Authoring Storyline-based Adaptive 3D Virtual Learning

Environments

Ahmed Ewais and Olga De Troyer

Web & Information Systems Engineering (WISE) Laboratory, Department of Computer Science,

Vrije Universiteit Brussel, Brussels, Belgium

Keywords: Authoring Tool, 3D Virtual Learning Environments, Virtual Reality, Adaptive, Storyline.

Abstract: Adaptive three-dimensional (3D) Virtual Leaning Environments are interesting for the e-learning domain as

they have great potential resulting from the capabilities offered by 3D Virtual Environments in combination

with the opportunities offered by adaptive systems. However, their breakthrough is hindered by the

difficulty of their development. This paper presents a development approach that allows course authors to

create adaptive 3D virtual learning environments without the need to be an expert in 3D or using

programming or scripting languages. In particular, the paper elaborates on the principles used for the

authoring approach, as well as on the different aspects that need to be supported, i.e. the pedagogical

aspects, the adaptation aspects, and the requirement to support the specification of an adaptive storyline

which should be followed by learners.

1 INTRODUCTION

3 dimensional (3D) virtual environments are defined

as three-dimensional (3D), multisensory, immersive,

real time, and interactive simulations of a space that

can be experienced by users via three dimensional

input and output devices (Burdea, 2006). This kind

of environment is often used for entertainment, but

can also be used in the context of education. For

certain subjects and for certain types of learners

using a 3D Virtual Learning Environment (3D VLE)

may be much more appealing and motivating than

the use of classical learning material, e.g., to

simulate the effect of physical laws (Cobb et al.,

1998); to simulate social environments and allow

people to practice social skills (Adams et al., 2008);

or to learn about history (Di Blas et al., 2003); Dede

et al., 2003). Equipping those 3D VLE with adaptive

capabilities could offer even more advantages

(Chittaro and Ranon, 2007). For instance, adaptivity

allows the 3D VLE to adapt dynamically (i.e. at run

time) to the individual learner and to the progress

that the learner makes during the learning process. It

could mediate the distinction between education and

entertainment, which could improve the learner’s

experience and motivate him better. Furthermore,

adaptivity may be used to prevent the learner form

being overwhelmed in the 3D VLE. For instance, it

may be more effective to guide a learner through the

3D VLE according to his or her background and

learning goals, or only show the learner the objects

that are relevant for his current knowledge level, or

adapt the environment to his learning style. Also

adaptivity may be used to decrease the risk that

learners are distracted too much and are therefore

not able to focus on the actual learning task (De

Troyer et al., 2010).

As the major goal of a 3D VLE is education,

involving teachers or educational-schooled people is

very important. However, with the current

development tools for 3D virtual environments, it is

very difficult to involve these people actively in the

development process, as they usually don’t have

background in programming or in 3D. One way to

overcome this drawback is to provide authors an

authoring tool that enables them to turn existing 3D

virtual environments into adaptive 3D virtual

learning environments. With such a tool, the authors

should be able to manipulate the 3D VLE contents to

fit it to the context of the course and to the learners,

and specify the required adaptive behaviour. The

usability of such a tool will be a critical factor in its

adoption.

Our work is based on the acquired experience in

the context of a EU FP7 STREP project GRAPPLE

Ewais A. and De Troyer O..

Authoring Storyline-based Adaptive 3D Virtual Learning Environments.

DOI: 10.5220/0004347700150024

In Proceedings of the 5th International Conference on Computer Supported Education (CSEDU-2013), pages 15-24

ISBN: 978-989-8565-53-2

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

(GRAPPLE, 2008). It aimed at the construction of a

generic adaptive learning environment that can be

used/accessed at home, school, work, or on the

move. GRAPPLE includes authoring tools that

enable teachers to specify adaptation strategies for

the contents and activities of their courses. The

project considered classical learning materials, as

well as simulations and virtual reality. Based on our

experience in the GRAPPLE project, we found it is

important to allow authors to provide a storyline for

an adaptive 3D VLE that learners can follow during

a course. This finding is supported by work done in

the context of educational games, where narration

helps the player to be more involved in the game

and, as a result, to increase the expected learning

outcomes (Salen and Zimmerman, 2003).

Furthermore, narrative aspects can play an important

role to deliver the instructional message and

appropriate goals (Laurillard, 1998).

The rest of the paper is structured as follow. In

section 2, we consider work related to authoring 3D

environments in the context of e-learning and

educational games. Section 3 explains the basics of

adaptive 3D VLEs. In section 4, our authoring

approach for adaptive 3D VLEs is explained, and

section 5 proposes the architecture for the associated

authoring tool. Section 6 presents conclusions.

2 RELATED WORK

This section reviews related work dealing with

authoring adaptivity in 3D virtual learning

environment and video games.

There exist different Virtual Reality (VR)

development environments that enable advanced

authors in VR to create 3D virtual learning

applications. Some of them are specific for certain

3D virtual environments. Examples are

(3Dxplorer®, 2002); (OpenQwaq, 2007); (Creator®,

2007). Another interesting example is the Sloodle

project (Livingstone and Kemp, 2008) which is

based on integrating learning and teaching across

SecondLife (SecondLife, 2003) which is an online

3D virtual world, and Moodle (Moodle, 2002) which

is a course management system. However, in the

context of authoring tools to deliver adaptive 3D

VLE, there is little available. For this reason, the

related work mentioned in this section will also

discuss narrative, adaptivity and authoring tools

developed in the context of educational games.

An authoring toolkit called <e-adventure3D> is

presented in (Torrente et al., 2008). The authoring

tool has a scene editor that supports authors to define

and configure size (width and depth), texture and

roughness settings of the 3D virtual environments.

Furthermore, the author is able to define the camera

and lights settings, which will be considered as user

guidance in the 3D virtual environment. The

authoring tool also provides an editor to define the

interactive elements inside the 3D virtual

environments. The end-user will be able to interact

with such predefined elements by carrying out some

tasks and actions. Also the editor allows the author

to place the interactive elements in a scene.

Furthermore, the author can scale, rotate and/or

translate them to fit correctly in the scene.

Additionally, a 3D avatar representing the learner is

available and which “can engage in interactive

conversations that can be graph-shaped if cycles are

needed or tree-shaped otherwise. The editor

includes a graphical facility to create these

structures easily.” (Torrente et al., 2008).

It should be noted that in <e-adventure3D>,

game authors are not able to specify which kind of

users interactions should be monitored, as this is

limited to what has been implemented in the game

API. Note that being able to monitor user

interactions is important for being able to adapt the

3D virtual environment at run time. Another

shortcoming is that adaptation resulted from

monitoring the user interactions can only be

performed in future execution of the game, not in

real time. Furthermore, the authoring tool is used for

only two predefined environments: closed

environments and connected virtual rooms.

In (Marchiori et al., 2012), WEEV (Writing

Environment for Educational Video games) is

explicitly considering three elements or tasks to

create narrative point-and-click educational games:

(1) the author needs to define the actors, which are

the interactive elements in the game. In other words,

actors are the 3D objects upon which the player can

perform an action and interact with them; (2) the

virtual game world or so-called gamespace should

also be defined. The virtual world includes the

different interactive elements and non-interactive

elements of the educational game; (3) the game story

must be edited separately to define the player’s

interactions and the game feedback. The WEEV

authoring tool is implemented as three editors to

support the author in defining the main three

elements: Simple Actor Editor, World Editor and

Story editor. A domain specific visual language is

used for the World and Story Editors. The Actors

Editor enables authors to view a list of all the actors

in the game in a specific panel. Furthermore, the

author is able to select a new actor form a “resource

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

library” containing different graphic resources. The

World Editor allows the author to define the game

world. On the other hand, the Story Editor is used to

define the story-flow of adventure games. In general,

a story is represented “…based on a state-transition

diagram, where each state represents a point in the

game story and each transition an interaction by the

user with the system, which moves the story along.”

(Marchiori et al., 2012). The basic elements of the

proposed game story flow language are: (a) state

which represents the status of the game, (b)

transition which is used to define a player action

which moves the state of the game to next one, (c)

feedback which is associated with the transition

element to display effects or feedback when a

transition happen. Furthermore, the proposed

domain specific visual language provides specific

features to integrate educational characteristics like

content adaptation. For instance, the author can

create an explicit representation of student

assessment, hints for guidance and adaptation of the

game story. Adaptation can be defined for the story-

flow of the game depending on the student profile.

The story-flow can have more than one initial state.

The initial nodes of the story-flow are associated

with different adaptation profiles in the game, for

instance, “easy”, “medium” and “hard” difficulty

levels.

Researchers in (Gaffney et al., 2010) described

an authoring tool (ACTSim) which is developed to

allow educators with non-technical background

authoring situational simulations. Situational

simulation focuses on teaching learners how to

perform targeted task (Alessi and Trollip, 2001).

Their approach designs adaptation aspects in two

ways. First, a tagging mechanism allows authors to

define multiple properties for so-called adaptive

dimensions (Wade, 2009) which include role of the

learner, learning outcomes, categorization of the

dialogue nodes and the related subjects. This is done

by providing a highlight function in the authoring

tool which allows the author to know how the

dialogue model will be adapted. Secondly, triggers

are used which allow the authors to define

adaptation depending on educational principles of

assessment, feedback and reflection. The ACTSim

authoring tool supports the two approaches to design

adaptive simulation.

 A good principle of the

approach is not adhering the authors to compose

complex rules in order to provide adaptation. An

evaluation was conducted for the proposed authoring

tool and the result was promising.

3 ADAPTIVE 3D VLE

In order to be able to discuss the principles used for

our authoring approach, we first need to present the

basics of adaptive 3D VLEs. First, we will discuss

the different components that make up a 3D VLE.

Next, we discuss the adaptation possibilities for 3D

VLEs.

3.1 3D VLE Anatomy

Before we discuss how to drive adaptation for 3D

VLE, it is important to know what are the 3D VLE

components on which adaptation can happen.

Conceptually, we can distinguish the following

components and associated functionality in a 3D

Virtual (Learning) Environment: (1) The virtual

scene that corresponds to the 3D space which will be

populated with the 3D virtual objects. (2) Virtual

objects are objects which have a visual

representation having colour and material properties,

a size, a position in the space, and an orientation. (3)

Object behaviours that are the behaviours associated

to the virtual objects. Behaviours may reflect real

life behaviours like rotation, walking, etc. (4) User

interactions as users are able to interact with the

virtual objects. For example, a user may pick up an

object and drag it to some other place in the space (if

the object is moveable). (5) User navigation which

is related to the way the user can move in the 3D

space, e.g., walking, running or flying. The user

navigates by a so-called avatar. The user’s avatar

can be represented explicitly (by an object) or

implicitly in which case the viewpoint of the camera

is used to show the user’s position. (6)

Communication as nowadays, more and more 3D

virtual environments are also collaborative

environments in which remote users can interact

with each other, e.g., talk or chat to each other or

perform activities together. In this research, we

concentrate on single-user environments. (7) Sound

can be important component in simulations to

enhance the feeling of reality or simply to simulate

some sound. Sound/speech can also be used as an

instruction and feedback mechanism during the

learning process. More explanation about 3D VLE

components is presented in (De Troyer et al., 2010)

and (Kipper and Palmer, 2000); (Bowman and

Hodges, 1999).

3.2 Possible Adaptation Techniques for

3D VLE

Different adaptation techniques and mechanism are

AuthoringStoryline-basedAdaptive3DVirtualLearningEnvironments

proposed to be applied to 3D virtual objects. For

instance, the work presented in (Chittaro and Ranon,

2007) presents adaptation techniques which are

basically limited to 3D contents representations. On

the other hand, the authors in (Dos Santos and

Osório, 2004) propose the so-called Intelligent

Virtual Agent to provide adaptive navigation

towards interesting 3D virtual objects. However,

adaptation can be applied beyond 3D material

representation or user navigation. In principle,

adaptation can happen for each component of a 3D

VLE. An adaptation can be limited to a single

component of the 3D VLE, but it can also involve

different components of the 3D VLE. We defined

two different adaptation categories (De Troyer et al.,

2010). The first category includes adaptations that

apply on a single component, i.e. possible

adaptations for objects, behaviours, interaction, and

for navigation. We call these adaptation types. For

instance, there are adaptation types related to 3D

virtual objects such as semi-display, changeSize,

changeMaterialProperties,

changeVRRepresentation, etc. For the moment, we

did not consider any adaptations types yet for the

scene or sound. Communication is also not

considered because we focus on single-user 3D

VLEs. The second category is related to more high-

level adaptations that involve more than one

component. These kinds of adaptations we called

adaptation strategies. For instance, adaptation

strategies that can be applied to group of 3D virtual

objects are filterObjects, markObjects,

displayAtMost, displayAfter, etc. More examples

about adaptation types and strategies can be found in

(De Troyer et al., 2010).

4 AUTHORING AN ADAPTIVE

3D VLE

In general, the purpose of an authoring tool is to

enable authors defining a course at a high level of

abstraction and without resorting to programming or

scripting languages. In the same way, the purpose of

our authoring tool is to allow authors to define an

adaptive 3D VLE without the need to be

experienced in 3D/VR and to know programming

languages. In principle, this involves authoring the

3D virtual environment, as well as the adaptivity and

the pedagogically relevant aspects of the course. We

will not consider the authoring of the 3D virtual

environment itself, as for this many tools are

available (e.g., Google SketchUp (GoogleSketchUp,

2000)), 3D Studio Max (Murdock, 2003)). The

purpose of our authoring is to add storyline,

adaptivity and pedagogical aspects to an existing 3D

virtual environment.

The approach taken for adaptivity is an author-

driven approach which means that during the design

of the adaptive course, the author needs to specify

explicitly when and how the content needs to be

adapted. The advantage of this is that it gives the

control to the author. On the other hand, the

disadvantage is that it requires the author to keep

track, at design time, of all possible adaptation

scenarios. However, we believe that supporting

authors with appropriated tools can overcome this

disadvantage.

In the following sections, we first present the

(high level) requirements that we formulated for our

authoring approach. Next, we discuss the models

and principles used for the authoring approach.

Finally, we provide more details on the three most

important steps of the authoring process.

4.1 Requirements

Based on our previous work done in the context of

the GRAPPLE project, we derived some important

features to be considered in authoring adaptive and

educational 3D virtual environments. The most

important features can be summarized as follows:

 Pedagogical-oriented: As we are dealing with 3D

virtual learning environments, we want to provide

means to take into consideration some pedagogical

aspects when specifying an adaptive 3D VLE (i.e.

the course). This is achieved by enabling authors

to create pedagogical relations between the

different learning concepts.

 Adaptive-specific: As we are dealing with adaptive

3D virtual learning environments, we want to

provide means to specify the adaptivity of a 3D

VLE using a high level of abstraction. The idea is

to provide a number of different (pre-defined)

adaptation types and strategies to allow the author

to express the adaptivity.

 Teacher-oriented: One of the major goals is

allowing teachers and pedagogically schooled

people to be directly involved in the authoring

process of the adaptive 3D VLE. This requires that

the authoring approach should be at a high level of

abstraction, and using terminology that is

understandable by this type of users.

 Storyline-oriented: We have observed (during

evaluations performed in the context of the

GRAPPLE project (Ewais and De Troyer, 2014))

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

that when using a 3D VLE, there is a need to guide

the learners through the learning material offered

in the 3D VLE in a way similar as this is done for

classical text-based learning material (where some

kind of sequence is enforced or advised). For

instance, the author may want to express that the

learners first should get a guided tour through the

virtual environment highlighting the most

important learning objects, next they should focus

on performing some activities related to one

learning concept, and so on. However, while for

classical learning material a personalized

“sequence” could be achieved by specifying

adaptation rules only, our previous work done in

the context of the GRAPPLE project has shown

that this is very difficult to be achieved for 3D

VLEs only with the use of adaptation rules. To

allow expressing such a “sequence”, we will use

the concept of storyline (taken from game

development).

 Web-based: A web-based authoring tool will allow

easy access to the tool independent of the

platforms or PCs that are used by the authors.

4.2 Models & Principles

In general, classical adaptive systems maintain

different kinds of information in different models

(Paramythis and Loidl-Reisinger, 2003). We also

follow such an approach. Authoring an adaptive 3D

VLE goes as follows. First a Domain Model (DM)

(Hendrix et al., 2008) describing the concepts that

should be considered in the course, needs to be

created. Learning resources and materials are

associated with these learning concepts.

We also use a User Model (UM), which is used

to maintain the learner’s characteristics, learning

background, and his learning progress.

Next, pedagogical relationship types should be

defined between the different learning concepts. An

example of a pedagogical relationship type is the

prerequisite relation that expresses that one concept

is a prerequisite to master the other concept. These

pedagogical relationships are defined in the

Pedagogical Model (PM). The PM also expresses at

a high-level how the User Model should be updated

(at runtime) based on what the learner does in the

learning environment. After that, the author defines

the actual adaptivity behaviours (Adaptation Model)

for the 3D VLE. For this a rule mechanism is used.

Adaptation rules are basically if-then rules. The if-

part specifies the conditions to be satisfied for

performing the adaptation, while the then-part

specifies the adaptation actions. These adaptation

actions are expressed in terms of the adaptation

types and strategies (see section 3.2) to be performed

on the 3D VLE components. In our approach, the

adaptation rules are defined in context of two sub-

models: Adaptive Storyline Model (ASLM) and

Adaptive Topic Model (ATM). We first explain the

purpose of these two models.

As indicated in section 4.1, to define the actual

course, the author will use a storyline. This storyline

defines the steps that the learner should follow

during his learning process. The storyline can be

adaptive, i.e. the actual flow can adapt at runtime

e.g., depending on the learner’s knowledge about

specific concepts. The storyline and how it should

adapt is defined in the Adaptive Storyline Model. A

storyline consists of number of topics; each topic is

dealing with a number of 3D virtual objects. The

topics are connected with so-called storyline

adaptation rules. These adaptation rules specify the

adaptive behaviour of the storyline. Figure 1 depicts

the sub models in the adaptation model.

Next to the adaptivity expressed for the storyline,

more adaptivity can be expressed for the topics. A

topic is defined by means of an Adaptive Topic

Model. In this model, the 3D virtual objects

involved in the topic are specified, as well as the

actual adaptation actions that have to be applied to

them. Moreover, it specifies when the adaptions

need to be applied. Also here, a rule mechanism is

used.

In the rules, the conditions are not only based on

the learner’s preferences or learning background and

learning progress (maintained in the User Model)

but also on the activities performed by the learner in

the 3D VLE (e.g., number of interactions with a

certain object, the behaviours performed). This type

of information, which may not always be directly

related to the learning, is added to be able to control

the learner’s behaviour in the 3D VLE, e.g., to avoid

that he wastes too much time by playing around or

to help him when he has problems in navigation or

interacting with objects. This information is stored

in the so-called 3D VLE Activity History Model. It is

important to mention that not all the user activities

inside the 3D VLE are monitored. But the more

information is kept about the activities of the learner

in the 3D VLE, the more this information can be

taken into account to adapt the 3D VLE to the

individual learner. Note that some data from this 3D

VLE Activity History need to be translated into User

Model data. For instance, it needs to be defined what

activities need to be performed in the 3D VLE by

the learner to raise the knowledge level of a certain

learning concept.

AuthoringStoryline-basedAdaptive3DVirtualLearningEnvironments

Figure 1: Two-level specification of the Adaptation

Model.

Maintaining the User Model and the 3D VLE

Activity History at runtime is done by the adaptive

delivery environment. Figure 2 depicts the different

models used in our approach to perform the

adaptation process inside a 3D VLE and how they

are related to the authoring process, respectively the

adaptive delivery process. It is actually the Adaptive

Engine that will figure out when to use which

adaptation rule. It is also the Adaptation Engine that

sends updating instructions that are related to the

User Model according to user activities performed

inside the 3D VLE. The Adaptive Engine is also

responsible for selecting the appropriated 3D

resources (based on the adaptation rules) and

sending them to the delivery environment (client

side). Actually, the exact content of the course is

composed on the fly by the Adaptation Engine.

Figure 2: Conceptual Framework for Adaptive 3D VLE.

In the next three sections, we will elaborate on

the three most important aspects of the authoring

process, i.e. authoring the Pedagogical Model, the

Adaptive Storyline Model, and the Adaptive Topic

Model.

4.3 Authoring Pedagogical Aspects

The Pedagogical Model (PM) is used to define the

pedagogical structure of the learning concepts, as

well as to indicate how the learner’s knowledge

(captured in the User Model) should be updated

based on the learner’s activities inside the 3D VLE.

A typical Pedagogical Relationship Type (PRT)

is the prerequisite. When a learning concept A is a

prerequisite for learning concepts B, this means that

the learner needs to study concept A before he can

start with concept B. Other PRTs like Defines,

Illustrates, Interest, Propagates_knowledge and

Update_knowledge can also be used to define

relationships between the different learning

concepts.

After defining Pedagogical Relationships Types

between learning concepts, the author can specify

the Pedagogical Updating Rule (PUR) for each

PRT. PURs are responsible for defining how the

User Model needs to be updated based on the

activities performed by the learner. PURs are

condition-action rules. At runtime, the PURs are

triggered on accessing (visiting) learning concepts.

The action part of a PUR defines the updates that

will happen. For instance, the interest-PRT has a

PUR that for visiting a 3D learning concept will

update the user’s interest level of the target 3D

learning concept. The following is an example of

Pedagogical Updating Rule associated to the

interest-PRT:

IF (UserID.3DLearningConceptA.

knowledge = ‘GOOD’)

THEN (UserID.3DLearningConceptB.

interest= ’HIGH’)

Note that, rather than requiring the authors to

specify the learning sequence for the learning

concepts in all possible cases, the authors only have

to provide the PRTs between the different learning

concepts and the adaptive engine will be able to

dynamically derive the required learning sequence.

4.4 Authoring Storyline Adaptation

As already explained, an adaptive storyline enables

the delivering of the topics in an adaptive sequence.

To achieve this, storyline adaptation rules are used.

A storyline adaptation rule is associated with a topic,

the source topic, and is used to control the transition

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

to the next topic, the target topic. Such a rule

consists of two parts: a condition part and an action

part. The condition part is used to specify when the

learner can be directed from the current topic (the

source topic) to the next topic (the target topic). The

condition is, in general, based on two aspects: the

learner’s current knowledge level about the current

topic and the suitability of the next topic for the

learner. To simplify tracking the learner’s

knowledge and topic suitability, every topic in the

storyline has two parameters: topic knowledge and

suitability. A learner can acquire knowledge about

the topic in a topic by navigating and interacting

with the 3D objects (concepts) related to the topic.

The knowledge level about the topic is incrementally

increased by acquired knowledge about the related

concepts. The author needs to define how the topic’s

knowledge level can be increased. Once the

learner’s knowledge about the current topic crosses

the specified threshold (given in the condition part

of the rule) and the target topic is suitable for the

learner then the learner will be able to progress to

the target topic. This is specified in the condition

part of the topic adaptation rule. For instance, in this

way, the author can specify that a beginner should

not be able to go to a topic for advanced learners.

In the action part of the storyline adaptation

rules, the author specifies the adaptation strategies

(see section 3.2) that should be applied on 3D virtual

objects (learning concepts) of the target topic. For

example, the author could specify the

interactionAtMost adaptation strategy in the action

part of a storyline adaptation rule to specify the

maximum number of times the learner can interact

with the 3D virtual objects (concepts) inside the

target topic.

Let us illustrate the adaptive storyline by the

following example. An author creates TopicX,

TopicY, TopicZ, TopicR, and TopicT. Those topics

consider the different issues to be discussed in the

course. Every topic should include a number of

learning concepts. The author can determine that

learners with basic knowledge (beginners) should

follow a larger number of topics than the learners

with good knowledge (advanced). For instance, two

possible storylines are envisaged. An advanced

learner should have the following flow of topics:

Topic X-> Topic Y -> Topic Z. After having

completed topic Z, he will finish the course. On the

other hand, learners with beginner level should

follow a larger number of topics, as they need to

master more concepts in order to master the topic:

Topic X-> Topic Y-> Topic T-> Topic Z. In the

same way as reading a course book, students with

advanced knowledge can skip sections or even

chapters. To realise this, the following topic

adaptation rule can be associated with Topic X:

IF(user.TopicX.knowledge>

Required_Value)&&

(user.TopicY.suitability=TRUE)

THEN navigationWithRestrictedBehaviour

<ConceptA,ConceptB,ConceptC>

This rule states that the learner can progress from

TopicX to TopicY when his knowledge about

TopicX is above “Required_Value” (defined by the

author) and his suitability for TopicY is TRUE. In

addition, the action part specifies that in TopicY, the

user can navigate through the environment but he

will not be able to see rotation behaviours that are

related to the ConceptA, ConceptB , and ConceptC.

4.5 Authoring Topic Adaptation

Inside a topic, the author can specify when and how

the content and structure of the 3D VLE needs to be

adapted. This is done using adaptation rules. For

instance, the author can define an adaptation rule to

change a 3D virtual object’s properties based on the

learner’s activities inside the 3D VLE. To achieve

that, an event-condition-action rule mechanism is

used. The event-part is related to specific events

related to the user’s activities inside the 3D VLE,

e.g., interaction with a 3D virtual object by mouse

clicks or navigating close to a 3D virtual object. The

event-part will be responsible for triggering the rule,

but the condition-part must be satisfied in order to

actually perform the action-part of the rule. The

condition-part can deal with the learner’s

preferences, his learning background, and progress

but also with previous activities performed by the

learner in the 3D VLE. These last types of

conditions are added to be able to control the

learner’s behaviour in the 3D VLE, e.g., to avoid

that the learner wastes too much time by playing

around. The following is an example of such a rule.

As an example, we explain an adaptation topic

rule called SemiDisplay3DObject between Concept

C and Concept D (see Figure 3). This rule will be

evaluated once the learner comes close to (the 3D

virtual representation of) concept C. As a result, the

condition part of the rule will be evaluated which is

in this case related to 3D VLE Activity History.

Remember that the condition part can refer to user

activities history inside the 3D VLE but also to his

preferences, knowledge level, etc. If the condition

part is true, then the so-called semidisplay adaptation

AuthoringStoryline-basedAdaptive3DVirtualLearningEnvironments

type will be applied to concept D.

Figure 3: SemiDisplay3DObject adaptation topic rule.

Such a rule could be defined to draw the learner’s

attention to another 3D object instead of wasting his

time by interacting with one single 3D virtual object.

5 ADAPTIVE 3D VLE

AUTHORING TOOL:

ARCHITECTURE

In this section, we present the architecture of the

authoring tool. The main modules are shown in

Figure 4. The Web services toolkit allows the

authoring tool to connect with other external models

such as the User Model and the Domain Model.

Moreover, it allows the author to use local and/or

external 3D modelling repositories used to store 3D

materials and resources.

The actual interface to the user (i.e. author) is

composed of two sub-environments. The Data

Models Environment facilitates the preparation

process for the required data models in the authoring

process. The other sub-environment is called the

Pedagogical and Storyline Visual Environment and

provides access to three editors, which use visual

languages to define respectively the Pedagogical

Model, the Adaptive Storyline Model and the

Adaptive Topic Model. Every sub-environment has

different editor tools which are linked together using

a unified user interface style. Furthermore, a

Visualization Tool (previewer) is integrated in the

authoring environment to allow authors to preview

3D VLE and to display actual effects of customized

adaptation types and strategies upon 3D VLE

components.

The Data Model Environment is composed of

five editor tools: the Pedagogical Relationship type

Editor to define PRTs; the 3D Learning Object

Editor to specify the required 3D learning objects;

the 3D Adaptation States Editor to create or modify

adaptation types and adaptation strategies that will

be applied to 3D objects; the 3D Learning Materials

Editor to associate adaptation types and strategies to

Figure 4: Overview of the 3D VLE authoring tool

Architecture.

specific 3D learning objects; and the VR Events

Editor to define or customize the VR events

specifications (for instance, the proximity area for

the ‘close to’-event can be set/changed with the VR

Events editor). It is important to note that these

editors are targeted to more VR experienced users.

Novice authors only need to use the 3D Learning

Materials Editor to indicate which 3D objects they

will use. The other editors in this environment are

needed to specify more details related to PRTs, 3D

virtual objects, adaptation types and strategies, and

VR events. But in general, defaults can be used. The

resulting data models are the input for the

Pedagogical and Storyline Environment.

The Pedagogical and Storyline Environment has

three visual editors, which are the Pedagogical

Model Editor, the Adaptive Storyline Editor and

Adaptive Topic Editor. We have opted for visual

modelling languages, as our authoring tool is

teacher-oriented and should be accessible by novice

authors. The languages themselves will be described

elsewhere.

Based on all information entered, the authoring

tool can generate a 3D course in an XML format that

can be parsed by an adaptive engine like AHA! (De

Bra et al., 2003) or the adaptive engine used in the

GRAPPLE project. However, this XML file can also

be translated by means of XSL and XSLT into some

other format.

6 CONCLUSIONS

Adaptive 3D virtual learning environments offer

many advantages for learning. However, their

breakthrough is impeded by the lack of authoring

tools. In this paper, we have presented an authoring

approach and associated tool architecture for these

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation

kinds of virtual learning environments. The

approach and tool is based on experience obtained

from our work done in the GRAPPLE project.

The approach is based on a number of principles

that have been discussed in the paper. We have

mentioned the different components of a 3D virtual

learning environment and discussed how they can be

adapted for satisfying the needs of a learner. We

have motivated the need for a pedagogical

component in the authoring process, as well as the

need for a storyline-based authoring approach. The

first requirement is achieved by enabling authors to

define pedagogical relationships between the

different learning concepts. To satisfy the second

requirement, they are able to define a storyline

inside the 3D VLE using a 2-layer approach.

Adaptation rules can be associated with both layers,

in each layer focussing on a different level of

adaptation.

We are currently working on developing a

prototype for the authoring tool. Furthermore, an

empirical evaluation will be considered to validate

the usability aspects of the authoring tool.

REFERENCES

3Dxplorer® (2002). 3Dxplorer®. 2002. Available from:

http://www.3dxplorer.com/. Accessed: November

2012.

Adams, W. K., Reid, S., Lemaster, R., Mckagan, S. B.,

Perkins, K. K., Dubson, M. & Wieman, C. E. (2008).

A Study of Educational Simulations Part I -

Engagement and Learning. Journal Of Interactive

Learning Research. 19 (3).pp. 397–419.

Alessi, S. M. & Trollip, S. R. (2001). Multimedia for

learning : methods and development. London: Allyn

and Bacon.

Bowman, D. A. & Hodges, L. F. (1999). Formalizing the

Design, Evaluation, and Application of Interaction

Techniques for Immersive Virtual Environments.

Journal of Visual Languages Computing. 10 (1). pp.

37–53.

Burdea, G. (2006). Virtual Reality Technology - An

Introduction. In: IEEE Virtual Reality Conference VR

2006. 2006, IEEE, pp. 307–307.

Chittaro, L. & Ranon, R. (2007). Adaptive Hypermedia

Techniques for 3D Educational Virtual Environments.

IEEE Intelligent Systems (SpecialIssue on Intelligent

Educational Systems). 22. pp. 31–37.

Cobb, S. V. G., Neale, H. R. & Reynolds, H. (1998).

Evaluation of virtual learning environment. In: the 2nd

European Conference on Disability, Virtual Reality

and Associated Technologies. 1998, Skovde, Sweden:

JI Lindstrom, pp. 17–23.

Creator®, V. (2007). Vastpark Creator®. Available from:

http://www.vastpark.com/. Accessed on november

2012.

De Bra, P., Aerts, A., Berden, B., De Lange, B., Rousseau,

B., Santic, T., Smits, D. & Stash, N. (2003). AHA!

The adaptive hypermedia architecture. Proceedings of

the fourteenth ACM conference on Hypertext and

hypermedia HYPERTEXT 03. 4 (1). pp. 81–84.

Dede, C., Ketelhut, D. & Ruess, K. (2003). Designing for

Motivation and Usability in a Museum-based Multi-

User Virtual Environment.

De Troyer, O., Kleinermann, F. & Ewais, A. (2010).

Enhancing Virtual Reality Learning Environments

with Adaptivity : Lessons Learned. In: The 6th

international conference on HCI in work and learning,

life and leisure: workgroup human-computer

interaction and usability engineering. Klagenfurt,

Austria: Springer-Verlag, pp. 244–265.

Di Blas, N., Hazan, S., Bearman, D. & Trant, J. (2003).

The SEE Experience: Edutainment in 3D Virtual

Worlds. In: Museums and the Web., Archives &

Museum Informatics.

Dos Santos, C. T. & Osório, F.S. (2004). An intelligent

and adaptive virtual environment and its application in

distance learning. Proceedings of the working

conference on Advanced visual interfaces AVI 04. pp.

362.

Ewais, A., & De Troyer, O, (2014) Authoring Adaptive

3D Virtual Learning Environments.In: International

Journal of Virtual and Personal Learning

Environments (IJVPLE), 5(1). [to be published]

Gaffney, C., Dagger, D. & Wade, V. (2010). Authoring

and Delivering Personalised Simulations - an

Innovative Approach to Adaptive eLearning for Soft

Skills. J.UCS. 16 (19). p.pp. 2780–2800.

GoogleSketchUp (2000). Google SketchUp. Accessed:

November 2012.

GRAPPLE (2008). GRAPPLE. Available from: http://

www.grapple-project.org/. Accessed: November 2012.

Hendrix, M., De Bra, P., Pechenizkiy, M., Smits, D. &

Cristea, A. (2008). Defining Adaptation in a Generic

Multi Layer Model: CAM: The GRAPPLE

Conceptual Adaptation Model P. Dillenbourg & M.

Specht (eds.). Times of Convergence Technologies

Across Learning Contexts. pp. 132–143.

Kipper, K. & Palmer, M. (2000). Representation of actions

as an interlingua. In: NAACLANLP 2000 Workshop

Applied Interlinguas Practical Applications of

Interlingual Approaches to NLP Seattle. pp 12-17.

Laurillard, D. (1998). Multimedia and the learner’s

experience of narrative. Computers & Education. 31

(2). pp. 229–242.

Livingstone, D. & Kemp, J. (2008). Integrating Web-

Based and 3D Learning Environments: Second Life

Meets Moodle. UPGRADE The European Journal for

the Informatics Professional. IX (3). pp. 8–14.

Marchiori, E. J., Torrente, J., Del Blanco, Á., Moreno-Ger,

P., Sancho, P. & Fernández-Manjón, B. (2012). A

narrative metaphor to facilitate educational game

authoring. Computers & Education. 58 (1). pp. 590–

599.

AuthoringStoryline-basedAdaptive3DVirtualLearningEnvironments

Moodle (2002). Moodle. Available from:

https://moodle.org/. Accessed: October 2012.

Murdock, K.. (2003). 3DS max 5 bible. Wiley Publishing.

OpenQwaq (2007). OpenQwaq. Available from:

http://code.google.com/p/openqwaq/. Accessed:

October 2012

Paramythis, A. & Loidl-Reisinger, S. (2003). Adaptive

Learning Environments and e-Learning Standards.

Electonic journal of ELearning. 2 (1). pp. 181–194.

Salen, K. & Zimmerman, E. (2003). Rules of Play: Game

Design Fundamentals. K. Salen & E. Zimmerman

(eds.). MIT Press.

SecondLife (2003). secondLife. Available from:

http://secondlife.com. Accessed: October 2012.

Torrente, J., Del Blanco, Á., Cañizal, G., Moreno-Ger, P.

& Fernández-Manjón, B. (2008). An Open Source

Authoring Environment for 3D Adventure Games in

Education. In: International Conference on Advances

in Computer Entertainment Technology ACE 2008,

ACM Press, pp. 191–194.

Wade, V. (2009). Challenges for the Multi-dimensional

Personalised Web. User Modeling Adaptation and

Personalization. pp. 3-3.

CSEDU2013-5thInternationalConferenceonComputerSupportedEducation