Sie sind auf Seite 1von 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/324673793

Accident Detection and Notification System using Android Smart Devices

Conference Paper · March 2018

CITATIONS READS

0 77

3 authors, including:

Jaturapith Krohkaew Pongpon Nilaphruek


Rajamangala University of Technology Thanyaburi Rajamangala University of Technology Thanyaburi
8 PUBLICATIONS   46 CITATIONS    7 PUBLICATIONS   3 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Accident Detection and Notification System View project

All content following this page was uploaded by Pongpon Nilaphruek on 21 April 2018.

The user has requested enhancement of the downloaded file.


ICDP
ICDP
Accident Detection and Notification System using
Android Smart Devices

Jaturapith Krohkaew Pongpon Nilaphruek


Department of Mathematics and Computer Science Department of Mathematics and Computer Science
AI2oT Research Unit, Faculty of Science and Technology, AI2oT Research Unit, Faculty of Science and Technology,
Rajamangala University of Technology Thanyaburi, Rajamangala University of Technology Thanyaburi,
Pathumthani Thailand Pathumthani, Thailand
kjatura@rmutt.ac.th pongpon_n@rmutt.ac.th

Nongluk Promthong
Department of Mathematics and Computer Science
AI2oT Research Unit, Faculty of Science and Technology,
Rajamangala University of Technology Thanyaburi,
Pathumthani, Thailand
Nongluk661@rmutt.ac.th

Abstract In Thailand, approximately ten thousand people die the smartphone much more frequently and the smartphones are
each year from 2007 to 2016 because of the accident on the road more frequently updated in software and even in hardware. As
side. One of the causes of death was the lack of access to the scene. we known that a lot of innovative technology and information
In this paper, we propose the Android application which exchanging has become more notable and the smartphones gives
automatically notifies the emergency or accident. When an a new style to the usage of mobile phones for the users. The
emergency or accident occurs to the people who install this smartphones also have additional sensors, advanced power
application on their phone, the application will be automatically processor and communication interfaces, which permits to
activated by the immediate change to zero of the acceleration of
develop traffic accident detection and notification system that is
the smartphone. It means that their moving was stopped suddenly,
based on sensor inputs to the smartphone without need to
then an emergency message which contains the location of the
scene that obtained from GPS sensor of the smartphone will be
interaction with a car or changing anything in the car. Usually
sent to the most available first emergency or accident responder. the smartphones are held with their owners. Therefore, an
There are two modes of emergency notification in this application. accident detection can be done without any installation of the
First, the message will be delivered via the Internet connection; if accident detection and notification system on the vehicle.
it is available. Second, it will be sent via SMS if there is no the On the other hand, when an accident occurs, the medicine
Internet connection. In view of the recipients of the emergency rescue units are notified by the communication radio or the
message, they can view the map and direction to the scene. This
telephone. Sometimes, the rescue team cannot reach the scene
application can help users to improve the fast and accuracy of the
severity of the accident by offering perfect facts to emergency
right on the time.
services as soon as possible and saves precious life of peoples. These facts are the ones that motivated us to take the
advantages of using the smartphone in the development of car
Keywords Accident Notification System, Smartphone, GPS, accident detection and notification system on smartphones.
Accelerometer
II. RELATED WORK
I. INTRODUCTION
Djajadi [3] proposed a communication system between cars
As the records of the Police Information System Center [1], to prevent and handle an accident. The system uses an Android
there are 101,752 cases of accidents on the road side in 2007. based smartphone as the main platform to detect some events,
Also, it is decreased continually to 84,405 cases in 2016. There such bumpy road, speed bump, and sudden brake. It utilizes the
are 12,492 persons were killed due to road accidents, in 2007. built-in sensors and employs Wi-Fi Direct for
The number is decreased continually to 8,369 persons in 2016. the fast and cheap inter-cars communication to other Android
Nevertheless, it is still approximately ten thousand people die devices in the surrounding areas.
each year from 2007 to 2016.
Sharma [4] proposed an automatic accident detection system
This Statista [2] forecast shows the number of smartphone based on smartphone to reduce the delay between accident and
users in Thailand from 2013 to 2022. In 2017, the number of emergency response. The algorithm reliably infers an accident
smartphone users in Thailand is estimated to reach 24.14 based on the data collected through various smartphone sensors
million, with the number of smartphone users worldwide despite their sensitivity and range limitations.
forecast to exceed 2.3 billion users by that time. The user renews
Senthilkumar [5] detected the occurrence of an accident and Most Android-powered devices have an accelerometer, and
finding the effective path to reach the accident spot. It detects the many now include a gyroscope. Motion sensors are useful for
accident occurrence and providing timely notification and timely monitoring device movement, such as tilt, shake, rotation, or
help to reduce the impacts of accidents by the use of the sensors swing. It can also reflect the physical environment in which the
built-in with the smartphones and the on-board sensors fitted device is sitting (for example, moving with you while you drive
with the vehicle. Both the distance and the traffic, to reach the your car) [16]. We derive their data to detect the occurrence of
accident spot is intimated to the relatives and to the nearby an accident.
hospital.
The Google Maps Distance Matrix API returns information
Chandran [6] built the smart helmet for detecting and based on the recommended route between start and end points,
reporting accidents. The accident detection system as calculated by the Google Maps API, and consists of rows
communicates the accelerometer values to the processor, which containing duration and distance values for each pair [17]. We
continuously monitors for erratic variations. When an accident use this API to optimize selecting the nearest and fastest rescue
occurs, the related details are sent to the emergency contacts by unit.
utilizing a cloud based service. The vehicle location is obtained
by making use of the global positioning system. The system IV. METHODOLOGY
delivers an information relating to the accident in real time. The methodology of this study can be divided into three
Singhal [7] has implemented the concept of IoT to leverage parts, as the following.
the nearby sensors to help the vehicles which have collided or A. Reviewing the related works
need any help. The collision in the vehicle which will be sensed
Related works were reviewed especially how to apply the
by the sensors. The crash sensors will measure and report the
smartphone in the accident detection and notification and the
intensity of collision. The strength of collision mapped on a scale
communication between the smartphone and server in order to
will then inform the respective nearby sensors in IoT, which can
find the appropriate solution to develop an accident detection
come out to help the victims.
and notification system. This step is explained as in previous
III. MATERIALS AND TOOLS sections.
The main materials and tools include the protocols, Android B. Investigating the requirements
Studio editor, Node.js, JSON, PHP, MySQL, and satisfaction The requirements were investigated from many emergency
evaluation form. They are explained in the following. first responder officers interviews as shown in the Fig. 1. The
The main protocols in this system, MQTT is a machine-to- main features are illustrated as Fig. 2-3. The roles in the
machine (M2M)/Internet of Things (IoTs) connectivity protocol proposed system are sender and receiver. After login to the
[8]. It was designed as an extremely lightweight publish and system, the sender can enable/disable the detection and
subscribe messaging transportation. MQTT was used for 2-ways notification, stop the unintentional detection and notification,
communication among clients and server. Also, HTTP protocol manually notification the occurrence of the accident and capture
[9] was used for data transmission among clients and server as the photo or record the VDO clip including in the notification,
well, because sometimes the data from smartphone should be submitted to the server. Once the request from the sender has
also recorded to the system, and HTTP is an easy way to deal been accepted by any rescue unit, the acknowledge screen will
with this task. be appeared in the smartphone.

Android Studio was used to develop an application on client


side. It is recommended to be a powerful tool for developing an
Android application [10].
Node.js was used inside the server because the system must
control the data transmission between client and server. It is the
server-side JavaScript space, which is a framework for
developing asynchronous task, high-performance, and
concurrent programs [11].
JSON is a popular format for data interchange between client
and server [12, 13]. JSON was used to be the format of data Figure 1. Requirements investigation
interchange in this study.
The receiver must login to the system as well as the sender.
MySQL was used to be the database of the system prototype Once the request was submitted to the server, the receiver can
because it is flexible enough to work well in various types of view the photo/video clip to make sure that it is the actual
application and it is also free [14]. accident or not. In additional, he can accept or decline the
For database management on the server, PHP server-side request. If the request is accepted, then the route and direction
script was used in the server side because it is one popular script from the current location of the rescue unit to the scene will be
for dealing with a database system [15]. displayed in Google Map.
clients real-time. HTTP also is applied to send clients data log
such as the locations, speed, status, and direction. JSON is the
main standard for data exchange.
Second, server is the service to manage shareable
information such as senders, receivers, and data logs. Node.js is
used to manage and filter the information from the
communication among server, and clients.
V. RESULT
The example screenshots of the application user interface are
illustrated as in the Fig. 5 Fig. 9.
Fig.5 shows the registration screen which all user, sender and
receiver, have register to the system about their information and
the mode of user, who can be either the sender (the drivers or
people who observe the accident) or receiver (the ambulances,
rescue units).
Fig. 6 illustrates that when the system detects the values from
the various sensors, which are likely to be the accidental. Then,
the emergency screen will be displayed immediately. In this
screen, there is a countdown timer waiting for the user to cancel
Figure 2. Use case diagram of Sender
the alert in the case that the notification is caused by other non-
accidental causes. Otherwise, the system will assume that the
event is a real accident and will notify the emergency to the
possible nearest and available rescue unit.

Figure 3. Use diagram of Receiver

C. Designing the communication framework and system Figure 4. System architecture


prototype development
Fig. 7 is the manual accident notification for people who are
According to the Fig. 4, the proposed communication living near the scene. They can use the application to report an
framework of the system prototype consists of two main parts emergency event or an accident. In this screen, the sender can
including clients (sender/receiver) and server. record the short VDO clip or take the photos of the event to
First, clients (sender/receiver) can be smartphones or tablets confirm with the receiver that the event is the real accident and
on Android platform. Clients are the components that installed it need an urgently help.
the application; GPS sensor of the client is used to collect the As Fig. 8 when someone has submitted the notification to the
global position; accelerometer sensor, compass sensor and system, the notification screen at the receiver-side will be
gyroscope sensor are used to detect the vehicle collision, and 4G displayed immediately. At this moment, the receiver can view
is used to enable the Internet connection. Clients have their own the photos or VDO clip attached with the notification to ensure
ID to tell the server when they wake up. Then they will connect that the accident is really occurred or not. In the case of
with the server using MQTT; after that, do the handshake. This unintentional notification, which has been submitted to the
channel is main path to exchange the data between server and receiver, who can decline the request by tapping the decline
button. However, if there is a real accident, the receiver can
accept the case by tapping the accept button and the direction
from the current location of the receiver to the scene is displayed
in the Google Map API at the bottom. We use Google Maps
Distance Matrix API to find the fastest route from the current
location to the destination location, where an accident occurred.

Figure 7. The screenshot of manual accident notification

The system prototype was also tested by general users related


to use an accident notification service in Pathumthani, Thailand
(N=30). After such users used the application, the result from
ction shows as in the Table 1 which

on 5 criteria.
The satisfaction levels are divided into five levels including:
very satisfied (4.5-5.0), satisfied (3.5-4.49), OK (2.5-3.49),
Figure 5. The registration screenshot unsatisfied (1.5-2.49), and very unsatisfied (0.0-1.49).
According to the Table 1, the average performance score of the
system prototype from general users is 3.95/5.00. It is in the
satisfied level.
The highest score is the robustness of the application and the
lowest score is the easy to use of the application. This is related
regularly said that the system
should support other platforms such as Windows Phone, iOS,
and on web browsers. However, the usability of the system is in
satisfied level; that means the system can be applied in real
condition.

Figure 6. The screenshot of automatic accident detection and nofication

Fig. 9 shows the screenshot at the sender-side after the rescue


unit accepted the accident occurrence notification. In this screen,
the sender can see the detail about rescue unit such as the name
and contact number. At the bottom of the screen, it is the map
direction that displays the incoming rescue unit. This map will
help the sender to know the progress of the rescue unit.

Figure 8. The screenshot of notification alert at the rescue unit


View publication stats

ACKNOWLEDGMENT
We would like to thank faculty of Science and Technology,
Rajamangala University of Technology Thanyaburi for
providing the infrastructure and funding to do this research.
REFERENCES
[1] Traffic accident on national highways in 2016
Annual report, Jun 2017.
http://bhs.doh.go.th/files/accident/59/report_accident2559.pdf
[2] Number of smartphone users in Thailand from 2013 to 2022
The Statistics Portal, 2017.
https://www.statista.com/statistics/467191/forecast-of-smartphone-users-
in-thailand/
[3] A. Djajadi, R. J. Putra Inter-cars safety communication system based on
android smartphone IEEE Conference on Open Systems (ICOS), 26-28
Oct. 2014, Subang, Malaysia, 2014.
[4] H. Sharma, R. K. Reddy and; A. Karthik S-CarCrash: Real-time crash
detection analysis and emergency alert using smartphone International
Conference on Connected Vehicles and Expo (ICCVE), 12-16 Sept 2016,
Seattle, WA, USA, 2016.
Figure 9. The screenshot of the acknowledge from the rescue unit [5] R. Senthilkumar, Gowtham. G; Hemalathaa. S and; Anirudh. S iCars:
Accident occurrence detection and effective path tracing using FCD
TABLE I. THE USERS SASTISFACTION OF THE PROTOTYPE SYSTEM Fifth International Conference on Advanced Computing (ICoAC), 2013
[6] S. Chandran. S. Chandrasekar and, N E. Elizabeth Konnect: An Internet
N = 30
No. criteria of Things(IoT) based smart helmet for accident detection and
Average S.D. Meaning notification IEEE Annual India Conference (INDICON), 16-18 Dec.
The response time from rescue is 2016, Bangalore, India, 2016.
1. 3.89 0.84 Satisfied
faster than traditional method. [7] Intelligent accident management
The optimization of selecting the system using IoT and cloud computing 2nd International Conference on
2. 4.00 0.85 Satisfied
most near rescue unit. Next Generation Computing Technologies (NGCT), 14-16 Oct. 2016,
Realtime monitoring incoming Dehradun, India, 2016.
3. 3.97 0.73 Satisfied
rescue process. [8] MQTT.org, "MQTT," http://mqtt.org/, 2014.
4. The robustness of the application. 4.06 0.88 Satisfied [9] T. Berners-Lee, HyperText Transfer Protocol, World Wide Web
Consortium, 1999.
5. The application is easy to use. 3.84 0.62 Satisfied
[10] DiMarzio and F. Jerome, Beginning Android Programming with
Total 3.95 0.79 Satisfied Android Studio, John Wiley & Sons, 2016.
[11] T. Stefan and S. Vinoski, Node. js: Using JavaScript to build high-
performance network programs, IEEE Internet Computing, vol. 14, no.
6, 2010.
VI. CONCLUSION AND FUTURE WORK
[12] G. Bruno and P. Trezentos, Impacts of data interchange formats on
This research proposes the application for automatically energy consumption and performance in smartphones, in Proceedings of
notifies the emergency or accident because the accidents cause the 2011 workshop on open source and design of communication, 2011.
approximately ten thousand people, who die each year from [13]
2007 to 2016 in Thailand. There are three main methods in this [14]
study including 1) reviewing the related works to find the
possible technologies and theories to solve the problems, 2) [15]
investigating the requirements from general users in the purpose 4.
of designing the functional system process which is represented [16] Motion Sensors
by use case diagrams, and 3) designing the communication https://developer.android.com/guide/topics/sensors/sensors_motion.html
framework to develop the client and server side for exchanging [17] Google Maps Distance Matrix API
the information among them. The system prototype was https://developers.google.com/maps/documentation/distance-
evaluated using the via In-App evaluation and matrix/intro
the result shows that
3.95/5.00 being in the satisfied level.
In the future work, according to
said that should bind the application with some social networks
such as Facebook, Twitter, and LINE in order to widely notify
the accident immediately. Also, the proposed system must be
used continuously to many areas, in Thailand; in the purpose of
precision evaluation, and improving the rescue process.

Das könnte Ihnen auch gefallen