Smart Parking System Based on Embedded

System and Sensor Network

1.1 Introduction

In today parking lots there are no standard system to check for parking spaces.
The system heavily relies on human interaction with the physical space and entity. This
leads to wastage of human manpower and also parking spaces at times. These parking
lots are dependent on Human-to-Human Interaction (HHI) which is not efficient.
Previously, various techniques have been proposed to overcome such problems like
Short Messaging Services, Zigbee, RFID, etc.

Most of the time when users go to malls and commercial complex, they
experience that there is a limited space for parking spots especially on prime hours.
Hence, there is a desperate need of a robust parking system that will enable us to
reserve the parking spots. For that it is necessary to build a centralized system to gather
all the information on parking spots of malls, commercial complexes, and multilevel car
parking systems. According to the International Organization of Motor Vehicle
Manufacturers (OICA), the number of cars produced in 2014 alone stands at 60 million.
With the explosive growth in automobiles, on-street parking will soon disappear, given
the constraints on road space. Hence there is a desperate need of multilevel car parking
and which should be smart also.

In urban areas, one-third of cars which have reached their destination and are
circling around looking for a parking space thus leading to problems like pollution and
traffic congestion. In a recent survey, researchers have found that in one year, car
cruising for parking created the equivalent of 38 times trips around the world, burning
1.7 lakh litre of fuel and producing 730 tons of CO2. So, it is essential to control the air
pollution using a robust parking system that will be used for the reservation of parking
spot as well as parking spot allocation in on spot resource allocation scenario.
Reservation can be made using multiplatform tools such as Android application, iOS
application, Windows applications or even Web Portal allowing user to have a hassle-
free parking experience. Recommender system will help user find the right parking lot
closest to his location.

With the help of recommender system predictability of nearest available slot

becomes easier. User location data is accessed in the mobile platform to suggest best
location for parking. The Google Map API uses users geo-location and location of
parking spot to determine the above.

1
1.2 Overview about the System

Many cities are currently undertaking parking information pricing programs to

decrease cruising. Our research team, with the aid of many partners, has developed a
parking information system for Pittsburgh that enables many of the results described in
this report. ParkPGH is a smart parking system that uses historical parking and event
data in a prediction model to provide real-time information on the availability of parking
within downtown Pittsburgh. The program enhances 19 existing off street parking
facilities by providing real time information using a host of information delivery methods
that includes an iPhone application, traditional and mobile website, text messaging and
an interactive voice response system. The primary goals of the program are to reduce
search time and search time variability when finding a parking space within the Cultural
District and to make the District a more desirable destination for patrons by reducing the
anxiety and uncertainties related to parking issues. In a study, included with this report
and partially funded by the T-SET UTC, we find that cruising for parking in downtown
Pittsburgh has decreased due to the information provided by ParkPGH.

The city of San Francisco has recently deployed a $20 million parking information
and pricing program called SFPark of which a member of the technical advisory
committee. This program is the first large scale deployment to combine both time of day
pricing and information aimed at reducing excessive searching for parking. This
program is expected to improve public transit service by reducing road congestion
caused by those searching for parking, reduce greenhouse gas emissions, improve
neighborhood livability, enhance San Franciscos economic vitality, and make parking
more convenient. Parking sensors will be installed in 6,000 on-street parking spaces
and 11,500 off-street parking spaces. The data generated by the sensors will enable the
thorough evaluation of new pricing structures. SFPark will distribute information about
parking to drivers before and during their trip. Real time parking information will be
made available via variable message signs, static wayfinding signs, web, PDSs, and
text messages (SMS). In a study, included with this report and partially funded by the T-
SET UTC, we estimate that cruising for parking has decreased in San Francisco due
the SFPark program.

While SFPark has led to decreases in road congestion caused by cruising for
parking, the program may be too expensive to implement for most cities. The goal of
this study is to investigate an inexpensive parking information and management
systems based on crowdsourcing. Vehicle-to-Vehicle(V2V) and Vehicle-to-
Infrastructure(V2I) communications is the enabling technology for the proposed crowd
sourced parking information system.

Our goal is to develop and analyze a cost effective decentralized crowdsourced

parking information system by leveraging the existing deployment of ParkPGH,org, a
centralized, parking information system.

2
1.3 Objective about the System

Locating a parking space in central city areas, especially during the peak hours,
is cumbersome for drivers. The issue arises from not having the knowledge of where
the available spaces may be at the time, even if known, many vehicles may seek very
limited parking spaces to cause severe traffic congestion. In this paper, the design and
implementation with a prototype of Reservation-based Smart Parking System (RSPS)
that permits drivers to effectively locate and withhold the vacant parking spaces in
mentioned. This system use cluster based algorithm which helps in periodically learning
the parking status from the sensor networks deployed in parking spaces, the reservation
service is influenced by the change of parking status. The drivers are allowed to access
this said cyber-physical system with their personal communication devices. The system
implemented is cost efficient smart parking system for multi-level parking facility using
WSN (IR Sensor) and develop an android based application, by cluster based allocation
method and performs automatic billing process. The system monitors the availability of
idle parking slots and guides the vehicle to the nearest free slot. Cost is minimized by
keeping the number of sensors low without sacrificing the reliability. Energy
consumption of each mote is kept in check by allowing the systems to sleep periodically
and by reducing their communication range. This systems reservation-based parking
policy has the potential to smoothen the operations of parking systems, as well as
mitigate traffic congestion caused by searching for parking.

1.4 Significance of the System

Parking lot problems difficulty in finding vacant spaces quickly finding a vacant
space in a multilevel parking lot is difficult if not impossible, especially on weekends or
public holidays. One study showed that 86% of drivers face difficulty in finding a parking
space in multilevel parking lots. Finding spaces during weekends or public holidays can
take more than 10 minutes for about 66% of visitors. Stadiums or shopping malls are
crowded at peak periods, and difficulty in finding vacant slots at these places is a major
problem for customers. Insufficient car park spaces lead to traffic congestion and driver
frustration Improper Parking If a car is parked in such a way that it occupies two parking
slots rather than one, this is called improper parking. Improper parking can happen
when a driver is not careful about another drivers rights. Sometimes improper parking
occurs when a driver parks on or a bit outside of the lines of a parking space. The driver
may notice his improper parking after leaving his car, but may not be willing to unlock
his car, restart it, and adjust it to be inside the lines. This matter annoys other drivers
and most of the time a driver who wants to park in a small leftover slot will give up and
feel frustrated.

3
1.5 Conceptual Framework

Parking Space Detection

The system works on real time parking information based on which it makes
reservation and on spot resource allocation of parking spaces for drivers. Several
sensing technologies available to monitor real time parking status like ambient light
sensor, Ultra sonic sound, etc. System works on IR led Sensors providing Energy
efficiency with less power consumption.
IR led keeps transmitting infrared rays up to some range when some object is
detected in the infrared range the IR waves hits the object and comes back at some
angle the photo diode next to IR led detects that infrared rays reflected from the object.
Thus, the object or the car standing in front of the IR sensors gets detected,
providing us the status or availability of a particular parking space.

V2I and I2V Communication

For the need of utilization of resources a reliable and effective two-way communication
is essential between the vehicle and the allocation center (infrastructure).
In this system V2I communication involves drivers sending their parking request
providing driver information, status confirming reservation and utilization of resources to
the system. In I2V communication the system reserves the space for user and the
application shows the driving directions to the reserved parking space with the help of a
directed map, the map contains the images of the path allocated to the driver by the
system.
Both I2V and V2I communication are implemented through an Android application &
other Windows or Web Application.

Reservation Guarantee

In case of an efficient parking system the resources allocated to a particular user must
not be designated to any other user. To successfully imply this idea, QR Code Generation
technology is applied.
The system does QR code verification at the entrance of the parking lot with the help of
a web camera which is placed on the system. User with its application reserves a parking
space, QR code gets generated on the reservation which provides unique object id for
every user on every transaction. This generated QR code should be presented in front

4
of the camera. The System scans and detects the unique object ID and then verifies the
ID from the database.

Optimal Allocation

The most powerful feature of the system is that it determines the nearest or closest
parking space for the user with the help of globally optimized list.
The Admin dashboard fires the query on the database through the application in such a
way that it returns only the unreserved resources which are globally optimized. On the
acquired list linear search is performed and it gives the closest parking slot from the list
which also the first element of the optimized list[O(1)].This helps in optimal allocation of
the resources through the system.

Backend

For the efficient working and communication of the server, application and database
backend devices use server-side languages to build an application and data storage tools
to find change or save data and serve it back to the user.
The system uses NoSQL database management system which contains unstructured
data. NoSQL also provides storing of images and videos. It also is scalable, instead of
using tables NoSQL stores data in the form of objects and key/value pairs tuples.
JSON (Java Script Object Notation) it is a light weight format used for data interchanging.
It is used as a web service response. Instead of using XML as a primary data format for
the transfer of data JSON is more preferred format due to its light weight property.
RESTful API (Representational State Transfer Application Programming Interface) is
used to ease development and simplify client adoption it does not follow a prescribed
standard except for HTTP. System uses four important methods viz POST, GET, PUT,
DELETE.

POST: POST is utilized to create new resources especially used to create

subordinate resources means POST to the parent. On creation, return HTTP
status 201.

GET: GET method is used to read or retrieve resource. It returns

representation in JSON and HTTP response code 200 and in case of error it
returns 404.

5
 PUT: PUT is utilize for update the resources it can also be used to create
resource when the resource id is chosen by client instead of server.

DELETE: It is used for deleting the resource. On deletion it returns HTTP status
200 with response body. It can also return HTTP status 204 with no response
body.

Sensors

Sensors are the devices which senses from the physical world and stimuli on the
environment. The system uses IR sensors to detect the vehicle above it.
Sensors respond to system to the availability of the vehicle above the IR sensors. If the
vehicle is present above the IR sensor, it will return value FALSE and the voltage goes
LOW, If the vehicle is not present above the IR sensor, it will return value TRUE and the
voltage remains HIGH.
The parking space is reserved but the vehicle is yet to arrive, the LED glows RED. In
another case vehicle is at the entrance of parking lot then LED glows BLUE, to notify the
driver the parking slot allocated to him. When a vehicle has already occupied a parking
slot the led will glow GREEN. The GREEN LED is used as a positive indicator to
showcase proper functioning of IR Sensors.

Data Analysis
Nearest resource allocation:

Admin fires the query to get the globally optimized list. The list contains only the
unreserved free parking spaces available in the parking lot. The list is natively
optimized so, the first element in the list is the nearest of all.
QR scanner:

Every time when user reserves a parking slot for him, a QR code is generated
along with the payment receipt with the help of unique object id. While entering in
the parking lot, user first has to scan his unique QR code. For that user has to
place his QR code i.e. smart phone in front of the webcam. The webcam will scan
the QR code and object id along with the name of reserved parking slot will get
displayed. Admin dashboard updates the logs and the event gets generate by
glowing the LED light BLUE. User already has a directed map showing the exact
location of parking space. Only after a successful authentication via QR code,
user is allowed to enter into the parking area.

6
 System Logs:

System log is produced on the admin dashboard showing all the activities which
are done in the past. It includes username, vehicle number, object id for every
transaction, slot number allocated to user, time of arrival and departure, date of
reservation etc.

Figure 3. Working state (Slot S2 & S4 Reserved and S1

occupied)

Platform

An Android application is created using Android studio. The android applications

are developed using the JAVA code. Using the JAVA compiler the source files are
converted to JAVA class files. The .dex file and the resources of an android application
are packed into an .apk (Android Package) file. The resulting .apk file contains all data
to run the Android application and can be deployed to an Android device using adb tool.
The Windows Application is made using Visual Studio 2012. Visual Studio allows
us to create windows forms and windows runnable applications. This application has its
minimum requirements according to the safety standards of Net Framework. This
platform generates a setup file for the Smart Parking and allows us to run the .exe for
creation of account to reserving parking.

7
 Using the slot allocation method these applications are developed for the Smart
parking. Slot reservation can be done using the slot allocation method. The request is
updated in the server and forwards it to parking area. Platform allows user to enter arrival
and departure time of the parking. Both platform application ensures that reservation of
parking spaces is allowed maximum two days before the actual reservation date. The
reserved spot will be available again only after three hours of departure time of previous
reserved vehicle, maintaining a buffer for consecutive reservation of same parking spot.

1.6. Analysis of the System

SYSTEM ARCHITECTURE

Figure 1. System Architecture

8
Raspberry Pi

The system is using Raspbian OS based on Debian OS running on raspberry pi 2B model

which has a configuration of ARM cortex-A7 quad core , 900MHz CPU . It also has 4
USB ports, 40 GPIO pins, Full HDMI port, Ethernet port, micro SD card slot. System has
a 16GB SD card of class 10 which is used to mount Raspbian OS. System is designed
in such a way that it will use Raspberry Pi as a web server. Raspberry Pi is mounted with
a custom designed shield that is used to connect 4 IR sensors using FRC cable.
Hardware system consists of 4 IR sensors which are connected to raspberry pi through
FRC cables.
The system uses 24 GPIO pins to interact with IR sensors and tricolor LEDs. They are
connected over a common VCC and GND.

IR Sensors

An infrared sensor is an electronic device which is used to sense light wavelength of its
surroundings by either emitting or detecting infrared spectrum. It is also capable of
measuring the heat being emitted by an object and detecting motion. LED will glow with
respect to the IR sensor detection. Every IR sensor has a transmitter and a receiver. For
every transmitter and receiver, there are total two registers and one more register is
present to maintain voltage of LED that will show the status of parking slot to customers.
When the IR does not detect any object in front of it, it returns the value TRUE, meaning
HIGH volt output from the sensor and FALSE for when a object is detected meaning
LOW volt output.

LED

A tricolor LED has a cut section at either of its sides. These LED have 4 pins. The longest
one is common cathode and the rest are represented as RGB.

Parse

Parse is an open source web server framework that allows the implementation of the
back end server. It is suitable for implementation of a prototype for various projects.
Parse works as a mediator between the Raspberry Pi and other Platform applications.
Parse handles the NoSQL database behind the system. It also facilitates the use of push
notifications through its services.

9
Admin Dashboard

Admin gets different kinds of data from various sources in JSON formats on its
dashboard. Admin dashboard shows the activity logs, reserved parking slots , free
parking spaces. It is also responsible for the parking spot allocation to different users.
Algorithm will allocate the nearest free parking space to various customers who dont
reserve their parking spaces.

Figure 2. Hardware Architecture

1.7 Output

MATHEMATICALNMODELBAND SIMULATION RESULTS:

S={C,Fme,S,F}

Let,

C={FL,OL,RL}

Where,

10
 FL be set of Free Lots { I1,I2,I3,In }

OL be set of Occupied Lots { I1,I2,I3,In } RL be set of Reserved Lots {

J1,J2,J3,Jn } Fme={LS,RS,Lo,Rr,Qr,OSr}

Where,

Let LS be the LED Light Status {BLUE(B), GREEN(G),RED(R)}

Where,
BLUE (B) is currently assigned slot.
GREEN (G) is currently occupied Lot.
RED (R) is currently reserved Lot.
Let RS be Reservation Status {NOT RESERVED (nRS),RESERVED(RS)}
Where,

NOT RESERVED (nRS) is Reservation not done.

RESERVED (RS) is Reservation done
Login (Lo) is User Login Function
Reservation Request (Rr) is Function to reserve specific lot
Qr is Generated Qr code after successful Reservation
On Spot Request (OSr) is to request nearest free parking lot without registration

Let S be Set of successful cases

S={Rp,OSp}

Where,

Park (Rp) is successful occupied lot

On Spot Park (OSp) is successful occupied lot F = Failure Cases where Qr
Code was not generated, Parking lot was not reserved

MOBILE BACK END AS A SERVICE (MBaaS)

It is a model for providing web and mobile application developers with a way to link their
applications to back end server which also can be used as a cloud storage. APIs exposed
by back end applications MBaaS also provides features such as user management, push

11
notifications, analysis of performance measures and integration with social networking
services, file storage, geo-location and ability to store custom objects. These services are
provided via the use of custom software development kits (SDKs) and application
programming interfaces (APIs). Backend as a Service eliminates the need for developers
to construct their own servers. MBaaS providers offer developers a ready-made,
customizable back end that's already outfitted with common back-end features. As a
result, proponents of the approach say it shifts developers' focus away from time-
consuming, complex server end development, and gives them more time to invest in front-
end work such as UI/UX. The scalability of mobile apps, which is a huge pain-point for
mobile app developers, is also handled by MBaaS providers; apps are scaled
automatically, in the cloud, according to variations in volume of users and traffic.

PROTOTYPE IMPLEMENTATION:

The proposed system has been partially implemented and tested in lab prototype. Client
devices have been connected via TCP/IP protocol to a parking database. The latter is
updated in real-time with the status of parking lots. Two kinds of client applications have
been considered for parking lots monitoring: mobile device application, for phones and
tablets, and desktop application for laptops and desktop computers.

Hardware Prototype

Hardware prototype contains 4 IR sensors as of now that will represent the 4 parking
spaces as shown in fig.2.According to fig1, these IR sensors are connected through wires
to Raspberry Pi which will also act as a server. A webcam is also attached to Raspberry
Pi to scan the QR code for the authentication of the driver at entry point. Internet is
provided to Raspberry Pi via LAN connection to send the real-time data to admin
dashboard.

Android Application Prototype

Android application contains only one parking area. That parking area contains only 4
parking slots viz. S1, S2, S3, and S4 which are present on the hardware. Fig.4 shows
the layout of parking slots as well as 3 legends to depict whether the particular slot is
Available, Unavailable/Reserved, User selected.

Fig. 5 shows the details of the reservation. It contains name of parking area, name of
parking slot, arrival time, departure time, date and duration. User must have to enter his
vehicle number to uniquely identify his vehicle in case a user has multiple vehicles.

12
Figure 4. Android app prototype shows S1 and S4 are unavailable or reserved
and only S2, S3 are available.

13
 Figure 5. Android app prototype shows the details of reservation before
confirming the slot

Windows Desktop Application:

On windows desktop application user has to enter arrival time, departure time, date, and
vehicle number. Then windows application will show the map and layout of all parking
lots which are similar to the android application. It will also show the available parking
slots so that user can enter the desired resource very easily. After clicking on the Book
button user will be greeted with a pop up window showing the QR code and other
relevant information about reservation.

14
PERFORMANCE EVALUATION

The authors conducted a number of experiments and simulations to quantitatively

investigate the impact of the proposed framework in terms of engineering and economic
efficiency. Three classes of performance metrics are considered. The first tab is to
check which database class the query is being fired upon. The 2nd tab suggests which
type of query the application is requesting. 3rd tab gives the total number of counts and
fourth calculated the slow percentage of query. The Median tab gives the avg. time
required to give the output of the query which is no longer than 7 milliseconds which is
much less that SQL query which takes appx. 5 sec on an average. The last tab gives
the maximum time required for a query to fetch data over all its performance.

15
8. COMPARISON

The figure below gives a comparison of the Smart Parking system which allows
reservations over the various system such as

1. Blind Search

2. Parking Information Sharing

3. Buffered PIS

4. Smart Parking

16
 Figure 9. Comparison of Smart Parking with other parking systems

5.

Over a given period of time the Smart Parking System proves efficient by providing real
time data and facility to book the parking.

1.8 Conclusion

Various system proposed by different authors helps us to cogently in reserving and also
annihilating the need of searching for parking spaces in private parking lots. Researchers
have acquired the systems which dynamically arrange the scheme for different drivers
as per their requirement, based on the real-time parking information. Thus, this concludes
that the paper simplifies the context for the researchers for innovating various techniques
to administrate and solve the problems faced by drivers on day to day basis.

In future, the system can be extended which is not only specific to a private parking like
Malls, Company parking, etc. but also can be implemented over various multiple
platforms such as public parking also extending the feature by giving parking information
based on cost in real time. This will make the management of the parking spaces more
efficient, by purging the need of human labor.

17
18