You are on page 1of 4

Concept, Design and Implementation of Sensing as a

Service Framework

Rischan Mafrur I Gde Dharma Nugraha Deokjai Choi


School of Electrical School of Electrical School of Electrical
Engineering Engineering Engineering
Chonnam National University Chonnam National University Chonnam National University
Gwangju, South Korea Gwangju, South Korea Gwangju, South Korea
rischanlab@gmail.com gdebig@gmail.com dchoi@jnu.ac.kr

ABSTRACT memory, bigger storage and etc. With this equipment, smart-
Today, personal data is becoming a new economic asset. The phone has capability to running complex application. Many
data that we generated from our smartphone, our interaction sensor also has embedded to the smartphone. With this
in social media, its like oil in the Internet. An increasing sensor and log capability of smartphone, we can develop
of personal data in the internet cause some issue such as many useful system or application in different domain such
privacy issue, complexity processing, and etc. That will as healthcare (elderly monitoring system [1–4], human fall
require a highly reliable, available, serviceable, and secure detection [5–8]) , transportation(monitoring road and traffic
infrastructure at its core and robust innovation. This paper condition) [9, 10], personal and social behavior [11–14], en-
propose a new concept, design and implementation of sensing vironmental monitoring(pollution, weather) [15–17] etc. To
as a service framework. This framework has three main develop such system, we have to collect the user personal
components: 1) personal data collector application; 2) web data and then analyse it. There are two ways to collect per-
portal that has user friendly interface and shows responsive sonal data from the users based on user involvement. First,
analysis result of personal data that can be accessed by the participatory sensing and then the second, opportunistic
users based on new web technology; 3) REST API feature sensing. Participatory sensing means the application still
and documentation for third party. The goals of this paper need user’s intervention to complete their task. The exam-
are: 1) to provide rich application as a service based on users ples for such application need user to taking text input for
personal data log; 2) to bridging the user and third party each time period, taking picture and etc. On the other hand,
such as developers in term of developing applications and opportunistic sensing means application does not need user’s
researcher for research based on user personal data; 3) to intervention to complete their task, users not involved in
develop reliable, available, serviceable and secure sensing as making decisions instead smart phone itself make decisions
a service framework. according to the sensed and stored data. In this paper, we
propose concept, design and implementation of sensing as
a service (S2 aaS) framework. The abbreviation of (S2 aaS)
Categories and Subject Descriptors coming from [18] which have present the challenge, solutions
J.4 [Computer Applications]: Social and Behavioral Sci- and future directions of sensing as a service.
ences; C.2.1 [Network Architecture and Design]: Net- Research about sensing framework has conducted by many
work communications researcher. Most of the research focus only for one goal.
The research about CRNTC+ [19] focus on health care ap-
General Terms plication. Most of sensing framework also based on par-
ticipatory method which need user intervention. UbiqLog
Design, Implementation
framework [20] is one example of the research based on par-
ticipatory method. In this paper, our contribution are: 1) we
Keywords focus on sensing framework development which is can collect
Sensing as a service, personal data analysis, mobile sensing as much as possible personal user data with high quality of
data. 2) The proposed framework is based on opportunistic
1. INTRODUCTION method for collecting data because we want to increase user
cosiness. For analyse the data, we focus on unsupervised
Nowadays, smartphone capability has increased signifi-
learning. 3) This framework can be used for many purposes,
cantly. Smartphone has equipped with high processor, bigger
everyone who need personal user data in their research or
their system can use this framework. This framework can be
Permission to make digital or hard copies of all or part of this work for bridge between user and application developer or researchers.
personal or classroom use is granted without fee provided that copies are Our framework consist of three main components: Personal
not made or distributed for profit or commercial advantage and that copies data collector application based on opportunistic sensing
bear this notice and the full citation on the first page. To copy otherwise, to method, analyse and visualize the data that can be accessed
republish, to post on servers or to redistribute to lists, requires prior specific by the users and REST API that can be used by third party
permission and/or a fee. Request permissions from Permissions@acm.org.
IMCOM ’15 January 08 - 10 2015, BALI, Indonesia for research or other authorized application, equipped with
Copyright 2015 ACM 978-1-4503-3377-1/15/01$15.00 documentation.
http://dx.doi.org/10.1145/2701126.2701173.
Figure 1: Framework design and architecture

Table 1: List of probes and time period of recording 2.2 Design of S2 aaS Framework
One of the goals of this framework is to provide rich appli-
No. Probes Interval,duration(s)
cation based on user’s personal data and become the bridge
#1 Location(GPS) 300
between user and third party. The concept and design of
#2 Wi-Fi 300
this framework can be seen on Figure 1. Figure 1A shows
#3 Bluetooth 300
how the framework works. First, user have to register to the
#4 Battery 300
web portal. The web portal provide the information about
#5 Call Log 86400
how to use the portal and the link to the senseF application.
#6 SMS Log 86400
senseF application is develop based on opportunistic sensing
#7 Application Installed 86400
method. So after it installed by the user, the application
#8 Hardware Info 86400
will collect the data automatically without intervention from
#9 Contact 86400
the user. This framework support real time analysis. After
#10 Browser Search Log 86400
some amount of user’s personal data is uploaded to the cloud
#11 Browser Bookmark 86400
data server, user can directly see the analysis result of their
#12 Light Sensor 120,0.07
personal data. Figure 1B shows the flow of S2 aaS framework,
#13 Proximity 120,0.07
the details explanation as follows:
#14 Temperature 120,0.07
#15 Accelerometer 120,0.07 1. Application will collect user’s personal data and store
#16 Magnetic Field 120,0.07 it in mobile storage.
#17 Pressure 120,0.07
#18 Activity Log 120,0.07 2. The stored personal data in user’s mobile storage will
#19 Screen Status 120,0.07 automatically uploaded to the cloud data server when
#20 Running Application 120,0.07 the user’s mobile is connect to internet through Wi-Fi.

3. S2 aaS framework provides REST API and documenta-


2. DESIGN AND IMPLEMENTATION tion for third party. This API and documentation can
be used by the developer or the researcher which have
registered to the portal and install the data collector
2.1 Personal Data Collector Application application. To use personal data that belongs to an-
Before collecting personal data, we need to define what other user, the developer or researcher have to ask the
kind of data that we have to collect. After we define the kind permission of the respective user.
of data, we can develop or decide the application accordingly.
In this research we did not develop our application from 4. Third party can register as developer to S2 aaS frame-
scratch. We use Funf library1 and we named the application work and then they can create new application for
as senseF. The probes or kind of data that we collected and accessing user’s personal data. The authorization be-
the time period of recording can be seen in Table 1. The size tween cloud data server with third party uses OAuth2.
of the data that we can collect is influenced by the activity The simple logic OAuth2 flow can be seen on Figure 2.
of the users. Based on setting in Table 1 users with high
activity can result bigger size data around 1GB each month 5. Third party can upload their application to application
and vice versa, the users with less activity will result smaller store.
size data. We use encryption method (hash) to sensitive data
such as phone number in contact, text message log, phone 6. Then that application can be search and installed by
call log, and etc. the user. After user gives the authorization to the
application, the application can access user’s personal
1
https://code.google.com/p/funf-open-sensing-framework/ data which had been stored in the cloud data server.
Figure 2: Simple logic of OAuth flow

7. We use R programming language to analyse user’s


personal data.
(a) list of probe and data (b) Activity summary in
8. To provide beautiful, intuitive, responsive and interac- in mobile browser mobile browser
tive report, we use shiny2 web application framework.
Shiny support and has responsive layout so the layout
will compatible for accessed by different type of devices
(PC, laptop, phone, etc).
9. User can access and see the analysis result from the
portal after login process by using mobile devices or
personal computer. (c) Activity summary in PC browser

As explained above, the authorization process based on Figure 3: Example of data report layout opened in
OAuth2. This authorization is between cloud data server and mobile and PC browser
third party. OAuth2 [21] provides a secure authentication to
enable client application to access the resources on behalf of
a resource owner. By using OAuth2, the user could authorize framework which support to turn data analysis into interac-
third party to access their resources (data) without shares tive web applications. Shiny is new package from RStudio
their credentials. OAuth2 was designed to working with that makes the development of interactive web applications
HyperText Transfer Protocol (HTTP) and allows access with R is incredibly easy. Figure 3a and 3b show the example
tokens to be issued by the third party using an authorization of web layout when opened by using mobile phone browser.
server. The third party application uses the token to access Figure 3c shows the example of web layout using PC browser.
the protected resources that hosted by the resource server Actually many system or research topics that can be pro-
(cloud data server). Figure 2. shows the simple logic of duced from user personal data, now we still focus on personal
OAuth2 flow. 1). User uses applications from third party activity and scheduling system such as when users open the
(original request for resource); 2). Redirect to cloud data page of daily activity they will see the message likes ”Today,
server for authorization; 3). Server domain response the until now (20:34PM) you have spent your time 3.12 hours
request by asking authentication from resource owner; 4). for facebook, kakao talk, and whatsapp application” or likes
Authenticate is done by the user; 5). Server issues the ”Hai.... you have spent one month in Gwangju, you just in
token to client application; 6). Client application confirms your dorm and your school, don’t you want to go to some-
the access. In figure, blue cloud means token exchange where else? I give you the recommendation places: Nami
and the red cloud means user credentials (username and Island (You’ve never been there), Jeju Island (You have 2
password). There is many vendor which has used OAuth2. times in there)”.
The vendors such as Facebook, Twitter, Google, Yahoo, etc.
To detail explanation of OAuth2 authentication process, we
uses TweetDeck application as an example. TweetDeck is the
3. CONCLUSION AND FUTURE WORK
popular third party of Twitter client application. When we This paper is ongoing research paper. We still have many
want to use TweetDeck, we have to authorize this application weakness and limitation. We divide the problem into two
to access our data in Twitter. This authorization process parts.
is based on OAuth2. And our S2 aaS framework is adopting
1. Personal data collector application: Currently, our per-
this authorization concept.
sonal data collector application only available for an-
2.3 Data Visualization and Analysis droid platform and limited into 20 probes. We have to
expand it into multi-platform and adding more probes.
To process and analyse the data, we use R language. R
But adding more probes will increase the storage, pro-
language is open source, powerful platform for data analysis,
cessing and energy consumption in mobile phone. So
excellent for visualization, large and enthusiastic community.
we have to research more deeply to collect the most
Our S2 aaS framework uses shiny, one of web application
valuable user’s personal data with minimum storage,
2
http://shiny.rstudio.com minimum processing and energy consumption in the
mobile phone. computing,” in Ubiquitous Intelligence and Computing,
2013 IEEE 10th International Conference on and 10th
2. S2 aaS framework : There are five main problems to International Conference on Autonomic and Trusted
solve in this area. The problems are user’s privacy, Computing (UIC/ATC), pp. 539–543, 2013.
data management and processing (big data issues),
[9] B. Hoh, T. Iwuchukwu, Q. Jacobson, D. Work, A. M.
data analysis algorithm, data visualitzation and more
Bayen, J. C. Herrera, M. Gruteser, and M. Annavaram,
applications based on personal data.
“Enhancing privacy and accuracy in probe vehicle-based
We can see this S2 aaS framework from two points of view, traffic monitoring via virtual trip lines,” Mobile
first from user’s view and the second from researcher and Computing, IEEE Transactions, vol. 11(5), pp. 849–864,
developers view. From the user’s view, this framework will May 2012.
increase user’s cosiness about privacy because their personal [10] P. Zhou, Y. Zheng, and M. Li, “How long to wait?:
data only uploaded to one place and make one agreement with predicting bus arrival time with mobile phone based
the cloud data server. When there is an application which participatory sensing,” in MobiSys ’12 Proceedings of
want to access their personal data, have to get authorization the 10th international conference on Mobile systems,
from the user. Moreover, this framework also provides the applications, and services.
analysis result of their personal data with dynamics, intuitive [11] R. D. OLIVEIRA, M. CHERUBINI, and N. OLIVER,
and interactive layout. From the researchers and developers “Influence of personality on satisfaction with mobile
view, this framework makes it easy for them to developing phone services,” ACM Transactions on
application or research the field related to personal data. Computer-Human Interaction, vol. 20(2), May 2013.
They do not need to define and make the application for [12] R. LiKamWa, Y. Liu, N. D. Lane, and L. Zhong, “Can
collecting personal data by themselves. They can register your smartphone infer your mood?,” in PhoneSense
and use our framework and then using the REST API to workshop, 2011.
request the data or another user’s personal data when they [13] G. Chittaranjan, J. Blom, and D. Gatica-Perez,
have get the authorization. “Mining large-scale smartphone data for personality
studies,” Journal of Personal and Ubiquitous
4. ACKNOWLEDGEMENT Computing, vol. 17(3), pp. 433–450, March 2013.
This research was supported by Basic Science Research [14] J.-K. Min and S.-B. Cho, “Mobile human network
Program through the National Research Foundation of Korea management and recommendation by probabilistic
(NRF) funded by the Ministry of Education (2012R1A1A2007014). social mining,” IEEE TRANSACTIONS ON
SYSTEMS, MAN, AND CYBERNETICSâĂŤPART B:
5. REFERENCES CYBERNETICS, vol. 41(3), June 2011.
[1] T. Faetti and R. Paradiso, “A novel wearable system [15] M. Bilandzic, M. Banholzer, D. Peev, V. Georgiev,
for elderly monitoring,” Advances in Science and F. BalagtasFernandez, and A. D. Luca, “Laermometer -
Technology, vol. 85, pp. 17–22, 2013. a mobile noise mapping application,” in inProceedings
[2] P. Pierleoni, L. Pernini, A. Belli, and L. Palma, “An of the 5th Nordic conference on Humancomputer
android-based heart monitoring system for the elderly interaction, pp. 415–418, 2008.
and for patients with heart disease,” International [16] M. Mun, S. Reddy, K. Shilton, N. Yau, J. Burke,
Journal of Telemedicine and Applications, vol. 2014, D. Estrin, M. Hansen, E. Howard, R. West, and
p. 11, 2014. P. Boda, “Peir, the personal environmental impact
[3] Y. Charlon, W. Bourennane, F. Bettahar, and report, as a platform for participatory sensing systems
E. Campo, “Activity monitoring system for elderly in a research,” in 7th international conference on Mobile
context of smart home,” Journal of Digital Technologies systems, applications, and services, pp. 55–68, 2009.
for healthcare, vol. 34(1), pp. 60–63, 2013. [17] N. Maisonneuve, M. Stevens, M. E. Niessen, and
[4] A. Ghose, P. Sinha, C. Bhaumik, A. Sinha, A. Agrawal, L. Steels, “Noisetube: Measuring and mapping noise
and A. D. Choudhur, “Ubiheld: ubiquitous healthcare pollution with mobile phones,” in Information
monitoring system for elderly and chronic patients,” in Technologies in Environmental Engineering,
UbiComp ’13 Adjunct Proceedings of the 2013 ACM pp. 215–228, 2009.
conference on Pervasive and ubiquitous computing [18] X. Sheng, J. Tang, and X. Xiao, “Sensing as a service:
adjunct publication. Challenges, solutions and future directions,” Sensors
[5] L. Tong, Q. Song, Y. Ge, and M. Liu, “Hmm-based Journal, IEEE, pp. 3733–3741, Oct 2013.
human fall detection and prediction method using [19] S. G, R. F, W. J, L. P, and A. O, “Crntc+: A
tri-axial accelerometer,” IEEE,Sensors Journal, smartphone-based sensor processing framework for
vol. 13(5). prototyping personal healthcare applications,” in 2013
[6] M. Mubashir, L. Shao, and L. Seed, “A survey on fall 7th International Conference on Pervasive Computing
detection: Principles and approaches,” Neurocomputing, Technologies for Healthcare (PervasiveHealth), 2013.
vol. 100, pp. 144–152, January 2013. [20] R. Rawassizadeh, M. Tomitsch, K. Wac, and A. M.
[7] O. Aziza, E. J. Parkc, G. Morid, and S. N. Robinovitch, Tjoa, “Ubiqlog: a generic mobile phone-based life-log
“Distinguishing the causes of falls in humans using an framework,” International Journal of Personal and
array of wearable tri-axial accelerometers,” Gait and Ubiquitous Computing, pp. Volume 17, Issue 4, pp
Posture, vol. 39(1), pp. 506–512, January 2014. 621–637, April 2013.
[8] J. H. C. Hu and Y. Li, “An autonomous fall detection [21] IETF, “The oauth 2.0 authorization framework,” Oct.
and alerting system based on mobile and ubiquitous 2012. http://tools.ietf.org/html/rfc6749.

You might also like