ROBOT BASED INTERACTIVE GAME FOR TEACHING ARABIC SPELLING

International Journal of Artificial Intelligence & Applications (IJAIA) Vol.10, No.6, November 2019
DOI: 10.5121/ijaia.2019.10602 15
ROBOT BASED INTERACTIVE GAME FOR
TEACHING ARABIC SPELLING
Ghada Alsebayel and Jawad Berri
Department of Information Systems, College of Computer and Information Sciences
King Saud University, Riyadh, Saudi Arabia
ABSTRACT
Game based learning is becoming a widespread technique used to enhance motivation, involvement and
educational experience of learners. Games have the potential to support educational curricula when
designed effectively. In this work, an educational game to teach Arabic spelling to children is proposed.
The game consists of two main parts; a robot and a desktop application. The robot is connected to the
desktop application to form the complete game. Our mere focus is to develop an interactive, adaptive game
to motivate students and let them interact joyfully in their environment while learning simple Arabic
spelling rules. The interaction was implemented through designing an interaction model between the user
and the robot, where the robot responds to user input with appropriate facial expressions and vocal
statements. On the other hand, adaption and intelligence of the game is done through utilizing the nutshell
of expert systems’ framework with some alterations. Our proposed game is based on the curriculum of
Saudi Arabia in elementary schools. It is anticipated that the deployment of robot-based games in the
classroom will advance students’ engagement and enthusiasm about learning Arabic spelling.
KEYWORDS
Game based learning, Arabic spelling, Arduino.
1. INTRODUCTION
Technology aided education means the application of technology to boost up knowledge
acquisition and retention among students. Nowadays, we need technology presence in each
classroom because it is the pen and paper of our time. Before exploring how to utilize such
techniques, it is worthy to mention that the development of any language skills is composed of
four gears; reading, writing, listening and speaking. The motivation of this research is to develop
a system that focuses on enhancing writing language skills for Arab children, specifically
spelling.
Arabic is a Semitic language that has many unique properties; it is very rich in vocabulary and
has a rich phonology system. The morphological pattern system is unique where words are
derived from triliteral roots. Words can always be traced back to their roots, which is then
transferred into all different patterns (awzaan) to give different shades of meanings from the
original meaning. Many words are syntactic. The diacritic marks in some cases control the
syntactic position of the word.[1]
In recent years, some studies noted that Arabic writing skills of students in different educational
levels are degrading. According to a field study with a sample of 300 students in Riyadh by
AlShaalan [2] designed to measure the level of linguistic achievements of students in the sixth
grade and the first year of intermediate school, it was found that there is an undeniable weakness

16
in spelling skills among students. Furthermore, based on personal observations, many users of
social media are making spelling and grammar mistakes when they express their thoughts through
writing in Arabic. Another image of abusing the language, is when college level students make
spelling or grammar mistakes in their written exams. Educators and language specialists are
worried about this noticeable phenomenon that requires some corrective measures to be
undertaken.
This research is a contribution to such corrective measures to enhance Arabic writing skills
among Arab children speakers with the assistance of technology. As technology is being
integrated more and more with every aspect of our lives, many technological tools and techniques
are employed in the educational process to make learning more efficient. One particularly
interesting method of teaching that could benefit from technology is; learning by playing. Playing
is often defined as an activity done for one’s own sake, characterized by means rather than ends.
As children invest time and energy in play, there are opportunities for learning when they do play
[3]. By assimilating the concept of ‘learn by play’ with technology, we aim to develop a fun,
interactive, adaptive game for teaching Arabic spelling. The nature of this research paradigm is
multidisciplinary, investigating many research areas such as; human computer interaction, user
experience, persuasion, gaming, artificial intelligence, robotics and education. The basic concept
is to build a robot based, interactive, adaptive game for teaching Arabic spelling. Robots have an
amazing ability to create instructional method motivating students, encouraging innovation,
creating more powerful interaction models, adding a joyful experience in a classroom, inspiring
students to think outside the box, and finally preparing students for future trends in IT.
The contribution of this research is to design and implement an interactive and adaptive
educational game to teach Arabic spelling. By designing the system to be adaptive, it will propose
questions relating to specific Arabic spelling rules based on the detected weakness a child might
have. The scope of this is based on the curriculum of Saudi Arabia for primary grades.
2. BACKGROUND AND LITERATURE REVIEW
Educational games are designed to help people understand concepts, learn domain knowledge,
and develop problem solving skills as they play games [4]. While educational games can be
designed to aid users in different age groups learn any field, this work is mainly focused on
educational games for young children. It is a noticeable phenomenon that children spend a lot of
time playing, especially computer-based games. If that time can be exploited for teaching them
some concepts, it is anticipated that their learning outcomes will increase drastically. However,
there is limited direct evidence of the effect of computer-based games on classroom learning [5].
In an attempt to bridge that gap, a randomized controlled trial by David J Miller and Derek P
Robertson, was conducted on a population of 634 primary school students ages between 10-11
from thirty-two schools across Scotland [5]. The sample was divided into two sub-groups;
controlled and experimental. Students in experimental schools were permitted to use game
console for 20 minutes each day running a brain training game while the controlled sample
continued their normal routine. The trail lasted 9 weeks. By the end of the trial, the following
benefits were found: i) Both groups have improved in terms of computational speed and accuracy,
however, the rate of improvement in the controlled group was significantly faster, and ii) There
was a minor but statistically substantial improvement in the attitude towards school among
students in the controlled group. Meaning that students in the controlled group were more
enthusiastic about attending school [5].These two findings shape the motivation of this research.
Even though, the previously mentioned study was directed to evaluate the computation skills of

17
students after using computer-based gaming, other evidence was found in the literature supporting
the same outcomes when examining language skills. A study by Emilija Stojmenova Duh and
Natasa Koceska proposed an application (Azbuka) developed to teach children how to write
Cyrillic letters. The research investigated many design challenges as well as an evaluation of the
impact of the newly developed tool on acquiring the skill of writing Cyrillic letters among the
sample understudy. It was found that the average time required to write a Cyrillic letter after
using the application was decreased when compared to the average time required to write a
Cyrillic letter using classical method [6].
2.1. Educational Games Design and Development Methodologies
Game design and game development are two terms that might be used interchangeably. However,
if we want to differentiate them, we would say that game design encompasses all the activities of
concept creation, storylines, mechanics…etc. While game development involves the steps of
making the game, animation, sounds, code…etc. When designing a game many challenges arise.
In addition to intense coding required for game engine, middleware, interface, sounds, animation
and game production, designers must also think about how to make the game engaging and
amusing. Such challenges are magnified when we talk about educational games since more
advanced design considerations are required as we try to combine both the educational factor and
the entertainment factor.
Despite the difficulties, many frameworks in the context of designing educational games were
proposed. According to Embi et al. [7], who proposed a theoretical Edutainment Environment
Framework, the difficulty for a child to learn a certain concept is composed of two elements;
motivation and psychological need. The motivation necessitates catching the child’s attention and
attracting him/her to spend time on some activity. While psychological need means triggering the
child’s cognitive ability to build the knowledge. Practices for game design includes utilization of
storytelling, animation, feedback and other game design methods when creating the educational
game by game designers. The educational game can then be used by educators to assist the
learning process either in collaborative or individual manner. Consequently, the “edutainment”
environment is created.
Another model for educational games development is the game object model GOM. GOM was
originally proposed by Amory et al. and Amory [8]. The model is loosely based on the object-
oriented concepts such as; polymorphism, encapsulation, inheritance, abstraction…etc.
Components of an educational game are usually related and dependent on each other, yelling
some complexity in the game design and development. As a solution, using the metaphor of
object orientation permits viewing the components of an educational game as correlated objects,
making the analysis and design easier.
Another contribution by Said [9] proposed the Engaging Multimedia Design Model for Children,
where he presented five design factors contributing to creating an engaging model for a child.
These factors include: i) Simulation interaction: permitting the child to react on behalf of
someone else like role playing, ii) Construction interaction: permitting a child to construct or
build something, iii) Immediacy: where every action by the child has an effect, and can be
reflected on the system, iv) Feedback: every action by the child has a reaction by the system, and
v) Goals: are obvious and are either set for the child or set by him/her when no external
motivation is present. In our implementation, we utilized immediacy and feedback in our design.
The following sections will elaborate on the proposed game design. The fact that educational
games have the potential to enhance learning outcomes, intrigues us to further explore the effect
of leveraging advanced technologies, such as; robotics as a part of the educational game
framework.

18
2.2. Robots in Classrooms
For fairly a long time now, robots have been known to serve in many domains such as the
industry, engineering, and rescue. However, the usage of robots has been extended to include
wider domains in recent years, including the education sector. The integration of robotics in the
educational system has opened-up many research areas as well as practical potentials. According
to Beran et al., children are now spending more time playing with advanced technological devices
during their free time [10]. As a result, studies were conducted to examine the influence of using
robotics in the development of cognition, language, interaction, social and moral skills of
children. In our review, we will be focusing on the literature addressing language skills
development using robots.
Based on Chang et al., robots can be used to create a fun, interactive, engaging experience for
children developing language skills [11]. This finding lines-up with the quantitative study
completed by Young et al. A sample of 68 young children in elementary school (Grade 3-4) were
under study; among them 6 students (2 boys and 4 girls) were selected as a focus group to
evaluate their attitude towards integrating tangible learning companions in learning English
conversation. The findings indicates that 95% of students had positive attitude towards tangible
learning technologies, which includes robots. It contributed positively to making students more
active and welling to engage in learning activities [12]. Another study by Sugimoto was applied
on 25 children in elementary school aging 11-12 years old, participants experimented story
creation with GENTORO robot and handheld projector. Result showed that children were highly
engaged and acted in a coordinated manner [13]. Additionally, Chen, Quadir and Teng conducted
anecdotal case studies on 5 elementary school students where they tested the impact of using a
robot, a computer and a book for teaching English language. They concluded that robots can
enhance the concentration in learning the language [14].
Commercially, Kubo is an application of educational robots for children. Kubo is a robot by
Tom’s Hardware Inc. aiming to teach young children basics of spelling and logic. Kubo is puzzle-
based teaching toy for children aged 3 years old and above. The packaging consists of a small
robot and card tags. The robot has two wheels to roll over the cards and read their content. The
idea behind Kubo is to utilize radio frequency identification RFID in the cards permitting the
child to assemble words then checking his/her spelling. Kubo has other features teaching logic
and programming, however, for the purpose of our research only its teaching capabilities for
languages were studied. Even though Kubo has potentials to accelerate language learning, its
interaction model remains limited. It does not incorporate speakers or headphones for vocal
communication. It does not provide extensive feedback to user input indicating whether user
input is correct or not due to the fact its targeted user group is too young.
2.2.1. Arduino for Robotics
Arduino is an open source platform that can be used for developing electronic projects. It consists
of physical circuit board that can be programmed and a software integrated development
environment IDE. The initial idea of this work was inspired by woodyrobo [15]. Woodyrobo is an
interactive game to teach spelling of simple English words. In the implementation, Arduino was
used with a mobile application. Even though the interactive aspect of the game can be enjoyable
and attracting to children, the system still lacks some key functionalities such as being expandable
and adaptive. Moreover, the intelligence aspect is completely neglected, it only supports 7 words.
We aimed to overcome these limitations in our implementation.

19
Next part of the literature review will be explicitly focusing on applications of different
technologies to implement educational games for teaching Arabic language and spelling
checking.
2.3. Educational Games for Teaching Arabic
Although the literature is very rich when it comes to game-based learning for teaching different
languages leveraging technology, there aren't many such learning tools for Arabic [16] specially
designated for young children.
Erradi et al. have published their experience designing a mobile multimedia supported platform to
teach Arabic spelling and vocabulary in an intuitive fun manner. The system was not designated
for young children. They took advantage of the touch capabilities in mobile and handheld devices
to make the game interactive [16]. Another paper was published by Belkhouche et al. [17],
addressing the design and implementation of an HTML5 based game to teach Arabic language
skills to young children. Firstly, they surveyed 100 university students about their outlook on
educational games, results showed that overall perception on educational games was positive;
educational games were found to be an important pillar of the educational curricula. In advanced
stages of their search, four major components contributing to game-based learning were
identified; content (curriculum), knowledge structure (concept mapping), learning stagey
(gameplay), and learning process (assessment). They analyzed primary school Arabic language
textbooks in UAE to generate the learning content. Conceptual mapping was then used to model
students’ cognitive structure, which Consists of the following levels: i) Level one: students are
taught the concept of alphabet, once they have mastered it they can move to level two, ii) Level
two: students are taught the concept of words, spelling and pronunciation, once they have
mastered it they move to level three, Level three: students are taught the concept of a simple
sentence. Our work extends their efforts to applying these principles to robot-based gaming.
2.4. Spelling Checkers as Knowledge Based Expert Systems
Expert systems are an application of artificial intelligence, the basic concept behind expert
systems is to mimic the abilities of a human expert in a specific field. Expert systems are
designed to have decision making abilities when fed by the required knowledge. Formally an
expert system is an intelligent computer program that uses knowledge and inference to solve
problems that are difficult enough to require significant human expertise to solve them. There are
three main components of an expert system; knowledge base, inference engine and user interface.
The knowledge base is a repository of specific knowledge and rules about a certain subject; it
holds objects associated with rules. Next component is the inference engine, which is the
processors of the system, leveraging knowledge of the knowledgebase to draw conclusion. There
are two types of inference engines: i) Deterministic inference engines: decisions by such
inference engines are finite, ii) Probabilistic inference engines: decisions by such inference
engines have uncertainty element associated with them [18]. One example found in the literature
for utilizing expert systems as spelling checkers is Hascheck. It was built on a dictionary of the
Croatian language. The initial dictionary counted less than 100,000 frequently used words.
Classification and tagging algorithms were then used on the words of the dictionary. The final
product was a spell checker open for all word processors and can analyze a passage of 100,000
words and response within one minute [19].

20
3. DESIGN METHODOLOGY
The utilization of technology and robotics has the potential to enhance the learning experience of
students. This research aims to leverage the advancement of technology to contribute to solve
weakness in Arabic spelling skills among Arab speaking youth. By integrating artificial expert
systems and robotics we aim to develop an interactive adaptive educational game. Throughout
this chapter, the implementation methodology of the proposed system will be presented.
The literature suggests that game-based learning is efficient in guiding young children to acquire
language skills. Furthermore, it advocated the integration of robotics applications in the
educational system. There exist in the literature some applications for using robotics as tools for
teaching language skills. However, they were not exploited to their full potentials in our the
Arabic context for the following reasons: i) Robotic tools presented in the literature were
experimental and mostly supporting English, ii) The adoption of technologically advanced tools
in classrooms can be costly, robots presented in the literature were relatively advanced, iii)
Robotics presented in the literature were not designated to teach Arabic spelling. Our proposed
model and system aim to produce an affordable robot-based game for teaching Arabic spelling to
bridge that gap.
There are many software development methodologies. The two most widely known are predictive
and adaptive. The following figure illustrates suitable software development life cycle based on
project characteristics. Most projects fall in a range between adaptive and productive SDLC [20].
In our implementation we followed an adaptive incremental approach. Hence, the game was be
developed in releases, where each release incorporates new functionalities added to the pervious
release. Each iteration was like a mini project. Within each iteration, several activities took place.
Including; planning, analysis, design, implementation, and deployment.
The reasons for this selection is because the technology that will be used is new and unfamiliar.
Moreover, incremental development methodology introduces many advantages including: i)
Generating working software as early as possible during software development life cycle SDLC,
ii) Being able to adopt to changes, change is not as costly as using predictive approach, iii)
Receiving feedback rapidly, which allows further enhancements to be made to each release, iv)
Prompting simplicity; simple solutions for complex situations. The game has been developed in
releases where the first release included a single lesson of Arabic spelling rules based on the
curricula of Saudi Arabia. Additional functionalities were added within each further system
releases.
4. SYSTEM DESIGN
Our proposed system includes two components: software and hardware. In the following section,
system analysis and design of both components is presented. Hardware design explains the
needed parts to build up the robot. While analyzing and designing the software aims to structure
the functional requirements that will be covered by the system, software design principles, and
use cases.
4.1. Game Scenario
The system’s hardware design was concerned with connecting the electrical parts then building a
shield to produce a functioning robot that could present encouraging gestures for the child based
on to his/her performance. On the other side, the system’s software component was mainly
focusing on developing the backend code to trace users’ spelling weakness and suggest words

21
accordingly. In the last stage of the implementation, these two major components were integrated
to produce the final product.
The robot/child interaction scenario is as follows: the robot pronounces a word for the child to
spell it. The child should spell the word using a keyboard. Once the child finishes inputting the
letters, he/she should press the enter key button to indicate the end of input. The robot responds to
user’s input with a facial and vocal expressions. The robot will utilize the servos to frown if the
answer is wrong and nod if the answer is correct. A suitable message will be spoken to the child
in both cases.
4.2. Robot Design
In designing the robot, we used affordable material to produce a prototype as a proof of concept.
The following hardware components were used: i) Arduino Uno, ii) 1 Standard Servo, iii) LED
Matrix (16x2), iv) Breadboard, wires, power supply.
Arduino Uno is a microcontroller board for building digital devices and interactive objects that
are used to form the basis of the robot. In addition, servos are used to create movements of the
robot’s head to encourage the child when he/she make correct answer. The interaction between
the child and the robot is done through keyboard and a screen. In the initial release we built a
desktop application to serve as means of communication between the robot and the user. The
same concept applies for a mobile application. The sequence of interaction is shown in the
following figure.
Figure 1: Interaction Model
The prime users of the software are Arab speaking children aiming to enhance their spelling
skills. Additionally, a knowledge acquisition subsystem can be developed in later releases to
obtain knowledge from teachers and extend the content of the game.
In our implementation we integrated the following design features to enhance the child
experience when learning:
 Interactivity: an educational game should be highly interactive. The reason is that the
higher the interactivity of the game is, the higher engagement of the user will be
accomplished, thus more enjoyment, more time spent on the game and ultimately, more
knowledge is gained.
 Adaptivity: and educational should attain high adaptivity. Adaptive teaching is a way of
informing users’ learning by personalizing and facilitating it in ways that meets every
individual’s needs.
 Guidance by goals and rules: educational games are always guided by goals and rules
complying with the subject understudy.
 Continuance feedback: the educational game should give instant feedback to the actions
of the user.

22
 Sensory stimuli: In physiology, a stimulus (plural stimuli) is a noticeable alteration in the
internal or external surroundings. The ability of an organism or organ to respond to
external stimuli is called sensitivity. [21]
When designing video game, we aim to trigger stimulus and imagination by adding voices,
animations and movements to make the game entertaining. The same is applicable for educational
games. Table 1 summarizes how these game design elements are integrated in the implemented
system.
Table 1: Game Design Elements
Game Design
Element
Ways Design Elements are Integrated in The Game
Interactivity 1- Robot Facial expressions corresponding to user input.
2- Robot encouraging messages in case of correctly spelled word.
3- Robot instructing messages in case of misspelled word
Adaptivity 1- System proposing words based of the performance of the child.
2- System advancing child to next levels when he/she meets some
criteria set by teacher.
Guidance by goals 1- The short-term goal is to spell the word correctly.
2- The long-term goal is to advance to the next level.
Continuance feedback 1- Robot facial expressions.
2- Verbal messages.
3- Reflecting child’s input on LED screen.
Sensory stimuli 1- Triggering both visual and verbal stimulus by voice and facial
expressions.
4.3. Framework
The framework used to develop the system includes the following components:
Knowledge base: the knowledge base serves as the dictionary of the game containing all the
words encountered in the game.
Inference engine: the inference engine is the component of our system that applies the spelling
rules on the knowledge base. In this release, we tackled one of the most basic and important
spelling rules in Arabic, the rule of sun and moon letters “ ‫القمرية‬ ‫و‬ ‫الشمسية‬ ‫الحروف‬ ‫قاعدة‬ ”. The
Arabic alphabet consists of 28 letters, which are classified into 14 sun letters and 14 moon letters.
The classification of the letters is based on how these letters influence the pronunciation of the
definite article at the beginning of the word.
In our implementation, a binary classifier was built to classify any word based on the following
classification rule:
Given the following two classes of Arabic letter:
SL = { ‫ت‬-‫ث‬-‫د‬-‫ذ‬-‫ر‬-‫ز‬-‫س‬-‫ش‬-‫ص‬-‫ض‬-‫ط‬-‫ظ‬-‫ل‬-‫ن‬ }
ML = { ‫أ‬-‫ب‬-‫ج‬-‫ح‬-‫خ‬-‫ع‬-‫غ‬-‫ف‬-‫ق‬-‫ك‬-‫م‬-‫هـ‬-‫و‬-‫ي‬ }
RULE Word-Classification
IF a word W begins with (‫ال‬) THEN // "‫"ال‬ is the article in Arabic
IF the third letter of W  SL THEN W is in class 1 // sun letters
ELSE W is in class 2; // in this case the third letter  ML and is a
(moon letters)

23
Binary Classification would generally fall into the domain of supervised learning since the
training dataset is labelled. Moreover, as the name suggests it is a special case where there are
only two classes. In Our implementation, the binary classifier is rule based.
User interface: in our system the user interface will be presented via a robot and a smart
application. The robot was integrated to maximize both amusement and interaction elements of
the game.
Question Generator: is an additional component which is responsible for producing questions to
the user in the form of “can you spell the word x?”. An important feature of the question
generator is its adaptivity. In other words, it proposes questions to the user based on his/her
former performance. For instance, if it was detected that the user had a weakness in a specific
spelling rule, he/she will be presented with a suitable explanation. Then other questions related to
the rule in question will be presented. A user will be able to progress to the next level only if a
user has met some specific criteria set by the teacher (for example 5 correct answers on the same
rule).
The user can respond to the question “Can you spell the word x?” with a query “Is my spelling of
x correct?”. The system responds to the query with an advice; either:
1- Your spelling is correct advance to the next level.
2- Your spelling could be improved you have to try again. The system keeps proposing
questions on the same spelling rule until some criteria is met.
Knowledge acquisition subsystem: this subsystem is intended to be used by the teacher to
extend the levels of the game by adding new spelling rules and lessons. It was not included in the
initial release. However, we intend to develop a knowledge acquisition subsystem in the
upcoming releases. The following figure conceptualizes system components.
Figure 8: Proposed Framework
Table 2 shows a sample system user based on a testing session.
Question
generated by
the system
User
response
System
interaction
advice
System action
‫كتابة‬ ‫تستطيع‬ ‫هل‬
" ‫الشمس‬ " ‫كلمة‬‫؟‬
Can you spell
the word “the
Sun”?
‫اشمس‬
Wrong
spelling of
the word
“the sun”
because it is
missing the
second letter
‫أخرى‬ ‫مرة‬ ‫حاول‬
Try again
The system detects that the child
is having difficulty with the sun
letters (‫الشمسية‬ ‫)الالم‬ therefore
proposes more questions on that
rule

24
" ‫الشمس‬ " ‫كلمة‬
Can you spell
the word “the
sun”?
‫الشمس‬
Correct
spelling of
the word
“the sun”
، ‫أحسنت‬ ، ‫ممتاز‬
‫صحيحة‬ ‫إجابة‬
Excellent,
Good job.
Correct
answer
The system detects that the child
has written the word correctly.
" ‫التاج‬ " ‫كلمة‬
Can you spell
the word “the
crown”
‫التاج‬
Correct
spelling of
the word
“the crown”
‫أ‬ ، ‫ممتاز‬، ‫حسنت‬
Excellent,
Good job.
Correct
answer
After the user completes a set of
words correctly, he/she can be
transformed to next level where
he/she is taught another rule.
The passing criteria (i.e. how
many words of a certain rule
should a child spell correctly
before moving to the next level)
is set by the teacher.
" ‫القمر‬ " ‫كلمة‬
Can you spell
the word “the
moon”
‫القمر‬
Correct
spelling of
the word
“the moon”
، ‫أحسنت‬ ، ‫ممتاز‬
Excellent,
Good job.
Correct
answer
The loop is repeated for the
moon letters ( ‫القمرية‬ ‫)الالم‬
The selected architecture of the system is layered. The following diagram shows the architectural
design of system components in the first release. The system is composite of 4 layers; the
presentation layer is represented by the robot and its facial expressions and vocal interactions.
The business layer consists of generators of these vocal and facial responses based on user input.
The persistence layer holds key modules; question generator which is responsible for proposing
spelling questions based on the child performance and the spelling rule he/she is currently
learning, spelling corrector which is responsible for determining whether user input is correct or
wrong, and deficiency detector which is responsible for determining the weakness a child might
have. This information is useful for the next iteration of questions, to address the issues he/she
might have. Lastly the knowledge base and inference engine containing dictionary and spelling
rules.
Figure 10: System Architecture

25
5. SYSTEM IMPLEMENTATION
This section presents in detail how the implementation of the system was carried on. Following
the iterative and incremental development methodology, our system was implemented in releases,
where each release focuses on building certain set of functionalities. After each component was
built and tested, integration of all components was done.
The following sections show the implementation of the presentation layer, where user interfaces
were executed. Furthermore, the implementation of the business layer is examined to show how
facial expressions of the robot were implemented. Additionally, and most importantly, the
persistence layer is introduced showing how each of the subcomponents, namely; question
generator, spelling corrector and deficiency detector are implemented and integrated. Lastly, the
dataset of the application is illustrated.
5.1. Graphical User Interface
In the initial release, we developed a desktop application as means for interaction between the
user and the robot. The software was implemented using Visual Studio as an integrated
development environment and C# as a programming language. Firstly, the interfaces were
designed stressing on simplicity and clarity since the system is meant to be used by kids. The
navigation has been designed to be clear and easy. Friendly icons and visual descriptors were
used. The following figures illustrates some of the application’s interfaces.
Figure 2: Main Interface Figure 3: Level one Spell a
Word
Figure 4: Positive
Feedback Interface
The main interface of the application (Fig. 11) displays a button to start the game as well as
playing voice in the background to welcome the player. The second interface (Fig. 12) displays
the player’s level. Additionally, a word will be pronounced for the player to spell. If the player
wants to listen to the word again, he/she can press the speaker’s icon. Using a keyboard, a player
can input the word then press check button. The third interface (Fig. 13) represents a positive
feedback to player’s input. In case the entered word is spelled correctly this interface will be
displayed as well as a positive response of the robot. The robot will node and show a smelly face,
this will be illustrated in the hardware implementation section.
5.2. Hardware Components
For the purpose of completing the implementation phase, a proof of concept prototype was
developed to represent the robot. Arduino Uno, which is a microcontroller board, was the center
of our prototype. It has 14 digital input/output pins (of which 6 can be used as PWM outputs), 6
analog inputs, a 16 MHz quartz crystal, a USB connection, a power jack, an ICSP header and a

26
reset button. Arduino was connected to two 8X8 LED matrix components to represent the eyes of
the robot. A standard servo was used to model neck movement of the robot.
Figure 5: Used Material
The following table illustrates how the assembly of the hardware was done:
Table 2: Arduino Assembly
Arduino LED Servo
Ground Ground Ground
V5 V5 V5
Pin 6 - Servo
Pin 11 CS -
Pin 12 CLK -
Pin 13 SDI -
Where SDI is the serial data input pin, CS is the chip select that turns on tristate outputs on data
pins, or signifies that inputs are valid and can be latched, stored or processed. CLK is a
connection for a square wave signal that synchronizes systems.
After connecting the hardware, Arduino 1.8.5. The open-source Arduino Software (IDE) was
used to develop a code to control the robot’s expression. A total of three facial expressions were
developed; happy expression, resting face and a sad expression. Furthermore, to make the robot
more realistic a blinking method was also implemented. Figure 15 shows the system components
and the graphical user interface.
Figure15: System User Interface.

27
5.3. Persistence Layer
5.3.1. Questions generator
This component is the first of the components in the persistence layer. The purpose of this
component is to propose a question for the player, word to spell. Table 4 shows the pseudo code
of the function.
Table 3: Questions Generator Pseudo Code
Questions Generator
Pre-condition None
Post- condition A word will be selected from the knowledge base and
pronounced for the player.
Pseudo code Begin
Index = Generate a random number;
Retrieve word [Index];
Retrieve Sound [Index];
Pass Sound [Index] to Speaker;
End
5.3.2. Spelling corrector
This is the second component in the persistence layer. The purpose of this component is to check
whether the entered word is correct. Table 5 shows the pseudo code of the function.
Table 4: Spelling Corrector Pseudo Code
Spelling Corrector
Pre-condition Player has entered a word in the textbox.
Post- condition The robot will show either positive or negative feedback
depending on player’s input.
Pseudo code Begin
Answer = Read player’s input;
If (Answer == word [Index])
{
Show positive feedback;
Move to next level;
}
Else
Detect deficiency;
End
5.3.3. Deficiency detector
This component is the third component in the persistence layer. It is a very essential component
which aims to detect what exactly is the deficiency the player is having. For this release of the
application the rule of sun and moon letters in Arabic “ ‫القمرية‬ ‫و‬ ‫الشمسية‬ ‫الحروف‬ ‫قاعدة‬ ” was
implemented. The following illustrates pseudo code for the function.

28
Table 5: Deficiency Detector Pseudo Code
Deficiency Detector
Pre-condition Entered word was spelled incorrectly by the player.
Post- condition The correct spelling will be presented to the player.
A new word will be generated for the player within the same
spelling rule detected.
Pseudo code Begin
Check if the word starts with (‫)ال‬
Check if the spelling error was in the third letter
{
Check if the third letter is one of the sun letters
Return 0;
Else
Return 1;
}
Else
Show correctly spelled word;
End
Each time a player makes a spelling error in one spelling rule, his/her error will be corrected and
for the next three iterations, new words of the same rule will be proposed. The reason for
selecting three trials was based on experts’ judgment of teachers in elementary schools.
5.4. Data Layer
In the first release of the game, the lexicon was built based on Saudi Arabia’s curriculum the first
lesson of 2nd
grader’s Arabic book)‫الخالدة‬ ‫(لغتي‬ Page#30, 28 words were selected and categorized
into two classes each of 14 words. The first class represents sun letters and the second class
represents moon letters.
The following table summarizes words currently available in the lexicon.
Figure 6: Game's Lexicon
# Class 1 (Sun letters) Class 2 (Sun moon)
1 ‫األمر‬ (the command) ‫الشمس‬ (the sun)
2 ‫البرتقال‬ ( the orange) ‫التفاح‬ (the apple)
3 ‫الجنة‬ (the heaven) ‫الثمرات‬ (the fruits)
4 ‫الحياة‬ (the life) ‫الدين‬ (the religion)
5 ‫الخس‬ (the lettuce) ‫الرمان‬(the pomegranate)
6 ‫العيون‬ (the eyes) ‫الزهور‬ (the flowers)
7 ‫الغيوم‬ (the clouds) ‫السماء‬ (the sky)
8 ‫الفارس‬ (the knight) ‫الصباح‬ (the morning)
9 ‫القائد‬ (the leader) ‫الضحى‬ (the morn)
10 ‫الكريم‬ (the generous) ‫الطين‬ (the mud)
11 ‫الموز‬ (the banana) ‫الظالم‬ (the dark)
12 ‫الهدهد‬ (the hoopoe) ‫الليل‬ (the night)
13 ‫الوعد‬ (the vow) ‫النهار‬ (the day)
14 ‫اليمين‬ (the right) ‫الذئب‬ (the wolf)

29
6. SYSTEM EVALUATION
In order to check whether the actual results meets the expected results many tests were conducted
to validate and verify the game. Unit testing was firstly conducted; the following section shows
test cases for each part of the system. Then, after integrating all components, integration testing
was performed. For the purpose of evaluating the game, the accuracy of the classifier was
measured.
6.1. Unit Testing
In this section the test cases examined during unit testing is presented. The testing process was
conducted in two phases; testing the software side and testing the hardware side. Tables 7 and 8
present some of the tests done.
Table 6: Unit Testing for the Software
I
D
Name of
the
function
Objectiv
e of the
function
Test
data
Test case
complexit
y
Expected
results
Actual result Pass
?
Cycle#1 Cycle#
2
Cycle#
3
1 Generate
Question
To
pronounc
e a
random
word for
the
player
Non
e
Simple A word is
pronounce
d for the
player
‫الليل‬
(the
night)
‫الشمس‬
(the
sun)
‫الهدهد‬
(the
hoopoe
)
Pass
2 Correct
Spelling
To
validate
player’s
input for
‫الشمس‬
‫الشم‬
‫س‬
Simple The word
should be
corrected,
screen
should
show
encouragin
g message
Positive
feedbac
k
interfac
e is
presente
d
Pass
3 Correct
Spelling
To
validate
player’s
input for
‫الشمس‬
‫الشم‬ Simple The word
should be
corrected,
screen
should
show
‫الشمس‬
‫الشمس‬
(the
sun)
Pass
4 Detect
Deficienc
y
To detect
what is
the
spelling
rule that
the
player is
having
difficulty
with
‫القر‬ Difficult Only show
the correct
spelling of
the word
since the
mistake is
not in a
spelling
rule
‫القمر‬
(the
moon)
‫الطين‬
(the
mud)
Pass

30
Additionally, unit testing was also conducted on the hardware side of the application. Test cases
were driven to check the responses of the robot. The table below shows some of the test cases.
Table 7: Unit Testing for the Hardware
ID Name of
the
function
Objective
of the
function
Test
data
Test case
complexity
Expected
results
Actual result Pass?
1 Generate
happy
face
To show
happy
facial
expression
and node.
None Medium Robot
nodding
and
showing
smiley face
Pass
2 Generate
natural
face
To show
natural
face
None Medium Robot
showing
natural face
Pass
6.2. Integration Testing
After all the components have passed unit testing, integration between hardware and software was
performed. The integration was done via connecting through serial port COM3. Table 9 shows
part of the code used to check the integration on both the software and hardware sides.
Table 9: Integration testing
Integration Code C# side Integration Code Arduino Side
serialPort1.Write(1); // for smiling
serialPort1.Write(0); // for frowning
if (Serial.available()){
int data = Serial.read();
if (data==1) HappyFace();
else SadFace();}
In addition, we intend to conduct observational studies to evaluate the player’s engagement,
enjoyment and enthusiasm about the game. Furthermore, test will be driven to evaluate
knowledge acquisition and retention for players. This is planned to be done in near future.
7. CONCLUSION
Integration of game-based learning and robotics into educational systems have the potential to
enhance knowledge acquisition and retention especially for young children. Throughout this
paper the analysis, design, implementation and testing of an educational game to teach Arabic
spelling was proposed. The proposed game has two main characteristics; adaptation and
interaction. By developing an adaptive game, the system shall be able to detect the weaknesses of
each individual child and propose questions accordingly. On the other hand, by adding the
interaction element, the game shall be fun and engaging for the child to keep playing, hence,
advance his/her knowledge.

31
For future work, we aim to design and execute users’ acceptance testing. Observation studies will
be performed to evaluate how children interact with the game in a classroom setting. Observation
of children behavior and interaction with the system will be recorded for future improvements.
Furthermore, more Arabic spelling rules are planned to be implemented and integrated into the
game’s engine. Also, as we implemented the first release of the system, we aim to gradually build
up to full release that include additional functionalities such as a dedicated module for teachers
allowing them to add words to the lexicon and set flexible evaluation criteria for the players.
REFERENCES
[1] Muhammad al-Sharkawi. “History and development of the Arabic language”,1st ed, Abingdon,
Routledge,2016.
[2] Alshalaan Rashed “Spelling Weakness ”, in Practical Methods to Treat Spelling Mistakes, 2ed,
Vol 85, Riyadh, Alobeikan, 2008.
[3] Smith, P.K. and Pellegrini, A., “Learning through play. Encyclopedia on early childhood
development”, PP.1-6, 2008.
[4] R. v. Horn, "Educational Games," The Phi Delta Kappan, vol. 89, (1), pp. 73-74, 2007.
[5] D. J. Miller and D. P. Robertson, "Educational benefits of using game consoles in a primary
classroom: A randomised controlled trial," British Journal of Educational Technology, vol.
42, (5), pp. 850-864, 2011.
[6] E. S. Duh, N. Koceska and S. Koceski, "Game-based learning: educational game Azbuka to help
young children learn writing Cyrillic letters," Multimedia Tools and Applications, 2016.
[7] C. E. Zarina and H. Hussain, "Analysis of Local and Foreign Edutainment Products - An Effort to
Implement the Design Framework for an Edutainment Environment in Malaysia," The Journal of
Computers in Mathematics and Science Teaching, vol. 24, (1), pp. 27-42, 2005
[8] A. Amory, "Game Object Model Version II: A Theoretical Framework for Educational Game
Development," Educational Technology Research and Development, vol. 55, (1), pp. 51-77, 2007.
[9] Said Norma, "An engaging multimedia design model," in 2004. DOI: 10.1145/1017833.1017873
[10] T. N. Beran et al, "Understanding how children understand robots: Perceived animism in child–
robot interaction," International Journal of Human - Computer Studies, vol. 69, (7), pp. 539-550,
2011.
[11] C. Chang et al, "Exploring the Possibility of Using Humanoid Robots as Instructional Tools for
Teaching a Second Language in Primary School," Journal of Educational Technology &
Society, vol. 13, (2), pp. 13, 2010
[12] S. S. Young, Y. H. Wang and J. R. Jang, "Exploring perceptions of integrating tangible learning
companions in learning English conversation: Colloquium," British Journal of Educational
Technology, vol. 41, (5), pp. E78-E83, 2010.
[13] M. Sugimoto, "A Mobile Mixed-Reality Environment for Children's Storytelling Using a
Handheld Projector and a Robot," IEEE Transactions on Learning Technologies, vol. 4, (3), pp.
249-260, 2011.

32
[14] N. Chen, B. Quadir and D. C. Teng, "A novel approach of learning English with robot for
elementary school students," in Anonymous Berlin, Heidelberg: Springer Berlin Heidelberg, 2011,
pp. 309-316.
[15] "DIY a Spelling Game with an Interactive Robot", Instructables.com, 2017. [Online]. Available:
http://www.instructables.com/id/DIY-a-Spelling-Game-with-an-Interactive-Robot/. [Accessed: 21-
Dec- 2017].
[16] A. Erradi et al, "ArabicTutor: A multimedia m-learning platform for learning Arabic spelling and
vocabulary," in 2012, . DOI: 10.1109/ICMCS.2012.6320220.
[17] Belkhouche, Boumediene & Saeed, Nafla & Darei, Al & Abdulla, Shamsa & Al Mazrouei,
Shamsa & Hamdan, Shaikha & Mandhari, Al & Ateeq, Mouza & Mehairi, Al. (2014). Learning
Arabic With Games. . 10.5176/2251-1679_CGAT14.03.
[18] S. Liao, "Expert system methodologies and applications—a decade review from 1995 to
2004," Expert Systems with Applications, vol. 28, (1), pp. 93-103, 2005.
[19] A. Dembitz, P. Kne evi and M. Sokele, "Developing a Spell Checker as an Expert System,"Journal
of Computing and Information Technology, vol. 11, (4), pp. 285, 2005.
[20] "Systems analysis and design in a changing world, 6th ed," Reference & Research Book News,
vol. 27, (5), 2012.
[21] "How does the brain process sensory stimuli? - Best Practice: Sensory", Best Practice: Sensory,
2017. [Online]. Available: http://sensory-processing.middletownautism.com/background/how-
does-the-brain-process-sensory-stimuli/. [Accessed: 17- Dec- 2017]
Authors
Ghada Alsebayel Ghada Alsebayel is a teaching assistant at King Saud University, Saudi Arabia. She
received both her B.Sc. and M.Sc. from College of Computer and Information Sciences CCIS at King Saud
University KSU. Ghada’s research interests include data mining, data management, knowledge
representation and reasoning, and their applications in health informatics. Ghada spent a year at National
Guard Health Affairs, where she worked as a system analyst for the clinical information management
department. In 2016, she co-founded the first summer training program at CCIS KSU, to enable young
Saudi women to be better represented in STEM. The program introduces areas like coding, data science,
artificial intelligence, and robotics.
Jawad Berri is an associate professor at King Saud University, Saudi Arabia. He
received his Ph.D. in Computer Science from Paris-Sorbonne University in France in
1996. Jawad’s research interests focus on context-aware web systems, learning
technologies and natural language processing. He has been involved in many projects
related to mobile learning, semantic web, automatic summarization, web information
filtering and mobile agents for web information discovery. He worked as a researcher
at the CNRS – the French National Research Center, the Computer Science Institute
at the University of Zurich – Switzerland and Sonatrach – the Algerian Petroleum
and Gas Corporation. Jawad taught in many universities around the world: The
United Arab Emirates University, Khalifa University of Science Technology and
Research and King Saud University. His contributions in research projects in the industry and academia led
to the publication of papers in numerous journals and conferences. Jawad is a senior member of the IEEE.

ROBOT BASED INTERACTIVE GAME FOR TEACHING ARABIC SPELLING

More Related Content

What's hot (19)

Similar to ROBOT BASED INTERACTIVE GAME FOR TEACHING ARABIC SPELLING (20)

More from gerogepatton (20)

Recently uploaded (20)

ROBOT BASED INTERACTIVE GAME FOR TEACHING ARABIC SPELLING