Big Data Fusion for eHealth and Ambient Assisted Living Cloud Applications
•
2 likes•763 views
The document describes a cloud-based system for monitoring senior citizens' healthcare using data from heterogeneous sensors. It discusses existing ambient assisted living solutions, challenges with data fusion from multiple sources, and proposed hardware and software components. The system would gather data from various physiological, environmental, and lifestyle sensors. Software components include a big data fusion platform running on a distributed cloud architecture. The goal is to enable applications for monitoring health conditions and daily activities.
1 of 17
Download to read offline
More Related Content
Big Data Fusion for eHealth and Ambient Assisted Living Cloud Applications
1. Big Data Fusion for eHealth and
Ambient Assisted Living
Cloud Applications
• George Suciu, Alexandru Vulpe and Razvan Craciunescu
Faculty of Electronics, Telecommunications and Information Technology,
University POLITEHNICA of Bucharest
• Cristina Butca and Victor Suciu
R&D Department, Beia Consult International
Constanta, Romania, 2015
2. Content
Short Biography
Introduction
State-of-the-art : Virtual Collaboration
Spaces
Proposed Cloud Acceleration Platform For
Innovation In Industry
Conclusions and Discussions
A. Vulpe, et.al. (2015)
3. Short Biography (1)
Graduated from the Faculty of Electronics,
Telecommunications and Information Technology at the
University “Politehnica” of Bucharest (UPB), Romania
(www.upb.ro)
Currently, Ph.D. Eng. Post-doc Researcher focused on the
field………Alex Vulpe…..
A. Vulpe, et.al. (2015)
4. Short Biography (2)
Projects – www.beiaro.eu / www.mobcomm.pub.ro
FP7 (2 on-going)
REDICT : Regional Economic Development by ICT
eWALL : Electronic Wall for Active Long Living
Cloud Consulting : Cloud-based Automation of ERP and CRM software for Small Businesses
ACCELERATE: A Platform for the Acceleration of go-to market in the ICT industry
H2020 (1 on-going)
SWITCH: Software Workbench for Interactive, Time Critical and Highly self-adaptive Cloud applications (ICT-9)
National (more than 10 past projects, 5 on-going)
MobiWay: Mobility Beyond Individualism: an Integrated Platform for Intelligent Transportation Systems of
Tomorrow
EV-BAT: Redox battery with fast charging capacity as a main source of energy for electric autovehicles
CarbaDetect: Imuno-biosensors for fast detection of carbamic pesticide residues (carbaryl, carbendazim) in
horticultural products
SARAT-IWSN : Scalable Radio Transceiver for Instrumental Wireless Sensor Networks
COMM-CENTER : Developing of a “cloud communication center" by integrating a call/contact center platform with
unified communication technology, CRM system, “text-to-speech” and “automatic speech recognition” solutions in
different languages (including Romanian)
A. Vulpe, et.al. (2015)
5. Introduction (1)
We describe a cloud-based approach for monitoring the healthcare condition of
senior citizens and the fusion of big data from heterogeneous information flows
coming from the sensors.
Solutions for Ambient Assisted Living Cloud Applications exists already and under
development:
eCAALYX - the monitoring system is implemented both inside and outside the home
through three main subsystems: the Mobile Monitoring System; the Home Monitoring
System and the Caretaker Site.
Persona - solution is very complex, with a sophisticated and up–to–date software
architecture implemented in the house.
eWALL - system has an architecture based on two main blocks: the sensing
environment and the eWALL cloud. The sensing environment is linked to the local
processing unit using a local gateway. The connection between the sensing
environment and the eWALL cloud is done via a cloud proxy. This component collects
all the data from the sensing environment and sends it to the cloud processing
components.
A. Vulpe, et.al. (2015)
6. DATA SOURCES AND DATA FUSION (1)
1. Data sources
Data sources can be in home metadata and external data sources.
In home metadata are generated by applying perceptual components on the
signals from diverse sensors and are temporarily stored in one database. The in
home metadata may span the following categories:
person - location, gender and age;
environment - humidity, illumination, temperature;
steps of the person monitored;
communication - Usage of phone, messaging services, social media;
sound - level, angle of arrival, speaker analytics;
sleep - Bed pressure & acceleration, sleep sounds.
External data sources can be social networks like Facebook, Twitter and
LinkedIn, entertainment and gaming sources such as YouTube, Video on Demand
(VOD) and Audio and Video on Demand (AVOD), games and gaming platforms.
A. Vulpe, et.al. (2015)
7. DATA SOURCES AND DATA FUSION (2)
2. Data fusion challenges
A single source it is not sufficient to detect a situation person - location, gender
and age
Combining audio and visual localization, an audiovisual localization system can
be:
Accurate, since now the location comes from two signals;
Persistent, since a person can continue being localized under adverse visual
conditions (occlusions) or audio conditions (noise, absence of speech.
Early fusion refers to the combination of the signals from different sources,
combining the unprocessed signals.
Late fusion is done when each data source is used independently to estimate
the state.
A. Vulpe, et.al. (2015)
8. DATA SOURCES AND DATA FUSION (3)
3. Fusion methods
Fusion of imperfect data
Bayesian estimator is one of the probabilistic methods, easy implementation and
optimality in a mean-squared error sense.
Monte Carlo method for fusion of imperfect data. This is very flexible because it
doesn't make any assumptions regarding the probability densities to be
approximated.
Fusion of correlated data
It is possible to design a fusion algorithm that takes into account correlated data,
and Covariance Intersection (CI).
The problem solving is made by formulation of an estimate of the covariance matrix
as a convex combination of the means and covariances of the input data.
One of the methods for fusion of correlated data is the Largest Ellipsoid (LE)
algorithm that is an alternative to CI and provides a tighter estimate of covariance
matrix by finding the largest ellipse that fits within the intersection region of the
input covariances.
A. Vulpe, et.al. (2015)
9. COMPONENTS OF THE SYSTEM (1)
We describe the hardware and software components of the proposed solution
for big data gathering, processing and fusion.
A. Vulpe, et.al. (2015)
10. COMPONENTS OF THE SYSTEM (2)
1. Hardware components (1)
The healthcare condition of the patient can be monitored through various
sensors, even by using traditional home automation systems.
Our proposed sensor devices measure:
pulse rate,
ECG,
body core temperature,
breathing rate,
blood pressure,
oxygen saturation,
glucose to home safety and smart home sensors and actuators that measure light,
ambient humidity,
CO2 level,
facial expression,
presence of other people
motion and lifestyle sensors.
A. Vulpe, et.al. (2015)
11. COMPONENTS OF THE SYSTEM (3)
1. Hardware components (2)
The TMP 102 temperature sensor is used for retrieving body temperature for the
person who wears it.
The sensor for ECG has a 1 inch diameter enclosure and is the one who combines
amplification, bandpass filtering and analog-to-digital conversion.
The pulse sensor is used for heart rate measurement for detection of human heart rate.
The Sensor for CO2 level detection can detect a concentration between 20 and 2000
ppm.
For temperature and humidity of the air it will be used a DHT11 sensor. This
sensor uses a thermistor to measure the surrounding air and then returns a digital signal
on the data pin.
For motion detection will be used a PIR sensor.
A. Vulpe, et.al. (2015)
12. COMPONENTS OF THE SYSTEM (4)
2. Cloud software components (1)
The system is based on distributed cloud computing software SlapOS.
The software components of the cloud platform are hosted on several server
nodes following an architecture based on the concept of Master and Slave
nodes
Master nodes act as a central marketplace of the cloud system
Slave nodes report their availability and costs of resources.
The software components that need to be installed on Slave nodes are Slapgrid and
Supervisord.
To understand the full extent of a person movement, multiple sensors sending
data about his activity are needed.
A. Vulpe, et.al. (2015)
13. COMPONENTS OF THE SYSTEM (5)
2. Cloud software components (2)
Beside the accelerometer, a PIR
sensor will be used to detect
the room in which the user is
in.
In Fig. we present the Daily
Function Monitoring
The PIR detects the room that
the user is in and the
accelerometer detects no
movement for several minutes.
The conclusion is that the user
is resting.
A. Vulpe, et.al. (2015
14. Conclusions
In this paper we described a cloud-based approach for monitoring the
healthcare condition of senior citizens.
We presented the components of this monitoring system, both software and
hardware.
The hardware component is made up of several sensors.
The software component is represented by a big data fusion software running
on a cloud system.
As future work we envision adding data sources from social networks in order
to enable other cloud applications such as games and unified communications.
A. Vulpe, et.al. (2015)
15. University “POLITEHNICA“ of Bucharest
Faculty of Electronics, Telecommunications & Information Technology
Any questions ?
The work has been funded by the Sectoral Operational Programme Human Resources Development 2007-2013 of the Ministry of
European Funds through the Financial Agreement POSDRU/159/1.5/S/134398.
A. Vulpe, et.al. (2015)
16. References (1)
Ford, Earl S., Janet B. Croft, David M. Mannino, Anne G. Wheaton, Xingyou Zhang, and Wayne H. Giles. "COPD
surveillance—United States, 1999-2011." CHEST Journal 144, no. 1, pp. 284-305, 2013.
H. Carvalho, W. Heinzelman, A. Murphy, and C. Coelho, “A general data fusion architecture”, Int. Conf. on Info. Fusion,
pp. 1465-1472, 2003.
D.F.M. Rodrigues, E.T. Horta, B. Silva, F.D.M. Guedes, and J.P.C. Rodrigues. "A mobile healthcare solution for ambient
assisted living environments." In e-Health Networking, Applications and Services (Healthcom), 2014 IEEE 16th
International Conference on, pp. 170-175, 2014.
A. K. Bourke, S. Prescher, F. Koehler, V. Cionca, C. Tavares, S. Gomis, V. Garcia, and J. Nelson, “Embedded fall and
activity monitoring for a wearable ambient assisted living solution for older adults”, Engineering in Medicine and
Biology Society (EMBC), Annual International Conference of the IEEE, 2012.
PERSONA Project, “Reference Architecture and information model for service infrastructure final version, January 2010,
last access on Feb. 2, 2015.
S. Kyriazakos, M. Mihaylov, B. Anggorojati, A. Mihovska, R. Craciunescu, O. Fratu, R. Prasad, “eWALL – An intelligent
caring home environment offering personalized context-aware applications based on advanced sensing”, Wireless
Personal Communications Journal, accepted
L.D. Stone, T.L. Corwin, C.A. Barlow, “Bayesian Multiple Target Tracking”, Artech House Inc., Norwood, MA, 1999.
M. Wimmer, B. Schuller, D. Arsic, G. Rigoll, and B. Radig. "Low-Level Fusion of Audio, Video Feature for Multi-Modal
Emotion Recognition." In VISAPP, vol. 2, pp. 145-151, 2008.
A. Doucet, N. de Freitas, N. Gordon, “Sequential Monte Carlo Methods in Practice (Statistics for Engineering and
Information Science)”, Springer, New York, 2001.
D. Crisan, A. Doucet, “A survey of convergence results on particle filtering methods for practitioners”, IEEE Transactions
on Signal Processing vol. 50, no. 3, pp. 736–746, 2002.
S.J. Julier, J.K. Uhlmann, “A non-divergent algorithm in the presence of unknown correlation”, Proc. of the American
Control Conference, pp. 2369–2373, 1997.
A. Vulpe, et.al. (2015)
17. References (2) Chen, P.O. Arambel, R.K. Mehra, “Estimation under unknown correlation: covariance intersection revisited”, IEEE
Transactions on Automatic Control 47, no. 11 pp. 1879–1882, 2002.
A.R. Benaskeur, “Consistent fusion of correlated data sources”, Proc. of IEEE Annual Conference of the Industrial
Electronics Society, pp. 2652–2656, 2002.
A.R. Benaskeur, “Consistent fusion of correlated data sources”, Proc. of IEEE Annual Conference of the Industrial
Electronics Society, pp. 2652–2656, 2002.
D. Patranabis, “Sensors and Transducers”, Second Edition, PHI, 2003.
Craciunescu, R.; Halunga, S.; Fratu, O., „Wireless ZigBee home automation system”, Proc. SPIE 9258, Advanced Topics
in Optoelectronics, Microelectronics, and Nanotechnologies VII, February 21, 2015, Constanta Romania
T. Sullivan, S. Deiss, G. Cauwenberghs, “A Low-Noise, Non-Contact EEG/ECG Sensor”, In Biomedical Circuits and
Systems Conference, 2007. BIOCAS 2007. IEEE, pp. 154-157, 2007.
S. Cook, M. Togni, M. C. Schaub et al., “High heart rate: a cardiovascular risk factor?”, Eur Heart Journal, vol. 27, no.
20, pp. 2387-93, Oct, 2006.
G. Suciu, S. Halunga, O. Fratu, A. Vasilescu, V. Suciu, "Study for renewable energy telemetry using a decentralized
cloud M2M system," Wireless Personal Multimedia Communications (WPMC), 16th International Symposium on, IEEE,
pp. 1-4, 2013.
The eWALL Consortium, “D3.2.1, Perception from sensors,” eWALL for Active Long Living FP7 project, Oct. 2014.
C. Tsai, Y. Bai, C. Chu, C. Chung and M. Lin ,”PIR-sensor-based lighting device with ultra-low standby power
consumption”, Consumer Electronics, IEEE Transactions on, vol. 57, no. 3, pp. 1157-1164, 2011.
F. Talantzis, A. Pnevmatikakis and A.G. Constantinides, “Audio-Visual Person Tracking: A Practical Approach”, World
Scientific Publication Company, Imperial College Press, Dec. 2011
G. Suciu, C. Cernat, G. Todoran, V. Suciu, V. Poenaru, T. Militaru, and S. Halunga. "A solution for implementing
resilience in open source Cloud platforms." In Communications (COMM), 2012 9th International Conference on, IEEE, pp.
335-338, 2012.
A. Vulpe, et.al. (2015)