ADAPTIVITY IN 3D VIRTUAL ENVIRONMENTS

FOR MULTI-USERS AND ITS APPLICATION

IN ADULT BASIC EDUCATION

Tassawar Iqbal

Institute of Software Technology and Interactive Systems, Vienna University of Technology, Vienna, Austria

COMSATS, Institute of Information Technology Abbottabad, Abbottabad, Pakistan

Klaus Hammermüller

Talkademy, An Open Learning (OLE) Language School, Vienna, Austria

A Min Tjoa

Institute of Software Technology and Interactive Systems, Vienna University of Technology, Vienna, Austria

Keywords: Three Dimensional (3D) Virtual Environments (VEs), Multi-users Virtual Environments (MUVEs),

Adaptive Systems, Adult Basic Education (ABE), Virtual Learning Environments, e-Learning.

Abstract: e-Learning applications have started to exploit web-based three dimensional (3D) Virtual Environments

(VEs) to augment learning experiences. Moreover adaptive support in these 3D VEs further endow potential

to empower the learner’s experiences and enhance effectiveness of learning process by presenting

personalized content and navigational support. Adaptivity in these 3D VEs however, sets the new

challenges for the researchers. Most of the techniques that are devised for adaptive content presentation and

navigation in 3D spaces are focused on single-user environment. Surprisingly few research efforts have

been yet dispensed for the adaptivity in multi-users VEs, however based on textual VEs. In this paper, we

present an adaptive approach for multi-users in 3D VEs. We customize the Adaptive Web 3D (AWE3D)

architecture to provide the adaptive support in 3D VEs for multi-users. We design an adaptive learning

scenario for Adult Basic Education (ABE) in 3D VE of OpenSim to explain, the proposed methodology.

1 INTRODUCTION

In the present era, e-learning applications have

started to exploit 3D VEs to further augment

learning experiences by providing sense of self-

presence, social-presence, immersive environment,

situated-learning and learning by doing. Educational

institutions use these modern technologies and

present learning scenarios for distance and blended

learning (Molka-Daneilsen, 2009). Adaptive support

in these 3D web-based VEs further empower the

learner’s experiences, enhance effectiveness of

learning process and interface usability (Chittaro and

Ranon, 2007a). Adaptivity in 3D VEs however, sets

the new challenges for the researchers. Most of the

techniques that are devised for adaptive content

presentation and navigation in 2D Web sites are not

applicable to build 3D adaptive Web sites because of

different technical requirements of these 3D VEs

(Chittaro and Ranon, 2007b).

Some research studies have presented the

approaches for personalized content presentation,

navigational support and interaction (Chittaro and

Ranon, 2002a; Nussbaumer et al., 2009; Santos and

Osorio, 2004) in 3D VEs. All of these efforts

however focused on single-user VEs. It is surprising

that few research efforts have been yet dispensed for

the issue of adaptivity in VEs for multi-users

(Dieberger, 1996). These efforts are focused on

textual virtual environments where text is used to

describe an imagined landscape.

213

Iqbal T., Hammermüller K. and Tjoa A..

ADAPTIVITY IN 3D VIRTUAL ENVIRONMENTS FOR MULTI-USERS AND ITS APPLICATION IN ADULT BASIC EDUCATION.

DOI: 10.5220/0003305802130218

In Proceedings of the 3rd International Conference on Computer Supported Education (CSEDU-2011), pages 213-218

ISBN: 978-989-8425-49-2

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

In this paper, we present an adaptive approach

for multi-users in 3D VEs. Secondly we customize

and exploit the Adaptive Web3D (AWE3D)

architecture (Chittaro and Ranon, 2007b) to provide

the adaptivity for multi-users in 3D VE. Finally we

present an adaptive scenario for Adult Basic

Education (ABE) designed in the OpenSim

(OpenSim, 2010) environment to explain how the

customized version of AWE3D architecture is

exploited in education domain.

2 RELATED WORK

Surprisingly, few efforts have been made for

adaptivity in 3D VEs. The focal point of these

studies has been adaptive content presentation and

navigational support in 3D VEs. Santos and Osorio

presented an adaptive VE for distance learning that

performs content insertion and removal in 3D space

and navigational support using intelligent agent

(Santos and Osorio, 2004). Apart from educational

domain, lot of work is witnessed on business side.

Virtual reality interfaces to e-commerce sites are

getting more common (Chittaro and Ranon, 2002b).

These e-commerce 3D environments are further

investigated for adaptive support to facilitate the

shopping experience by providing most relevant

product in stores and navigational support on the

basis of preferences and interests of buyers (Chittaro

and Ranon, 2007b). In 2002, Chittaro and Ranon

proposed a generic architecture for adaptive 3D Web

sites named Adaptive Web 3D (AWE3D) that is

responsible for generating and delivering adaptive

VRML content (Chittaro and Ranon, 2002a). This

client-server AWE3D architecture is exploited for e-

commerce case studies and favorable user’s

responses to the system were observed (Chittaro and

Ranon, 2002a). Furthermore, adaptivity using

competence-based Knowledge Space Theory is

presented by Nussbaumer in 3D space of Second

Life (SL) (Nussbaumer et al., 2009). All these

research studies on adaptivity in 3D VEs are focused

towards single-user environment.

For adaptivity in multi-user environments, an

approach is presented in (Dieberger, 1996).

However this approach is developed for textual VEs

(environments base on imagined landscape where

locations, objects, users and their interactions are

described solely through text instead of metaphors).

This approach exploits the concept of ‘read wears’

to provide adaptive support in multi-user textual VE,

where description of objects is changed in imagined

landscape according to frequency of their usage

(e.g., usage frequency of room exits and posting on

bulletin boards).

3 CHALLENGES

OF ADAPTIVITY IN 3D

SPACES

3D environments are more complex than 2D web

sites due to differences in presentation space,

content presentation, organization, navigational

support and users’ actions (Chittaro and Ranon,

2007b). For these reasons, techniques and tools used

for adaptivity in 2D environments cannot be directly

applied for content presentation and navigation in

3D spaces (Chittaro and Ranon, 2007b).

In 2D spaces, optional-fragments and altering-

fragment techniques are common for content

presentation (Chittaro and Ranon, 2007b), whereas

hyperlinks are used to navigate across web pages

(Hughes, Brusilovsky and Lewis, 2002). For content

presentation in 3D space, same techniques are used

where metaphors instead of text (as in 2D) are

selected and 3D spaces are created; however special

care is required to avoid overlapping of metaphors

(Chittaro and Ranon, 2007b). Whereas for

navigational support in 3D space, intelligent agents

(Chittaro, Ieronutti and Ranon, 2004; Chittaro and

Ranon, 2000) and patterns of interaction from usage

data (Celentano and Pittarelo, 2004) are known

approaches instead of simple hyperlinks (as in 2D

space).

All adaptive efforts in 3D VEs focused on single-

user; unfortunately no application is reported in

literature that considers adaptivity in 3D spaces for

multi-users (Chittaro and Ranon, 2007b). In single-

user 3D environments, it is relatively trivial to

provide personalized content and navigational path

according to the user’s needs because there is no

concept of space sharing. On the contrary, adaptivity

in 3D space for multi-users is a complex job. In

multi-users 3D VEs, as the same 3D space is shared

by more than one user simultaneously, adaptive

content presentation for all users in the same 3D

space raises conflicts (like overlapping of contents)

in content presentation that leads towards new

challenges. Similarly navigational support cannot be

customized for an individual user because other

users are sharing the same 3D space and

demarcation in navigational cues for each individual

user is a difficult job (Chittaro and Ranon, 2007b).

CSEDU 2011 - 3rd International Conference on Computer Supported Education

214

4 METHODOLOGY

FOR ADAPTIVITY IN 3D

MUVE

In 3D Multi-Users Virtual Environment (MUVE) as

more than one user share the same 3D space and it

causes conflicts in adaptive content presentation, and

navigation as discussed in Section 3. We propose an

approach that reduces the conflicts by optimizing the

best common adaptation strategy, where a group of

users (or learners) able to share a portion of 3D

space, customized according to their preferences. In

order to optimize best common adaption strategy,

we divided the 3D space in multiple zones (called

learning- zones for application of ABE) based on

language, domain, level (Beginner or moderate), and

difficulty level in each beginner or moderate levels.

The Figure 1 describes division strategy of 3D

space, where 3D space is divided into sixteen

different zones. Two centric circles set the difficulty

level in 3D space, the outer one is level one (lower

difficulty level) and inner one is level two (higher

difficulty level). In the circular shaped 3D learning

space, the left half of circle is for English language

and the right half of the circle is for Urdu language.

Similarly the upper half of the circle represents a

domain D1 and the lower half of the circular space

forms domain D2. Two lines at diagonal positions

from the upper left quadrate to the lower right

quadrate and from the upper right quadrate to the

lower left quadrate divides 3D learning space for

alphabet learning (for beginners) and vocabulary

(for moderates) learning with respect to language

domain and difficulty level. Each unit of this 3D

learning space is called Learning-Zone as shown in

the Figure 1. Each learning-Zone consists of stage

setting proposed in the Section 6.

Figure 1: Division Plan of 3D Space.

For adaptivity in 3D space for multi-users, we

propose an algorithm. This algorithm exploits the

proposed enhanced AWE3D architecture. When a

learner logins in the 3D space for learning, the

system reads his/her profile. On client-side Linden

Scripting Language (LSL) script initiate a request

for contents (for each learner when he/she logins)

and forward it to the server-side, where the

Personalization module reads the learner profile and

accordingly formulate a query and determines

through the content ontology, what the learner needs

to learn and forward these results (list of LOs) back

to client-side. The OpenSim content Presenter on

client-side finds whether these LOs already available

in some Learning-Zone, if yes, the adaptive system

teleport the learner to the target Learning-Zone, in

case of no, it creates new Learning-Zone with

desired LOs and teleport the learner to the newly

created Learning-Zone, where personalized contents

are presented to the learners and learners’ profiles

are updated accordingly. However if a learner during

a session learns all LOs and more Learning-Zone are

still unlearned, in this case adaptive request within a

session initiated through LSL, forwarded to server-

side application and in response receives description

of imminent LOs, if these LOs already available in

some Learning-Zone then the leaner is teleported to

that zone otherwise new Learning-Zone is created

with selected LOs and learner is teleported there.

The adaptive system also allows learners to leave the

session anytime, in this case their profiles are

updated by the system and they can logout the

system.

The proposed approach is able to assign names at

run time to the different partitioned spaces

(Learning-Zones) and to select relevant Learning-

Zone according to user profile; however static

division may reduce the performance lags.

The selected 3D VE of OpenSim is capable to

fulfill all the technical requirements to develop the

proposed adaptive system. A region in the OpenSim

version 0.6.3.873 allows 45000 prims (OpenSim,

2010). A prim is short version of primitive, means

3D shape. The proposed method exploits less than

hundred prims in all 16 regions that is much lesser

than allowed limit. This consideration of limited

prims avoids performance lags in selected 3D spae

of OpenSim. The reason for the selection of

OpenSim is its support for 1) content scripting, 2)

editable avatars, 3) available for educational

purposes or easily modifiable, 4) capable to run as

standalone application or as collaborative world, 5)

in-world 3D content creation etc (de Freitas, 2008),

and also the proposed methodology is equally

applicable to other 3D spaces offer the same support.

ADAPTIVITY IN 3D VIRTUAL ENVIRONMENTS FOR MULTI-USERS AND ITS APPLICATION IN ADULT BASIC

EDUCATION

215

5 ENHANCED ADAPTIVE WEB

3D (AWE3D) ARCHITECTURE

The architecture we proposed for adaptivity in 3D

VEs for multi-users is based on the Adaptive Web

3D (AWE3D) (Chittaro and Ranon, 2002a). We

customize and use this AWE3D architecture to

provide adaptive support in 3D VEs for multi-users

(see Figure 2). The customized AWE3D consist of

following eight modules;



Figure 2: Customized AWE3D for multi-users adaptive

Support in 3D VE.

Usage Data Sensing Module - This is a client-

based module and it is responsible to collect data

about the users’ interaction in the OpenSim and

sends the collected data through HTTP protocol to

the server-side. This module uses built-in functions

of Linden Scripting Language (LSL) for sensing the

user’s interaction in the OpenSim. This users’

interaction data includes information about user’s

region in which his/her avatar is teleported, current

position of user’s avatar, walking-status of user’s

avatar, flying-status of user’s avatar, sitting-status of

user’s avatar, away-status, touches to metaphors in

3D space and distance from learning object (LO).

The adaptive system collects this usage data using

sensors designed in LSL. Script for sending this data

is also written in LSL.

Usage Data Recorder Module - This module is

responsible for receiving data sent by the Usage

Data Sensing module from OpenSim and stores it

into the Buffer. It is located on the server-side. For

its implementation, we use PHP scripts to deal with

the usage data sent by client-side. This module

receive HTTP request with POST method from the

client-side sending script and it first stores the data

into a buffer and finally transfers and saves it into

the User Model database.

Buffer - The Buffer is a part of server-side

application and its purpose is to reduce the number

of accesses to relational database. It temporarily

stores all the data into a text file, sent by the Usage

Data Recorder module and when stored data exceeds

the set limit of the buffer, it transfers into the User

Model.

Personalization Module - The Personalization

Module is a core component, located on server-side.

It interacts with three components of the architecture

1) User Model, 2) Content Ontology and 3)

OpenSim Content Presenter. It provides session-

based adaptive support. For personalizing the

contents for the users in 3D space, it first takes the

user’s data and usage data from the User Model,

formulate a SPARQL query from this information

and forward it to the Content Ontology. The

Personalization module receives results against the

query. These results (include text-based description

of LOs) are sent back to the OpenSim Content

Presenter Module, against the HTTP request made

by client-side application. The Personalization

module is written in PHP that integrate ontology

model, mySQL database and SPARQL query

execution.

For adaptive navigation support, the

Personalization module finds the location where

relevant contents for the user are available in the 3D

space and teleport the user to that particular location

in 3D space. This personalized navigational support

also addresses the issue of usability (inadequate user

assistance in 3D space) at the cost of user’s control

in 3D space.

User Model - The User model keeps track of

users’ data and usage data sent to it. We use,

MySQL to develop the User Model database.

Information storage and retrieval in the user model

database is made possible by scripts written in PHP.

OpenSim Content Presenter

- This is a client-

based module receives the responses against HTTP

request from the Personalization module and present

the personalized content to the users in the

OpenSim. The OpenSim Content Presenter is

developed using LSL.

Metaphor Inventory -

The Metaphor Inventory

is located on client-side. It is responsible to store 3D

contents in the OpenSim environment. All the

metaphors required for the complete 3D space are

uploaded beforehand and stored in this inventory.

Content Ontology -

Domain ontology for content

serves as contents’ repository on server-side. This

ontological model defines contents of a course. It

consists of two main classes, Learner and Content.

Learner class provides information about the

learner’s domain (to whom he/she belongs) through

specialized subclass named Domain. Instance of the

subclass Domain of class Learner specifies learner’s



CSEDU 2011 - 3rd International Conference on Computer Supported Education

216

domain e.g., farming, office-worker, kitchenware

etc. The Content class defines contents and it consist

of four subclasses; Alphabet, Common-Vocabulary,

Numeracy and Domain-Specific-Vocabulary. The

Personalization module exploits this ontology for

contents selection according to the learners’ profiles

through SPARQL query and specified rules provide

results to the Personalization module.

6 AN APPLICATION IN ABE

In this section, we describe the reasons for the

selection of learning scenario for Adult Basic

Education (ABE) and use of 3D spaces. Then brief

introduction of learning scenario is presented and

finally personalization support for contents and

navigation is explained through this scenario.

6.1 Why ABE and Why 3D MUVE?

Due to bad experiences from schooling in past,

adults do not prefer to go to literacy programs

(Eberle and Robinson, 1980). Thus educationist and

technologist are striving to offer technology-based

literacy program to augment learning and to attract

these illiterates to the literacy (Hopey, 1998).

Unfortunately, main theme of most of techno-

literacy solutions is text-based, that is not suitable

for absolute illiterates. According to Visual Literacy,

these text-based learning and communication

applications are not acceptable in the present era of

visual literacy and there is need to provide image

dominant learning environments (Aanstoos, 2003).

For effective learning, Gardner’s theory of Multiple

Intelligences recommends to exploits multiple

intelligences in learning environments (Gardner,

1983). Furthermore, features such as sense of self-

presence, social presence, embodied environment,

learning by doing, situated learning, and

collaborative learning, recommended in learning

theories (Dale, 1969; Koh et al. 2007; Kolb, 1984;

Lave and Wenger, 1990) and considered important

for learning are achievable by exploiting 3D

MUVEs.

6.2 Scenario for ABE in 3D Space

We design learning scenarios in the OpenSim for

ABE. In this 3D learning space each learner is

represented by an avatar (an embodied object of a

learner) that can walk, fly, sit, speak and chat under

the control of learner. The learning scenarios are

intended to educate the illiterate learners (target

users of the proposed environment) and these

scenarios are categorized into 1) linguistic scenario,

2) numeracy scenario and 3) game-based scenario.

This paper focuses on the linguistic learning and

game-based evaluation scenarios in order to validate

the multi-users adaptive support in 3D space using

the customized AWE3D architecture. The proposed

scenario serves for learning and evaluation

simultaneously, where a learner may learn

alphabetic characters and commonly used domain

specific words in 3D space with adaptive support of

content presentation and navigation. These scenarios

offer learning material in two languages English and

Urdu (national language of Pakistan), where

alphabetic characters and words are offered for

learning in each language with different difficulty

levels (level I is easier one and level II is difficult

one). Furthermore selection of these learning

contents (alphabetic character/Word and their

metaphors) depends on the learner’s working

domain (scenarios presents different content for

different domains, for instance, house worker,

farmer etc). In 3D space, a stage setting with five 3D

objects presents the learning contents to the learners

as shown in the Figure 1. The larger object on the

left displays alphabetic character/word (written text)

where as four smaller objects on the right present

metaphors (against the alphabetic character/word on

left object). Out of these four metaphoric answers,

only one is a correct answer. Learner is required to

select one metaphoric answer through mouse click.

The adaptive system evaluates whether the learner’s

response is correct or incorrect and update learner’s

profile accordingly.



6.3 Personalized Content Presentation

and Navigation Support in 3D

Learning Space

User data (both usage and user data of learner)

serves as basis for adaptivity in 3D space. For

adaptive content presentation, we consider four

parameters; these are language, domain, learner’s

level (moderate or beginners), and LO last-learned.

These parameters are used to formulate the query to

find out the forthcoming contents for learning. The

Personalization module formulates this query to

personalize content that is executed on ontology.

The results obtained from this query consist of a list

of forthcoming personalized contents for a group of

learners. These results consist of textual information

that is forwarded to the OpenSim Content Presenter

Module to present 3D content in the space.

ADAPTIVITY IN 3D VIRTUAL ENVIRONMENTS FOR MULTI-USERS AND ITS APPLICATION IN ADULT BASIC

EDUCATION

217

Adaptive navigation support is achieved through

the teleporting facility offered by the proposed

system, where learner’s avatar is teleported at

desired location in the 3D space after the

recommendation of user model.

7 CONCLUSIONS

In this paper, we presented an approach that

describes how adaptivity in 3D VE for multi-users is

realized. A customization in the AWE3D

architecture is proposed to achieve adaptivity in

multi-users VEs. We presented application of

proposed approach in 3D learning scenarios for ABE

designed in 3D VE of OpenSim. Using presented

learning scenarios, we described the implementation

strategy of multi-user adaptivity in 3D spaces.

Future work will focus on investigation of

learners’ behavioral data in collaborative learning

scenarios.

REFERENCES

Aanstoos, J. 2003. Visual literacy: An Overview, Proc. of

32nd Applied Imagery Pattern Recognition Workshop

(AIPR'03), pp.189-193.

Celentano, A., Pittarello F. 2004. Observing and Adapting

User Behavior in Navigational 3D Interfaces. Proc. of

7th International Conf. on Advanced Visual Interfaces,

Gallipoli, Italy, ACM Press, pp. 275-282.

Chittaro, L., Ieronutti, L., Ranon, R. 2004. Navigating 3D

Virtual Environments by Following Embodied gents: a

Proposal and its Informal Evaluation on a Virtual

Museum Application, PsychNology Journal, Vol. 2,

No. 1, pp.24-42.

Chittaro, L., Ranon, R. 2000. Adding Adaptive Features to

Virtual Reality Interfaces for Ecommerce, Proc. of the

1st International Conf. on Adaptive Hypermedia and

Adaptive Web-based Systems, Trento, Italy, Springer-

Verlag, Berlin Heidelberg New York, pp.86-97.

Chittaro, L., Ranon, R. 2002a. Dynamic generation of

personalized VRML content: a general approach and

its application to 3D E-commerce, Proc. of 7

International Conf. on 3D Web Technology, Tempe,

Arizona, USA ACM Press, New York, pp.145-154.

Chittaro, L., Ranon, R. 2002b. New Directions for the

Design of Virtual Reality Interfaces to E-Commerce

Sites, Proc. of the 5th International Conf. on

Advanced Visual Interfaces, Trento, Italy, ACM Press,

New York, pp. 308-315.

Chittaro, L., Ranon, R. 2007a. Adaptive Hypermedia

Techniques for 3D Educational Virtual Environments,

IEEE Intelligent Systems, Vol. 22, No. 4, pp. 31-37.

Chittaro, L., Ranon, R. 2007b. Adaptive 3D Web Sites, In

Brusilovsky, P.,Kobsa, A., Nejdl, W. (eds.), The

Adaptive Web - Methods and Strategies of Web

Personalization, Lecture Notes in Computer Science

Vol. 4321, Springer-Verlag, Berlin, pp. 433-464.

de Freitas, S. (2008). Serious Virtual Worlds: a Scoping

Study, Prepared for the JISC e-Learning Program,

http://www.jisc.ac.uk/media/documents/publications/s

eriousvirtualworldsv1.pdf.

Dieberger, A. 1996. Browsing the WWW by interacting

with a textual virtual environment - A framework for

experimenting with navigational metaphors, Proc. of

the 7th Conf. on Hypertext, Washington DC, USA,

ACM Press, pp. 170-179.

Eberle, A., Robinson, S. 1980. The adult illiterate speaks

out: personal perspectives on learning to read and

write, Washington, D.C. National Institute of

Education, U.S. Dept. of Education.

Gardner, H. 1983. Frames of mind: A theory of multiple

intelligences, New York, Basic Books.

Hopey, C. E. 1998. Making Technology Happen in Adult

Education, in: Hopey, C. E., Technology, Basic Skills,

and Adult Education: Getting Ready and Moving

Forward, Ohio, ERIC Clearinghouse on Adult,

Career, and Vocational Education, Ohio State

University. pp. 3-23.

Hughes, S., Brusilovsky, P., Lewis, M. 2002. Adaptive

navigation support in 3D ecommerce activities. Proc.

of Workshop on Recommendation and Personalization

in eCommerce at AH, pp.132-139.

Koh, J., Young-Gul, K., Butler, B. & Bock, G-W. 2007.

Encouraging Participation in Virtual Communities,

Communications of the ACM, Vol. 50, No. 2, pp. 69-

73.

Kolb, David A. 1984. Experiential Learning: Experience

as the Source of Learning and Development, Prentice-

Hall, Inc., Englewood Cliffs, N.J.

Lave, J., & Wenger, E. 1990. Situated Learning:

Legitimate Peripheral Participation, Cambridge, UK,

Cambridge University Press.

Molka-Danielsen, E. J., Deutschmann, M. 2009. Learning

and Teaching in the Virtual World of Second Life.

Tapir Academic Press.

Nussbaumer, A., Gütl, C., Albert, D., & Helic, D. 2009.

Competence-based adaptation of learning

environments in 3D space. Proc. of the 4

International Conf. on Interactive Mobile and

Computer Aided Learning, Amman, Jordan, pp.103-

108.

OpenSim. 2010. Open Source 3D Virtual Environemtn,

http://opensimulator.org/wiki/Main_Page, Last

Accessed August 23, 2010.

Santos, C. T., Osório, F. S. 2004. An intelligent and

adaptive virtual environment and its application in

distance learning. Proc. of the Working Conf. on

Advanced Visual interfaces, Gallipoli, Italy, ACM

Press, New York, pp.362-365.

CSEDU 2011 - 3rd International Conference on Computer Supported Education

218