Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Iot based car parking management system with an application program

interface
To cite this article: prosper philip et al 2019 J. Phys.: Conf. Ser. 1228 012066

View the article online for updates and enhancements.

This content was downloaded from IP address 181.214.20.41 on 03/06/2019 at 14:04

International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Iot based car parking management system with an

application program interface
prosper philip1, sathish kumar selvaperumal2, ravi lakshmanan3, chandrasekharan
nataraj4
1,2,3,4,
Faculty, of Computing, Engineering and Technology, Asia Pacific University
E-mail:1pnguma@yahoo.com,2dr.sathish@apu.edu.my, 3ravi@apu.edu.my,
4
chander @apu.edu.my

Abstract—With the population on the rise, urban areas congestion continues to become a
major transportation problem. As number of cars increases, the demand for parking spaces
rises, thus causing insufficient availability of parking spaces. This paper presents a new cloud
based parking management system implemented on the Internet-of-Things technology network
architecture that increases user convenience and parking management.Drivers are able to view
parking availability online, and proceed to reserve a parking space. The parking system is
designed based on zone control architecture using NodeMCU, thus allowing easy management
and implementation. An eight hexadecimal character is generated for both entry and exit access
codes, and sent to the driver's email and mobile phone via SMS. Parking information are
relayed and stored in the server using RESTful API implemented using Laravel PHP
technology. Parking spot information like entry and exit times can be requested through the
API.

Index Terms—Internet, smart, system.

1. INTRODUCTION
Population growth is a great concern in urban areas. As the population grows, the number of car also
increases and the demand for parking spaces rises. Even though more parking facilities are built, there are still
several problems in utilization. A major problem involves the time wasted searching for parking places, which
not only causes road congestion and irate drivers searching for empty parking spaces, but also has a big negative
impact on the environment.
Clearly, a solution is needed that will be hassle free for the parking users. The solution proposed here
allows the driver to make a reservation online. The system will be monitored in a real time and will allow users
to make reservations for parking as well as make payments online. [1], [2]
For example, in Israel in 2016, the average time spent looking for parking is twenty (20) minutes each
time they park, fifty-five hours (55) is the average time a driver wastes each year looking for parking. Not only is
time wasted searching for parking, but thirty (30) percent of urban traffic is caused by drivers searching for
parking and four point eight billion (4.8) hours are wasted in traffic. [6]
In addition to the time spent searching for parking, three point nine (3.9) billion gallons of fuel every
year is wasted and this in turn causes pollution to our environment. [6] In a current survey, researchers have
found that in one year, searching for parking has created the equivalent of thirty-eight (38) trips around the
world, burning 1.7 lakh litre of fuel and producing 730 tons of CO 2. [9]
According to [6] about thirty-eight thousand and three hundred (38,300) drivers are killed, four point
four (4.4) million are seriously injured searching for parking space.
Parking systems based on plate number recognition have major problems considering that plate
numbers are different in terms of size, colour and type depending on countries. These systems are restricted to
seven alphanumeric characters, of which four are numbers and 3 letters, and can only analyse the plate number
with single row plate number. [8] Hence, there is a need to develop a system that is not dependent on the plate
numbers to recognize the vehicle.
Parking systems that utilize image-processing techniques are expensive because cameras are costlier
than sensors. They also have larger storage and processing requirements and more specifications. [3] These
systems are also susceptible to weather conditions, which confuse the images; however, this problem can be
addressed by filtering the images in a high-quality transform. [1], [2] Therefore, there is need to design a system
that is robust, affordable and that is not impacted by changing weather conditions.
Most of the existing parking systems have a management structure that uses ticket payment and vehicle guiding,
which results into employing many full-time management personnel and high maintenance and management
cost. [7]

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Many people to tend forget their parking space location, especially when they park their car in
unfamiliar lot or when the parking lot is big and this causes frustration and inconvenience to the parking user. [7]
It is always a problem for parking managers to decide on a system implementation because of fear of increased
cost, time and integration issues with their current system. This result in them taking a slower approach to
replacing old systems, which prolongs the problem [4]. This system aims to minimise integration issues and help
foster acceptance at a comfortable pace, hence properly managing the changeover process.
There is also wasted space in some current car parking locations. In these locations, there are more
parking spaces per vehicle, and most are empty most of the time. The need, location, and cost must be considered
and convenience-driven to increase sustainability [5].
Although smart parking solutions have been successful, there are no citywide solutions across the disjointed
public and private parking suppliers, thus making them uncoordinated [10]. If occupancy data exists, they lack
standardization or accessibility and do not have an integration API that would allow access of these data. Hence,
there is the need of having a universal API to make data more accessible and standardized [11], [12].

2. PROPOSED METHODOLOGY
2.1. Block Diagram
The designed system is composed of five levels, these are: sensor, network, middleware, application and
mechanical. The sensor and mechanical level exists in the parking lot, where they’re used to detect a vehicle, and
control entry and exit. Communication and transmission of information to the servers is done wirelessly, and the
middleware act as a bridge between the two systems. Fig.1 depicts critical functionalities of each sub-system.
The client parking system primarily handles vehicle detection, uploading of lot status to server, and configuring
the LEDs. The parking system server receives these information and processes it to determine the total available
parking in the specified location. The information is also saved as historical data, of which can be retrieved using
the API.

Fig.1: Block Diagram of the system implemented

The reservation system allows the user to make reservation, as it assigns an available parking spot,
along with entry and exit access codes to be used at the entrance gate. Confirmation of the reservation is sent via
SMS and email to the specified user mobile number and registered email.
The gate control system monitors and allows input of the access codes using a keypad. Verification and
validation of the code is done by querying the database server. A result is returned back to the controller; if the
status result is “1” the boom gate opens, else if “0” the gate remains closed whilst prompting a message to the
user.

2.2 Constructional details

This section describes the production and communication protocol processes for the project. In order to
ensure that all components and system algorithm is working, the system is designed and simulated using
PROTEUS. After verification and validation of the simulation process, the system is then assembled and tested
on a breadboard. When all system components and tests are finalized, the system is then implemented on a PCB
and final connectivity and continuity tests are conducted.
The server application developed in an Ubuntu 16.04 Server using Laravel Framework and is hosted via
Cloud9. It is designed using the following use-case scenarios: user is unique, email is used for registration, user
can make reservations, phone number is required to make reservation, entry and exit access codes are generated
upon successful reservation, access codes are sent via SMS & email.

2
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

User Registration & Login

The registration page implements a minimalist approach, where only crucial details like name, email,
password and acceptance of the T&C are collected to complete the registration. Upon successful registration, the
user is automatically logged in and can proceed to use the system.If the user has already registered, can directly
proceed in logging in with ease. Authentication is done to ensure the user’s credentials and access to the system
as shown in Fig. 2.

Fig. 2: User login authentication failed

Parking Space Reservation

In order to make a successful reservation, the user needs to enter their mobile phone number, check in and out
times, and the parking location zone, as shown in Fig. 3. Failure to do so, results in errors, that the user is
prompted to fix.

Fig. 3: Reservation of parking lot

Access Control
The access control system is comprised of an LCD, 4x4 keypad and servo motor that is attached to a
microcontroller that captures input from the user and forwards to the network module that performs verification
of the access code from the server.
There are two sets of gates, i.e. main entry and exit, and, reservation entry and exit gates as shown in Fig. 4. The
main entry and exit gates allow the user to park in the normal parking lots, and are required to proceed to the
reservation parking lots in order to enjoy the reservation.

Fig. 4: Access gate configuration

In the current setup, the system is always active, thus no need to be activated when a vehicle is present. The
system will automatically reset if it is idle for ten seconds after detecting input, and after either a successful or
failed attempt to access the gate.

3
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

The system is configured to capture a password length of eight characters. When less characters are
detected, it automatically issues an error, and dismisses the rest if more than eight. The number of permutations
for generating an eight character HEX password at random is as follows:

Safety mechanisms are in place to prevent use of the same access code on gate sets, i.e. when an entry access
code is used to enter the main entry gate, it can’t be used again for the main entry gates, but can be used in the
reservation entry gates. However, once the entry access code is used at both main and reservation gates, the
access code is purged and can’t be used again. Fig. 5 shows the function spec requirements used in the access
control system.

Fig. 5: Access control construction

Vehicle Detection at Parking Space

This subsystem is designed so as to detect when a vehicle is parked underneath the parking lot. It
employs a two-step verification system, which requires two set of conditions to be satisfied. An ultrasonic and
Hall Effect sensor are used to accurately detect the presence of a vehicle when their threshold have been met.
The Hall Effect sensor is placed underneath the lot, whilst the ultrasonic sensor is placed on the ceiling,
perpendicularly above the parking space. The sensing distance reduces with increase in vehicle height as shown
in Fig. 6. The ultrasonic sensor acquires the height of the vehicle as follows:
………………………………(0.1)

Fig. 6: Parking sensors layout

When the ultrasonic sensor detects a wide object at a height above the threshold, and when the Hall effect sensor
is triggered, the system records the input as a vehicle present, and proceeds to change the LED colour to RED.
Table 1: LED Configuration

Wireless Network Nodes Colour Changeover Description

These nodes are critical to the project, as they Red indicates that the parking lot is
provide a TCP/IP connection protocol to the Red Green, Blue
occupied and not available
servers. Each device has a unique MAC address Green Red
Green indicates that the parking spot is
which is transmitted along with the data. These available for parking by any user
Blue indicates that the parking spot is
devices, called nodes, are preregistered by the Blue Red available for parking by special users, e.g.
admin and are tied to specific parking zones. Valet, or disabled.
There are two types of network nodes, i.e. the Yellow indicates that the parking spot is not
parking and gate access node. The parking node Yellow Yellow available for parking and that it is under
maintenance.
handles the communication of parking sensors and

4
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

LEDs to the servers. The gate access node captures the access code, transmit it to the servers, and waits for a
response, which is then sent back to the individual gate controller to allow or deny access.
The current setup is designed optimum performance and efficiency, thus only two parking spots and gate access
controllers are connected to their respective nodes. The designed software program is generic since the node can
be reconfigured easily by changing a header identifier.

2.3 Working Principle

The system is composed of four sub-systems, i.e. the user interface application, web server, gate access,
and parking space, which work together in order to achieve the objectives, as shown in Fig. 7.
In order to be a complete IOT system, all configurations for the system are made in the web server. The
gate access and parking space systems request updated configurations from the server to initialize their
functionality. The user is required to valid credentials in order to make a reservation. The parking space sensor
system updates the parking availability to the web server, which is responsible to assign a parking space to a
user.
If a parking space is available, entry and exit access codes are generated and sent via email and SMS to
the registered user’s email and phone number respectively which are used to access the gates. Validation is done
on the server end, and a response is sent to the gate access system to allow or deny access.
Fig. 8 shows the reservation process for the system. The parking location and zone, check in and out times, and
users’ mobile number is required to complete the reservation. Once submitted, the system backend checks for
available parking based on the provided parameters. Access codes are generated if a parking is found, and are
sent to the users via SMS and email. If no parking is found, a response is sent to the user

Fig. 7: Flowchart of the protection system

Reservation

5
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Fig. 8: Reservation process flowchart

Entry and exit gate access

A keypad is used to capture the access codes in the system. A timeout resets the system if user fails to
enter the password at a specified time (7 seconds). Once the password is captured, it is transmitted to the gate
node, which transmits it to the server for verification. The server responds back with information regarding
accessing the gate. A response 1 means success, and the gates will be open, else they’ll remain closed. Fig. 9
shows the flowchart describing this process.
Fig. 10 shows the flowchart of vehicle detection and server update. An interval timer of three (3)
seconds is used to check and update the parking spaces status. Readings from the ultrasonic and hall-effect
sensor is collected to detect if a car is parked. If the vehicle height is above threshold, the LED will change to
indicate a car parked. The sensor readings are then uploaded to the server for storage.

Fig. 9: Entry and exit access gate flowchart

Vehicle detection at parking spot & Server update

Fig. 10: Vehicle detection at parking space and server update

Information request using API
Integrators and users with valid credentials can request parking lot details/information for their own use.
A request is composed by the user, which includes the API token/key, information location, zone and time range.
RESTful/POST request is used to relay the information to server. Fig. 11 shows the flowchart of the server, as it
performs validation and decoding on the request, and processes it. If all details in the request is valid, the server
responds back to the user with all requested information.

6
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Fig. 11: Information request using API

3. SIMULATION AND HARDWARE RESULTS

Fig. 12 shows the simulation circuit for the gate access system. It is composed of the ATmega328P,
LCD, keypad, boom gate and serial interface. There are two I2C modules are interfaced with the LCD and
keypad, so as to reduce the number of input pins.

Fig. 12: Simulation of gate access system

Fig. 13 shows the parking sensor system. It is composed of a microcontroller, ultrasonic sensor, and
hall-effect sensor. An ATmega328p was used instead of the NodeMCU in simulation, as the controller is not
available in the current release of PROTEUS. A serial terminal is used to view the output of the microcontroller
for easy troubleshooting.

Fig. 13: Simulation of parking sensors

Fig. 14 and Fig. 15 show the system hardware implementation. They both show the parking lot layouts with
LEDs indications of the state of occupancy for the car park.

7
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Fig. 14: Prototype system awaiting input from user

Fig. 15: Prototype system showing gate vehicle parked at parking lot

4. TESTING RESULTS
4.1Gate Access system
Simulation results of the gate access system is shown in Fig. 16 yields a success rate total average of 94
percent.The keypad capture input feature scored the lowest success rate of 80 percent. After troubleshooting, it
was found the simulation was not running in real time, as the computer was overloaded.
The LCD simulation was able to show all outputs but failed on one occasion, thus achieved a 90 percent success
rate. The remainder of the system like system reset after entering the access code, opening and closing of gates,
and system timeout after incomplete input achieved a 100 percent success rate.

Fig. 16: Gate Access System Simulation Results

4.2 Vehicle Detection system

The simulation results of the vehicle detection at parking space was tested repeatedly for both the
ultrasonic and hall-effect sensor and the data is visualized in Fig. 17. The average accuracy and success rate
found for the system operation is 91.67 percent. When the vehicle height was 13 cm, the system was able to
accurately detect the correct height 80 percent of the time, whilst the ultrasonic achieved 100 percent
accuracy.This test is done to demonstrate the usability of the system to register a new unique user who can
proceed to make a reservation.

8
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Fig. 17: Vehicle detection parking space results analysis

4.3 User registration and login
This test is done to demonstrate the usability of the system to register a new unique user who can proceed
to make a reservation. Table 2 shows that the system is capable of registering the user easily and accurately,
Validation is done to ensure that the registered user is unique, through the supplied email. The registration process
is easy and simple for anyone to use.
Table 2: User registration and login usability test results
Setting Expected Result Actual Result
New user Successful registration, and PASSED. User
with correct user is automatically successfully registered,
input logged into the system and login to the system.
Registration Fail to register. Error PASSED. User fails to
with existing messages shown register when email exits
email

4.4 Parking Reservation

This test aims to determine the usability and performance of the reservation system. It tests issuance of
parking spaces when available, generation of access codes, and, sending SMS and email to the user.
Table 3: Parking reservation usability test results

Table 3 show that the parking reservation system is

Setting Expected Result Actual Result
User does not An error appears, PASSED. When the user capable of assigning a parking to a user, provided that
fill the form outputting messages fails to fill in the forms it is available. When a parking is available, access
properly of validation results. correctly, an error message codes are generated and are sent via email and SMS
is displayed to the user. When a parking is not available, the user
All user Invoice displayed PASSED. When a parking
details filled detailing parking space is available in the is redirected back to the reservation page with an
correctly, space, entry and exit zone, the system error informing them that no parking is available. It
parking is access codes. SMS automatically assigns the was found that the results were consistent over
available and email sent to user one, generates access repeated testing with 100 percent accuracy.
user codes, and, sends them
All user Error message is PASSED. When no
details filled displayed telling parking space is available
correctly, no user that no parking in selected location zone,
parking is is available user is informed, and
available prompted to select another
zone.
4.5 User at access gate
The aim of this test is to check the functionality of the gate access system. It tests the time it takes from
when the user enters the correct code to the time the gate opens. It also tests the systems usability and hardware
consistency.
Fig. 18 and Fig. 19 shows the observed angles of the boom gate when they are opening and closing
respectively. It can be clearly observed that the error rate is of sinewave, i.e. the boom gate have slippage in the
gears and damping. On the first run, the opening angle was 92° and closing angle was 0°, however, on the angles
fluctuate whilst opening and closing. This is mostly probably due to the weight of the boom gate and the
opening/closing speeds, which causes an offset.

9
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

Fig. 18: Boom gate opening angles

Fig. 19: Boom gate closing angles

The R-squared mean, also known as the coefficient of determination, is computed to determine how
close the data are to the fitted regression line. For opening and closing boom gates the R-squared are 0.0036 and
0.0236 respectively. The overall average angles for opening and closing is 92.65° and 0.95° respectively.

Fig. 20: Performance test of gate access system data analysis

Fig. 20 shows the results of the performance tests. It is noticed that test run seven (7) took the longest
time of 8.2 seconds to execute a full cycle of the test. The second slowest test run was eight (8) and nine (9),
which both took 8 seconds to complete. T3 is the longest wait time, since it waits for an average of 3.12 seconds
till the gate close. On average, the total time it took for the system to complete this test was 7.76 seconds. T1, T2,
T3 and T4 took an average of 1.61, 1.55, 3.12, and 1.48 seconds respectively.
In summary, no errors were observed in the usability and performance tests. The system is capable of
accepting input form the user via the keypad, process and transmit it to the server, receive a response and open
the boom gates as designed

5. CONCLUSION
Thus, an IOT based car parking management system with API is designed. The CPMS is composed of
WSN that are used to detect and collect information of vehicles at parking spots back to a centralized server,
hosted using cloud-based services. All vehicle and parking information can be easily viewed and acquired using
a web app and API, respectively.
The system data/information allows users to view available parking lots, and proceed to reserve them.
Users are required to register using their emails and can reserve using any internet-enabled device. The system
automatically generates two access codes, for entry and exit, when a parking space is available, and then sends

10
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012066 doi:10.1088/1742-6596/1228/1/012066

them to the user’s email and phone (SMS). The results show that the system is capable of operating at an
accuracy of 100 percent.
Once the user reaches the car park zone, they’re required to enter the entry access code at the gate
access terminal. The gate access terminal captures the input and transmits it to the server for validation and
verification. Upon successful verification, the user can enter the parking lot. The same is repeated upon exit, thus
increasing security and integrity of the system. It takes roughly less than 1.61 seconds for verification with the
server and gates to open.
At the parking space level, the vehicle detection system, which is composed of an ultrasonic sensor and
hall-effect sensor, accurately detects a height below 20 cm with 99.79 percent. It can 100 percent detect a vehicle
with a ride height below 1cm. This shows that the combination of these techniques greatly improves vehicle
detection at a parking space.

References
[1] H. Al-Kharusi Intelligent Car Parking Management System Palmerston North New Zealand Massey
University 2014
[2] H. Al-Kharusi andI Al-Bahadly Intelligent Parking Management System Based on Image Processing
Scientific Research 2014.
[3] P Almeida L S Oliveira E J Silva A. J Britto, and A. Koerich Parking Space Detection using Textural
Descriptors International Conference on Systems Man, and Cybernetic 2013.
[4] M. Chen C Hu and T Chang The Research on Optimal Parking Space Choice Model in Parking Lots
3rdInternational Conference on Computer Research and Development (ICCRD) 2011.
[5] L L Elena Polycarpou Smart parking solutions for urban areas 2013.
[6] A. Israel Intergated Mobility. Transport Redefined ITS America 2016.
[7] L. Meng D. Cheng and Z WeiminThe Research of Intelligent Parking System based on Internet of Things
Technology International Journal of Computer Applications 124(6) 2015.
[8] Muhammad Mahbubur Rashid, Abiodun Musa Aibinu, MD. Ataur Rahman Automatic Parking Management
System and Parking Fee Collection Based on Number Plate Recognition International Journal of Machine
Learning and Computing 2 2012.
[9] F I Shaikh P J Jadhav and S P. Bandarkar Smart Parking System Based on Embedded System and Sensor
Network International Journal of Computer Applications 140(12) p. 51 2016.
[10] R. Yusnita and M N Fariza A Secure Parking Reservation System Using GSM Technology International
Journal of Computer and Communication Engineering 2(4) p. 518-520 2013.
[11] J. Anitha Y Thoyajakshi A Ramya V SravanuPrashant Kumar Intelligent Parking System using Android
application International Journal of Pure and Applies mathematics 114(7) p. 165-174 2017.

11