International Conference on Innovative Mechanisms for Industry Applications

(ICIMIA 2017)

Crowd-Sourcing based Android Application for

Structural Health Monitoring and Data Analytics of
Roads using Cloud Computing

Vinaydeep Kaur1, Akriti Tyagi1, M Kritika1, Sanket Salvi2

Pratima Kumari1 Assistant Professor, Department of Information Science and
UG Students, Department of Information Science and
Engineering
Engineering
Nitte Meenakshi Institute of Technology
Nitte Meenakshi Institute of Technology
Bangalore, India
Bangalore, India
[email protected]
[email protected]#1, [email protected]#2,
[email protected]#3, [email protected]#4

Abstract—In this paper, we propose Crowd-Sourcing based which takes place at different levels and across various
Android Application to detect damaged roads and generate a industries.
report which will assist the vehicle driver for safer driving by
providing an early warning system to warn the driver of any Crowd Sourcing with respect to finding collective solution
abrupt discontinuities on the road. We have designed and for common problems is gaining popularity these days. In our
implemented a model using layered approach that consists of project we have tried to explore the domain of Structural
sensor data collected from the users, cloud based data analytics Health Monitoring Systems which uses the collection of data.
and a response model for the mobile which is integrated with Implementing a classification technique and damage
Google Maps. This paper discusses about the review for previous identification for engineering structures is known as Structural
works in the related domain, our proposed architecture, its Health Monitoring (SHM). Damage is changes (to the
implementation, obtained results and its detailed inference. boundary conditions and system connectivity) to the material
and/or physical attributes of a structural system, which affect
its durability and efficiency. It involves careful scrutiny of a
Keywords—Crowd Sourcing, GSM, GPS, Android Phone
Sensors, Cloud Computing, Data Analytics.
system for a period of time, the collection, and the analysis of
the attributes to elaborate the live status of system health using
I. INTRODUCTION frequently sampled random response calculations from a sensor
array.
In the world with increasing number of Smarter Devices,
the amount of data collected is becoming exponentially big. To make the Crowd Sourcing model successful the amount
But instead of thinking of it as a problem, the larger amount of of data collected should be large but cost efficient. Hence,
data can be used for making few deductions which were not Cloud Computing. Using Cloud Computing, we can exploit its
possible earlier. This gave rise to the concept of ‘Crowd key benefits of scalability and cost efficiency. It can provide a
Sourcing’. The collective mobilization, in a ‘crowd’ for a virtual machine with desired storage. This storage can be used
common goal -, problem solving, or efficiency -often for performing data analytics. According to Wikipedia “Big
innovation - whether with ideas, time, expertise, or funds — to Data is an all-encompassing term for any collection of data sets
a project or cause, is Crowd Sourcing. Its key advantages are as so large and complex that it becomes difficult to process using
follows: on-hand data management tools or traditional data processing
applications.” The applications being processing the data,
x access to new ideas and solutions analyzing the trends, anticipating them in the future and taking
x opportunities for co-creation advantage of it.

x deeper consumer engagement For providing an ubiquitous access to the Application as

well as to take sensor input we thought of building it on the
x Task optimization most widely used and open source mobile platform i.e,
Android. Android Mobile Application Development controls
x reduced costs
mobile devices via Google enabled libraries and is based on
Our growing connectivity & the media brought codes written in Java language. The platform is helpful in
organizations closer to their stakeholders, making them developing the mobile applications using the Android SDK
collaborate & collectively contribute in creating value together provided by Google.

978-1-5090-5960-7/17/$31.00 ©2017 IEEE 354

 International Conference on Innovative Mechanisms for Industry Applications
(ICIMIA 2017)

Nowadays all Smartphones are equipped with the number with contextual and description information problem. In order
of sensors (motion, position sensor, etc.). Hence, Smartphone to increase their independence in food shopping and other
users can be engaged in crowd-sourced based computing product recognition, they implemented an Android application
service which contribute in solving complex problems. In this called UCap to assist the blind in this visual challenge problem.
paper, we propose to utilize cloud computing and data analytics They used Infrastructure-as-a-Service (IaaS) on MS Azure
to store data related to structural health of roads to monitor and cloud server for the initial system testing and evaluation. In [6]
develop an early warning system in accurately detecting abrupt The SHM uses building vibration readings from deployed
discontinuities in terrain (manholes and bumps) which can be sensors to analyze and measure probable structural deformities.
used by pedestrians as well as people travelling in automobiles. Common data mining methods, like support vector machine
and data clustering, fail to classify these live data streams
This paper is organized in following order. Section II, directly. Their results show that the damage can be successfully
provides concise information which we have gathered based on detected by training the proposed SHM algorithms.
earlier work in the domain of Structural Health Monitoring
using Crowd Sourcing. We have also discussed about III. PROPOSED METHOD
advantages, disadvantages and features or concepts that had
inspired our framework from corresponding papers. In Section
III, we have explained the detailed proposed framework and
the algorithms. Section IV provides with detailed information
about the requirements, purpose and experimental setup.
Section V gives the experimental results gathered post
implementation and its justification. Finally, Section VI,
highlights the major takeaways of the project, conclusion and
future scope.
II. RELATED WORK
In [1], a simple and precise way to solve the problem of
identification of damaged roads is discussed. It uses typical
smartphones with GPS sensors and accelerometers to quantify
the abrupt fluctuations sensed by the car when driving over a
bad road surface. The study of this paper helped us to
understand and develop our proposed model.
In [2], the authors investigated discontinuities in a lake in
Bangalore, India over three months and found that there was
double digit percentage change in several chemical and
physical parameters, confirming that frequent monitoring of
lakes would be valuable. The app also measures contaminant
concentration levels of pH, Hardness, Alkalinity, Copper, and Figure 1: Block Diagram
Nitrites by detecting color changes in chemical test strips. All
data collected by the app is uploaded to a cloud platform that Figure 1 shows a detection system that aims at detecting,
enables spatiotemporal visualizations of health parameters of warning and generating a report about the potholes and road
multiple lakes. This work has helped us to understand the humps in the path. The user interacts with the mobile
details with respect to data modelling and cloud storage. application. The mobile app consists of GPS, Gyroscope and
Database. Figure 4: The data collection class is built using
Paper [3], authors had designed detection model and alert android studio. It consists of three sub modules i.e. Sensor class
model to provide vital ways for public safety in the of disaster. to manage the gyroscope readings, Database Helper to store the
They proposed a crowdsourcing-based disaster management data temporarily in the mobile device and class Location is
using fog computing (CDMFC) model in IoT. Their proposed used to manage the GPS readings. Figure 8: Illustrates how the
CDMFC model and its data offloading mechanism can identify data is stored in the device.
real-time disasters and produces early suggestions for public
safety as compared to the conventional cloud computing based The application has two phases. First phase is data
disaster management models. In [4] Users can share collection. The app uses Gyroscope sensor present in the
information with location-based systems. This system will be mobile device to record the jerks and GPS to obtain the geo
use to report any crime where and when it happened. The location of the mobile device. User uses the pervasive interface
proposed system will be very useful for users when traveling to that is it enters the sensory data. For example, GPS is used to
unfamiliar areas, where they have little knowledge about the get the location and direction which can then determine traffic
foreign environment. This model has helped us to optimize the congestions. Google traffic updates are obtained using the
real time data acquisition model such that only the filtered data pervasive approach. This android application reads the current
with potential valid information will be stored in the cloud. location of the user and also takes the gyro sensor readings if
this reading is either greater than or equal to 1 or equal to -1
In [5], the author has provided with a noble solution for the then that location and gyro readings are recorded as a object in
problems faced by blind for consumer product identification mobile node. Upon connection with the internet through Wi-Fi

978-1-5090-5960-7/17/$31.00 ©2017 IEEE 355

 International Conference on Innovative Mechanisms for Industry Applications
(ICIMIA 2017)

or hotspot the data about the pothole locations is uploaded into E. CVM:
a remote database that stores location in form of latitude and Cloud Virtual Machine will be used for administering and
longitude and gyroscope reading (X-axis, Y-axis, Z-axis). managing the cloud storage and data analytics system.
Here Amazon RDS is used to host the data on the
cloud.The frequency of the approximately close readings of the
gyroscope sensor will be counted and through a code it will be
identified whether that location requires marker of pothole on
the map or not.The remote database provides input to the data
analytics module.
The data analytics module is responsible for merging the
data with the processed data storage after applying data
analytics techniques. Hence output of this module is the mean
value of the gyroscope readings for a given location.
The processed data is stored in the processed data store
located in the cloud. Another mode in which the app operates is
interactive mode. This mode interacts with the user and provide Figure 2: Gyroscope Reading (Data not
notification and warnings about the damaged roads well in inserted)
advance to the user by the Google voice interaction API. The
user is required to enter the destination. This mode obtains the
geo coordinates of the destination and finds the current location
of the user. The two location objects are used by the
application to find the shortest and most optimal path for the
driver. This is done by sending the location coordinates to
navigation API. It returns an array of location objects that are
present in the path. The object array is fed into a search
algorithm that determines if rerouting is required or not based
on the road condition of the shortest route between the source
and the destination. If rerouting is required, an alternate path
short path is suggested from the source to the destination. It
provides navigation and pothole locations marked on the map
with red dots.
IV. REQUIREMENT AND IMPLEMENTATION
A. GPS:
Figure 3: Gyroscope Reading (Data
Global Positioning System can be defined as a network of Inserted)
orbiting satellites that sends very precise details of their
position in the space back to the earth. In the application Figure 1: This detection system that aims at detecting,
proposed in this paper, data about potholes fetched from the warning and generating a report about the abrupt
sensors is stored in terms of x, y, z parameters. The start and discontinuities in the path. The user interacts with the mobile
end of a bad road is indicated by [(x1, y1, z1), (x2, y2, z2)]. application. The mobile app consists of GPS, Gyroscope and
The data received about the location of the potholes present Database. (Figure 2). The data collection class is built using
can be shown on the GPS map. Depending on the severity of android studio. It consists of three sub modules i.e. Sensor class
the pothole it can be depicted with different brightness. to manage the gyroscope readings, Database Helper to manage
the data collected in the mobile device and Location class to
B. GYROSCOPE SENSOR: manage the GPS readings. (Figure 6). Illustrates how the data
These are devices that sense angular velocity and detect the is stored in the device. (Figure 6). The application has two
stability of the mobile node. phases.

C. GSM: First phase is data collection. The app uses Gyroscope

sensor present in the mobile device to record the jerks and GPS
Global system for mobile communication: it digitizes and to obtain the geo location of the mobile device. User uses the
compresses the data and sends it down the channel for pervasive interface that is it enters the sensory data. For
transmission across the communication channel. example GPS which can be used to determine the vehicle’s
D. DATA ANALYTICS: speed and direction and hence traffic congestion. Traffic layers
in the Google maps are obtained using the pervasive approach.
Data mining technique to handle the road traffic big data. This android application reads the current location of the user
CLOUD STORAGE: and also takes the gyro sensor readings if this reading is either
It is a model of data storage in which the data is stored on greater than or equal to 1 or equal to -1 then that location and
remote servers accessed from the internet or cloud.

978-1-5090-5960-7/17/$31.00 ©2017 IEEE 356

 International Conference on Innovative Mechanisms for Industry Applications
(ICIMIA 2017)

gyro readings are recorded as a object in mobile node. Upon Android Studio: Android Studio is the official integrated
connection with the internet through Wi-Fi or hotspot the data development environment (IDE) for Android platform
about the pothole locations is uploaded into a remote database development. Android Studio is freely available under the
that stores location in form of latitude and longitude and Apache License 2.0.
gyroscope reading (X-axis, Y-axis, Z-axis). The remote
Sensing Subsystem :
database provides input to the data analytics module.
This subsystem is mainly used for fetching data which
The data analytics module is responsible for merging the
consists of the location and severity of the pothole.
data with the processed data storage after applying data
analytics techniques. Hence output of this module is the mean
value of the gyroscope readings for a given location.
The processed data is stored in the processed data store
located in the cloud.

Figure 5: Cloud database

Figure 4: Class Diagram

Another mode in which the app operates is interactive

mode. This mode interacts with the user and provide
notification and warnings about the damaged roads well in
advance to the user.
The user is required to enter the destination. This mode
obtains the geo coordinates of the destination and finds the
current location of the user. The two location objects are used Figure 6: Sample Database
by the application to find the shortest and most optimal path for
the driver. This is done by sending the location coordinates to
navigation API. It returns an array of location objects that are
present in the path. The object array is fed into a search
algorithm that determines if rerouting is required or not bases
on the road condition of the shortest route between the source
and the destination.
If rerouting is required, an alternate path short path is
suggested from the source to the destination. It provides
navigation and pothole locations marked on the map with red
dots.

978-1-5090-5960-7/17/$31.00 ©2017 IEEE 357

 International Conference on Innovative Mechanisms for Industry Applications
(ICIMIA 2017)

periodically. The similar kind of information will be available

to the government agencies like BBMP, PWD etc.
REFERENCES
[1] L. C. Lima, V. J. P. Amorim, I. M. Pereira, F. N. Ribeiro and R. A. R.
Oliveira, "Using Crowdsourcing Techniques and Mobile Devices for
Asphaltic Pavement Quality Recognition," 2016 VI Brazilian
Symposium on Computing Systems Engineering (SBESC), João Pessoa,
Paraíba, Brazil, 2016, pp.144-149. doi: 10.1109/SBESC.2016.029
[2] S. Pingali, "Cloud Computing and Crowdsourcing for Monitoring Lakes
in Developing Countries," 2016 IEEE International Conference on
Cloud Computing in Emerging Markets (CCEM), Bangalore, India,
2016, pp. 161-163. doi: 10.1109/CCEM.2016.037
[3] A. Rauniyar, P. Engelstad, B. Feng and D. V. Thanh, "Crowdsourcing-
Based Disaster Management Using Fog Computing in Internet of Things
Figure 7: The marker set near Mechanical Department Paradigm," 2016 IEEE 2nd International Conference on Collaboration
and Internet Computing (CIC), Pittsburgh, PA, USA, 2016, pp. 490-494.
doi: 10.1109/CIC.2016.074
This system detects the jerks using the hardware available
[4] F. I. Maulana, A. S. Prihatmanto and C. S. Kim, "Design and
in the android device that is gyroscope sensor, which generates implementation the concept of crowdsourcing on a web portal
readings in all the three axes. It also uses GPS (in-built android Crime," 2016 6th International Annual Engineering Seminar (InAES),
hardware) to track the location of the device and record its Yogyakarta, Indonesia, 2016, pp. 51-55. doi:
latitude and longitude readings. The app creates an object of 10.1109/INAES.2016.7821906
two readings from GPS and gyroscope. If the value of the any- [5] A. Hoonlor, S. P. N. Ayudhya, S. Harnmetta, S. Kitpanon and K.
axis reading goes above 1or below -1, the data gets inserted Khlaprasit, "UCap: A crowdsourcing application for the visually
impaired and blind persons on Android smartphone," 2015 International
into the database. Computer Science and Engineering Conference (ICSEC), Chiang Mai,
2015, pp. 1-6. doi: 10.1109/ICSEC.2015.7401406
V. CONCLUSION [6] X. Li, W. Yu and S. Villegas, "Structural Health Monitoring of Building
In this work, we are proposing Crowd-sourcing based Structures With Online Data Mining Methods," in IEEE Systems
mobile app for detection of potholes, road humps and roads Journal, vol. 10, no. 3, pp. 1291-1300, Sept. 2016. doi:
10.1109/JSYST.2015.2481380
under repair. The app will be installed in the vehicles of taxi
[7] X. Zhang, L. Shangguan and Y. Yuan, "A Crowd Wisdom Management
services. as the number of these vehicles is nearly 2.5L in a city Framework for Crowdsourcing Systems," in IEEE Access, vol. 4, no. ,
like Bangalore. The app will detect the potholes, road humps pp. 9764-9774, 2016. doi: 10.1109/ACCESS.2016.2581298
and roads under repair and transfer the information like type of [8] R. Meng; L. Chen; Y. Tong; C. Zhang, "Knowledge Base Semantic
obstacle, GPS location and the time of detection to the cloud. Integration using Crowdsourcing," in IEEE Transactions on Knowledge
The cloud will process the data based on sliding window and and Data Engineering , vol.PP, no.99, pp.1-1 doi:
the processed information will be uploaded to the end user app 10.1109/TKDE.2017.2656086

Figure 8: Activity Diagram

978-1-5090-5960-7/17/$31.00 ©2017 IEEE 358