ISWC '14 ADJUNCT, SEPTEMBER 13 - 17, 2014, SEATTLE, WA, USA
A Smart Scarf for Pulse Signal
Monitoring Using a Flexible Pressure
Nanosensor
Dongzhan Chen
Ting Zhang
Abstract
Mobile Computing Center,
iLab, Suzhou Institute of Nano-
University of Science and
Tech and Nano-Bionics,
Technology of China.
Chinese Academy of Sciences
Chengdu,China
Suzhou,
[email protected]
Jiangsu China
Real-time, long-term pulse signal monitoring plays a
significant role in monitoring chronic diseases for elder
people, pregnant women etc. For users with weak pulse,
a carotid pulse signal monitoring system can be
appropriate. However, till today no unobtrusive solution
is available. Therefore, we suggest a novel flexible and
planar pressure nanosensor weaved in a smart scarf
system for pulse signal monitoring with better user
experience called Smart-SP (smart scarf for pulse
signal monitoring system). To our knowledge, this is
the first time applying a flexible pressure nanosensor
on a wearable carotid pulse monitoring system. To
meet the needs of the application, a method is
proposed to design the size of the sensor. An interface
allows third-party analysis software and provides raw
data. A database of pulse signal for diagnostic purposes
is set-up. Digital low-pass filter improves the signal
accuracy.
[email protected]
Michael Lawo
TZI – Center for Computing and Yang Gu
Communication Technologies,
iLab, Suzhou Institute of Nano-
Universität Bremen,
Tech and Nano-Bionics,
Bremen, German
Chinese Academy of Sciences
[email protected]
Suzhou,
Jiangsu China
Yu Zhang
[email protected]
Mobile Computing Center,
University of Science and
Dongyi Chen
Technology of China.
（Corresponding author）
Chengdu,China
Mobile Computing Center,
[email protected]
University of Science and
Technology of China.
Author Keywords
Chengdu,China
wearable, carotid pulse monitoring, pressure
nanosensor, interface design
[email protected]
Permission to make digital or hard copies of part or all of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. Copyrights
for third-party components of this work must be honored. For all other
uses, contact the owner/author(s).
Copyright is held by the owner/author(s).
ISWC '14 Adjunct, September 13-17, 2014, Seattle, WA, USA
ACM 978-1-4503-3047-3/14/09.
237
Introduction
In recent years due to the ageing population [1] and
increasing healthcare costs wearable long-term
healthcare monitoring plays a more and more
significant role for disease prevention in daily life.
ISWC '14 ADJUNCT, SEPTEMBER 13 - 17, 2014, SEATTLE, WA, USA
Monitoring and diagnosis of some chronic disease like
cardiac disease, hypertensive disease and chronic
obstructive pulmonary disease (COPD) require artery
pulse signals [2]-[5]. Contour, duration and amplitude
of the pulse signal provide important information [6].
Users suffering diseases like cold syndrome or bad
blood circulation have pulse signals hard to detect,
including radial artery, ulnar artery but except carotid
artery signal. The carotid is an aorta easily detected
and signals there make sure that the patient is alive.
However, as the neck is sensitive and flexible a carotid
pulse monitoring system is difficult to design.
Figure 1. Detection of carotid pulse.
Several methods of measuring carotid pulse signals
have been proposed. The most common way is using
some photoelectric principle [7]. Light emitted by LED
is reflected and transmitted by tissue, muscles, skin
and blood vessels before received by the photodiode.
To avoid the photodiode influenced by other illuminant
the sensor must be pressed onto the users’ body. For
the sensitive neck it is obviously an uncomfortable
method. Paper [8] presents a method to detect the
carotid pulse signal based on electromagnetic induction.
The carotid pulse is in [9] detected using the principle
of the Bragg probe. Yu-Pin Hsu and Darrin J. Young
used a MEMS pressure sensor [10] to detect the blood
pressure waveform by skin surface coupling. The
papers propose new but infeasible methods for
wearable systems, as the sensors used are hard, rigid,
bulky and heavy, which make them unfeasible to be
applied at the flexible and sensitive neck, as any
sudden movement may hurt the user. Thus, all the
above systems are inappropriate for any long-term
wearing. But without long-term data system is required
for diagnostic purposes. For elder, weak, bad blood
circulated, cold syndrome users, monitoring regions
238
such as the wrist and finger do not deliver reliable data.
In conclusion, building a long-term carotid pulse
monitoring wearable system is desirable.
This paper proposes a carotid pulse signal monitoring
system -- SmartSP, which is based on a new-type of
soft pressure nanosensor[11][12]. SmartSP can
provide the users the feeling of unrestrained and
relaxed to do indoor or outdoor activities while
constantly monitoring the carotid pulse. A user
centered design approach is proposed for the size of
the sensor meeting the needs of comfort and
measurability. Beyond this, SmartSP is designed as a
common scarf, users are wearing SmartSP will not feel
weird or even stigmatized. SmartSP cannot only suit for
critical patient, but might become fashionable. The
flexible, light-weight, comfortable and non-irritant
sensor is adhered on the skin-surface to get the carotid
pulse, and the PCB is weaved on the other side of the
scarf, so that users will not directly contact with hard
components. Because of the ultra-high sensitivity of the
sensor, the carotid pulse signal can be measured
without constraint. The data is collected in a general file
system integrated in SmartSP can be checked on mainstream operating system platforms. This interfaces
third-party analysis software. The data can be merged
with other physical data for improved diagnosis.
This paper explores and implements the new-type
nanosensor in the application of wearable carotid pulse
signal monitoring significant also for other applications
of soft sensors.
System design
Characters of the sensor and the detecting principle
The flexible pressure nanosensor was fabricated with
the method published in the previous work[11], which
WORKSHOP: WOSG
is constructed as a sandwich structure containing two
single-walled carbon nanotubes(SWCNTs) conductive
layers on micro-patterned polydimethylsiloxane(PDMS)
substrates, where the SWCNTs layers contact with each
other. Microstructured PDMS film is the most popular
flexible substrate to integrate sensitive nanomaterials
for the sensing applications due to its excellent
elasticity and biocompatibility, and SWCNTs networks
can be stretched and bended without large changes in
conductivity. Because of these advantages, the sensor
is soft with little effect on the users’ daily life when
applied. The pattern on the PDMS substrate will deform
when applied tiny pressure, so the contact resistance
between two SWCNTs conductive layers will change,
which allows to detectable pressure as low as 0.6 Pa
Carotid
Artery
and have a 1.8 kPa
(a) Side view of carotid artery
A
B
D
E
sensitivity under 300
Pa .
Besides, the size of the sensor can be defined as we
need. Through testing, these characters make the
sensor perfect to detect carotid pulse signal.
C
Carotid
Artery
1
Figure 1 shows the principle of detecting carotid pulse
by using the sensor. Carotid pulse causes shockwaves.
When a pulse wave is spreading to the skin-surface
through the muscle and tissue and then received by the
sensor, the resistance of the sensor will change
depending on the strength of the wave. The resistance
is converted into a voltage signal. With an electronic
amplifier and a low-pass filter the signal-to-noise ratio
is improved.
(b) Cross section of neck
Figure 2. Demonstration of
carotid artery position.
The design of sensor size
The sensor will be directly adhered onto the skin of the
neck. To identify the right position one might recall that
the carotid pulse signal should be detected on the side
of the neck as indicated in figure 2. A proper size of the
sensor should ensure covering the activity area. Figure
239
2(a) shows that the carotid artery is vertical
distribution in the neck. So, whether the sensor could
capture the signal just depends on the horizontal rather
than vertical position. Though there are different neck
perimeters, the persons’ carotid artery position is
individual with more or less similar proportion. Actually,
as the carotid pulse signal will spread, the section A-B
is affected as figure 2(b) shows. Although the sensor
can detect a wider section, for simplicity, this segment
AB is used as target section for putting the sensor on
the skin of the neck.
In the figure 2(b), the points C, D and E divide the
perimeter of the neck into 3 parts. The cylindrical
projection is shown in figure 3. Point C as the center of
the front side of the neck is easy to be located.
Although it is hard to locate the accurate point D, every
position of segment DE should doubtless belong to
the back of the neck. The sensor has to be in segment
CD
rather than segment
DE . Two extreme
conditions should be considered for an appropriate
position. If the sensor is placed like M or N in figure
3, the length of the sensor should be longer
than   Max( BD, AC ) . If so, the sensor can
certainly cover the segment AB and detect the carotid
pulse. Assuming that  : CD   and CD :   
(with  as the perimeter of neck), the length of the
sensor can be designed as follows:
  
(1)
The proportion of each segment of this model is
AC : AB : BD : DE  1:1:1: 3 . That is    
1
.
3
ISWC '14 ADJUNCT, SEPTEMBER 13 - 17, 2014, SEATTLE, WA, USA
Figure 3. Analysis model of dimension
of sensor.
Normal persons’ perimeter of neck is less than 400mm.
Substitute into the formula results to   45mm . After
consumption and the poor reliability of wireless
communication.
adding necessary package, the length of the sensor
should be     15  60mm .
Figure 5(a) shows the hardware of the system. Modules
of SmartSP are weaved in a common scarf and the wire
is weaved as well. Figure 5(b) left shows the on-body
appearance. As right of figure 5(b) shows a smart
phone or watch can be used to check the data in realtime. Users could adjust the system by themselves
according to the data.
Unless the users’ perimeter of the neck is less than

 180mm , the sensor will be usable. In this way

the sensor is suitable for most of users. In this case,
users who have wider neck perimeter can still locate
the active section accurately after simple practices.
(a) Hardware block diagram.
Carotid pulse signal
Digital Low-pass filter
Data storage
Data translate
(b)Digital signal processing stage.
Figure 4. Overview of the SmartSP.
System architecture
SmartSP meets the primary needs of wearable mobile
monitoring. Figure 4 shows the overview of the whole
system including hardware block diagram (a) and
digital signal processing stage (b). Signals acquired by
the sensor will be pre-processed by an AC coupling
circuit to obtain high noise-signal ratio outputs. And
then, signal is acquired with a frequency of 500Hz by
the 12-bit ADC which is integrated in the
STM32F103C8T6 microprocessor. Li-battery and its
management unit, Mini-USB (for charging battery) and
T-Flash Card interfaces are also designed in the MCU
module. The collected data will be processed by a
digital low-pass filter first. The preprocessed data is
sent via circumscribed ultra-low power consumption
Bluetooth module (HC-05) and stored onto a Micro SD
card through the SPI interface. The SmartSP integrated
with a FAT File System (FATFS) also store the data into
“.txt” files. Thus third-party developers and the users
can obtain the data easily. Storing data into Micro SD
card directly ensures the reliability and integrity of the
data and due to the band width of Bluetooth, power
240
Evaluation and Discussion
100 participants were invited to evaluate the SmartSP
(53 males and 47 females). The ages of the
participants arranged from 18 to 65 (average age is
about 39.6 and the standard deviation is about 12.4)
and education background arranged from junior high to
Ph.D degree. The distribution of age is that 23
participants in the 18-27 year old, 16 participants in
the 28-37 year old, 25 participants in the 38-47 year
old, 21 participants in the 48-57 year old, 15
participants in the 58-67 year old. The distribution of
education background is that 16 participants in junior
high degree, 25 participants in senior high degree, 29
participants in bachelor degree, 20 participants in
master degree, 10 participants in Ph.D degree. The
evaluation had two parts to check whether the SmartSP
was easy to use and the user experience.
To analyze if the SmartSP could capture carotid signal
with fuzzy location, the users were first instructed to
use the SmartSP by a direction-handbook of SmartSP.
The handbook explained what SmartSP was, how to use
it, its functions, what the sensor looked like and where
the correct position was to adhere to the sensor. To
enlighten the users, pictures and descriptions were
included in the direction-handbook. When a user was
WORKSHOP: WOSG
(a)
(b)
Modules of the system.
Wearing the SmartSP and display
based on Android platform.
Figure 5. Wearable monitor system.
At the second part, to investigate the user experience,
participants were asked to wear SmartSP for an hour.
Then, they were asked to describe their experiences. A
21-year-old male undergraduate said:” It just looks like
a common scarf. Nobody noticed that I was wearing an
electronic product. Even for me, I occasionally forgot
that it’s not just a scarf.” A 20-year-old female
participant said:” It’s amazing that I can see my heart
beat whenever I want to.” For a 41-year woman, she
thought her father would love it. “It’s a wonderful
wearable device. It's comfortable to wear.”, a 34-yearold teacher said. A 59-year-old worker asked where he
could buy a scarf like this.
As mentioned above, a low-pass filter is used to obtain
the signal with higher signal to noise ratio. Figure 6(a)
shows the raw data and figure 6(b) shows the signal
processed by the low-pass filter. As a result, low-pass
filter successfully restrain the noise signal. To following
analysis and research, a signal with better features is
necessary.
241
1 .9
Raw Carotid Pulse Waveform
1 .8 5
A m p litu d e (V )
1 .8
1 .7 5
1 .7
1 .6 5
1 .6
1 .5 5
1 .5
0
500
1000
1500
2000
2500
S a m p le
3000
3500
4000
4500
5000
1 .9
Filtered Carotid Pulse Waveform
1 .8 5
1 .8
A m p litu d e (V )
equipped with SmartSP, we consider the participant
using the system correctly if pulse signal was captured
by SmartSP. Participants with lower education
background and elders tend to use the SmartSP in a
wrong way. As a result, only 3 people could not
correctly use the SmartSP at the first time (The
background of the 3 persons are: 63, female, junior
high; 64, male, junior high; 58, female, junior high).
They could not understand the direction-handbook,
neither how to used SmartSP. After explanations and
training, they finally learned to use SmartSP
successfully. It means that this design ensures
SmartSP to be friendly to use.
1 .7 5
1 .7
1 .6 5
1 .6
1 .5 5
1 .5
0
500
1000
1500
2000
2500
s a m p le
3000
3500
4000
4500
5000
Figure 6. Carotid waveform processed by low-pass filter.
A disease cannot accurately be diagnosed by merely
one physical signal. Furthermore, different diseases
need different interpretations of the recorded data. In
case a specific algorithm is integrated into the system,
the application of this system will be limited. We
therefore set aside this interpretation of signals.
Conclusion and further work
A new method to monitor carotid pulse was presented.
A flexible nanosensor is used in this wearable
monitoring system called SmartSP. Users can
experience this comfortable, light-weight and
fashionable SmartSP. It meets the needs of real-time
wearable carotid pulse monitoring. The using of digital
low-pass filter produces output signal with high signal
noise ratio. As a result, relatively pure carotid pulse
signals were obtained.
In the future, the way to eliminate the influence of neck
motion, speaking and cough will be researched and
added. Further more, probably, other physical signals
like swallowing or breathing can be gathered
ISWC '14 ADJUNCT, SEPTEMBER 13 - 17, 2014, SEATTLE, WA, USA
automatically in one day for improving the accuracy in
disease analysis.
electromagnetic induction, Proceedings of 2012 ICME
International Conference on Complex Medical
Engineering”, IEEE, Japan, July 2012, pp. 441-444.
Acknowledgements
[9] C Leitão, L Bilro, N Alberto, et al. “Feasibility
studies of Bragg probe for noninvasive carotid pulse
waveform assessment”, Portugal, Journal of biomedical
optics, January 2013, 18(1): 017006.
We gratefully acknowledge MCC staff and volunteers,
and the study participants. This research is a part of
HealthWear@AAL(An Intelligent Wearable Solution for
Home and Outdoor Health Monitoring and Diagnosis),
which is funded by Sino-German Center (GZ817).
Reference
[1] World Bank, World Development Indicators 2012,
World Bank Publications, US, 04.2012
[2] A Sudden Drop in Heart Rate During Exercise.
http://getfit.jillianmichaels.com/sudden-drop-heartrate-during-exercise-2288.html
[3] Rapid or irregular heartbeat,
http://www.anxieties.com/21/panicstep1b#.Uq51LLLsWTU
[4] Heart Disease and Abnormal Heart Rhythm
(Arrhythmia), http://www.medicinenet.com/
arrhythmia_irregular_heartbeat/article.htm
[5] H L Quan, C L Blizzard, J E Sharman, et al. “Resting
Heart Rate and the Association of Physical Fitness With
Carotid Artery Stiffness”, American journal of
hypertension, US, January 2014: pp. 65-71.
[6] B. Karnath, W. Thornton, “Precordial and cordial
pulse palpation, Review of clinical signs, Hospital
Physician, Karnath&Thornton”, Palpation, July 2002, pp.
20-24.
[7] Yunjoo Lee, Hyonyoung Han, Kim Jung. “Influence
of Motion Artifacts on Photoplethysmographic Signals
for Measuring Pulse Rates”, International Conference on
Control, Automation and Systems 2008, Korea, October
2008. pp. 962-965
[8] H Hirano, T Fukuchi, Y Kurita, et al. “Development
of a palpable carotid pulse pressure sensor using
242
[10] Y P Hsu, D J Young. “Skin-surface-coupled personal
health monitoring system, Sensors”, IEEE, US, 2013
[11] X Wang, Y Gu, Z Xiong, et al. “Electronic Skin: Silk
‐Molded Flexible, Ultrasensitive, and Highly Stable
Electronic Skin for Monitoring Human Physiological
Signals”, Advanced Materials, 2014, 26(9): 1309-1309.
[12] C Pang, C Lee, K Y Suh. “Recent advances in
flexible sensors for wearable and implantable devices”,
Journal of Applied Polymer Science, 2013, 130(3):
1429-1441.
[13] G Yang, X Su, L Zhao, et al. “Research of portable
community-oriented health monitoring terminal”,
Proceedings of the 8th World Congress on Intelligent
Control and Automation (WCICA), IEEE, July 2010,
China, pp. 2979-2984.
[14] C Y Lim, K J Jang, H Kim, et al. “A wearable
healthcare system for cardiac signal monitoring using
conductive textile electrodes”, Engineering in Medicine
and Biology Society (EMBC), 2013 35th Annual
International Conference of the IEEE, IEEE, Japan,
2013, pp. 7500-7503.
[15] A Alzahrani, S Hu. “An effective solution to0 reduce
motion artefact in new generation reflectance pulse
oximeter”, 2013 Saudi International Electronics,
Communications and Photonics Conference (SIECPC),
IEEE, 2013, pp. 1-5.
[16] C Y Wang, Y H Yang, K T Tang, et al. “A wireless
pulse oximetry system with active noise cancellation of
motion artifacts”, Biomedical Circuits and Systems
Conference (BioCAS), IEEE, 2012, pp.148-151.