders a30 operators).

Team Leaders A Team Leader manages a group of up to six Opeto: i) train and assist Operators when problemsii) perform assembly tasks when required (duefeedback regarding quality and performance toOperators To perform a set of assembly tasks in one or moreworkstations with a high quality in a givenperiod of time. Job rotation is performed by assigning an Operator to different work stations.

Trainee

(continued )

Position Roles and responsibilities Current involvement in training

VOLVO Design Engineers To design parts of new or modied products and develop assembly instructions Supporting assembly sequenceplanning

Production Engineers/Introduction Engineer

To create/modify paper-based assembly instruction when a new product is developed orchanges in the plant occur (SPRINT) and identify best practices.

Planning and deciding assemblysequences

Virtual ManufacturingEngineers

To manage, support and provide overview of ways to work with virtual systems. N/A

Technical PreparationEngineer

To coordinate and steer of all product changes to the manufacturing line. Improving assembly sequence

Team Leaders A Team Leader manages a group of up to 4 stations (up to 6 Operators per work station). ATeam Leader appoints a number of Operators in his team as Key Operators who will trainoperators in his team and reports any unresolved issues to the Group Manager whoforwards the information to the Production Engineer.

Oversee training in his team

Pilot Plant Team To participate in virtual and physical builds and to provide training for operators fromassembly line on a new/update vehicle built.

Train key operators at the early stage ofproduct launching (at the pilot line)

Operators/KeyOperators

Operators: To perform a set of assembly tasks in one or more work stations with a highquality in a given period of time. There are up to 4 work cells in each work station. Jobrotation is performed by assigning Operators to either different work stations or work cells.Key Operators: An experienced Operator that is appointed by a Team Leader to assumeadditional roles. There are three types of Key Operator: i) Teacher e train Operators in themain line; ii) Hand One to resolve assembly problems at the Adjustment station; iii) Safety-to enforce safety within their team.

Key Operators e trainee at the pilot lineand trainer at assembly lineOperators e Trainee

Appendix C. Rened elicited user requirements for assembly tasks.

Theme Sub theme(s) Stakeholders Importance Identied from

Key skills/knowledgeincluded

Assembly sequence, parts number andappearance, tools, end location of parts/tools, critical part of the assemblyoperations

OPEL-Team Leaders, ManufacturingSystem Engineers, Core Team; VOLVO-Production Engineers, Pilot Plant Team

High Hierarchical Task Analysis

Self-inspection of own work Medium Hierarchical Task AnalysisHow to read and understand SPRINT-forVOLVO

Medium Hierarchical Task Analysis

Cause and effect of assembly operation Medium Hierarchical Task AnalysisStart locations of parts/tools High Hierarchical Task AnalysisTool and lifting device handling Medium Hierarchical Task Analysis

Performancemeasures

Cost and time to execute training All High Thematic AnalysisQuality (number of correct assemblysequence)

High Thematic Analysis

Time (performing correct sequencewithin the time given)

High Thematic Analysis

Appendix D. Suggested potentially benecial or worthy investigation features for virtual training.

Theme Sub theme(s) Stakeholders Importance Feasibility Identied from

Capture and Feedback ofassembly issues

Yes All High High Link AnalysisShould also be used to captureknowledge/experience of Operators

Medium High Link Analysis

Relationship with existing ITsystem

Should allow linkage to enterprise dataof the existing IT system

All, except Team Leadersand Operators

High High Thematic analysis

User Interface Easy to use All High High Thematic AnalysisNo additional training is required High High

Access to the system Local (Operators: main line) Team Leaders, Operators High High Thematic AnalysisGame like approach Yes All High High Thematic AnalysisRelationship with current

training on pre-series vehicleUsed in conjunction with training onpre-series vehicle (for new product)

All High High Timeline Analysis

Used to train new variants ormodication of existing productswithout pre-series vehicle

High High Timeline Analysis

High realism Simulate parts behaviour for exibleparts e.g. wiring

OPEL (System Owner,Manufacturing SystemEngineers, DesignEngineers, VirtualEngineers, Launch Team,Core Team), VOLVO (Virtual

High High Thematic Analysis

Simulate assembly process in differentworkstations or work cells

High High Thematic Analysis,Hierarchical TaskAnalysis

Simulate appearance of parts High High Thematic Analysis

(continued on next page)

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e157 155

olde

actutionEngeam

ead

S. Hermawati et al. / Applied Erg156References

Abate, A.F., Guida, M.G., Leoncini, P., Nappi, M., Ricciardi, S., 2009. A haptic basedapproach to virtual training for aerospace industry. J. Vis. Lang. Comput. 20,318e325.

Abe, N., Zheng, J.Y., Tanaka, K., Taki, H., 1996. A training system using virtual ma-chines for teaching assembling/disassembling operation to novices. In: Pro-ceedings of 1996 IEEE International Conference on Systems, Man, andCybernetics, vol. 3, pp. 2096e2101.

Agarwal, R., Sambamurthy, V., Stair, R.M., 2000. Research report: the evolvingrelationship between general and specic computer efcacy-an empiricalassessment. Inform. Syst. Res. 11 (4), 418e430.

Albinsson, P.-A., Morin, M., Fransson, J., 2003. Finding information needs in militarycommand and control systems using exploratory tools for communicationanalysis. In: Proceedings of the Human Factors and Ergonomics Society AnnualMeeting, vol. 47, pp. 1918e1922 (17).

Amiri, Dezfooli, Mortezaei, 2012. Designing and ergonomics backpack for studentaged 7-9 with user centred design approach. Work J. Prevent. Assess. Rehabil.41 (1), 1193e1201.

Anastassova, M., Burkhardt, J.-M., 2009. Automotive technicians' training as acommunity-of-practice: Implications for the design of an augmented realityteaching aid. Appl. Ergon. 40 (4), 713e721.

Artman, H., Zallh, S., 2005. Finding a way to usability: procurement of a taxidispatch system. Cogn. Technol. Work 7 (3), 141e155.

Atkinson, C., Eldabi, T., Paulm, R.J., Pouloudi, A., 2001. Investigating integrated socio-technical approaches to health informatics. In: Proceedings of the 34th HawaiiInternational Conference on System Sciences.

Barnett, B., Helbing, K., Hancock, G., Heininger, R., Perrin, B., 2000. An evaluation ofthe training effectiveness of virtual environments. In: Proceedings of theInterservice/Industry Training, Simulation & Education Conference.

Beal, S.A., 2009. Instructor-Facilitated vs. Stand-alone Tactical Game Training. USArmy Research Institute for the Behavioural and Social Sciences, Arlington.

Bluemel, E., Hintze, A., Schulz, T., Schumann, M., Stuering, S., 2003. Virtual envi-ronment for the training of maintenance and service. In: Proceedings of the2003 Winter Simulation Conference, pp. 2001e2007.

Boud, A.C., Haniff, D.J., Baber, C., Steiner, S.J., 1999. Virtual reality and augmentedreality as a training tool for assembly tasks. In: Proceedings of the 1999 Inter-national Conference on Information, pp. 32e36.

Brooks, A., 1998. Ergonomics approaches to ofce layout and space planning. Fa-

(continued )

Theme Sub theme(s) Stakeh

ManufProducDesignPlant T

Simulate multiple Operators (up to 2e6Operators) e for VOLVOSimulate tolerances

Simulate effect of weight and centre ofgravitySimulate force feedbackAccommodate feel of the partSimulate time pressure on the main line(OPEL)

Team L

Simulate moving assembly line duringassembly operations (OPEL)cilities 16 (3/4), 73e78.Brough, J.E., Schwartz, M., Gupta, S.K., Anand, D.K., Kavetsky, R., Pettersen, R., 2007.

Towards the development of a virtual environment-based training system formechanical assembly operations. Virtual Real. 11, 89e206.

Chapanis, A., 1959. Research Techniques in Human Engineering. John Hopkins Press,Baltimore.

Chi-Wai, R.K., So-Ning, T.M., Wing-Kuen, K.L., Hui, S.S.-C., Ka-Shun, P.W., Wong, C.C.-K., 2011. Can mobile virtual tness apps replace human tness trainer?. In:Proceedings of 5th International Conference on New Trends in InformationScience and Service Science, pp. 56e63.

Coleman, G.W., Gibson, L., Hanson, V.L., Bobrowicz, A., McKay, A., 2010. Engagingthe disengaged: how do we design technology for digitally excluded olderadults. In: Proceedings of the 8th ACM Conference on Designing InteractiveSystems, pp. 175e178.

Colquitt, J.A., LePine, J.A., Noe, R.A., 2000. Toward an integrative theory of trainingmotivation: a meta-analytic path analysis of 20 years of research. J. Appl. Psy-chol. 85, 678e707.

Compeau, D.R., Higgins, C.A., 1995. Computer self-efcacy: development of ameasure and initial test. Manag. Inf. Syst. Q. 19, 189e211.

Constant, D., Kiesler, S., Sproul, L., 1994. What's mine is ours, or is it? A study ofattitudes about information sharing. Inform. Syst. Res. 5 (4), 400e421.

Corbeil, P., 1999. Learning from the children: practical and theoretical reections onplaying and learning. Simul. Gaming 30, 163e180.Crosby, B., 1992. Stakeholder Analysis: a Vital Tool for Strategic Managers. USAID,Washington DC.

Davis, F.D., 1989. Perceived usefulness, perceived ease of use and user acceptance ofinformation technology. MIS Q. 13 (3), 319e339.

Dondi, C., Edvinson, B., Moretti, M., 2004. Why choose a game for improvinglearning and teaching process? In: Pivec, M., Koubek, A., Dondi, C. (Eds.),Guidelines for Game-Based Learning. KM-Druck, Germany, pp. 20e76.

Egeneldt-Nielsen, S., 2005. Beyond Edutainment: Exploring the Educational Po-tential of Computer Games. Doctoral thesis. IT-University of Copenhagen,Denmark.

Engel, K., Langkamer, K.L., Estock, J.L., Orvis, K.L., Salas, E., Bedwell, W., Conkey, C.,2009. Investigating the effectiveness of game-based approaches to training. In:Proceedings of the 31st Annual Interservice/Industry Training, Simulation andEducation Conference.

Frankola, K., 2001. Why online learners drop out. Workforce 80 (10), 52e60.Franzosi, R., 2004. Content analysis. In: Lewis-Beck, M.S., Bryman, A., Liao, T.F. (Eds.),

The Sage Encyclopaedia of Social Science Research Methods, vol. 1. SAGE Pub-lications, Inc., California, pp. 186e189.

Garris, R., Ahlers, R., Driskell, J.E., 2002. Games, motivation and learning: a researchand practice model. Simul. Gaming 33 (4), 441e467.

Gavish, N., Gutierrez, T., Webel, S., Rodriguez, J., Tecchia, F., 2011. Design guidelines forthe development of virtual reality and augmented reality training system formaintenance andassembly tasks. In:Proceedings ofBIOWebofConferences, vol.1.

Gee, J.P., 2003. What Video Games Have to Teach Us about Learning and Literacy.Palgrave Macmillan, New York.

Gerbaud, S., Mollet, N., Ganier, F., Arnaldi, B., Tisseau, J., 2008. GVT: a platform tocreate virtual environments for procedural training. In: Proceedings of 2008IEEE Virtual Reality Conference.

Gopher, E., Weil, M., Bareket, T., 1994. Transfer of a skill from a computer gametraining to ight. Hum. Fact. 36 (3), 387e405.

Gutierrez, T., Rodriguez, J., Velaz, Y., Casado, A., Suescun, A., Sanchez, E.J.,2010. IMA-VR: a multimodal virtual training system for skills transfer inindustrial maintenance and assembly tasks. In: Proceedings of 18th IEEInternational Symposium on Robot and Human Interactive Communica-tion, pp. 428e433.

Hajarnavis, V., 2012. Process automation in the automotive industry. In: Liptak, B.G.(Ed.), Process Software and Digital Networks. CRC Press, Boca Raton FL,pp. 809e820.

Harviainen, J.T., Lainema, T., Saarinen, E., June 2012. Player-reported impediments togame-base learning. In: Proceedings of 2012 Digra Nordic. University of Tam-

rs Importance Feasibility Identied from

ring Engineers,Engineers,ineers, Pilot)

High Medium Hierarchical TaskAnalysis

Medium (VOLVO), Low(OPEL)

Low Thematic Analysis

High Very low Thematic Analysis

High Very low Thematic AnalysisHigh Very low Thematic Analysis

ers, Operators High High Thematic Analysis

Medium (OPEL), Low(VOLVO)

Low Thematic Analysis

onomics 46 (2015) 144e157pere, Tampere, Finland. Available at: http://www.digra.org/wp-content/uploads/digital-library/12168.02279.pdf (accessed 10.04.14.).

Jong, M.S.Y., Shang, J., Lee, F., Lee, J.H.M., 2008. Harnessing computer games ineducation. Int. J. Dist. Educ. Technol. 6, 1e9.

Kirwan, B., Ainsworth, L.K., 1992. A Guide to Task Analysis. Taylor & Francis, London.Krammer, P., Neef, D., Plapper, P., 2011. Advanced manufacturing technologies for

general assembly. In: Proceedings of the SAE 2011World Congress & Exhibition.Kraus, D.C., Gramopadhye, A.K., 2001. Effect of team training on aircraft mainte-

nance technicians: computer-based training versus instructor-based training.Int. J. Indust. Ergon. 27, 141e157.

Lam, A., 1996. Engineers, management and work organization: a comparativeanalysis of engineers' work roles in British and Japanese electronic rms.J. Manag. Stud. 33 (2), 183e212.

Lapan, S.D., Quartaroli, M.T., Riemer, F.J., 2012. Qualitative Research: an Introductionto Methods and Designs. Jossey Bass, San Francisco.

Larsen, C.R., Soerensenm, J.L., Grantcharov, T.P., Dalsgaard, T.D., Schroeder, T.V.,Ottesen, B.S., 2009. Effect of virtual training on laparoscopic surgery: rando-mised controlled trial. Br. Med. J. 338.

Leger, P., 2006. Using a simulation game approach to teach enterprise resourceplanning concepts. J. Inform. Syst. Educ. 17, 441e447.

Leonard, V.K., Jacko, J.A., Yi, J.S., Saintfort, F., 2006. Human factors and ergonomicsmethods. In: Salvendy, G. (Ed.), Handbook of Human Factors and Ergonomics.John Wiley & Sons, Inc., New Jersey, pp. 292e321.

Lili, Y., Xueqing, L., Jiayao, C., 2009. Research and implementation of virtual as-sembly training system. In: Proceedings of 2009 IEEE International Symposiumon IT in Medicine & Education, vol. 1, pp. 641e647.

Lin, F., Ye, L., Duffy, V.G., Su, C.-J., 2002. Developing virtual environments for in-dustrial training. Inf. Sci. 140, 153e170.

Maguire, M., Bevan, N., 2002. User requirements analysis: a review of supportingmethods. In: Proceedings of IFIP 17th World Computer Congress, pp. 133e148.

Minocha, S., Reeves, A.J., 2010. Design of learning spaces in 3D virtual worlds: anempirical investigation of Second Life. Learn. Media Technol. 35 (2), 111e137.

Moseley, A., Whitton, N., Culver, J., Piatt, K., 2009. Motivation in alternative realitygaming environments and implication for education. In: Proceedings of the 3rdEuropean Conference on Games Based Learning, pp. 279e286.

Neary, H.T., Sinclair, M.A., 1998. A case study of job design in a steel plant. In:Hanson, M.A. (Ed.), Contemporary Ergonomics. Taylor and Francis, London,pp. 218e222.

Nemeth, C., 2005. Human Factors Methods for Design. CRC Press LLC, Florida.Neuman, W.P., 2004. Production Ergonomics: Identifying and Managing Risks in the

Design of High Performance Work System. Doctoral thesis. Lund TechnicalUniversity, Lund.

Nielsen, J., 1993. Usability Engineering. Academic Press, San Diego.Nof, S.Y., Wilhelm, W.E., Warnecke, H.J., 1997. Industrial Assembly, rst ed. Chapman

& Hall, London.O'Hagan, R.G., Zelinsky, A., Rougeaux, S., 2002. Visual gesture interfaces for virtual

environments. Interact. Comput. 14, 231e250.Oliveira, D.M., Cao, S.C., Hermida, X.F., Rodrguez, F.M., 2007. Virtual reality system

for industrial training. In: Proceedings of 2007 IEEE International Symposiumon Industrial Electronics, pp. 1715e1720.

Peniche, A., Diaz, C., Trefftz, H., Paramo, G., 2011. An immersive virtual reality trainingsystem for mechanical assembly. In: Proceedings of the 4th International Con-ference on Manufacturing Engineering, Quality and Production Systems.

Prensky, M., 2001. Digital Game-based Learning. McGraw-Hill, New York.Rittle-Johnson, B., Siegler, R., Alibali, M.W., 2001. Developing conceptual under-

standing and procedural skill in mathematics: an iterative process. J. Educ.Psychol. 93 (2), 346e362.

Sawasdchai, N., Poggenpohl, S., 2002. User purposes and information-seeking be-haviours in we-based media: a user centred approach to information design onwebsites. In: Proceedings of the DIS'02 Conference on Designing InteractiveSystem: Processes, Practices, Methods and Techniques, pp. 201e212.

Seymour, N.E., Gallagher, A.G., Roman, S.A., O'Brien, M.K., Bansal, V.K.,Andersen, D.K., Satava, R.M., 2002. Virtual reality training improves operatingroom performance. Annu. Surg. 236 (4), 458e464.

Squire, K., 2004. Replaying History: Learning World History through Playing Civi-lization III. Doctoral thesis. University of Indiana, USA.

Stanney, K.M., Mollaghasemi, M., Reeves, L., Breaux, R., Graeber, D.A., 2003. Usabilityengineering of virtual environments (VEs): identifying multiple criteria thatdrive effective VE system design. Int. J. Hum. Comput. Stud. 58, 447e481.

Stanton, N.A., 2004. The psychology of task analysis today. In: Diaper, D.,Stanton, N.A. (Eds.), The Handbook of Task Analysis for Human-ComputerInteraction. Lawrence Erlbaum Associates, Mahwah, pp. 569e584.

Stedmon, A.W., Stone, R.J., 2001. Re-viewing reality: human factors of synthetictraining environments. Int. J. Hum. Comput. Stud. 55, 675e698.

Stone, R.J., 2008. Human Factors Guidelines for Interactive 3D Game-based Training Sys-temsDesign.UniversityofBirmingham.http://www.hdtc.com/ (accessed23.02.12.).

Tagliaventi, M.R., Mattarelli, E., 2006. The role of networks of practices, valuesharing and operational proximity in knowledge ows between professionalgroups. Hum. Relat. 59 (3), 291e319.

Thorstensson, M., Axelsson, M., Morin, M., Jenvald, J., 2001. Monitoring and analysisof command post communication in rescue operation. Saf. Sci. 39 (1), 51e60.

Topolski, R., Leibrecht, B., Cooley, S., Rossi, N., Lampton, D., Knerr, B., 2010. Impact ofGame-based Training on Classroom Learning Outcomes. US Army ResearchInstitute for the Behavioural and Social Sciences, Arlington.

Vora, J., Nair, S., Gramopadhye, A.K., Duchowski, A.T., Melloy, B.J., Kanki, B., 2002.Using virtual reality technology for aircraft visual inspection training: presenceand comparison studies. Appl. Ergon. 33 (6), 559e570.

Wang, Q.H., Li, J.R., 2004. A desktop VR prototype for industrial training applica-tions. Virtual Real. 7, 187e197.

Wang, S., Noe, R.A., 2010. Knowledge sharing: a review and directions for futureresearch. Hum. Resour. Manag. Rev. 20, 115e131.

Webel, S., Bockholt, U., Engelke, T., Gavish, N., Olbrich, M., Preusche, C., 2013. Anaugmented reality training platform for assembly and maintenance skills.Robot. Auton. Syst. 61 (4), 398e403.

Xia, P., Lopes, A.M., Restivo, M.T., Yao, Y., 2012. A new type haptics-based virtualenvironment system for assembly training of complex products. Int. J. Adv.Manuf. Technol. 58 (1), 379e396.

Zhang, D., Zhao, J.L., Zhou, L., Nunamaker, J.F., 2004. Can e-learning replace class-room learning? Commun. ACM 47 (5), 75e79.

S. Hermawati et al. / Applied Ergonomics 46 (2015) 144e157 157

Understanding the complex needs of automotive training at final assembly lines1 Introduction2 Methods2.1 Participating companies2.2 Interviewees2.3 The semi-structured interview and observation2.4 Data analysis2.5 Identification of training needs, training requirements and potential features of game-based virtual training

3 Results3.1 Stakeholder analysis3.2 Timeline analysis3.3 Link analysis3.4 Hierarchical Task Analysis (HTA)3.5 Thematic analysis

4 Discussion5 ConclusionsAcknowledgementsAppendix A Level of education, technology awareness, language skills and the likelihood in embracing new technology for ide ...Appendix B Stakeholders' roles, responsibilities and current involvement to training.Appendix C Refined elicited user requirements for assembly tasks.Appendix D Suggested potentially beneficial or worthy investigation features for virtual training.References