A SYSTEM FOR GENERATING PEDAGOGICAL SCENARIOS

FOR SERIOUS GAMES

Aarij Mahmood Hussaan

Universit

e de Lyon, CNRS, Universit

e Lyon 1, LIRIS, UMR5205, F-69622, Lyon, France

Karim Sehaba

Universit

e de Lyon, CNRS, Universit

e Lyon 2, LIRIS, UMR5205, F-69676, Lyon, France

Keywords:

Adaptive scenario, Serious games, Users interaction traces, Users proﬁle, Handicap.

Abstract:

Serious games are often used in education as they provide an excellent opportunity for learning. This pa-

per presents our proposition of a system capable of generating adaptive learning scenarios for serious games.

Thus, we have presented the architecture of this system. The domain knowledge is organized in three layers.

These layers include the domain concepts, the pedagogical resources and the serious game resources. The ap-

proach we propose is capable to keep in account the speciﬁcities of serious games, the user proﬁle and his/her

goals while generating scenarios. Furthermore, it also uses the interaction traces as knowledge sources in the

adaptation process. To test the applicability of our system we are working on the Project CLES (Cognitive

and Linguistic Element Stimulation). This project targets producing serious games for users with cognitive

disabilities. We’ve presented an example of how our system will generate a pedagogical scenario.

1 INTRODUCTION

Serious games are deﬁned as ”a mental contest,

played with a computer in accordance with speciﬁc

rules, which uses entertainment to further government

or corporate training, education, health, public pol-

icy, and strategic communication objectives” (Zyda,

2005). Hence, the idea to learn while playing games

is very attractive for most of the users.

In the context of systems capable of generating

pedagogical courses and serious games. The objective

of this study is to propose a system that can generate

dynamically adaptive pedagogical scenarios keep-

ing into account the following properties:

• The ability to be utilized in a variety of serious

games taking into account their speciﬁcities.

• Use of users’ interaction traces as knowledge

sources in the adaptation process.

• The ability to be presented to cognitively handi-

capped users.

Indeed, a serious game like any other game has

many elements namely game characters, narrative,

game world, game world objects, goals, rules, ped-

agogical elements, etc. A review about the design of

different serious games can be found in (Dondlinger,

2007) (Michael, David R. And Chen, 2005). To create

pedagogical scenarios that can be utilized in many se-

rious games, it is necessary to keep in account various

aspects of serious game design that are common in a

vast amount of games. This will help in keeping the

scenarios usables with different games. Furthermore,

it is also necessary to make sure that the games can be

easily playable for handicapped users.

Another requirement is the use of modeled users’

interaction traces as knowledge sources for the adap-

tation process. A user’s interaction traces can be de-

ﬁned as ”a history of learner’s actions collected, in

real time, from his/her interaction with a computer

system” (Sehaba et al., 2009). So these traces will

also help the system in keeping an account of the his-

tory of user’s actions i.e. the evolution of the user’s

competence can be detected and be used by the sys-

tem. Furthermore, their modeling will help us in de-

tecting the state in which the user ﬁnds himself in

while playing the game, enabling the system to pro-

vide help (hints, etc) to the user in case of need. Fur-

thermore, the level of the game can be adjusted ac-

cording to the user’s performance. Furthermore, the

system can also make sure that the same exercises are

246

Mahmood Hussaan A. and Sehaba K..

A SYSTEM FOR GENERATING PEDAGOGICAL SCENARIOS FOR SERIOUS GAMES.

DOI: 10.5220/0003336002460251

In Proceedings of the 3rd International Conference on Computer Supported Education (CSEDU-2011), pages 246-251

ISBN: 978-989-8425-49-2

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

not repeated many times to a user.

The problem of generating pedagogical scenarios

for a user is not new and is addressed previously by

many authors like (Brusilovsky, 1993) (Ullrich and

Melis, 2009) (Specht et al., 2001) (Karampiperis and

Sampson, 2005) . These works focuses only on gener-

ating pedagogical scenarios. Therefore, they do not

take into account the speciﬁcities of serious games.

Consequently, they cannot be used easily with them.

Furthermore, none of these systems neither model the

users’ interaction traces nor use them as knowledge

sources in their adaptation process.

Similarly, the idea of using serious games in ed-

ucation is not new and is being employed by many

systems like (Morenoger et al., 2007) (Torrente et al.,

2009) (Carron, Thibault And Marty, Jean-Charles

And Heraud, 2008). However, most of the time these

systems lacks the notion of a pedagogical scenario.

Furthermore, almost all of the systems neither model

users’ interaction traces nor use them for adaptation.

(Carron, Thibault And Marty, Jean-Charles And Her-

aud, 2008) do use the interaction traces but their sys-

tem neither is usable with different games nor their

scenarios are adaptive.

The remainder of this article is organized as fol-

lows, the next section will present the application’s

context for our work. In section 3 the architecture of

our system will be presented. This is followed by the

presentation of our course generator in section 4. In

section 5 we presents an example of scenario genera-

tion of our system. This article is then concluded by

presenting a brief conclusion in section 6.

2 APPLICATION’S CONTEXT

This work concerns the project CLES (Cognitive and

Linguistic Element Stimulation). CLES is a project

funded by the French Ministry. It is conducted in part-

nership with the laboratories d’

Etude des M

ecanismes

Cognitifs (EMC)

, Laboratoire des Usages en Tech-

niques d’Information Num

eriques (LUTIN)

and the

society GERIP

. This project will create a ”Serious

Game” based environment with integrated tools based

on educational game for the reeducation and the cog-

nitive simulation of children, adults and seniors. This

project will cover all the themes of speech (spoken

language and written language), memory, calculation

disorders, attention disorders, cognitive impairments,

assessment of the cognitive skills, voice disorders and

http://recherche.univ-lyon2.fr/emc/rubrique-2-

Presentation.html

http://www.lutin-userlab.fr/ pages/english/

http://www.gerip.com/about/presentation.html

Figure 1: Conﬁguration screen of ”Long-Vue”.

audio and visual perception disorders. Therefore,

for every pathology, this project aims to develop a

game that is focused on the speciﬁc deﬁciencies while

maximizing, through techniques employed in video

games, their cognitive ergonomics..

Based on these components, our contribution, in

this project, consists in developing a module capable

of generating personalized courses according to the

challenges to be faced and progress of each patient.

This module should therefore keep in account

• The speciﬁcities of the serious game

• What the practitioner has prescribed his patients

• The knowledge base of the treatments available

for the pathology

• Results of the previous exercises of the user

This module will also be able to help the prac-

titioner in monitoring the progress of his patients in

each learning session. The latter, once designed, will

be subject to a validation process including ”clinical

tests” that will be conducted with project partners.

The games that have been already developed in

this area contains games testing different cognitive

abilities of a person. The ﬁgure 1 presents the screen

shot of a game called Long − Vue. This game tests

the working memory capacity of its users. It asks the

user to recall a word or image after showing a list of

words or images. The parameters of the game let the

user to adjust the length of the list of words or images,

the time a single word is displayed on the screen, font

sizes, size of characters in the words of the list, etc.

3 SYSTEM ARCHITECTURE

Figure 2 shows the general architecture of the system.

The different components of the system are as fol-

lows : domain model, resource model, game model,

presentation model, user proﬁle and course genera-

tor. The domain model consists of the concepts re-

lated to the pedagogical domain (the domain to be

A SYSTEM FOR GENERATING PEDAGOGICAL SCENARIOS FOR SERIOUS GAMES

247

taught by the system) and the concepts related to the

physical and cognitive capacities of the user. The do-

main concepts are related to each other with one or

more relations. These relations are didactic in nature.

We’ve identiﬁed some relations that are used in the

system. These relations are: Has-Parts, Required, Or-

der, Type-Of and Parallel. Some of these relations

are used by other authors as well (Karampiperis and

Sampson, 2005) (Duitama et al., 2005).

Has Parts: HP = (x, y

, y

. . . y

) : the concept x

is composed of the sub-concepts y

, y

. . . y

To learn x it is necessary to learn all the concepts

, y

. . . y

Requires: R = (x, y) : to learn x it is necessary to

have sufﬁcient knowledge of y.

Order: O = (x, y) : it is preferable to present con-

cept x before y

Type-Of: TO = (x, y) : concept y is a type of con-

cept x. This relation can be considered as a Spe-

cialization relation.

Parallel: P = (x, y) : the concept x and y are parallel

i.e. if one is selected then the other should also be

selected during a course generation

The resource model contains all the pedagogical

resources for e.g. deﬁnitions, examples, tests etc.

These resources are related to one or more domain

concepts. The game model contains all the resources

of serious games for e.g. bed, lamps, paintings,

non-player characters (NPC), drawers etc. These re-

sources uses the pedagogical resources to be pre-

sented to the user.

The three kinds of knowledge (pedagogical do-

main, pedagogical resources and serious game re-

sources) are kept separate in the system. This sepa-

ration is necessary to make reduce dependency of one

type of knowledge on the other. This helps in mak-

ing the system more generic i.e. usable for different

domains. Furthermore, this also helps in making the

system usable with a variety of serious games.

The purpose of the presentation model is to or-

ganize the pedagogical resources presented to the

user. In fact, this model contains two sub models

namely ’scenario model’ and ’test model’. The sce-

nario model deﬁnes the structure of the scenario for

e.g. a scenario starts by presenting two deﬁnitions fol-

lowed by an example and an exercise. The test model

describes the system’s behavior on test type resources

for e.g an easier test is presented after each failure.

The selection of these models can either be made by

the user or by the teacher (expert) for the student. The

structure of the scenario model can be of the form as

deﬁned in (Ullrich and Melis, 2009).

Figure 2: General architecture of our system.

The user proﬁle contains the user’s information.

This information includes the user’s personal details

and the progress made by the user. The traces of his

interaction with the system is also kept in the pro-

ﬁle. These traces help in keeping an account the con-

cepts the user has already visited and user’s mastery

of those concepts. The pedagogical and serious game

resources s/he has already seen. The order in which

the user has visited the resources. The evaluation of

his progress. Traces help in making sure that the same

resources are not repeated unnecessarily to the user.

This repetition can be discouraging for the user.

The course generator is where all the generation

of courses take place. The working of this module is

discussed in detail in the following section.

4 COURSE GENERATOR

This is the heart of our system. The main purpose of

the Course Generator (CG) is to generate a scenario

according to the proﬁle of the user and his objectives.

Furthermore, during the interaction this module is ca-

pable of adapting the scenario according to the users’

traces. The process of this module is shown in the ﬁg-

ure 3. In this article we don’t give the details of the

implementation of the algorithms used by the course

generator. We give here only the general principle of

its functioning.

In the 1

step the learning goals (selected by the

users or by the system according to the user’s pro-

ﬁle) in the form of target concepts are provide to

CSEDU 2011 - 3rd International Conference on Computer Supported Education

248

Figure 3: Course Generation process.

the course generator. Based on these learning goals

and the user’s proﬁle a list of result concepts are se-

lected. These result concepts are necessary to teach

the learning goals. This generation is done by Algo-

rithm 1. The working of this algorithm goes as fol-

lows: for every concept in the learning goals the al-

gorithm consults the user’s proﬁle to verify whether

or not the concept is sufﬁciently known by the user.

If it is known by the user this concept is ignored. If

the concept is not known then the algorithm searches

for related concepts (recall that a concept can be re-

lated to other concepts via relations deﬁned earlier).

If there are related concepts that are not known by the

user then these concepts are added to the list of re-

sult concepts. The algorithm is called recursively for

every concept in the result concepts list. The work-

ing is also described in ﬁgure 4. After the generation

of the concepts, to be taught, a linearization phase

is applied to them. This linearization is described

in (Vassileva, 1995) and it is utilized by (Capuano

et al., 2002) among other authors. The objective of

this phase is to order the generated concepts accord-

ing to the Order relation. Furthermore, the algorithm

can generate a concept multiple times; this phase will

help to eliminate this anomaly.

In the 2

step, these concepts are sent as input to

another algorithm Algorithm 2. The purpose of this

algorithm is to generate pedagogical resources based

on the generated concepts (1

step), the scenario and

the test model selected by the user. This is done by the

algorithm by querying the ”pedagogical and serious

game resource repository” keeping into the account

the user’s proﬁle and the selected models. In return

the repository returns to the algorithm the queried re-

sources.

In the 3

step the selected resources are given

to the algorithm Algorithm 3. The purpose of this

algorithm is to associate the selected pedagogical

Figure 4: Working of the Algorithm 1.

resources with appropriate serious game resources.

This is done by selecting an appropriate serious game

model and then querying the resource repository for

appropriate resources keeping into account the user’s

proﬁle. The result of this query is a set of serious

game resources initialized by pedagogical resources.

In the 4

, these resources are presented to the appli-

cation in the form of an scenario.

The user starts to interact with the serious game

and as a result the user’s interaction traces are gen-

erated. In the 5

step, these traces are transfered to

the user proﬁle. In the user proﬁle, these traces are

used as knowledge sources to update the proﬁle and

modify the scenario if necessary.

5 ILLUSTRATION

The application’s context (Project CLES) has been

mentioned and described earlier in section 2 . Here

we’ll give an example of the scenario generated by

our system for this application.

Project CLES has created a serious game based

environment. This game is an adventure game. The

main protagonist of this story is a treasure hunter

called Tom O’Connor. His objective is to ﬁnd a hid-

den relic. The story places Tom in one of several

rooms. Each room is connected to other rooms. Each

room represents one of the cognitive domains (de-

scribed above). Each room has multiple game objects

like (bed, lamps, paintings etc). Hidden behind some

of these objects are challenges. These challenges are

in the form of games. A screen shot of one of these

games is shown in ﬁgure 1. The user is required to

click one of these objects to access the challenges hid-

den behind them. Then the user is required to solve

these challenges in order to progress to other rooms.

A screenshot of the CLES’s environment is shown in

ﬁgure 5.

A SYSTEM FOR GENERATING PEDAGOGICAL SCENARIOS FOR SERIOUS GAMES

249

Figure 5: Screenshot of the game.

Our contribution in this project is the generation

of pedagogical scenarios given the pedagogical ob-

jectives and the proﬁle of the user. This domain has

eight main concepts and then these main concepts are

further divided. These main concepts are 1) percep-

tion, 2) attention, 3) memory, 4) visual-spatial 5) log-

ical reasoning 6) oral language 7) written language,

8) transversal competencies. The sub structure of all

of these eight domains is quite large. Hence, for the

sake of space we’ll show the modeling of only one of

the structures in the ﬁgure 6.

Figure 6: Domain Model of the concept Attention’. The

arrows labeled H.P. are Has-Parts relations.

Each and every atomic concept is associated with

a number of pedagogical resources. In this case these

resources are only of the type ’test’. These tests vary

in difﬁculty. For e.g. an element is hidden in a com-

plex image and it is required from the user to ﬁnd that

element as quickly as possible, many pair of glasses

are hidden in a decorated room and the user is re-

quired to ﬁnd these glasses, etc. Furthermore, se-

lected pedagogical resources are used by selected se-

rious game resources to be displayed to the user.

The scenario model deﬁnes that each concept

should be treated separately i.e. the user is presented

one concept at a time. Initially user is assigned with

one of the predeﬁned proﬁle (Stereotype) based on

his age. This proﬁle will be evolved according to the

user’s over time. Furthermore, the proﬁle is also used

in the system to provide the adaptation.

Now, let us suppose that the target concept to be

learned by the system is attention visual. Also sup-

pose, that the knowledge of the user about the con-

ception attention visual is declared 40% in the proﬁle.

Furthermore, the level of knowledge the user wants to

acquire is 50%. The user selects the appropriate pre-

sentation models (scenario and test model) and seri-

ous game model. These models along with the target

concept are given as input to the course generator.

Firstly, the CG, using the Algorithm 1 will generate

the list of concepts require to teach the goal concept.

For the target concept attention visual the selected

concepts are: 1.Barrage → 2.Sequence → 3.Total.

Now let us suppose that the selected scenario

model have the following structure: 1.Test → 2.Test

→ 3.Test.

For the concept Barrage, based on this scenario

model and the user’s proﬁle the Algorithm 2 will give

the following resources :

• ’Images’ of Barrage (difﬁculty 50%)

• ’letter’ of Barrage (difﬁculty 50%)

• ’number’ of Barrage (difﬁculty 50%)

Supposing that we have a repository of serious

game containing the resources like Chair, Table,

Statue. Based on the game model and the user’s pro-

ﬁle the Algorithm 3 will give the following serious

game resources :

• ’Images’ is hidden behind Bed

• ’letter’ is hidden behind Painting

• ’number’ is hidden behind Drawer

These resources are presented to the user via the

serious game. The user interacts with the serious

game and consequently traces are generated. These

traces are stored in the user’s proﬁle. Furthermore,

the proﬁle is updated based on these traces and the

scenario is modiﬁed (if necessary). For example, if

the user fails the 1

test another test with a difﬁculty

of 40% is presented to the user. This exercise will

be selected dynamically while the user is interacting

with the 1

exercise. This selection is done based on

the test model.

CSEDU 2011 - 3rd International Conference on Computer Supported Education

250

6 CONCLUSIONS

In this paper, we presented a system that is able to

generate pedagogical scenarios while keeping into ac-

count the speciﬁcities of serious games and users’ in-

teraction traces. We also presented the modeling of

this system along with a worked example of how the

system will generate a scenario.

There are systems that are designed to generate

pedagogical scenarios. However, these systems only

focuses on pedagogy and do not take into account the

speciﬁcities of serious games. Hence, they can’t be

used with serious games. Furthermore, they do not

make use of modeled users’ interaction traces in their

adaptation process. They are limitations of these sys-

tems from our point of view.

Similarly, there are some serious game systems

that are used to provide education. Some of these sys-

tems have a notion of pedagogical scenario but these

scenarios are static in nature and are not modiﬁed ac-

cording to the user’s proﬁle. Like course generators

these systems also do not take into account the users’

interaction traces for adaptation purposes.

The system we proposed in this paper addresses

these problems. The solution is presented in the form

of a system. The architecture of this system is pre-

sented and detailed. This proﬁle includes the infor-

mation about the user, his competence and the traces

of his/her interaction with the serious game. These

traces are used as knowledge sources in the adaptation

process. The domain knowledge is divided into three

layers. These layers include the domain concepts,

the pedagogical resources and the resources of seri-

ous games. Furthermore, the working of the Course

Generator is also presented. This course generator

is responsible for the generation of the serious game

scenario.

The context of our work’s application is presented

in this paper. We are creating a prototype of the sys-

tem that implements this architecture. Then this sys-

tem will be tested on a real time system with real stu-

dents. These students will have a cognitive impair.

These tests will present us with an opportunity to val-

idate our approach. After the validation of this ap-

proach we’ll test our system in other domains to vali-

date the general applicability of our system.

REFERENCES

Brusilovsky, P. (1993). Task sequencing in an intelligent

learning environment for calculus. In Seventh Inter-

national PEG Conference, pages 57–62.

Capuano, N., Gaeta, M., Micarelli, A., and Sangineto,

E. (2002). An integrated architecture for automatic

course generation. In Proceedings of the IEEE Inter-

national Conference on Advanced Learning Technolo-

gies (ICALT 02), number Section 4, pages 322–326.

Citeseer.

Carron, Thibault And Marty, Jean-Charles And Heraud, J.-

M. (2008). Teaching with Game Based Learning Man-

agement Systems : Exploring and observing a peda-

gogical. Simulation & Gaming, 39(3):353—-378.

Dondlinger, M. (2007). Educational video game design: A

review of the literature. Journal of Applied Educa-

tional Technology, 4(1):21–31.

Duitama, F., Defude, B., Bouzeghoub, A., and Lecocq,

C. (2005). A framework for the generation of adap-

tive courses based on semantic metadata. Multimedia

Tools and Applications, 25(3):377–390.

Karampiperis, P. and Sampson, D. (2005). Adaptive

learning resources sequencing in educational hyper-

media systems. Educational Technology & Society,

8(4):128–147.

Michael, David R. And Chen, S. L. (2005). Serious Games:

Games that educate, train, and inform. Muska &

Lipman/Premier-Trade.

Morenoger, P., Sierra, J., Martinezortiz, I., and Fernandez-

manjon, B. (2007). A documental approach to adven-

ture game development. Science of Computer Pro-

gramming, 67(1):3–31.

Sehaba, K., Encelle, B., and Mille, A. (2009). Adaptive

TEL based on Interaction Traces. In In AIED 09

(14 International Conference on Artiﬁcial Intelligence

in Education) workshop on ”Towards User Modeling

and Adaptive Systems for All (TUMAS-A 2009): Mod-

eling and Evaluation of Accessible Intelligent Learn-

ing Systems”.

Specht, M., Kravcik, M., Pesin, L., and Klemke, R.

(2001). Authoring adaptive educational hypermedia in

WINDS. Proceedings of ABIS2001, Dortmund, Ger-

many, 3(3):1–8.

Torrente, J., Moreno-Ger, P., Fern

andez-Manj

on, B., and

del Blanco, A. (2009). Game-like Simulations for

Online Adaptive Learning: A Case Study. In Pro-

ceedings of the 4th International Conference on E-

Learning and Games: Learning by Playing. Game-

based Education System Design and Development,

page 173. Springer.

Ullrich, C. and Melis, E. (2009). Pedagogically founded

courseware generation based on HTN-planning. Ex-

pert Systems with Applications, 36(5):9319–9332.

Vassileva, J. (1995). Dynamic courseware generation: at

the cross point of CAL, ITS and authoring. In Pro-

ceedings of ICCE, volume 95, pages 290–297.

Zyda, M. (2005). From visual simulation to virtual reality

to games. Computer, 38(9):25–32.

A SYSTEM FOR GENERATING PEDAGOGICAL SCENARIOS FOR SERIOUS GAMES

251