SIMULATING CHARACTERS FOR OBSERVATION

A. Albin-Clark

, T. L. J. Howard

and B. Anderson

School of Computer Science, The University of Manchester, Manchester, U.K.

Faculty of Education, Liverpool Hope University, Liverpool, U.K.

Keywords:

Education, Simulation, Virtual Human.

Abstract:

Observation of young children is common in educational settings but student practitioners have infrequent ac-

cess so limited amounts of observable activity can ever be captured. We have developed a software application

using research-based models of child development to support observations made by pedagogical practitioners.

The prototype system – Observation – which we believe is the ﬁrst to employ 3D interactive computer graphics

for visualising early childhood play, is available for download and evaluation.

1 INTRODUCTION

Observation of young children is common in early

childhood settings to plan and assess activities ac-

cording to the developmental needs of the individ-

ual child. It is difﬁcult for an observer to be a dis-

tant onlooker because children expect adults present

to supervise and help them. Making notes, record-

ing a video, or taking photographs, can also have an

impact on the way children play and behave. Fixed

cameras can only capture activity within one area and

may need editing to remove periods of inactivity but

an edited recording is not a true representation of

events. Hand-held cameras are more ﬂexible but are

intrusive. An observer may miss something interest-

ing. It is also quite hard to walk and ﬁlm at the same

time (Bruce and Meggitt, 2006). For student practi-

tioners, access to children is infrequent, and observ-

able activity may be limited, so observation sched-

ules are used with textual case studies, perhaps with

supporting images and/or video. Early work demon-

strated that the use of sound/colour ﬁlm modules,

combined with questions posed from a computer ter-

minal, increased the observational abilities of the stu-

dents (Durrett and Richards, 1976). However, there

has been limited work attempting to simulate the ed-

ucational environments of young children. Support-

ing the decision-making process of the adult is usually

the underlying theme. The activities are often adult-

initiated, more typical with children after pre-school,

with static 2D graphics of learners. These systems

do not show any activities occurring; it is left to the

user’s imagination.

2 OUR APPROACH

We have developed a cross-platform sandbox applica-

tion, using C++ and OpenGL, to support observations

made by pedagogical practitioners (Figure 1).

Figure 1: Characters engaged in blockplay on their patch.

A mixture of graphical and textual information

is provided for the observer, via an intuitive inter-

face. The observer is not represented as a 3D charac-

ter themselves, but viewing is designed to be through

the eyes of the observer, including: walking; ﬂying;

and following a character. New characters are ran-

domly generated or deﬁned manually along three di-

mensions: blockplay stage (Wellhousen and Kieff,

2000); social play stage (Parten, 1932); and egocen-

trism (Kesselring and M

uller, 2011). Character ap-

501

Albin-Clark A., L. J. Howard T. and Anderson B..

SIMULATING CHARACTERS FOR OBSERVATION.

DOI: 10.5220/0003838805010502

In Proceedings of the International Conference on Computer Graphics Theory and Applications (GRAPP-2012), pages 501-502

ISBN: 978-989-8565-02-0

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

pearance is deliberately abstract, a physical manifes-

tation of engagement in play, rather than ﬁgurative.

Clicking somewhere in the scene adds the character

at that position and allocates a colour-coded patch (1

metre by 1 metre) provided it is not: outside the ter-

rain; on bumpy terrain; on another patch; or on areas

with objects. Once added, a character may relocate

their patch (at any time) if the location is not suitable

for them, based upon their sociability. Each charac-

ter has their own inventory to hold a database of cur-

rently owned objects. If displayed, it shows each type

of object and its respective quantity. To keep the ter-

rain free from clutter, in addition to returning unused

items to their inventory, characters periodically claim

objects within their patch and place them in their in-

ventory, removing them from view. Characters may

join a group; in some groups, individual inventories

may be shared. The well-being for a character is a

simpliﬁed representation of the ‘Leuven scale’ used in

the process-oriented child monitoring system (Laev-

ers et al., 2002), and uses the same ﬁve-points: ex-

tremely high; high; moderate; low; extremely low.

The well-being can be determined visually by looking

at their vest, where the tone ranges from white (high-

est), to black (lowest) using a continuous represen-

tation of the scale (Figure 1). Characters pathﬁnd to

obtain the shortest route through the world and around

obstacles using our implementation of the A* search

algorithm, and local steering helps characters avoid

one other. Skeletal animations conform to the Bio-

vision Hierarchy motion capture standard. Charac-

ters are selected within the scene to display static and

dynamic data in the character information pane. A

timestamped event log shows both archived and real-

time events, ﬁltered on characters, if needed. The ter-

rain is interactively edited (height, textures) to create

different types of landscape. Characters follow the

terrain, even while it is being edited, and they will

avoid steep areas. The corners of a character’s patch

are locked and cannot be moved. If a character relo-

cates patch, those corners are unlocked. Objects are

custom shapes deﬁned in terms of other primitives,

or 3D models; these are dropped into the scene and

interactively manipulated. The physics library Bullet

has been integrated to give added realism to objects

in motion and at rest. Various visualisation options

expose the underlying functionality of the simulation.

3 USER EXPERIENCE

A pilot evaluation with student practitioners and in-

structors in early childhood studies at Liverpool Hope

University was promising. The student practitioners

felt it was helpful: to have an inﬂuence on the well-

being of the characters; to be present among the char-

acters; to adjust their viewing position; and to deﬁne

the behaviour of the characters. Some students re-

ported increased understanding in the stages of block-

play. Several were enthusiastic about changing the

environment by modifying the terrain and applying

textures, such as ﬂooring. They found setting the

scene to be useful because this is something that must

be done in real life; “A practitioner provides things,”

said one of them. Another student said, “Observation

would be perfect for developing observational skills

in childcare courses, for those lacking in conﬁdence

to go into settings and just do it.” Larger scale trials

are planned.

4 CONCLUSIONS

The pilot evaluation was encouraging and suggests

that computer graphics simulation of young children

has value for student practitioners. However, we do

recognise some limitations with our work. The well-

being representation is very much a simpliﬁcation,

and does not take into account indicators such as body

posture and facial expression, which are used in the

Leuven scale. In addition, the uniform appearance

and animations of the characters imply no speciﬁc

age, which may be too open-ended when considering

the development of the child. These, and other is-

sues, are being addressed in our continuing research.

Observation is available for download and evaluation

from http://www.cs.man.ac.uk/∼aac/observation/.

REFERENCES

Bruce, T. and Meggitt, C. (2006). Child Care and Educa-

tion. Hodder Arnold, fourth edition.

Durrett, M. E. and Richards, P. (1976). Computer-Based

Simulation for Observing Child Behavior. Home Eco-

nomics Research Journal, 5(2):88–91.

Kesselring, T. and M

uller, U. (2011). The concept of ego-

centrism in the context of Piaget’s theory. New Ideas

in Psychology, 29(3):327–345.

Laevers, F., Vandenbussche, E., Kog, M. and Depondt, L.

(2002). A Process-Oriented Child Monitoring System

for Young Children. Experiential Education. Centre

for Experiential Education.

Parten, M. B. (1932). Social participation among pre-school

children. The Journal of Abnormal & Social Psychol-

ogy, 27(3):243–269.

Wellhousen, K. and Kieff, J. (2000). A Constructivist Ap-

proach to Block Play in Early Childhood. Wadsworth

Publishing, ﬁrst edition.

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

502