Interactive Learning Environments

ISSN: 1049-4820 (Print) 1744-5191 (Online) Journal homepage: https://www.tandfonline.com/loi/nile20

Ambient intelligence-based smart classroom

model

Vitomir Radosavljevic, Slavica Radosavljevic & Gordana Jelic

To cite this article: Vitomir Radosavljevic, Slavica Radosavljevic & Gordana Jelic (2019):
Ambient intelligence-based smart classroom model, Interactive Learning Environments, DOI:
10.1080/10494820.2019.1652836

To link to this article: https://doi.org/10.1080/10494820.2019.1652836

Published online: 13 Aug 2019.

Submit your article to this journal

Article views: 4

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=nile20
INTERACTIVE LEARNING ENVIRONMENTS
https://doi.org/10.1080/10494820.2019.1652836

Ambient intelligence-based smart classroom model

Vitomir Radosavljevic , Slavica Radosavljevic and Gordana Jelic
Department of Communication Technologies, ICT College of Vocational Studies, Belgrade, Serbia

ABSTRACT ARTICLE HISTORY

This paper introduces the smart classroom learning model based on the Received 31 March 2019
concept of ambient intelligence. By analyzing a smart classroom, the Accepted 23 July 2019
ambient intelligence system detects a student and determines their level of
KEYWORDS
fatigue based on the data about their previous daily academic activities. This Smart classroom; adaptive
information is then used to assign the student the appropriate learning learning; ambient
strategy. The paper describes relevant factors for developing the model. The intelligence; learning
model was tested on a sample of 80 students. By analyzing the information strategies; technology-
available through ambient intelligence, it was possible to utilize the smart supported learning
classroom to provide students with the adequate learning strategy in environments
accordance with the criteria compatible with the expected leaning
outcomes. The research results have shown positive effects of the application
of the ambient intelligence-based smart classroom model on learning results.

1. Introduction
Learning is a set of constructive processes through which a person becomes activate, individually or
in a group, and elaborates, builds, and organizes knowledge (Seidel & Shavelson, 2007). A smart class-
room uses technology to familiarize and make easier the learning process for a student and the
teaching process for the teacher, as well as to make the transmission of knowledge efficient and
effective (Guinard, Fischer, & Trifa, 2010). It should enable the presentation of various teaching
materials, provide the elements necessary for personalized learning group learning, mobile and
virtual learning, as well as provide support with adaptive learning, student-centered learning, and
all other learning-related activities (Li, Kong, & Chen, 2015). In order to adjust the learning process
to the individual needs of students and teachers in the smart learning environment, it is necessary
to implement elements which analyze the context of learning and adjust the individual process of
learning to the student. Ambient intelligence is the element that is used to connect the smart class-
room and the adaptation of the learning process.
This paper introduces the smart classroom learning model based on the concept of ambient intel-
ligence. The paper is divided into several chapters. After the introduction, Chapter 2 presents an over-
view of relevant scientific terms and bibliography which served as the basis for the model. Chapter 3
describes the smart classroom learning model. The scientific research which uses the described
model is presented in Chapter 4. This chapter also covers the results of the research. At the end of
the paper, Chapter 5 discusses the results and covers research conclusions.

2. Background study
The smart classroom learning model described in this paper relies on smart classrooms, ambient
intelligence, and learning strategies.

CONTACT Vitomir Radosavljevic vitomir.radosavljevic@ict.edu.rs

© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 V. RADOSAVLJEVIC ET AL.

2.1. Smart classrooms

A smart classroom can be defined as an intelligent environment equipped with a wide spectrum of
various hardware and software equipment: projectors, cameras, sensors, speech and face recognition
modules, RFID readers, and similar elements, the main goal of which is to support the educational
program which needs to be realize (Guinard et al., 2010). Smart classrooms are a technology-sup-
ported form of learning environment. Technology-supported learning environments (TSLEs) are
instructional systems which make use of technology which helps the student perform learning-
related activities while also providing teachers and other parties involved in the learning process
with additional support (Wang & Hannafin, 2005).
Aside from technologically advanced hardware equipment, smart classrooms are equipped with
sensor networks. The variety of sensors provides basis for a large number of concepts which can be
realized. Based on its settings or data from the network, the sensor network can start an application or
provide information to an individual which they will use to perform an action (Twomey, Diethe, Crad-
dock, & Flach, 2017). Usually, sensor networks keep track of the smart classroom, providing an image
of the physical parameters of the learning environment. The physical classroom environment has an
impact on both the teacher and the student, and it may also affect the academic results of the
student (Yang & Huang, 2015). Most studies take into consideration temperature, noise, lighting,
and the quality of air as the parameters for measuring the quality of the learning environment (Al-
Hemoud et al., 2017; Castilla, Llinares, Bravo, & Blanca, 2017; Mihai & Iordache, 2016; Sala &
Rantala, 2016; Uzelac, Gligorić, & Krčo, 2018; Vilčeková, Kapalo, Mečiarová, Burdová, & Imreczeová,
2017). By connecting different actuators, it is possible to develop a system which independently regu-
lates parameters in an effort to create optimal working physical conditions in smart classrooms (Guo,
Zhang, Wang, Yu, & Zhou, 2013). The challenges faced when developing sensors and sensor systems
include the development of technically and energetically efficient sensors which fit the environment
with their minimalistic design and dimensions. Sensors usually contain small batteries and a limited
capacity for storing and processing data (Gómez, Huete, Hoyos, Perez, & Grigori, 2013).
The potential offered by smart classrooms as a technology-supported environment is great. In his
research, Shen, Wu, and Lee (2014) wanted to examine and determine the potential of information
technologies within a smart classroom. The smart classroom used in his study was equipped with LED
displays, all-in-one computers with multi-touch displays, mobile devices, near field communication
(NFC) technology. The real-time system for automatic detection of students who are present
proved to be an efficient solution for classes attended by large numbers of students. The same
system was used to deliver students’ questions to the teacher during the lecture. The system utilized
students’ smartphones’ NFC technology or NFC cards to register lecture attendance. The system
communicated with the database which stores all relevant information. In addition to this, students
sent their questions to the teacher after every explained presentation slide. Research shows that this
approach is especially useful for students from Asia who are hesitant to express their opinions during
lectures (Liu & Littlewood, 1997).
The efficient use of smart classrooms was proven in other projects as well. The use of smart class-
rooms in 4G systems for the purposes of recording the lecture and distributing those recordings
through a learning platform was the turning point of Alelaiwi’s research (Alelaiwi et al., 2015). The
goal of the project which followed the research was to establish two-way communication between
the student and the teacher during the lecture. Using the real time streaming protocol (RTSP),
teacher and student sessions were controlled and adjusted. Students listened to lectures and had
the option to send the teacher a question by circling the part of the presentation they found
unclear and sending that screenshot, along with their question, to the teacher. The research high-
lighted the importance and usefulness of this approach to learning. A project conducted in 2009
(Suo, Miyata, Morikawa, Ishida, & Shi, 2009) examined the possibility of expanding the smart class-
room-based learning systems and making them scalar. The goal was to determine the effects of con-
necting two smart classrooms over the internet. For the realization of this concept, it was necessary to
 INTERACTIVE LEARNING ENVIRONMENTS 3

make the architecture and the interface of the smart classroom accessible to anyone. As a part of this
study, two smart classrooms (one in the city of Tsinghua, China, the other one in Kyoto, Japan) were
connected to each other. Both smart classrooms were equipped with cameras, and those recordings
were livestreamed in the other classroom. The projectors showed the livestream from the classroom
in the other city. A teacher in one city showed their presentation in both classrooms simultaneously
using the smart board.
Smart educational environments are suitable for the implementation of adaptive forms of edu-
cation (Brusilovsky & Peylo, 2003).

2.2. Ambient intelligence

Ambient intelligence is a pervasive computing concept in which the working environment is aware of
the presence of users. Such a concept can be achieved by integrating sensors, cameras, microphones,
and other devices which can generate signals based on the changes in the environment and the
intelligent software system. Systems which utilize the ambient intelligence concept should detect
the presence of a user in the working environment, identify the user, recognize the activities in
the environment, analyze them within a context and perform the right change to the environment
in accordance with the user’s needs (“Ambient intelligence - The ultimate IoT use cases | IoT for all,”
2019; Bravo, Cook, & Riva, 2016).
The ambient intelligence concept implements elements of pervasive computing, ubiquitous com-
puting, people profiling, context awareness, and human-centric computing (“Ambient intelligence,”
2019). Ambient intelligence systems contain a large number of devices and sensors which are inte-
grated into the user’s working environment. Most often, these include sensors for lighting, tempera-
ture, noise, pressure, object position, face and speech recognition technology bio-signal reading, GPS,
RFID and other technologies. The following and detection of users take place using motion sensors
which detect users’ movements, but cannot detect the cause of those movements. Aside from
motion sensors, the system utilizes RFID tags which are placed on the user and which are read in
certain places, or i-Buttons computer chips placed in a 16-mm circular metal box which should be
placed on a compatible reader. The challenges which ambient intelligence systems face, and
which have to do with sensors, include the managing of large quantities of information generated
over long periods of time, the autonomy of power charging (batteries in wireless sensors), the poten-
tial inability to identify a user, as well as the need for processing unconventional types of information
gathered from sensors. The data from sensors often exceed conventional types of information, and as
such require specific forms of analysis. This information is multidimensional and susceptible to sig-
nificant deviations due to the presence of impeding factors in the environment (which can collec-
tively be called white noise). Depending on the type of sensors and the purpose of information,
there might emerge a need for real-time data processing, which in turn additionally burdens the
data-processing computer system (Cook, Augusto, & Jakkula, 2009).
The connection between intelligent algorithms used to recognize the user and their activities and
the environment the user is in can be viewed through the activities performed by the intelligent
environment. In order to adjust the algorithms to the needs of the user, the intelligent system has
to make decisions. These decisions are based on:

. The way the user is modeled in the system,

. User activity prediction and recognition,
. The defined decision-making process, and
. Space–time reasoning.

User modeling is based on user behavior. The ambient intelligence system should recognize
behavior patterns based on the information gathered from the network of sensors in order to
predict future behavior of the user. During this process, navigating the ambient intelligence
4 V. RADOSAVLJEVIC ET AL.

system should be intuitive and in accordance with the conclusions reached by the field of communi-
cation between humans and computers (Cook et al., 2009; Cottone, Maida, & Morana, 2014).
The ambient intelligence model developed by (Ramos, Augusto, & Shapiro, 2008) places the user
in the center of the system (Figure 1). The user communicates with the ambient intelligence system
and the ambient intelligence environment.
The ambient intelligence system processes data received from the environment. The ambient
intelligence environment sends information about the presence of users in the ambient intelligence
system in one or more ways, such as sensor reading, RFID technology, face and speech recognition
technology, GPS, and other technologies. The ambient intelligence system processes this data on two
levels – operational and intelligent. The operational level takes information from the ambient intelli-
gence environment and adjusts it to different technologies and principles necessary to transmit that
information to the intelligent level. The intelligent level takes various types of information from the
operational level and applies calculations to define the so-called “system intelligence”. The result of
these calculations is the appropriate action in the ambient intelligence environment which has an
impact on the user.
The ambient intelligence concept can be used in different areas of life, such as transport, smart
house development, education. In the field of education, there have been projects which used tech-
nologically enriched working environments for tracking students’ progress in the process of learning,
as well as their activities. The conclusion reached in these studies is that technology-supported learn-
ing systems (which includes ambient intelligence) can have a positive effect on the learning process
(Chen, 2014; Santana-Mancilla, Echeverría, Santos, Castellanos, & Díaz, 2013; Shen et al., 2014; Tibúrcio
& Finch, 2005).

2.3. Learning strategies

Learning strategies can be defined as cognitive operation sequences which take student from under-
standing teaching materials, to asking questions, to providing answers to the given question.

Figure 1. Ambient intelligence model developed by (Ramos et al., 2008).

 INTERACTIVE LEARNING ENVIRONMENTS 5

Learning strategies include various learning techniques and procedures used by the student with the
appropriate goal, in coordination with the context, in order to process information, and with the goal
of meaningful learning. Learning strategies can also be described as mental operations which
the student conducts during the learning process, which means that the choice of the learning strat-
egy impacts the student’s success in learning (Muelas & Navarro, 2015). Learning strategies can be
viewed as:

. Support strategies,
. Process strategies, and
. Personalization strategies.

Support strategies are those learning strategies which have to do with the motivation of the
student and their personal attitude towards learning (working with the teacher). Process strategies
include the organization of the learning process, the choice of learning materials, and the method
of processing information (re-reading the materials, extracting and re-writing notes). Material perso-
nalization strategies involve creative and critical thinking, as well as the transmission of knowledge
(group work, mnemonics, mind maps) (Muelas & Navarro, 2015; Soliman, 2017).
The adequate choice of the learning strategy depends on the type of materials the student is
learning, their motivation, and the way they organize knowledge into mental cognitive models
(metacognition) (Donker, de Boer, Kostons, Dignath van Ewijk, & van der Werf, 2014).
A study by (Persky, 2018) makes a connection between the choice of the learning strategy with
stress and fatigue during a semester and over the course of studies. This study emphasized the
fact that the most common learning strategy used by students during their four-year studies is re-
reading teaching materials. Aside from this learning strategy (which was marked as primary), over
the course of their studies students started to implement additional (secondary) learning strategies
such as study groups, taking notes, rewriting of materials, making mind maps. This shift in learning
strategies was explained with two facts. Firstly, as the semester passed, the element of fatigue was
becoming more prominent. And secondly, as the semester was coming to an end, students were
under more stress to tackle all of their obligations and learn the materials before the exam. The
authors of this study mentioned that it is still unclear whether implementing a new learning strategy
increases the quality of learning or simply speed up the progress towards the goal of passing the
exam.
A study conducted by (Dunlosky, Rawson, Marsh, Nathan, & Willingham, 2013) analyzed ten learn-
ing strategies against four criteria: the materials used, the learning conditions, student characteristics,
and method of realization. The strategies which yielded the best results were practice testing and
distributed practice. Student characteristics which were considered relevant for the learning strategy
analysis were age, previous knowledge, working memory capacity, verbal competence, interests,
intelligence, motivation, previous achievements, and self-efficiency. For the development of the
model in this study, it was important to establish the characteristics of the working memory capacity,
as it is closely connected to fatigue.
Working memory capacity determines the amount of mental activity, i.e. cognitive load which an
individual can be exposed to. Different learning strategies impact the cognitive load of a student
during the process of learning differently (Kanjug & Chaijaroen, 2012; Kassim, Nicholas, & Ng,
2014). The decrease in cognitive functions is caused by fatigue, which is a natural and inevitable
occurrence in any process of working, including learning (Nota, 2011). Fatigue significantly
impacts productivity, the occurrence of errors, and efficiency when working or learning. Fatigue
can be mental or physical. Physical fatigue involves a decrease in the ability to perform physical activi-
ties due to previous physical activities (Sedighi Maman, Alamdar Yazdi, Cavuoto, & Megahed, 2017).
Mental fatigue is a psycho-biological condition caused by long periods of demanding cognitive
activity (Marcora, Staiano, & Manning, 2009). It has been established that people with fatigue have
issues concentrating and focusing on tasks (Boksem, Meijman, & Lorist, 2005; Qi et al., 2019). In
6 V. RADOSAVLJEVIC ET AL.

order to perform tasks while fatigued, it is necessary to additionally motivate a person or to increase
their control and engagement in the working process. Observing the learning process, it has been
concluded that a fatigued student needs additional motivation to learn or an alternative learning
strategy through which the teacher or the learning platform will increase the student’s control
over the learning process (Müller & Apps, 2019).
Learning strategies such as re-reading call for a student’s individual engagement in the organiz-
ation and acquisition of knowledge. The student learns at a pace which they determine themselves.
Although this is not a commonly recommended strategy or a strategy regarded as efficient, research
has shown that it is the one most commonly used in the learning process (Dunlosky et al., 2013;
Persky, 2018). By re-reading the materials, the student independently creates a mental image of
the text. The creation of the mental image requires memorizing data, as well as connecting new
and preexisting knowledge. It is recommended to re-read the text several times in order to create
a correct mental image (Callender & McDaniel, 2009). This strategy requires significant mental
engagement, which is why it should be avoided when a student is fatigued.
Study groups are a learning strategy in which students come together in order to exchange knowl-
edge and experiences through communication, as well as to successfully reach conclusions and gain
knowledge about the thematic unit they need to study (Roscoe & Chi, 2008; Sibley & Parmelee, 2008).
Students within a group ask each other questions related to the materials they find unclear. They also
provide answers to those questions which they assume they know the answers to. The presentation
of knowledge to other participants in the system is good both for those who are presenting and those
being presented to (Fiorella & Mayer, 2013). Students who present what they have learned through
group engagement deepen their knowledge. This can be explained with the fact that this process
requires a connection between preexisting and new knowledge. On the other hand, the students
being presented to do not increase their level of engagement, but only adopt new information
(Roscoe & Chi, 2007). Together, students organize knowledge and information they need to adopt.
An individual adjusts to the group’s learning pace. The drawback of this approach is the potential
misunderstanding of the materials (van Leeuwen, Janssen, Erkens, & Brekelmans, 2015).
Working with the teacher is a learning strategy which relies on the traditional face-to-face
approach. Through direct communication with the student, the teacher conveys the materials to
the student and analyzes problems. The teacher organizes knowledge and information they
convey to the student, while the engagement is minimal. The teacher determines the learning
pace. The role of the teacher in this strategy is dominant. Research (Erdogan & Campbell, 2008)
has shown that the engagement of the teacher in individual work with a student has a positive
effect on the student’s motivation to work and achieve better results. By asking questions and
offering positive feedback on the student’s progress, the teacher stimulates cognitive learning activi-
ties and better material comprehension, encourages the student to be more active, and improves the
student’s intrinsic motivation (Chin, 2006; Deci, Koestner, & Ryan, 1999; Jurik, Gröschner, & Seidel,
2014). Through communication with the student, the teacher builds their own attitude towards
the individual characteristics of the student. Based on that attitude, they can adjust their communi-
cation and interaction, ask appropriate questions, and provide useful guidelines for learning and per-
forming tasks (Furtak & Kunter, 2012).

3. Ambient intelligence model in a smart classroom

Keeping in mind previous research, the authors developed a learning model in smart learning
environments which uses elements of ambient intelligence to optimize the learning process. The
model provides the student with the appropriate learning strategy which corresponds to their esti-
mated level of fatigue. The fatigue of students is estimated based on their engagement in the aca-
demic activities they have attended during the day. By entering the smart classroom, the student
activates the smart classroom system which recognizes and identifies him or her. The information
about the student is then sent to the ambient intelligence system which extracts information from
 INTERACTIVE LEARNING ENVIRONMENTS 7

the database about the student’s daily schedule and other academic activities. Based on this infor-
mation, the student’s level of fatigue is estimated. The data on the student’s level of fatigue
adapts the learning process by choosing the appropriate learning strategy for the lecture.
The model is comprised of the following five parts:

. User
. Smart classroom
. Ambient intelligence system
. Database
. External system integration block (Figure 2).

A smart classroom can detect the presence of a student using sensors, video cameras, RFID
systems, and other systems integrated into the smart classroom. The smart classroom is also com-
prised of actuators which manage the physical and working environment of the classroom. Managing
the physical environment involves the ability to control parameters such as temperature, air quality
through a ventilation system, lighting etc. Managing the working environment involves communi-
cation between the smart classroom and devices such as smart boards, projectors, LMSs. The
smart classroom communicates in both directions with the ambient intelligence environment. It
sends information about the presence of users in the classroom. The ambient intelligence system
combines this information with the information from the database and other information from exter-
nal systems (other e-learning platforms, m-learning platforms, augmented reality, etc.). Based on the
information available to the ambient intelligence system and the way the user and their behavior are
modeled, the system recognizes behavior patterns and generates information which it then sends to
the smart classroom. Through actuators, the smart classroom manages the processes of adapting the
working and physical environment to the user. This model supports four characteristics crucial for the
development of the ambient intelligence concept. According to (“Ambient intelligence - The ultimate
IoT use cases | IoT for all,” 2019), these include usability, technical feasibility, trust and confidence, and
social and economic impacts. The usability of the model is reflected in the organization of relevant
parts and users within the system with the goal of automatically providing a requested service to the
user. Technical feasibility requires a precisely designed, reliable, and efficient system. Clearly defined
users and communication protocols between them make the system highly reliable. The model is
user-centered and provides services in a way that makes the system’s decisions trustworthy for
the user. Through the decisions made by the ambient intelligence system, the model provides
support for social interaction between users and increases user productivity.

Figure 2. Ambient intelligence system model in a smart classroom.

8 V. RADOSAVLJEVIC ET AL.

4. Research
The model presented in this paper was used in a study which examined the effects of applying the
ambient intelligence model to the smart classroom on students’ success in reaching adequate learn-
ing goals. The learning goals are defined as skills, knowledge, or attitudes which should develop as a
result of learning (Kumpas-Lenk, Eisenschmidt, & Veispak, 2018). In this study, the ambient intelli-
gence system detected a student based on the student’s QR code and then used information
about the student’s curricular activities during the day to assess the student’s level of fatigue.
Based on that assessment, the ambient intelligence system determined and assigned learning strat-
egies to every individual student (Figure 3). At the end of the lecture, the students took a test which
measured their success in reaching learning goals.
The study was conducted at the ICT College of Vocational Studies in Belgrade, Serbia. Eighty stu-
dents from the Telecommunications department taking a course in Digital Telecommunications par-
ticipated in the study. They were divided into two random groups, control and experimental, with no
statistical differences.

4.1. Research plan

Students in the control group entered the classroom at the beginning of the lecture, set in front of
their computers, after which the teacher called their names from a list in order to record attendance.
After this, the students accessed the Moodle e-learning platform which contained teaching materials
which they were supposed to learn within 45 minutes. All students in the control group used the
usual strategy of re-reading the materials. After the time ran out, the students took a test on the
Moodle e-learning platform.
The experimental group had attendance recorded through the smart classroom using individually
generated QR codes on students’ mobile phones. That is, every student had their own QR code sent
to them via e-mail several days before the experiment (Figure 4).
The teacher records attendance using the application Orca Scan (“Orca Scan: Cloud Sheets,”
2018) by reading every student’s QR code. The application Orca Scan was chosen among a
large number of QR reading applications because, aside from code reading, it offers an array of
other options such as uploading data to the application’s cloud platform, synchronizing infor-
mation in real time with other cloud platforms such as Google Spreadsheets, and providing
access to data to multiple users. In addition to this, the application is compatible with both
Android and iOS devices and navigating the application doesn’t require additional training. The
scanned information can be transferred from a mobile phone to a Google Spreadsheet account,
stored in the Orca Scan cloud platform and then exported or automatically integrated into external
cloud platforms (Figure 5).

Figure 3. Information flow diagram.

 INTERACTIVE LEARNING ENVIRONMENTS 9

Figure 4. QR codes for identifying students.

After the information about a student were read and transferred from Orca Scan to Google Spread-
sheets, the ambient intelligence system compared the collected information with the information
about class schedules in the database. The ambient intelligence system was realized by forming
lookup tables and writing appropriate scripts in the Google Spreadsheet platform. In this way, the infor-
mation about the students’ daily activities was gained. According to the Digital Telecommunications
teacher’s previous experience, it was noted that students with a higher amount of curricular activities
during the day had different levels of fatigue. Usually, before this class students can have 0, 1–3, or 4–6
other classes. Table 1 shows the relationship between the number of classes which impact the level of
fatigue and the recommended learning strategy assigned to a student (Table 1).
The ambient intelligence system generated an e-mail notification (Figure 6.) and sent it to every
student informing them of the learning strategy they should apply. Students accessed the e-learning
Moodle platform and, over the next 45 minutes, studied in accordance with the learning strategy they
had been assigned. After this, the students took a test on the Moodle e-learning platform.

Figure 5. The Orca Scan application on a desktop (“Orca Scan: Cloud Sheets,” 2018).
10 V. RADOSAVLJEVIC ET AL.

Table 1. Distribution of learning strategies according to students’ class schedules.

The number of classes a student had before the class in which the experiment was conducted Learning strategy
4 or more Work with teacher
1–3 classes Group work
0 Rereading materials

Study materials were distributed among students using the platform Moodle. All students (control
group and experimantal group) were tasked with studying the same unit consisting of two parts:

. Theory presented in the form of textual materials and multimedia content (images and
animations).
. Practical examples presented in the form of video recordings and problem-solving tasks concern-
ing the study materials. This part of the unit helps the student construct a mental model of knowl-
edge by providing answers for and comments on various questions.

5. Results
This study had the goal of determining whether the ambient intelligence concept-based smart
classroom model has a positive impact on learning outcomes. Students in the experimental group

Figure 6. The generating of the e-mail notification about learning strategies.

 INTERACTIVE LEARNING ENVIRONMENTS 11

Figure 7. Final test results.

were assigned learning strategies in accordance with their estimated level of fatigue. Fatigue was
assessed based on their previous academic activities as specified in their class schedules. The
success in reaching learning goals was measured with a test which students in both groups took
after the class.
By comparing the results in the final test between the two groups (Figure 7), it was determined
that the students in the experimental group performed better. The highest possible grade in the
test was 10. The average grade in the control group was 7.10, while the average grade in the exper-
imental group was 8.70. The results of the t-test showed the significance value .000 in both tests (p <
0.05), which confirmed that the mean values of the points achieved in both groups were statistically
different. The results of the research were processed using the IBM SPSS software.
Aside from the analysis of final test results, a comparison was made between grades and
the amount of daily curricular activities of students in both groups. Students from both groups
were divided according to their previous academic activities during the day in the form of the above-
mentioned number of classes (4 or more classes, 1–3 classes, or no classes). The numbers of students
in the control and experimental group in all of these categories varied to the minimal extent.
(Figure 8).
The greatest difference in performance was visible in students who had 4 or more classes, where
students in the experimental group had a significantly higher average grade than students in the
control group (with the same amount of previous activities) who used their usual learning strategy
of re-reading (Figure 9).

6. Discussion and conclusion

The ambient intelligence concept-based smart classroom model has the goal of showing the impact
of the right choice of a learning strategy on the student’s success, based on the information ambient
intelligence processes and distributes through the smart classroom system.
The results reached in this study can be viewed from two angles:

. By analyzing the information available through ambient intelligence, it is possible to increase stu-
dents’ performance in learning and realizing goals;
. By analyzing the information available through ambient intelligence, it is possible to utilize the
smart classroom to provide students with the adequate learning strategy in accordance with
the criteria compatible with the expected leaning outcomes.
12 V. RADOSAVLJEVIC ET AL.

Figure 8. Number of students per group according to previous daily academic activities.

Figure 9. Student performance in the final test depending on their previous daily academic activities.

Disadvantages of the model mainly have to do with defining adequate learning strategies in
accordance with teaching materials. The choice of strategies can be made based on other criteria,
as well as based on the mutual influence of several criteria. This study took into consideration only
the students’ level of fatigue as a result of their previous academic activities during the day. In
 INTERACTIVE LEARNING ENVIRONMENTS 13

addition to this, the sample of the study was relatively small. The organization of the smart classroom,
as well as the accessibility of information which ambient intelligence can properly analyze, vary from
one educational institution to another.
The model uses information about the students’ level of fatigue determined based on realized
daily academic activities, which served as the parameter for choosing an appropriate learning strat-
egy. For the purposes of testing the model in the study, three learning strategies were chosen in
accordance with cognitive activities adjusted to the students’ level of fatigue. Further development
of the model should investigate the effects of using other learning strategies (such as learning
through practical use or the flipped classroom model) and other criteria for choosing the appropriate
strategy.
This paper is only one way of presenting the model which uses the information from the environ-
ment gathered and processed by the ambient intelligence system to create learning strategies which
can be realized within the smart classroom environment. Future research should test the model using
other learning strategies, investigate different criteria for student fatigue estimation and test possible
integration with other learning platforms and systems.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributors
Vitomir Radosavljevic received his BS and MS degree in Electrical engineering and telecommunication technology from
the Faculty of Electrical Engineering, University of Belgrade, in 2006. He is a PhD candidate at the Faculty of Organis-
ational Science at the Information systems and quantitative management program, University of Belgrade. Since 2007
he has been working at ICT College in Belgrade as a lecturer. His work and interests are related to digital multimedia,
new educational technology design, knowledge management.
Slavica Radosavljevic received her PhD at the Faculty of Organisational Science at the Information systems and quanti-
tative management program, University of Belgrade. Her BS and MS degree in Postal and telecommunication traffic and
network from the Faculty of Transport and Traffic Engineering, University of Belgrade, in 1997. She works at ICT College in
Belgrade as a lecturer since 2002. Her work and interests are related to postal technology, augmented reality and new
educational technology.
Gordana Jelic received her MA and PhD degrees from the Faculty of Philology, University of Belgrade. She works at the
ICT College in Belgrade since 2001 as a professor. Her research interests are methodology of English language teaching,
new educational technologies, discourse analysis, academic writing, metadiscourse.

ORCID
Vitomir Radosavljevic http://orcid.org/0000-0003-0597-1095
Slavica Radosavljevic http://orcid.org/0000-0002-7285-0210

References
Al-Hemoud, A., Al-Awadi, L., Al-Rashidi, M., Rahman, K., Al- Khayat, A., & Behbehani, W. (2017). Comparison of indoor air
quality in schools: Urban vs. Industrial ’oil & gas’ zones in Kuwait. Building and Environment, 122, 50–60. doi:10.1016/j.
buildenv.2017.06.001
Alelaiwi, A., Alghamdi, A., Shorfuzzaman, M., Rawashdeh, M., Hossain, M., & Muhammad, G. (2015). Enhanced engineering
education using smart class environment. Computers in Human Behavior, 51, 852–856. doi:10.1016/j.chb.2014.11.061
Ambient intelligence. (2019). Retrieved from https://en.wikipedia.org/wiki/Ambient_intelligence
Ambient intelligence – The ultimate IoT use cases | IoT for all. (2019). Retrieved from https://www.iotforall.com/ambient-
intelligence-ami-iot-use-cases/
Boksem, M., Meijman, T., & Lorist, M. (2005). Effects of mental fatigue on attention: An ERP study. Cognitive Brain Research,
25(1), 107–116. doi:10.1016/j.cogbrainres.2005.04.011
14 V. RADOSAVLJEVIC ET AL.

Bravo, J., Cook, D., & Riva, G. (2016). Ambient intelligence for health environments. Journal of Biomedical Informatics, 64,
207–210. doi:10.1016/j.jbi.2016.10.009
Brusilovsky, P., & Peylo, C. (2003). Adaptive and intelligent web-based educational systems. International Journal of
Artificial Intelligence in Education, 13, 156–169.
Callender, A., & McDaniel, M. (2009). The limited benefits of rereading educational texts. Contemporary Educational
Psychology, 34(1), 30–41. doi:10.1016/j.cedpsych.2008.07.001
Castilla, N., Llinares, C., Bravo, J., & Blanca, V. (2017). Subjective assessment of university classroom environment. Building
and Environment, 122, 72–81. doi:10.1016/j.buildenv.2017.06.004
Chen, H. (2014). Intelligent classroom attendance checking system based on RFID and GSM. Advanced Materials Research,
989-994, 5532–5535. doi:10.4028/www.scientific.net/amr.989-994.5532
Chin, C. (2006). Classroom interaction in science: Teacher questioning and feedback to students’ responses. International
Journal of Science Education, 28(11), 1315–1346. doi:10.1080/09500690600621100
Cook, D., Augusto, J., & Jakkula, V. (2009). Ambient intelligence: Technologies, applications, and opportunities. Pervasive
and Mobile Computing, 5(4), 277–298. doi:10.1016/j.pmcj.2009.04.001
Cottone, P., Maida, G., & Morana, M. (2014). User activity recognition via Kinect in an ambient intelligence scenario. IERI
Procedia, 7, 49–54. doi:10.1016/j.ieri.2014.08.009
Deci, E. L., Koestner, R., & Ryan, M. R. (1999). A meta-analytic review of experiments examining the effects of extrinsic
rewards on intrinsic motivation. Psychological Bulletin, 125, 627–668.
Donker, A., de Boer, H., Kostons, D., Dignath van Ewijk, C., & van der Werf, M. (2014). Effectiveness of learning strategy
instruction on academic performance: A meta-analysis. Educational Research Review, 11, 1–26. doi:10.1016/j.edurev.
2013.11.002
Dunlosky, J., Rawson, K., Marsh, E., Nathan, M., & Willingham, D. (2013). Improving students’ learning with effective learn-
ing techniques. Psychological Science in the Public Interest, 14(1), 4–58. doi:10.1177/1529100612453266
Erdogan, I., & Campbell, T. (2008). Teacher questioning and interaction patterns in classrooms facilitated with differing
levels of constructivist teaching practices. International Journal of Science Education, 30(14), 1891–1914. doi:10.
1080/09500690701587028
Fiorella, L., & Mayer, R. E. (2013). The relative benefits of learning by teaching and teaching expectancy. Contemporary
Educational Psychology, 38, 281–288.
Furtak, E., & Kunter, M. (2012). Effects of autonomy-supportive teaching on student learning and motivation. The Journal
of Experimental Education, 80(3), 284–316. doi:10.1080/00220973.2011.573019
Gómez, J., Huete, J., Hoyos, O., Perez, L., & Grigori, D. (2013). Interaction system based on Internet of things as support for
education. Procedia Computer Science, 21, 132–139. doi:10.1016/j.procs.2013.09.019
Guinard, D., Fischer, M., & Trifa, V. (2010). Sharing using social networks in a composable web of things. 2010 8th IEEE inter-
national conference on pervasive computing and communications workshops (PERCOM Workshops). doi:10.1109/
percomw.2010.5470524
Guo, B., Zhang, D., Wang, Z., Yu, Z., & Zhou, X. (2013). Opportunistic IoT: Exploring the harmonious interaction between
human and the internet of things. Journal Of Network And Computer, 36(6), 1531–1539. doi:10.1016/j.jnca.2012.12.028
Jurik, V., Gröschner, A., & Seidel, T. (2014). Predicting students’ cognitive learning activity and intrinsic learning motiv-
ation: How powerful are teacher statements, student profiles, and gender? Learning and Individual Differences, 32,
132–139. doi:10.1016/j.lindif.2014.01.005
Kanjug, I., & Chaijaroen, S. (2012). The design of web-based learning environments enhancing mental model construc-
tion. Procedia – Social and Behavioral Sciences, 46, 3134–3140. doi:10.1016/j.sbspro.2012.06.025
Kassim, H., Nicholas, H., & Ng, W. (2014). Using a multimedia learning tool to improve creative performance. Thinking Skills
and Creativity, 13, 9–19. doi:10.1016/j.tsc.2014.02.004
Kumpas-Lenk, K., Eisenschmidt, E., & Veispak, A. (2018). Does the design of learning outcomes matter from students’ per-
spective? Studies in Educational Evaluation, 59, 179–186. doi:10.1016/j.stueduc.2018.07.008
Li, B., Kong, S., & Chen, G. (2015). Development and validation of the smart classroom inventory. Smart Learning
Environments, 2(1). doi:10.1186/s40561-015-0012-0
Liu, N., & Littlewood, W. (1997). Why do many students appear reluctant to participate in classroom learning discourse?
System, 25(3), 371–384. doi:10.1016/s0346-251x(97)00029-8
Marcora, S., Staiano, W., & Manning, V. (2009). Mental fatigue impairs physical performance in humans. Journal of Applied
Physiology, 106(3), 857–864. doi:10.1152/japplphysiol.91324.2008
Mihai, T., & Iordache, V. (2016). Determining the indoor environment quality for an educational building. Energy Procedia,
85, 566–574. doi:10.1016/j.egypro.2015.12.246
Muelas, A., & Navarro, E. (2015). Learning strategies and academic achievement. Procedia - Social and Behavioral Sciences,
165, 217–221. doi:10.1016/j.sbspro.2014.12.625
Müller, T., & Apps, M. (2019). Motivational fatigue: A neurocognitive framework for the impact of effortful exertion on
subsequent motivation. Neuropsychologia, 123, 141–151. doi:10.1016/j.neuropsychologia.2018.04.030
Nota, G. (2011). Risk management trends. London: InTech.
Orca Scan: Cloud Sheets. (2018). Retrieved from https://cloud.orcascan.com/
 INTERACTIVE LEARNING ENVIRONMENTS 15

Persky, A. (2018). A four year longitudinal study of student learning strategies. Currents in Pharmacy Teaching and
Learning, 10(11), 1496–1500. doi:10.1016/j.cptl.2018.08.012
Qi, P., Ru, H., Gao, L., Zhang, X., Zhou, T., Tian, Y., … Sun, Y. (2019). Neural mechanisms of mental fatigue revisited: New
insights from the brain connectome. Engineering. doi:10.1016/j.eng.2018.11.025
Ramos, C., Augusto, J., & Shapiro, D. (2008). Ambient intelligence—The next step for artificial intelligence. IEEE Intelligent
Systems, 23(2), 15–18. doi:10.1109/mis.2008.19
Roscoe, R. D., & Chi, M. T. H. (2007). Understanding tutor learning: Knowledge-building and knowledge-telling in peer
tutors’ explanations and questions. Review of Educational Research, 77(4), 534–574. doi:10.3102/0034654307309920
Roscoe, R. D., & Chi, M. T. H. (2008). Tutor learning: The role of explaining and responding to questions. Instructional
Science, 36, 321–350. 10.1007/s11251-007-9034-5
Sala, E., & Rantala, L. (2016). Acoustics and activity noise in school classrooms in Finland. Applied Acoustics, 114, 252–259.
doi:10.1016/j.apacoust.2016.08.009
Santana-Mancilla, P., Echeverría, M., Santos, J., Castellanos, J., & Díaz, A. (2013). Towards smart education: Ambient intelli-
gence in the Mexican classrooms. Procedia - Social and Behavioral Sciences, 106, 3141–3148. doi:10.1016/j.sbspro.2013.
12.363
Sedighi Maman, Z., Alamdar Yazdi, M., Cavuoto, L., & Megahed, F. (2017). A data-driven approach to modeling physical
fatigue in the workplace using wearable sensors. Applied Ergonomics, 65, 515–529. doi:10.1016/j.apergo.2017.02.001
Seidel, T., & Shavelson, R. (2007). Teaching effectiveness research in the past decade: The role of theory and research
design in disentangling meta-analysis results. Review Of Educational Research, 77(4), 454–499. doi:10.3102/
0034654307310317
Shen, C., Wu, Y., & Lee, T. (2014). Developing a NFC-equipped smart classroom: Effects on attitudes toward computer
science. Computers in Human Behavior, 30, 731–738. doi:10.1016/j.chb.2013.09.002
Sibley, J., & Parmelee, D. (2008). Knowledge is no longer enough: Enhancing professional education with team-based
learning. New Directions for Teaching and Learning, 2008(116), 41–53. doi:10.1002/tl.332
Soliman, A. (2017). Appropriate teaching and learning strategies for the architectural design process in pedagogic design
studios. Frontiers of Architectural Research, 6(2), 204–217. doi:10.1016/j.foar.2017.03.002
Suo, Y., Miyata, N., Morikawa, H., Ishida, T., & Shi, Y. (2009). Open smart classroom: Extensible and scalable learning system
in smart space using web service technology. IEEE Transactions on Knowledge and Data Engineering, 21(6), 814–828.
doi:10.1109/tkde.2008.117
Tibúrcio, T., & Finch, E. (2005). The impact of an intelligent classroom on pupils’ interactive behaviour. Facilities, 23(5/6),
262–278. doi:10.1108/02632770510588664
Twomey, N., Diethe, T., Craddock, I., & Flach, P. (2017). Unsupervised learning of sensor topologies for improving activity
recognition in smart environments. Neurocomputing, 234, 93–106.
Uzelac, A., Gligorić, N., & Krčo, S. (2018). System for recognizing lecture quality based on analysis of physical parameters.
Telematics and Informatics, 35(3), 579–594. doi:10.1016/j.tele.2017.06.014.
van Leeuwen, A., Janssen, J., Erkens, G., & Brekelmans, M. (2015). Teacher regulation of cognitive activities during student
collaboration: Effects of learning analytics. Computers & Education, 90, 80–94. doi:10.1016/j.compedu.2015.09.006
Vilčeková, S., Kapalo, P., Mečiarová, L., Burdová, E., & Imreczeová, V. (2017). Investigation of indoor environment quality in
classroom – case study. Procedia Engineering, 190, 496–503. doi:10.1016/j.proeng.2017.05.369
Wang, F., & Hannafin, M. (2005). Design-based research and technology-enhanced learning environments. ETR&D, 53(4),
5–23. doi:10.1007/bf02504682
Yang, J., & Huang, R. (2015). Development and validation of a scale for evaluating technology-rich classroom environ-
ment. Journal Of Computers In Education, 2(2), 145–162. doi:10.1007/s40692-015-0029-y