Mobile Application for Smart Parking Systems

based on Arduino Uno

Danny Santoso Muhammad Sudais Hakim Muhammad Naufal Rachfian

Information System Department, Information System Department, Djamhur
School of Information System School of Information System Information System Department,
Bina Nusantara University Bina Nusantara University School of Information System
Jakarta, Indonesia 11480 Jakarta, Indonesia 11480 Bina Nusantara University
[email protected] [email protected] Jakarta, Indonesia 11480
[email protected]
Muhammad Luthfi Hidayat Harco Leslie Hendric Spits
Information System Department, Warnars
School of Information System Computer Science Department, BINUS
Bina Nusantara University Graduate Program – Doctor of
Jakarta, Indonesia 11480 Computer Science
[email protected] Bina Nusantara University
Jakarta, Indonesia 11480
[email protected]

Abstract— Smart parking systems is an application that can a developer. Therefore it is possible to create a system that is
help user to get information about the available parking space more efficient in a parking system. Especially with the current
in someplace. Within smart parking systems, there are technological devices that are increasingly cheap every day.
implications of several technologies that can solve problems that
occur in parking systems. The technologies involved in smart In the parking system that we wrote, we provide our
parking systems, include neural network and internet of things. premise regarding the analysis in designing this parking
The use of the internet of things is done by connecting the system. As well as providing details about available parking
scattered sensor devices in each parking space to the internet spaces in a parking area that is visualized through a mobile
network then the data from each sensor will be stored on the application that can be accessed by everyone.
server. To connect sensors to the internet network requires
component devices such as esp8266-01. This tool helps Arduino In the development of smart parking systems, it is
to connect to the internet network. so that, it can facilitate necessary to integrate sensor devices in one parking area with
communication that occurs between sensor devices and servers the internet or the term we know as the internet of things.
to be able to store parking space data into the database. which Besides using the internet of things technology. We can
will be accessed by the user to obtain information about the also apply neural network modeling technology as an
parking space. In addition to the use of neural networks, it is
additional feature in this smart parking system. The neural
useful to know how many vehicles visit the same place visited by
users.
network model in this system is useful for predicting or
estimating the number of visitors who are driving to the same
Keywords— Smart Parking System, Neural Network, Internet place as the intended user.
of Things, Arduino Uno, esp8266-01 Therefore this can help users to find out which places are
I. INTRODUCTION being targeted by other users. So that users can find alternative
places that are less visited to go to. This additional feature
The problem that often occurs in a parking system is the might be useful for users in estimating availability parking
lack of details regarding the description of available parking space. Due to the existence of data regarding the location of
space. Although the parking system is now quite helpful in the empty parking space and data of users who are driving
providing number of available parking space. however, some towards the intended location, the user can estimate the
people are still confused about where is the available parking opportunity to obtain the parking space.
space and which floor there is available parking space. So this
makes some people seek to every basement floor and every II. PREVIOUS RESEARCH
basement alley to look for the available parking area. This lack In here we will discuss the previous research that has been
of efficiency makes the problems that arise in the parking conducted by the previous researchers which helps us in
system today. completing the research we are going to do with Smart parking
To make an efficient parking system very possible with system.
current technology, because the current technology has The previous research that has been done by Galih Raditya
developed very widely as quoted by Mark Weiser in a paper in his paper entitled Arduino Uno-based smart parking is
[1] which provides a vision of the development of an almost similar to the research that we made, but in the research
environment that is adapted to help humans in doing an he made, he used photodiode sensors and infrared. Where the
activity. Where the vision in question is ubiquitous photodiode sensor is a sensor that is sensitive to the light.
computing. Where the workings of the detection carried out in this study
Current technological developments have given breadth to by utilizing the sensitivity of light that covers this sensor
computing machines in carrying out an action or order device caused by objects or vehicles that are above it which
regarding a particular task that has been designed or given by results in resistance when receiving light intensity. This allows
the sensor to provide data to Arduino that there are objects that B. Arduino Uno Project Board
are above it [14]. This is different from the research that we Arduino UNO is a computing tool or minicontroller that is
conducted where the research that we conducted used open-source as a simple input or output board and can
ultrasonic sensors. implement processing languages with the tool that is
In addition, there is also research that has been carried out connected to it [4]. In Arduino Uno, we can use USB to
by Achmad Muhajir in his paper where they make a connect to the computer for uploading the sketch or code to
monitoring system using ultrasonic sensors but in the the Arduino Uno [19]. This tool is usually used to help
implementation carried out still has a limitation, where to develop a project in the field of IT or technology. Arduino can
send data obtained by the sensor requires the help of a be connected with various tools such as sensors, displays,
computer to store data which will be stored on the database. drives and so on.
Then with this computer, data is sent to social media networks In the case of this smart parking system, Arduino is useful
such as Twitter by using the internet to monitor parking as a receiver of sensor data produced by sensor devices that
locations from a distance [15]. This is different from the detect or identify the availability of parking space.
research that we made where we will apply additional
components to Arduino Uno, namely ESP8266 which is a tool C. Sensor PING HC-SR04
to connect Arduino Uno to the internet without the help of PING sensor is an ultrasonic sensor that can identify or
computers that send data to a server via the internet network detect an object at a certain range. How these tool works use
where users can access via mobile application. Our application the principle of reflecting a sound wave with a certain
involves the Internet of Things technology where ESP8266 is frequency [6]. In addition, this tool is similar to RFID because
one of the component that supports this Internet of Things to detect an object, this tool emits ultrasonic waves to an
technology. object. then the wave reflection is detected against the object.
In addition, there is also research conducted by Bakhtiyar Where the resulting data from the wave reflection is
Arasada which explains the use of Arduino in determining the transmitted to the Arduino Uno project board.
distance in a room by using an ultrasonic sensor or HC-SRF04 In addition, ultrasonic waves on the PING sensor are
sensor. Where the sensor used in determining the distance of created by an object called piezoelectric. Piezoelectric
an object is the same as the sensor we used in our study. The produces an ultrasonic wave with a frequency of around 40
research he did help our research in understanding the kHz to detect objects around it [10][16].
concepts carried out by ultrasonic sensors in determining
distances by giving waves to an object. Where in the research D. Arduino IDE
he did provide a formula that can be used in determining the Arduino IDE (Integrated Development Environment) is a
distance on an object carried out by an ultrasonic sensor [16]. software that is useful for giving commands to the Arduino
This helps us in determining the distance of an object which project board in managing input or output on a device
we specify as a specific distance for an object as an indicator connected to it. In addition, the Arduino IDE is an open-source
that a parking space is filled or not. software, so we can easily contribute to developing this
software and can easily get the source code that was deployed
All the previous research that has been conducted by
by Arduino developers [18].
various researcher help us to improve the system parking that
is not efficient to be more efficient. With previous research, it Where this command can be done by writing the code on
is easier for us to determine what tools and components need the Arduino IDE. This encoding in the Ardiuno IDE is called
to be used in improving the parking system today to be better. sketch. Where Arduino IDE is made from the JAVA
programming language.
III. THEORETICAL BASIS
Besides that, the Arduino IDE is equipped with many
A. Internet of Things libraries that were formed of C or C++ language also known
IoT also known as the Internet of Things is a set of as “wiring” which is useful for making input and output
technologies that can access data that is collected by various operations easier and more controlled.
devices or objects connected via wireless or internet [2]. With
this IoT, all activities are carried out by using the internet [9]. E. Neural Network
So with the existence of the Internet of things, all data that is Neural Network or commonly referred to as an artificial
in an object or object connected to the internet can be accessed neural network is a computational method that can be used to
through our mobile devices. predict or estimate something. Where this artificial neural
network has the same concept as the way of the human brain
In this smart parking system, the sensor that is useful in works in processing information in solving a problem or task
providing data by identification of parking space must have [5]. This neural network is inspired by the neural network of
integration to the internet. So, the parking space data can be the human brain, where each neuron in the human brain is
accessed by users through the application. Where this sensor interconnected with other neurons. That way the Neural
device identifies objects or obstacle that are around it with Network has a level of complexity in each layer like that of
ultrasonic waves. the human nervous system [7].
As stated by Kevin Asthon in a journal "RFID and sensor F. Deep Belief Network
technology enables computers to observe, identify, and
understand the world without the limitations of data entered Deep Belief Network or commonly known as DBN, was
by humans" [3]. introduced by Hinton [8]. Is a probability of various layers.
DBN is formed using Greedy Layer-Wise (GLW) which is
used to measure the complexity and depth of network layers.
DBN is used to extract from the input layer. DBN usually
consists of many numbers to be included in a more complex From the description of the network diagram above, the
formula. workings of this system are that every available parking space
in each basement will be given an ultrasonic sensor or PING
G. Matlab sensor, this sensor will detect the distance of an objects in the
Matlab or stands for Matrix Laboratory is an advanced area by emitting waves or ultrasonic frequencies which will
mathematical programming language and a program that is detect reflections produced by the object. then connected with
useful for analyzing and computating numeric data. Where the Arduino Uno project board to transmit data obtained by
matlab is formed based on the nature and shape of the matrix. the ultrasonic sensor hardware. Then from this data, it is
Matlab also provide tools for design, implement and analyzed as parking space that are available and entered into a
stimulate neural network for various purpose. That can be database that can be accessed by the user. To send sensor data
used to create artificial neural networks for the mindset or to the database requires the involvement of the esp8266-01
paradigm to the system in carrying out a task [13]. component, to connecting sensor device to the network of the
server exist. this makes the data from the sensor can be
H. ESP8266-01 transferred to the database.
ESP8266-01 is a serial wifi transceiver module that used To be able to access data in this system the user needs to
to integrate an object that is connected with esp8266 into an use a smartphone. Where the smartphone requires an internet
internet network so as to facilitate the integration of an object connection in order to access the parking space data available
into the internet network. This module can give access to the in the system.
internet connection because, in this module it has an
integration with TCP/IP that can give access microcontroller B. Activity Diagram
to the Wifi network [17].
In this paper we use this esp8266 as an Arduino connection
to the internet network that allow the sensor to transmit data
to the database server that has a same connection with the
arduino. in configuring esp8266 we need to initialize the name
of the access point and also the password so that it can be
connected to the wifi router on a network [11][12].
With ESP8266-01, it makes it easier for an object or sensor
to transmit data over the internet, so ESP8266 can be a very
important tool in developing the Internet of Things [20].
IV. RESEARCH METHOD
To get the data that supports this writing, we conducted
several research methods, first we observed several parking
lots inside a shopping mall and office buildings such as how
they provided information about the available parking space
and the parking system that was carried out. second we
conduct litelatur studies or litelatur review related to parking
systems, sensor devices and ways of thinking in the neural
network.
V. RESULT AND DISCUSSION
To give a clear picture of this parking system we provide
several diagrams, models and user interfaces as an illustration Fig. 2. Activity Diagram for Smart Parking System
of how the parking system that we wrote works. This design
analysis is based on our premise taken from several literature The diagram above is from the user's perspective when
references that are available online or in journals. searching for available parking through a mobile application
that we develop later, where the user just needs to enter the
A. Network Diagram application and look for the location or area they want to visit,
but the place that the user wants to visit must have integrated
SENSOR
with our application or support our application. After that the
ULTRASONIC
user can search for locations by writing the location in the
search textfield that has been provided.
INTERNET
Then the system will look for an available location in the
FIREWALL ROUTER SWITCH APPLICATION SERVER
database. If the location is not available or full, the system will
display a notification that the parking space is full or there is
no empty space. But if the system finds available space the
system will display the available basement floor first, which
later the user will choose the floor and the system will display
DATABASE
MOBILE
a more specific location.

Fig. 1. Network Diagram for Smart Parking System

C. ER-Diagram

Fig. 3. ER-Diagram for Smart Parking System

From this ER-diagram it explains that the user can see

some parking logs that have been updated or entered
automatically by the sensor hardware. where parking log data
that is provided including location and parking space.
D. User Interface

Fig. 5. User Interface for Choose Floor Menu

After that the system will display the available parking

space and the amount based on the existing floor. To get
specific information the user must choose one of the available
floors.

Fig. 4. User Interface for Find Destination Menu

This is a display for the user when entering the search

menu to determine the destination location of the user wants
to go. where the user needs to enter the address of the
destination that you want to go to in the textfield search above
and then maps will display the location. The button will be
able to be clicked if the location intended by the user is
supported or integrated with this smart parking system
application.

Fig. 6. User Interface for Specific Space Menu

In the last menu, it provides more detailed parking space

information according to the floor chosen by the user. Where
the red colored box means that the space is not available or
has been filled by another vehicle. And the green box means
that the space is available.
 Where the available and filled parking space data is board gives the command to turn off the LED as an indicator
obtained from the ultrasonic sensor which has detected or that the location is filled and gives the value to the variable
identified the parking area. The identification results from the space that is not available. that is later stored to the database
ultrasonic sensor are then transmitted to the Arduino project with the involvement of the esp8266-01 tool which is useful
board for processing and stored into a database. so that the data to help Arduino connect to the internet to be able to access
provided by the ultrasonic sensor can be visualized into a the server which is useful for storing data in the database that
mobile application. Where the user can access the details of will be accessed from the mobile application to display a
parking space in the form of an image such as in the user visualization of which locations are available and not
interface display above.
available.
The user also can mark their location by clicking or
tapping the green or red box in order to mark their location
and then the system will change the box color that was clicked
or tapped by the user into blue color.Tthis can be an indicator
for the user where they parked their vehicle. So user can know
or recognize where they parked their vehicle by looking at the
application, and it also helps the user to remind if they forget
where they parked their vehicle.
With all the functions or features that exist in these
applications make this system have a big advantage that can
be applied in the current parking system to make it more
efficient and useful for users, so that it can satisfy the users or
visitors that want to go some place or location.
E. Ardiuno Code Fig. 8. Output from ultrasonic sensor

In the picture above, Where Arduino is looping data

collection every 1 second, in order to get real-time or fast and
accurate data. Then stored in the database in the form of
written or string values in a variable or attribute that is useful
as an indicator or status of the location or space available or
not available.

Fig. 9. Development Sensor Ultrasonic with Arduino Board

In the picture above, it is explained that if an ultrasonic

sensor is not blocked by an object or obstacle the LED will
light up, but if the ultrasonic sensor is blocked by an object the
LED will automatically turn off, according to the code written
in the Arduino IDE.
Where this ultrasonic sensor takes data every 1 second,
like the one governed in the previous code where by using
function delay (1000) which means to delay for 1000
Fig. 7. Source Code to get sensor data milliseconds or 1 second.
Here is the Arduino source code to get data from an
ultrasonic sensor, where in this code insert an ultrasonic
sensor as input and led as output and provide one variable to
store location status. In this code, the distance between the
sensor and the vehicle in the parking space will be
determined. If a vehicle is at a specified distance, the Arduino
F. Neural Network Model

Fig 12. Ouput from Artificial Neural Network

From the picture above is an example of the output

Fig. 10. Neural Network Model for Smart Parking System produced by each GPS coordinate that is passed by the user, if
this neural network model used to predict the number of we see the output results are still numeric but if we pay more
users who are driving to some destination. Where in this attention to that number, these number are probability number
model the parameters used to be input are GPS coordinate produced by the artificial neural network, where if we see the
point that has been passed by the user. probability number which has a number 1 or close to 1 it
indicates true or appropriate. And these row of table indicates
Where on the way to a place, the system has determined the specific location to which it was intended.
which coordinates will be taken if a user passes that point.
Then the coordinates of the point that has been passed by this For example, we categorize row 1 towards Bina Nusantara
user will be used as input to be included in this neural network University, line 2 goes to Mall Kelapa Gading, and row 3 goes
model. to Central Park Mall. So if we look at the output from the 6
input previous GPS coordinate point locations.
The output produces 0.9 or close to 1 in the first row,
where we have determined the first row as the location of Bina
Nusantara University. So from the artificial neural network
above, the system can be sure that the user goes to Bina
Nusantara University. Because numbers close to one are the
right or correct values.
Fig. 11. Coordinate GPS point that has been obtain

For the example, these data are obtained from the user
when driving through several GPS points that have been
determined, then the coordinate point will be taken from each
point that has been passed by the user as an input to be
managed into the mindset or paradigm of the artificial neural
network.
Later in the hidden layer will be included a formula or
algorithm to estimate the location target. Then the output
produced is a probability value to give the value of accuracy
or suitability to the final location.
This hidden layer is used to improve the ability to predict
and solve problems for reading on the input layer, the
consequences in using this hidden layer are difficult
management or better known as Deep Belief Network.
In the hidden layer, there is a deep belief network that
contains specific formulas, which will determine the mindset
or paradigm of the location of the destination or the name of
the destination being targeted by other users. This is done by
giving a probability value in each name of the destination
available at the output layer, where the row that has a
probability value of 1 or close to 1, then the answer from that
column is right or is being addressed by another user.

Fig. 13. display of UI that has implemented a Neural Network

This is the user interface that has been added to the neural
network model as an additional feature where the system will
provide a little information under the button regarding
information of people who are driving to a destination. So that
users can consider visiting these locations.
 VI. CONCLUSION [9] R. Khan, S. U. Khan, R. Zaheer and S. Khan, "Future Internet: The
Internet of Things architecture possible applications and key
The conclusions obtained from the development of this challenges", Proc. 10th Int. Conf. FIT, pp. 257-260, 2012.
application system are based on our premise. That our mobile [10] D. S. Vidhya, D. P. Rebelo, C. J. D'Silva and L. W. Fernandes,
applications for smart parking systems are very efficient for “Obstacle Detection using Ultrasonic Sensors,” International Journal
users when searching for parking space at the destination that for Innovative Research in Science & Technology, vol. 2, no. 11, pp.
316–320, 2016.
is being targeted. This could be an improvisation of the current
parking system to be a more efficient parking system and more [11] Y. H. Lin, “Novel smart home system architecture facilitated with
distributed and embedded flexible edge analytics in demand-side
implicate technology that is currently developing such as IoT management”, International Transaction on Electrical Energy Systems,
and the neural network. Where users only need to be oriented pp. 1-21, 2019.
on our application in finding parking lots with capitalized [12] D. Frisch, S. Reißmann, C. Pape, "An over the air update mechanism
mobile devices on each user. for ESP8266 microcontrollers", Proc. 12th Int. Conf. Syst. Netw.
Commun. (ICSNC), pp. 12-17, 2017.
REFERENCES [13] J. L. Epiphany, “Artificial Neural Network Design Flow for
[1] M. Weiser, “The Computer For 21st Century,” ACM SIGMOBILE Classification Problem using Matlab,” IJARBEST, vol. 1, no. 6, pp.
Mobile Computing and Communications Review, pp. 1, 1999. 22-25, 2015.
[2] J. Gubbi, R. Buyya, S. Marusic and M. Palaniswami, “Internet of [14] G. R. Pradana, “Smart Parking Berbasis Arduino Uno,” E-Jurnal Prodi
Things (IoT): A vision, architectural elements, and future directions,” Teknik Elektronika Edisi Proyek Akhir D3, pp. 1-9, 2015.
Future Generation Computer Systems, vol. 29, pp. 1645-1660, 2013. [15] S. A. Muhajir and S. Amini, “Sistem Monitoring Tempat Parkir
[3] K. Ashton, “That 'Internet of Things Thing,” RFID Journal, pp. 1, 2009. Dengan Sensor Ultrasonik Berbasis Arduino Uno pada Cibinong City
Mall,” SENIATI, pp. 350-355, 2016.
[4] P. Jamieson, "Arduino for teaching embedded systems. are computer
scientists and engineering educators missing the boat?", Proc. FECS, [16] B. Arasada and B. Suprianto, “Aplikasi Sensor Ultrasonik Untuk
pp. 289-294, 2011. Deteksi Posisi Jarak Pada Ruang Menggunakan Arduino Uno,” Jurnal
Teknik Elektro, vol. 6, no. 2, pp. 137-145, 2017.
[5] A. K. Zyen and R. H. Kusumodestoni, “Prediksi Kecepatan Angin
Menggunakan Model Neural Network Untuk Mengetahui Besar Daya [17] S. Gore, S. Kadam, S. Mallayanmath and S. Jadhav, “Review on
Listrik Yang Dihasilkan,” DISPROTEK, vol. 6, pp. 53-59, 2015. Programming ESP8266 with Over the Air Programming Capability,”
International Journal of Engineering Science, vol. 7, no. 4, pp. 6684-
[6] S. Adarsh, S. M. Kaleemuddin, D. Bose and K. Ramachandran, 6686, 2016.
"Performance comparison of Infrared and Ultrasonic sensors for
obstacles of different materials in vehicle/robot navigation [18] G. Lockridge, B. Dzwonkowski, R. Nelson and S. Powers,
applications", IOP Conference Series: Materials Science and “Development of a Low-Cost Arduino-Based Sonde for Coastal
Engineering, vol. 149, 2016. Applications”, SENSORS, vol. 16, no. 4, 2016.
[7] A. Y. Prathama, A. Aminullah and A. Saputra, “Pendekatan ANN [19] M. Ichwan, M. G. Husada and M. I. Arrasyid, “Pembangunan Prototipe
(Artificial Neural Network) Untuk Penentuan Prosentase Bobot Sistem Pengendalian Peralatan Listrik Pada Platform Android,” Jurnal
Pekerjaan Dan Estimasi Nilai Pekerjaan Struktur Pada Rumah Sakit Informatika, vol. 4, no. 1, pp. 13-25, 2013.
Pratama,” TEKNOSAINS, vol. 7, pp 1-82, 2017. [20] M. Meht, “ESP 8266: a breakthrough in wireless sensor networks and
[8] G.E. Hinton, S. Osindero, and Y.W. Teh, “A Fast Learning Algorithm internet of things”, International Journal of Electronics and
for Deep Belief Nets,” Neural Computation, vol. 18, no. 7, pp. 1527- Communication Engineering & Technology (IJECET), vol. 6, no. 8,
1554, 2006. pp. 07-11, 2015.