This article has been accepted for inclusion in a future issue of this journal.
Content is final as presented, with the exception of pagination.
IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS
Wearable Noncontact Armband for
Mobile ECG Monitoring System
Vega Pradana Rachim and Wan-Young Chung, Member, IEEE
Abstract—One of the best ways to obtain health information
is from an electrocardiogram (ECG). Through an ECG, charac-
teristics such as patients’ heartbeats, heart conditions, and heart
disease can be analyzed. Unfortunately, most available healthcare
devices do not provide clinical data such as information regarding
patients’ heart activities. Many researchers have tried to solve this
problem by inventing wearable heart monitoring systems with a
chest strap or wristband, but their performances were not feasible
for practical applications. Thus, the aim of this study is to build
a new system to monitor heart activity through ECG signals. The
proposed system consists of capacitive-coupled electrodes embed-
ded in an armband. It is considered to be a reliable, robust, and
low-power-transmission ECG monitoring system. The reliability
of this system was achieved by the careful placement of sensors in
the armband. Bluetooth low energy (BLE) was used as the protocol
for data transmission; this protocol was proposed to develop the
low-power-transmission system. For robustness, the proposed sys-
tem is equipped with analysis capabilities–e.g., real-time heartbeat
detection and a filter algorithm to ignore distractions from body
movements or noise from the environment.
Index Terms—Armband, bluetooth low energy, capacitive-
coupled, ECG monitoring, wearable device.
I. I NTRODUCTION
M -HEALTH or mobile health [1] is a term that is used to
describe healthcare services with the assistance of mo-
bile devices. The M-health monitoring and diagnostic medical
devices market is currently in its infancy but is expected to grow
rapidly due to market demand for remote monitoring systems.
According to a new report from Transparency Market Research
[2], the market for mobile health monitoring and diagnostics
will grow at a compound annual growth rate of 43.3 percent
from 2013 to 2019. This will put the market at $8 billion in
2019. Increased sports activities and awareness about health
and fitness are some of the factors that are anticipated to fuel the
growth of this market [3]. Unfortunately, high demand does not
automatically guarantee that the devices available in the market
are reliable and trustworthy.
Manuscript received February 23, 2015; revised October 11, 2015 and
December 10, 2015; accepted December 21, 2015. This research was
supported by the Basic Science Research Program through the National
Research Foundation of Korea (NRF) funded by the Ministry of Education
(2012R1A1B3004140). This paper was recommended by Associate Editor
S. Leonhardt.
The authors are with the Department of Electronic Engineering, Pukyong
National University, Busan 608739, South Korea (e-mail: veggapradana@
yahoo.com; wychung@pknu.ac.kr).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TBCAS.2016.2519523
1932-4545 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1
Various studies have been conducted related to the devel-
opment of remote healthcare systems, especially heart rate
monitoring systems. Most of these studies have focused on
four main topics: sensor technology, wearable systems, signal
processing, and mobile monitoring systems. Researchers have
tried to develop sensors that are able to sense bio-signals with-
out generating side effects or distracting users, but these were
still unable to provide a high level of accuracy. The standard
procedure in a hospital to measure an electrocardiogram (ECG)
is to use a 12-lead ECG. In a conventional 12-lead clinical ECG
system, electrodes are affixed to specific parts of the chest,
arms, or hands and legs. Even though this promises highly
accurate results, it often requires a great deal of preparation
and an expert to attach the electrodes to the patient’s body.
These electrodes also require skin preparation and conduction
gel to reduce contact impedance [4]. The main problem with
this method is that it cannot be used for long-term measurement
because of the limitation of the gel and complicated wiring.
Because of this, other techniques have been provided to over-
come these weaknesses, such as using dry electrodes [5], [6] or
noncontact electrodes [7]–[13].
Dry electrodes consist of a metal with no electrolyte or con-
ductive gel between the electrode and the skin. Instead, sweat
or moisture on the skin will reduce the impedance between
the skin and the electrodes. Various applications have been
proposed using the dry-electrode technique, but these required
direct contact with the skin. Thus, researchers have invented a
noncontact or capacitive-coupled ECG. The capacitive-coupled
ECG was first introduced by Lopez and Richardson [7].
Then, researchers expanded its use in various environments.
Lim et al. [8] and Yama et al. [9] developed a heart monitoring
system in office chairs and mattresses using flexible fabric
electrodes. Leonhardt et al. [10] introduced the idea of an
insulated electrode implemented in a car system. Lee et al. [11]
proposed the use of thin and flexible electrodes for a wearable
ECG system to build a system that can measure an ECG
in remote areas. Oehler et al. [12] proposed an integrated
15-capacitive-electrode array and combined this array with a
personal computer tablet.
Different placements of sensors have been proposed by var-
ious researchers. Some have suggested using belt-type ECG
monitoring systems with the implementation of conductive
fabric and active electrodes [14], [15]. Others have suggested
using bio-clothes for vital sign monitoring [16], [17]. Different
placements would provide different shapes of the ECG signal;
thus, they tried to find the best place from which to measure an
ECG based on demand. Nevertheless, researchers also focused
on developing a mobile monitoring system. This system uses
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
2 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS
Fig. 1. Hardware architecture with implementation in armband.
bio-signals to analyze health conditions with a mobile device
as an interface and an analysis device. Several applications
have been researched and proposed by other researchers in
[18]–[25]—for example, to monitor HRV for vehicle drivers,
detect arrhythmia, observe sleep disorders, and monitor driver
fatigue.
Therefore, capacitive-coupled technology is used in the pro-
posed system to overcome these problems. It is a promising
method to measure an ECG in remote areas without inconve-
niencing or disturbing the user. The advantages of this tech-
nique are that it can measure an ECG without direct contact
and provides better accuracy than other methods, especially for
measuring stress. The proposed monitoring system is impor-
tant for monitoring exercise intensity, estimation of maximal
oxygen uptake and energy expenditure and early detection
and in helping keep persons healthy by being able to track
their heart activities at any time. Currently, most systems that
are available in the market use a chest strap or a wristband,
but this system is uncomfortable and requires direct contact
with the skin. Consequently, an armband was chosen as an
alternative to this inconvenient system. The proposed arm-
band ECG is equipped with a variety of features. Considering
mobile monitoring, it is equipped with an Android mobile
application and Bluetooth low energy (BLE) data transmission.
Real-time heart rate detection was added to the Android appli-
cation, and this was tested in various scenarios, from sitting in
a chair, standing, and walking to jogging or running in place.
The results show that this monitoring can function in all of
these scenarios.
II. P ROPOSED H ARDWARE P LATFORM
An overview of the design for the data acquisition module is
shown in Fig. 1. As shown in the figure, conductive electrodes
act as sensors to obtain electrical signals from the human body.
The conductive electrodes are flexible, soft, and bendable. This
Fig. 2. Block diagram of analog filter.
prevents the user from feeling uncomfortable when using the
device. It is made of silver-plated polyester that can be stretched
in both directions with a thickness of 0.45 mm and a surface
resistivity of < 1 Ω/ sq. The proposed electrodes are sewn
into the armband, and the device contains plus electrodes,
minus electrodes, and ground electrodes, as shown in Fig. 1(c).
The electronic circuitry in the proposed system consists of a
Bluetooth 4.0 module, an ECG module, and a microcontroller,
as shown in Fig. 1(b). The Bluetooth 4.0 module was chosen
to build a low transmission power system with active mode TX
of approximately 24 mA and active mode RX of 19.6 mA. The
area of the module is approximately 1810 mm2 with a diameter
of 48 mm and a round shape. It has a cutoff frequency in the
range of 0.04–40 Hz. Finally, LilyPad Arduino is mainly used
for analog-to-digital conversion (ADC).
A. Analog Circuit Design
Based on the AHA document recommendation, the low fre-
quency cutoff is 0.05 Hz for routine filters, but this requirement
could be relaxed to 0.67 Hz or below [26]. In addition, the
ANSI/AAMI standard recommends a high frequency cutoff of
at least 150 Hz for all standard 12-lead ECGs; however, 40 Hz
as a high frequency cutoff is sufficient, although it is invalid
for amplitude measurement for diagnostic classification [26]. In
the proposed system, 0.04 Hz for a high-pass filter and 40 Hz
for a low-pass filter were used. Also, a notch filter was used to
eliminate the power line noise of 60 Hz.
Two sensor electrodes and one driven-right-leg electrode
are used in the proposed study. A driven-right-leg circuit [27]
was designed as an additional reference electrode to suppress
the interference caused by the finite common-mode rejection
ratio (CMRR) of the instrumental amplifier. The overall block
diagram of the analog filter is shown in Fig. 2.
B. Active Electrodes
The sensors consist of two active electrodes with a contact
area of 3 cm × 3 cm. The coupling capacitance is formed
between the subject arm and the input of the unity gain am-
plifiers A1 and A2. The integrated unity gain amplifiers ensure
the required high input impedance of the measurement. The
electrode was directly connected to a high input impedance pre-
amplifier, as shown in Fig. 3. An op-amp OPA 124 (Burr Brown
Corp., Dallas, TX, USA) was employed as a pre-amplifier.
The input resistance and input capacitance of OPA 124 were
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

RACHIM AND CHUNG: WEARABLE NONCONTACT ARMBAND FOR MOBILE ECG MONITORING SYSTEM 3

Fig. 4. Filtering and RR detection.

Fig. 3. Active electrode system configuration.

1013 Ω and 1 pF, respectively. As shown in (1), the coupling

capacitance depends on the thickness d, the dielectric constant
εR of the cloth, and the subject’s skin. Assuming values of
d = 0.4 mm and εR = 1, the coupling capacitance between the
body and the electrodes is
εo εR A
Cc = = 20 pF. (1) Fig. 5. Setup of real-time monitoring environment.
d
The electrodes need to be actively shielded to suppress the via inverted amplifier A3 to a conductive plane. Resistors RA
noise from the environment, mainly the static charge on the and RF adjust the amplifier’s gain
coupling capacitance. A static charge is formed if two objects
come into contact with each other, and interaction between 2RF
G= = 1000. (3)
the atoms (adhesion) will appear at the boundary layer. De- RA
pending on the different electron affinities, one object may
The driven ground plane circuit actually drives a very small
deliver electrons, and another may accumulate electrons on the
amount of current (< 1 uA) into the right leg to equal the
surface. An electric double layer will appear with a contact
displacement current flowing in the body. The body therefore
voltage between both material surfaces. When a circuit consists
becomes a summing junction in a feedback loop, and the
of only a charged capacitor and a resistor, the capacitor will
negative feedback from this circuit drives the common-mode
discharge its stored energy through the resistor. Thus, it may be
voltage (VCM) to a low value.
discharged over Rbias . The value of Rbias needs to be carefully
adjusted because it will lead to the cutoff frequency of the
D. Peak Detection
electrodes. A value of 100 GΩ for Rbias was selected. Hence,
the proposed electrode cutoff frequency is In the proposed system, Pan & Tompkins algorithm to
process the ECG data from the sensor and detect the heartbeat
1
fc = = 0.08 Hz. (2) from a particular innovative ECG signal is implemented. The
Cc Rbias 2π processing steps of the algorithm [28] are as follows (Fig. 4):
(i) band-pass filtering composed of cascaded high- and low-
C. Common-Mode Rejection Ratio pass filters; (ii) differentiation, which is a standard technique
for finding the high slopes (distinguishing the QRS complex
Because of the finite CMRR of the instrumental amplifier,
from the other ECG waves); (iii) squaring to make all data
the ECG signal experiences interference, and this common-
positive; (iv) moving-window integration (in our case, with a
mode noise needs to be suppressed. For this purpose, a driven-
sampling rate of 100 Hz, the size of the window is 15 samples);
right-leg circuit with an additional reference electrode is used
(v) determining the threshold to identify the peaks of the signal;
in most conventional ECG measurements [27]. In noncontact
and (vi) detecting the QRS complex. A peak must be recognized
measurements, the same concept of the driven right leg is called
as the QRS complex in the integrated and filtered waveforms.
a driven ground plane system. A driven ground plane electrode
is connected to the instrument ground and driven by the inverse
III. M EASUREMENT S ETUP
of the common-mode noise component. Resistors RA and RF
average the voltage of the differential electrode pair to sense The proposed system serves as a smart, noncontact health-
VCM . The pre-amplifier A3 in Fig. 3 amplifies and inverts this care monitoring system based on mobile phones that is a
voltage and then feeds it back to the body via the reference convenient, noninvasive system. A setup scenario of the mea-
electrode. The sum of the electrode’s output signal is feedback surement is shown in Fig. 5. This figure shows a user wearing
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
4 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS
Fig. 6. ECG signal result after band-pass filter from right arm (left) and left
arm (right).
Fig. 7. Signal output of monitoring system: Standard system (top) and pro-
posed system (bottom).
a running jacket with the armband in his left hand. From the
experiment shown in Fig. 6, the amplitude of the ECG signal
that comes from the left arm is higher than the signal from
the right arm. This signal intensity difference is due to the
distance difference from signal origin, heart. Left arm is slightly
near from heart and the ECG signal at the left armband shows
stronger than the signal from right arm. This result also proved
the previous conclusions in previous studies regarding single-
arm ECG measurement [30], [31].
While wearing the armband, the subject carried out daily
activities to evaluate the performance of the armband. The
armband receives the ECG data, filters it, and then transmits
the filtered signal to the Android smartphone via Bluetooth.
IV. VALIDATION E XPERIMENTS
A. Sensor Validation
Experiments were performed to validate the proposed arm-
band device. These compared the output signal with the stan-
dard wet-electrode system and verified the output signal while
performing daily activities. As shown in Fig. 7, the signal
output of the proposed system is comparable with the signal
that comes from a clinical standard lead 2 wet-electrode system.
The main part of the ECG that is used for the monitoring system
is the R wave, and the result shows that the QRS wave is quite
clear and is identical to the standard system signal.
B. Sensor Testing With Different Cotton Wool Shirt Thickness
Variations of the cotton wool shirt as the dielectric between
the electrodes and the skin are examined. This shows that
the ECG signal obtained from a user wearing thicker clothes
has more noise than that obtained from a user wearing
thinner clothes, especially in the PT wave of the ECG signal.
Fig. 8. Comparison of the ECG signal with different cotton wool shirt thick-
nesses: 0.5 mm (top), 1 mm (middle), and 3 mm (bottom).
Although this particular experiment shows the effect of the shirt
thickness on the obtained ECG, the system still shows a clear
peak or QRS wave, which is the most important information
in the ECG signal. The overall results of this experiment are
shown in Fig. 8.
C. Peak Detection Algorithm Testing With Different Activities
Multiple activities were performed by the subject wearing
the armband, including sitting, hand movements, jogging, and
running in place. The PCs mainly received the raw data from
the proposed application through the Internet; thus, Matlab
was used to calculate the performance of the recorded signal.
In Matlab, the recorded data were filtered using the same
algorithm that was implemented in the Android application. In
this step, the result of the performance evaluation was referred
to the Matlab application proposed in [29]. Basically, there are
two processing parts: the filtering part and the peak detection
part. Therefore, the results were compared with those of other
standard devices, and a missed peak calculation was performed
to evaluate the performance of the proposed system.
Even though we focus on presenting the device as an alterna-
tive ECG monitoring tool for sports activity measurement and
awareness about health and fitness during excise or daily activ-
ity, and not for precise medical use like arrhythmia detection at
hospital, we checked the performance of our redesigned QRS
detection algorithm. The performance of the QRS detection
algorithm in PC was tested using a standard arrhythmia, that is,
MIT/BIH. The performance of the algorithm for No. 106 and
No. 119 from MIT/BIH database produced 0.2% false negative
beats and 0.3% false positive beats for a total detection failure
of 0.5%. The performance of the algorithm on database is not
the same as to its utility in a clinical environment, however
it can provide a means of comparing the performance of one
algorithm to others.
Fig. 9 shows the filtered and QRS detection results from the
ECG signal from the hand movement and jogging scenarios. As
shown in the figure, the signal peak was correctly detected in
the original data. The green line in the graph shows the signal
level, while the black line shows the noise level. There is no
error in the peak detection data when the moving scenario is
applied. However, when the algorithm tried to detect the signal
peak in the jogging situation, some of the peaks were wrongly
identified. The results in Fig. 9 (bottom) show that two peaks
are wrongly identified from the proposed algorithm.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

RACHIM AND CHUNG: WEARABLE NONCONTACT ARMBAND FOR MOBILE ECG MONITORING SYSTEM 5

Fig. 9. QRS detection in filtered ECG signal for various scenarios.

D. Peak Detection Algorithm Performance Analysis TABLE I

C OMPARISON OF P ROPOSED S YSTEM P ERFORMANCE
In the previous section, the graph that shows the performance W ITH C OMMERCIAL D EVICE
of the output signal was described; thus, in this section, the
performance of the output signal as a percentage is examined.
The experiment was performed by measuring the ECG signal
with two devices at the same time. In this experiment, the
signal output from the armband as the proposed system and
the commercial PulseSensor was calculated. In this experiment,
10 subjects with 75 kg of average weight have been involved.
Each subject did each activity for 5 min and 10 times repetition,
and 200 sampling peaks were taken in each 5 min of the activity
after 30 seconds. According to our survey on the relationship was jogging; the lowest error rate was 2.7% while the subject
between body weight and ECG signal detection performance, was sitting. For further experiments, we checked the missed
we found there are no linear relationship between two factors. peak from each activity; the results in Table II show that as the
Instead the signal performance was affected by the thickness of level of activity intensities, the higher the probability that the
the shirts or the material of the clothe under the armband. proposed system will miscalculate the peak.
Thus we summarized the data as can be seen in Table I. The
signal from PulseSensor strapped at the wrist [32] was used
E. Real-Time Monitoring System
as the ground truth of performance, so the error rate can be
calculated by comparing the signal. As shown in Table I, the As mentioned in the previous section, the proposed system
highest error rate was approximately 8.05% while the subject may obtain an ECG signal without direct contact with the skin;
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
6 IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS
TABLE II
M ISSED P EAK C ALCULATION
Fig. 10. Real-time monitoring with the subject wearing a jacket with good
(left) and bad (right) contact comparison.
Fig. 10 shows two kinds of signals that might be received from
the user. A bad signal might be caused by a wandering baseline
and muscle tremors. Baseline wandering occurs when there is
poor contact and movement of the cable due to respiration;
thus, muscle tremors may indicate that the proposed armband
is too tight.
V. C ONCLUSION
A new wearable device for a healthcare monitoring system
was proposed in this paper. The device was implemented in
an armband to achieve a non-obstructive system. An armband
was chosen as an alternative to previous wearable devices that
are strapped to the chest, which can be inconvenient for some
users. The technology used in this system uses capacitive-
coupled electrodes. The capacitive-coupled electrodes enable
the system to measure bio-signals through the clothes so that
the weaknesses from wet or dry electrodes can be reduced.
The bio-signal that was measured in this system is an ECG
signal. Although the ECG signal that is received from the arm is
quite small, the proposed system was smart enough to overcome
noise and detect useful information from the recorded signal.
Experiments were performed to evaluate the performance of
the system. Hardware testing was presented to show the output
signal of the proposed system in different scenarios. The results
from these experiments show that the proposed system can still
function even with different clothing thicknesses between the
sensors and the skin and can also still function when the user
carries out various daily living activities.
In the proposed system, an Android application was also
created to achieve a mobile healthcare system. The proposed
application can show an ECG signal in real time on a graph and
then analyze the data to make smart decisions. The proposed
algorithm was tested on a PC to validate the system. The results
indicate that the proposed heart rate calculation error rate is less
than 10% compared with a standard system.
R EFERENCES
[1] mHealth. [Online]. Available: http://en.wikipedia.org/wiki/MHealth
[2] mHealth Monitoring Diagnostic Medical Devices Market. [Online].
Available: http://www.transparencymarketresearch.com/mHealth-
monitoring-diagnostic-medical-devices.html
[3] mHealth Monitoring and Diagnostic Medical Devices Market Expected to
Reach USD 8.03Billion Globally in 2019 Transparency Market Research.
[Online]. Available: http://www.digitaljournal.com/pr/1877980
[4] H. A. Miller and D. C. Harrison, Biomedical Electrode Technology.
New York, NY, USA: Academic, 1974.
[5] Scarle and L. Kirkup, “A direct comparison of wet, dry and insulating
bioelectric recording electrodes,” Phys. Meas., vol. 21, no. 2, p. 271,
2000.
[6] Y. M. Chi, T. P. Jung, and G. Cauwenberhs, “Dry-contact and noncon-
tact bio potential electrodes: Methodological review,” IEEE Rev. Biomed.
Eng., vol. 3, pp. 106–119, 2010.
[7] J. Lopez and P. C. Richardson, “Capacitive electrocardiographic and
bioelectric electrodes,” IEEE Trans. Biomed. Eng., vol. 16, p. 99,
Jan. 1969.
[8] Y. Yama, A. Ueno, and Y. Uchikawa, “Development of a wireless capac-
itive sensor for ambulatory ECG monitoring over clothes,” in Proc. 29th
Int. Conf. IEEE EMBS, Lyon, pp. 5727–5730, Aug. 2007.
[9] Y. G. Lim, K. K. Kim, and K. S. Park, “ECG measurement on a chair
without conductive contact,” IEEE Trans. Biomed. Eng., vol. 53, no. 5,
pp. 956–959, May 2006.
[10] S. Leonhardt and A. Aleksandrowicz, “Non-contact ECG monitoring for
automotive application,” in Proc 5th Int. BSN Conf., Hong Kong, China,
Jun. 2008, pp. 183–185.
[11] S. M. Lee, K. S. Sim, K. K. Kim, Y. G. Lim, and K. S. Park, “Thin and
flexible active electrodes with shield for capacitive electrocardio-
gram measurement,” Med. Biol. Eng. Comput., vol. 48, pp. 447–457,
2010.
[12] M. Oehler, V. Ling, K. Melhorn, and M. Schiling, “A multichannel
portable ECG sytem with capacitive sensors,” Phys. Meas., vol. 29,
pp. 783–793, 2008.
[13] T. Torfs, Y. H. Chen, H. Kim, and R. F. Yazicioglu, “Noncontact ECG
recording system with real time capacitance measurement for motion
artifact reduction,” IEEE Trans. Biomed. Circuits Syst., vol. 8, no. 5,
pp. 617–625, Oct. 2014.
[14] D. D. Rossi, F. Carpi, F. Lorussi, A. Mazzoldi, R. Paradiso, E. P. Scilingo,
and A. Tognetti, “Electroactive fabric and wearable Biomonitoring de-
vices,” AUTEX Res. J., vol. 3, no. 4, pp. 180–185, Dec. 2003.
[15] M. Pacelli, G. Loriga, N. Tacccini, and R. Paradiso, “Sensing fabrics
for monitoring physiological and biomechanical variables: E-textile so-
lutions,” in Proc. 3rd Int. Conf. IEEE-EMBS, Sep. 2006.
[16] S. H. Lee, S. M. Jung, C. K. Lee, K. S. Jeong, G. Cho, and S. K Yoo,
“Wearable ECG monitoring system using conductive fabrics and active
electrodes,” in Human—Computer Interaction Handbook, J. A. Jacko, Ed.
Boca Raton, FL, USA: CRC Press, 2009, pt. III, pp. 778–783.
[17] K. M. Lee, S. M. Lee, and K. S. Park, “Belt-type wireless and non-contact
electrocardiogram monitoring system using flexible active electrode,” Int.
J. Bioelecromagn., vol. 12, no. 4, pp. 153–157, 2010.
[18] S. Gradl, P. Kugler, C. Lohmuller, and B. Eskofier, “Real-time ECG mon-
itoring and arrhythmia detection using Android-based mobile devices,”
in Proc. 34th Int. Conf. IEEE EMBS, San Diego, CA, USA, Aug. 2012,
pp. 2452–2455.
[19] J. Behar, A. Roebuck, M. Shahid, J. Daly, A. Hallack, N. Palmius,
J. Stradling, and G. D. Clifford, “SleepAp: An automated obstuctive sleep
apnoea screening application for smartphones,” IEEE J. Biomed. Health
Inform., pp. 2168–2194, 2013.
[20] V. P. Rachim, S. C. Kang, W. Y. Chung, and T. H. Kwon, “Implementation
of extended kalman filter for real-time noncontact ECG signal acquisi-
tion in android-based mobile monitoring system,” J. Sens. Sci. Technol.,
vol. 23, no. 1, pp. 7–14, 2014.
[21] P. C. Yang, J. H. Cheng, M. S. Tu, and C. H. Tseng, “A smartphone-based
heart rate variability analysis system for vehicle drivers,” in Proc. 12th
Int. Conf. ITS Telecom., 2012, pp. 827–831.
[22] S. J. Jung, H. S. Shin, and W. Y. Chung, “Driver fatigue and drowsiness
monitoring system with embedded electrocardiogram sensor on steering
wheel,” IET Intell. Transp. Syst., vol. 8, no. 1, pp. 43–50, 2014.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.
RACHIM AND CHUNG: WEARABLE NONCONTACT ARMBAND FOR MOBILE ECG MONITORING SYSTEM
[23] E. M. Fong, T. H. Kwon, and W. Y. Chung, “Seamless and non-contact
health monitoring system in cloud computing,” in Proc. IEEE Int. Conf.
Consumer Electronics, 2013, pp. 530–531.
[24] M. Patel, P. K. Gakare, and A. N. Cheeran, “Real time ECG feature
extraction and arrhythmia detection on a mobile platform,” Int. J. Comput.
Appl., vol. 44, no. 23, pp. 40–45, Apr. 2012.
[25] W. T. Sung, J. H. Chen, and K. W. Chang, “Mobile physiological measure-
ment platform with cloud and analysis functions implemented via IPSO,”
IEEE Sensors J., vol. 14, no. 1, pp. 111–123, Jan. 2014.
[26] P. Kligfield, L. S. Gettes, and J. J. Bailey, et al., “Recommen-
dations for the standardization and interpretation of the electro-
cardiogram: Part I: The electrocardiogram and its technology a
scientific statement from the American Heart Association Electrocar-
diography and Arrhythmias Committee, Council on Clinical Cardi-
ology; The American College of Cardiology Foundation; And the
Heart Rhythm Society,” Circulation, vol. 115, no. 10, pp. 1306–1324,
2007.
[27] K. K. Kim, Y. K. Lim, and K. S. Park, “Commong mode noise cancel-
lation for electrically non-contact ECG measurement system on a chair,”
in Proc. IEEE Engineering in Medicine and Biology 27th Annu. Conf.,
pp. 5881–5883, Sep. 2005.
[28] J. Pan and W. J. Tompkins, “A real-time QRS detection algorithm,” IEEE
Trans. Biomed. Eng., vol. 32, no. 3, pp. 230–236, 1985.
[29] H. Sedghamiz, Complete Pan Tompkins Implementation ECG QRS De-
tector. [Online]. Available: http://www.mathworks.co.kr/matlabcentral/
fileexchange/45840-complete-pan-tompkins-implementation-ecg-qrs-
detector
[30] H. C. Yang, T. F. Chien, S. H. Liu, and H. H. Chiang, Dept. Elec. Eng.,
Southern Taiwan University, Study of Single-Arm Electrode for ECG
Measurement Using Flexible Print Circuit.
[31] Application Note #291491: Single Arm ECG Measurement Using EPIC,
Plessey semiconductor.
[32] PulseSensor. [Online]. Available: http://pulsesensor.myshopify.com/
7
Vega Pradana Rachim received the B.E. degree in
electrical engineering from Diponegoro University,
Semarang, Indonesia, and the M.E. degree in elec-
tronic engineering from Pukyong National University,
Busan, South Korea, in 2012 and 2015, respectively.
Currently, he is working toward the Ph.D. degree
in engineering in the Department of Electronic Engi-
neering, Pukyong National University. His research
interests include wearable devices, capacitive cou-
pling, and ECGs.
Wan-Young Chung (M’02) received the B.Eng.
and Master’s degrees in electronic engineering from
Kyungpook National University, Daegu, South Korea,
in 1987 and 1989, respectively, and the Ph.D. de-
gree in sensor engineering from Kyushu University,
Fukuoka, Japan, in 1998.
From 1999 to 2008, he was an Associate Pro-
fessor at Dongseo University, Busan, South Korea.
Currently, he is a Professor in the Department of
Electronic Engineering, Pukyong National Univer-
sity, Busan. His research interests include wireless
sensor networks, ubiquitous healthcare and automobile applications, smart
light-emitting systems with visible light communication, and embedded sys-
tems. He is the team leader for next-generation u-healthcare technology devel-
opment, which is supported by the Brain Korea 21 (BK21) PLUS project.