Patient Monitoring based on Mobile Sensor Network
Eliasz Kańtoch
AGH University of Science and Technology, Krakow, Poland
Keywords: Ubiquitous Computing, Telemedicine, Mobile Sensors, Body Area Networks, Mobile Health System,
Computer Aided-diagnosis.
Abstract: Mobile sensor networks offer opportunities to monitor state of health without constraining the activities of a
wearer. These mobile systems are now realizable due to miniaturization of integrated circuits, low power
microcontrollers and wireless communications. This paper presents a design of mobile sensor network
patient monitoring system based on several health status and movement monitoring sensors including ECG,
GPS and accelerometer. Key functionalities of the software include calculating and displaying the values of
heart rate variability parameters and movement patterns. Detection of QRS is based on Pan-Tompkins
algorithm. Algorithms detect non-standard situation and in case of emergency sends an SMS to a selected
number. System uses Android smartphone as a gateway to forward patient electronic health record to
medical web server, where data are available via graphical web-based user interface.
1 INTRODUCTION
Medical professions and practitioners in health care
as well as engineers in the area of the information
and communication technology have shown recently
great interest in body sensor networks (BSN) or
WBAN(wireless body area network).
A WBAN consists of one or more wearable
network nodes, each of them capable of sensing, and
processing one or more physiological signals (e.g.,
heart rate, blood oxygen saturation, physical activity
(e.g., body orientation, type and level of activity),
and environmental parameters (location, light,
atmospheric pressure). These nodes are placed on
the human body as tiny patches or attached to users’
clothes (Jovanov et al., 2005; http://ieee802.org/).
There are several examples of commercial
systems based on WBAN. The most common
application is monitoring of cardiac patients.
Corventis System (Corventis, 2012) consist of
wearable device that captures ECG data and a
mobile transmitter. It offers continuous surveillance
of symptomatic and asymptomatic cardiac
abnormalities to help physicians diagnose and treat
cardiac arrhythmias. When an arrhythmia is
detected, system automatically transmits the ECG
via a wireless data transmitter device to the
Monitoring Center. Another example of cardiac
monitoring system is CardioNet (CardioNet, 2012)
which monitors the patient via the small sensor
during normal daily routine. As events occur, patient
activity is automatically transmitted to the
Monitoring Center for analysis and response.
CardioNet is focused on helping physicians diagnose
and treat patients with arrhythmias.
This paper presents a WBAN-based health
monitoring system which integrates wearable and
battery-operated ECG, movement (accelerometer)
and GPS sensors which send data to mobile phone
via wireless Bluetooth network. Algorithms were
developed that process and analyze signals in real
time in order to calculate heart rate and locate the
patient. The main advantage of the system is
algorithms optimization for real time data
processing.
The rest of the paper is organized as follows.
Section 2 overviews the system design. Section 3
presents hardware used to build a prototype. Section
4 describes principles of vital signal processing.
Section 5 discusses the paper and describes ideas for
the future work. Section 6 concludes the paper.
2 SYSTEM DESIGN
The architecture of the proposed health monitoring
system consists of a mobile wireless ECG,
acceleration activity(ACC) and GPS sensors that is
295
Ka
´
ntoch E..
Patient Monitoring based on Mobile Sensor Network.
DOI: 10.5220/0004125302950298
In Proceedings of the International Conference on Signal Processing and Multimedia Applications and Wireless Information Networks and Systems
(WINSYS-2012), pages 295-298
ISBN: 978-989-8565-25-9
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
placed on the user body and a monitoring gateway.
The sensor samples, processes and sends the
information about user ECG, ACC signals and GPS
location to the monitoring gateway via Bluetooth.
Received data is analyzed by custom built
algorithms and forwarded to Internet database,
where medical data is accessible to a doctor. Figure
1 presents the system design.
Figure 1: System design.
System was designed to face the following
requirements:
wearable,
use miniature and low power sensors,
wireless communication using Bluetooth,
battery operated,
allow monitor patient ECG signal,
easy to use,
automatic fall detection,
calculation of heart rate,
low-cost,
secure,
data remote access interface.
3 MOBILE SENSORS AND
MONITORING GATEWAY
Monitoring system integrates Aspel Aspekt 500
ECG signal transmitter and Android based
smartphone with GPS used as monitoring gateway to
perform analysis and forward data to dedicated
medical web server. Aspekt 500 (Fig. 2) is a digital
unit designated for wireless ECG signal
transmission. It is equipped with a ten electrodes
cable. ECG signal is sampled at the frequency of
500 Hz. The transmitter allows a free patients
movement up to 10 m from monitoring unit. Aspekt
500 is a portable sensor (dimensions 130×96×30
mm).
Figure 2: Aspel Aspekt 500 ECG signal transmitter
The MMA7341L is a low power, miniature
(3mm x 5mm x 1mm) capacitive micromachined
accelerometer featuring signal conditioning, low
current consumption (400uA), temperature
compensation and self test. The typical application is
tilt and motion sensing.
Figure 3: Accelerometer MMA7341L.
Smartphone HTC Desire (119 x 60 x 11.9 mm) -
monitoring gateway is used to acquire monitoring
signal and forward it to the medical web-server. It is
equipped with 1 GHz Scorpion CPU, GPS and has
576 MB RAM, what makes it very powerful
processing unit. Smartphone runs Android OS v2.2
(Eclair).
Figure 4: HTC Desire.
I International onerence on ireless Inormation etworks and ystems
All this data are forwarded using the Internet via
UMTS or GSM network to a purposely-designed
medical web server where it is accessible to a doctor
or selected people through easy to use web-based
graphical interface. In case of emergency an action
is automatically taken.
4 VITAL SIGNAL PROCESSING
The most challenging task was to design the
software that will cope which huge data flow from
the monitoring device and real time signal analysis
at the same time. As a result algorithms were
designed and implemented.
Key functionalities of the software include
calculating and displaying the values of heart rate
and locating patient based on GPS data. All data can
be saved in the file for further statistical analysis.
One of the most popular and often cited QRS
detection algorithms that works in the time domain
is the Pan and Tomkins algorithm that was proposed
in 1985 (Tompkins and Pan, 1985). The QRS
detection algorithm is based on analysis of the slope,
amplitude and width of the QRS complex which
refers to the depolarization of the right and left
ventricles.
First, in order to reduce noise, the ECG signal
passes through a digital bandpass filter composed of
cascaded high-pass and low-pass filters. The next
process after filtering is differentiation, followed by
squaring, and then moving window integration.
Algorithms were implemented in Java.
Therefore, they can be used by a wide range of
smartphone devices and operating systems. One of
the project goals is to locate the patient. Tracking is
based on GPS data that are forwarded to a server and
accessible through web based interface which uses
Google Maps API(http://code.google.com/) in order
to mark patient location on the map (Fig. 5).
Positionning data (longitude, lattitude and height
above the sea level) are used to calculate the speed
and inclination and with assumption of active
motion and knowing the subject’s body weight
determine the total energy required. This value is
then correlated with the heart rate variability in order
to determine the correctness of its acceleration in
response to a physica load and deceleration in the
rest phase. The respective factors, although not yet
widely accepted by cardiologists are provided by the
system when everyday activities are used as a safe
alternative to a regular stress test.
Fall detection is perform by algorithm is
responsible for analyzing data from accelerometer.
Figure 5: Localization of patient via web page interface.
The absolute sum of tree axis accelerometer data is
calculated. If obtained value is higher than
experimentally set trigger point, the alarm module is
switched on.
Existing fall detection solutions analyzes
acceleration to detect falls. In Ref. (Mathie et al.,
2001) a single, waist-mounted, tri-axial
accelerometer is used to detect falls. Lindemann
(Lindemann et al., 2005) integrated a tri-axial
accelerometer into a hearing aid housing, and used
thresholds for acceleration and velocity to decide if
falls happen. In Ref. (Jantaraprim et al., 2010) a fall
detection algorithm utilizing two thresholds for the
resultant acceleration in 1.5-s window segments was
presented. The results, tested on 300 sequences show
that falls can be distinguished from activities of
daily living with 100% sensitivity and more that
93% specificity.
5 DISCUSSION AND RESULTS
Testing the QRS detection with MIT/BIH
Arrhythmia Database resulted in 94.50% sensitivity.
Through analysis of the 20 recorded simulated falls,
85% were correctly identified.
WBAN is one of the key components of the
future e-health initiative that could make significant
improvements in patient care and monitoring. The
application of the WBAN technology could make
some of the specialist treatments more accessible
and efficient as well as cost-effective for the service
delivery point of views (Otto et al., 2006; Yang,
2006).
Sensor-based measurements and monitoring
techniques have been widely used in electronic
Patient onitoring based on obile ensor etwork
patient care systems for a long time. The concept of
sensor-based patient monitoring using wireless body
area network (WBAN) will bring revolutionary
changes in health care systems. WBAN allows
flexibility in providing location independent and
seamless patient monitoring without affecting the
lifestyle of patients.
6 CONCLUSIONS
This paper presents an overview of wearable
WBAN-based patient monitoring system. Potential
applications include early detection of abnormal
conditions and supervised cardiac rehabilitation.
Automatic integration of collected information and
user’s inputs into research databases can provide
medical community with opportunity to search for
personalized trends and group patterns, allowing
insights into disease evolution, the rehabilitation
process, and the effects of drug therapy.
The achieved results are satisfactory for the
monitoring purposes. However, more tests are
needed to develop system that will focus on
prevention and early detection of health conditions.
ACKNOWLEDGEMENTS
This work was supported by the National Research
Center as a research project No. 2011/01/N/ST7/
06779.
REFERENCES
W. J. Tompkins and J. Pan, A Real-Time QRS Detection
Algorithm, Biomedical Signal Analysis, IEEE Press
(1985), Vol. BME-32.
Jovanov, E., Milenkovic, A., Otto, C., and de Groen, P.,
“Awireless body area network of intelligent motion
sensors for computer assisted physical rehabilitation.”
J NeuroEng Rehabilitation. 2:6, 2005, March 1, 2005.
C. Otto, A. Milenkovic, C. Sanders, and E. Jovanov,
"System Architecture of a Wireless Body Area Sensor
Network for Ubiquitous Health Monitoring," Journal
of Mobile Multimedia, vol. 1, 2006, pp. 307-326.
Yang, G-Z., Body Sensor Networks, Ed. Springer, 2006.
Bonato, P.,“Wearable Sensors and Systems.” IEEE Eng.
Med. Biol. Mag. 29(3):25–36, May–June 2010.
Aspel ASPEKT 500 technical documentation:
http://www.aspel.com.pl/index.php?lang=pl&pg=372
IEEE 802.15 WPAN Task Group 6 Body Area Networks
(BAN), http://ieee802.org/15/pub/TG6.html.
Google Maps API http://code.google.com/apis/maps/
index.html
M. J. Mathie, J. Basilakis, and B. G. Celler, A system for
monitoring posture and physical activity using
accelerometers, engineering in medicine and biology
society, 2001, Proceedings of the 23rd Annual
International Conference of the IEEE (2001), Vol. 4,
pp. 3654–3657.
U. Lindemann, A. Hock, M. Stuber, W. Keck, and C.
Becker, Evaluation of a fall detector based on
accelerometers: A pilot study. Med. Biol. Eng.
Comput.43, 548 (2005).
P. Jantaraprim, P. Phukpattaranont, C. Limsakul, and B.
Wongkittisuksa, Improving the accuracy of a fall
detection algorithm using free fall characteristics,
Electrical Engineering/Electronics Computer
Telecommunications and Information Technology
(ECTI-CON), 2010 International Conference, May
(2010), pp. 501–504.
Corventis, http://www.corventis.com/ (Feb 2012)
CardioNet, http://www.cardionet.com/ (Feb 2012)
Chen, X., Ho, C.T., Lim, E.T., Kyaw, T.Z.: Cellular Phone
based online ECG Processing for Ambulatory and
Continuous Detection. Computers in Cardiology 34,
653-656 (2007).
Sixsmih, A., Johnson, N.: A smart sensor to detect the
falls of the elderly. IEEE in pervasive Computing 3(2),
42-47 (2004).
Sun, H., De Florio, V., Gui, N., Blondia, C.: Promises and
Challenges of Ambient Assisted Living Systems. In:
Sixth International Conference on Information
Technology: New Generations, ITNG 2009, pp. 1201-
1207 (2009).
I International onerence on ireless Inormation etworks and ystems
