A Sleep Monitoring System Using Force Sensor and

an Accelerometer Sensor for Screening Sleep Apnea

Shongpun Lokavee Visasiri Tantrakul
Materials Science and Engineering Programme and Center of Sleep Disorder Center and Pulmonary and Critical Care
Nanoscience, Faculty of Science, Mahidol University Ramathibodi Hospital, Mahidol University
Bangkok 10400, Thailand Bangkok 10400, Thailand
[email protected] [email protected]
Janejira Pengjiam Teerakiat Kerdcharoen*
Sleep Disorder Center and Pulmonary and Critical Care Department of Physics, Faculty of Science, and Research
Ramathibodi Hospital, Mahidol University Network of NANOTEC at Mahidol University
Bangkok 10400, Thailand National Nanotechnology Center
[email protected] Bangkok 10400, Thailand
*Corresponding author: [email protected]
2021 13th International Conference on Knowledge and Smart Technology (KST) | 978-1-7281-7602-4/21/$31.00 ©2021 IEEE | DOI: 10.1109/KST51265.2021.9415835

Abstract— Obstructive sleep apnea (OSA) is the essential for the diagnosis and analysis of sleep apnea. It can also
cause of sleep-related breathing disorders in which breathing analyze other related sleep disorders, such as insomnia.
repeatedly stops and starts during sleep. Sleep-disordered Although it represents the best diagnostic tool, this method
breathing can affect our overall health, safety, and quality of has limitations including cost expensive, long time to wait,
life. This study was focused on the detection of sleep posture limited availability, and inconvenience. To overcome these
and sleep breathing disorder, by measuring the sensing problems, new techniques of unconstrained sensing of the
response of a piezoelectric sensor installed on the mattress, physical and biokinetics movement are necessary to be
pillowcase, and sleep apnea positional-therapy device. In this developed.
work, we have developed a sleep monitoring system based on
an unobtrusive method. The results of sleep efficiency, sleep The purpose of this study is to construct a sleep
pose, and sleep apnea was compared with the gold standard of monitoring system based on pillowcase and bedsheet form
sleep monitoring in-laboratory polysomnography (PSG). factors [8]. The force-sensing resistor (FSR) sensors were
Linear regression of all results reveals strong correlations embedded in the fabric developed inside the bedsheet and
between our smart sleep system and the PSG device. This novel pillowcase. This system can determine sleeping and awake
system has high efficiency as an alternative not only to monitor times based on the frequency of posture and gesture
sleep but also to identify and treat abnormal breathing movements [9]. A secondary aim was to evaluate the
patterns by training to sleep on your side positions. It can help
symptoms of SRBD from multi-biokinetic signals and to
improve the quality of sleep and reduce snoring and sleep
compare the sleep disturbance (movement arousal)
apnea.
associated with snoring caused by apnoea and hypopneas.
Keywords— sleep-related breathing disorders, sleep apnea,
polysomnography, sleep efficiency, respiratory pattern. II. SYSTEM DESIGN AND ARCHITECTURE
Our in-lab developed sleep monitoring system consists of
I. INTRODUCTION multi-biokinetic sensing interfaced with LabVIEW software
Sleep-related breathing disordered (SRBD) describes a for sleep analysis. We have designed the system as fabric-
group of disorders that involve difficulty in breathing during compatible to be embedded into the pillowcases and
sleep as classified by abnormal breathing patterns (e.g. the bedsheets. The system consists of an FSR sensor array which
presence of apneas or hypopneas). Particularly, obstructive is the important part that acquires the gesture and movement
sleep apnea (OSA) is a condition in which the obstructive data during sleep, as shown in Fig. 1(a, b). In addition to the
airway occurs during sleep, leading to hypoxemia and sleep monitoring device, we have developed a training
hypercarbia. The most common cause of OSAS is an device for OSA positional Therapy to help treat sleep
enlargement of the tonsils and adenoids. While external disorder breathing, as shown in Fig. 1(c).
factors of sleep apnea can be observed from abnormal facial The overall architecture of the sleep tracking system is
contours and obesity [1-3]. This is an important primary risk comprised of three key components: (1) an FSR sensor
factor for OSAS as the prevalence of obesity is an increase. embedded in a fabric envelope inside pillowcases and
Many researchers believe that these symptoms may occur bedsheet; (2) data acquisition hardware module integrated
hundreds of times with cessation of airflow through the into three different microcontroller boxes, namely the
respiratory tract during sleep, thus leading to sleep-related pillowcase, bedsheet, and OSA positional Therapy system;
issues and cardiovascular diseases [4]. Sleep-related (3) programs to classify the human body language and
breathing disordered (SRBD) covers a range from frequent quality of sleep. for analysis of the quality of sleep, including
loud snoring to obstructive sleep apnoea (OSA). The sleep latency, respiratory rate, and sleep efficiency. In detail,
prevalence could reach 57% of men 40% of women among subsections describe the components mentioned above. More
patients who are obese or snoring repeatedly during sleep details on the data acquisition and the measurement accuracy
[5,6]. In sleep laboratories, polysomnogram (PSG) or sleep of each sensor are presented by calibration in the next
study is the gold standard for diagnosing diseases that may section.
occur while sleeping [7]. PSG is cited as the most accurate

978-1-7281-7602-4/21/$31.00 ©2021 IEEE

208
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.
 2(b). The input devices include three types of sensors,
namely, the FSR sensor, triple-axis accelerometer sensor
(ADXL345), and microphone amplifier (MAX9814). The
OSA PT is packaged in a waist bag that has a fabric pouch to
hold the device inside. The platform can analyze body
position, respiratory rate, snoring, and sleep apnea. Besides,
it can help to alarm the user in real-time to change his/her
sleep position. When the patient is snoring and the noise is
(a) Pillowcase systems (b) Bedsheet systems
higher than the standard, which may disturb the partner.
Therefore, this device is very useful for training the user to
maintain sleep in the appropriate position.

(c) The OSA positional Therapy-prototype

III. DATA ACQUISITION AND MEASUREMENT ACCURACY
In developing our prototypes, we applied Arduino
technology for a smart sleep system. The Arduino is a ready-
made microcontroller board based on ATMEL's ARM family
of controllers. The advantage of the microcontroller board is
the open-source issue that can be developed into any device.
It is also possible to add a Boot Loader to the ARM, making
it easier to upload code to the board. In this work, the
controller’s system can support two working modes; Firstly,
it worked with stand-alone. Next, it can connect to a
computer/Raspberry Pi via USB cable. The data acquisition
(d) Pillowcase and Bedsheet and OSA PT-prototype
board can receive different types of sensor data, including
the pressure distribution, the 3-axis accelerometer data, and
Fig. 1. Schematic of hardware architecture and design for a smart sleep snoring sound.
system.

In this section, we will show the design of the systems for A. Characterization of the triple-axis accelerometer sensor
sleep monitoring and therapy prototype, as shown in Fig. 2. response
In this section, we will describe the sleep therapy device
(a) embedded into a belt around a person’s chest or abdomen.
The platform includes the body position sensor that has an
ADXL345 module for monitoring five different patient
positions, as illustrated in Fig. 3.

(a)

Supine position Left lateral Prone position Right lateral Supine position
(b) position position

Zˊ
(b) (c)
Z Zˊ Z
Y Y
θ
Xˊ
θρ
X X
ρθ
Xˊ

Three-axis accelerometer
Three-axis accelerometer

Fig. 2. The architecture of the sleep monitoring and therapy prototypes; (a) (d) (e)
Fritzing's breadboard view for the pillowcase and bedsheet monitoring
systems, (b) Fritzing's breadboard view for the OSA Positional Therapy
system.
ϕ
The pillowcase and bedsheets sensor systems can be used
for monitoring the gesture movements during sleep to
300-450
diagnose symptoms or disorders of breathing, but it cannot
do real-time therapy of the disorder. Accordingly, we have
Fig. 3 Example responses of the body position accelerometer (a) four
developed an OSA positional Therapy device for the latter different types of lying positions, (b) acceleration reading that represents
purpose. The architecture of the prototype is shown in Fig. supine to left lateral recumbent; (c) acceleration reading that represents

209
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.
supine to right lateral recumbent, (d) four common sleep positions, and (e) B. Characterization of the FSR sensor response
Fowler’s position
This section will describe a method to calibrate force
The triple-axis accelerometer (ADXL345) is a complete sensors that are used for measuring biokinetic signals of
3-axis acceleration measurement system. The dynamic body movement. The FSR sensors can measure forces
acceleration data can be measured through body motion and between 0-100 N with a resolution of 1 N. Analog voltage
positional change during sleep, which allows the system to divider is used to obtain signal output, with a fixed reference
be used as a tilt sensor. In this case, the ADXL345 sensor resistor (Rf) of 10 KΩ. The voltage output (Vout) of the FSR
module was used to measure the person's position in four circuit is shown in Fig. 5(a)., it is used to convert a resistive
different positions. The sensor module embedded in the change into a voltage that can be measured using an ADC.
controller box based on the acceleration along three The graph depicts the sensor values and the voltage outputs
dimensions; including pitch (ρ) and roll (ϕ) is defined as the that are increasing with increasing forces.
angle of the x-axis and y-axis relative to the ground, theta (θ)
is the angle of the z-axis relative to gravity as shown in Fig. The force-sensitive resistor is a sensor that lowers its
3(b), (c). In Fig. 3(b), we can also characterize and monitor resistance when there is a force upon it. If there is no force
the accuracy of the body position in the horizontal movement acting on it, the resistance (RFSR) is ~10MOhms. The force
of the user and forward/backward movement. The horizontal range can be extended by reducing the voltage input, Vin, or
x-axis can rotate completely at 360 degrees (the angle (ρ) is the fixed referent resistor, Rf. Conversely, the sensitivity can
00 to 3600 ), including different the following four common be increased for measurement of lower forces by increasing
sleep positional: lying on the supine (back sleepers), lying on Vin or Rf. In Arduino microcontrollers, the analog reading
the left lateral decubitus, lying on the right lateral decubitus, goes from an integer value (Sensor value) between 0 and
and lying on the prone (stomach sleepers). The vertical (y- 1023 (increments 1024) which converts to a voltage value
axis) is rotated maximum at 135 degrees, capturing Fowler’s between 0 and 5V. An equation of the sensor value for the
position (the lying positions) as the angle (ϕ) is 00 to 450 as voltage output is shown in (1). Then the voltage output
shown in Fig. 3(e). After the angle (ϕ) start from 460 to 1350, between the fixed reference resistor (Rf) and the variable
thus capturing Fowler’s position was to represent the sitting FSR (RFSR) is used as the output of the FSR as shown in
pose or standing (2,3).
15
(a)
……….….. (1)
10 X-Axis
5
0
-5 ………………………… (2)
-10
-15
12 Y-Axis (b)
9 ……………………….... (3)
6
3
0
-3
15 …………….……………………….... (4)
Z-Axis (c)
10
5
0
-5
Where VOUT is the FSR output voltage; Vin is the input
-10 voltage to the FSR; Rf is a fixed referent resistor, and RFSR is
Stand/sit
(d) the variable resistance from the FSR.
Prone
Right To read the signal of the FSR from the voltage divider
Left
Supine
circuit, the sensor output signal is mentioned above. The
15:14:20 15:15:30
Time
15:16:40 15:17:50 results of the sensor value and the voltage output were
Fig. 4. Comparison of raw acceleration signals together with the obtained for the correlation of the electrical conversion with
corresponding decoded body position. the increased pressure. From Figure 5, the slope of the
trendline is determined by the response of the FSR sensor.
The method of sleep position analysis in the horizontal The trendline curve in Fig. 5(a)., shows the sensor value of
movement describes the degrees of rotations of the x-axis the FSR sensor and the voltage output obtained by pressure
change: when the patient lies in supine positions, the angle of changes from 1 N to 100 N.
the pitch (ρ) can rotate forward and backward at a maximum
value around 45 degrees relative to the ground (the z-axis is a
positive value). Then the patient changed positions from
supine to right lateral; the box changed toward their right
side, thus resulting in the angle of the pitch that can rotate up
to 135 degrees relative to the ground. In contrast, the patient
changed positions from supine to left lateral; the box
changed toward their left side, thus resulting in the angle of
the pitch that can rotate up to 135 degrees (450 ≤ (ρ) ≥ 1350)
relative to the ground. Finally, the prone position is opposite
in the supine direction which the z-axis is a negative value

210
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.
 (a)

(b)

Fig. 5. Experimentally measured; (a) the relationship between the curve of

force-sensor value and force-voltage outputs, (b) the relation between force-
resistance and force-conductance curve of the FSR406.

Figure 5(b) shows the linearity in the initial range to 5.6

Kg, after which the response of the conductance (C) with the
pressure from the digital force gauge was less responsive.
Fig. 6. Respiration sound monitoring during the night (a), respiration sound
Therefore, the sensitivity of the pressure response after 5.6 in a minute (b).
kg is not suitable for use due to the slight change observed in
Fig. 5(a). It has been demonstrated that the force (F) applied This microphone module system work with The OSA
to the sensor is linearly related to the conductivity (C) signal. PT-prototype, snoring sound can help to change the sleeping
Moreover, the pressure response to the conductivity signal pose when the person sleeps in the supine position and
(C) can be analyzed by a separate plot as shown in Fig. 5(b). snores loudly than the threshold level. However, the fixed
The conductivity output has a linear approximation up to threshold is not good because the level of snoring sound in
0.94 μS. However, the increased conductivity after 0.94 μS each patient is different that it has a strong influence on the
onwards was not used in the sleep testing system due to the accuracy of the result.
low-pressure response.
The Figure clearly shows that the sensor still produced a IV. DATA VISUALIZATION AND INTERFACE
highly linear response between 0.1 and 5.6 kg of force but The smart sleep platforms are consisting of a force sensor
after exceeding 5.6 kg to 10 kg of force, linearity decreases and controller module that the platform can read and upload
remarkably. Therefore, the resulting sensitivity and the non- the signal to LABVIEW programming. This system can
linear range might be the limitation of the FSR sensor that measure head and body movement. The system extracts the
has a slightly internal resistance. From the curves in Fig. user action and biokinetics signal from the acquired data in
5(b), it is shown that the linearity within the 0.1–5.6 kg real-time monitoring.
range. The R2 value for the logarithm fitting-line height is
0.9985, which demonstrates a strong linear correlation. It
was determined that the FSR sensor can be used to apply in
the smart sleep system. However, the reference resistor under
the voltage divider circuit can be adjusted to achieve a wide
range of pressure applications. For example, when we
increase the reference resistance in the voltage divider, the
sensor is more responsive to low pressure. Which is suitable
for use in systems with a low acting force range. Although
the FSR sensor has a little response after the force rang 5.6
kg, the measurement is a summation of the sensor array that
it pushes down on the pillow and mattress. In this case, the
force can affect sensors in their surroundings which makes
the weight spread in all directions.

C. Characterization of the MAX9814 sensor response

The high-quality microphone amplifier is present, the
output features of the MAX9814, including variable gain
amplifier (VGA), low-noise amplifier, automatic gain control
(AGC), supply voltage range at 2.7 V to 5.5 V. A snoring
signals is used as the voltage output is shown in Eq. (1).
Where sensor value is the variable resistance from the
MAX9814 sensor. Snoring results are shown in Fig. 6., Fig. 7. Screenshot of the smart sleep application for monitoring and
treatment sleep relate-breathing disorder.
usually occurs during inspiration, but can also occur
during expiration. The purpose of the smart sleep application is to let users
find out about their sleep and lifestyle. The smart sleep
monitoring system was used to monitor many of our sleep-

211
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.
related features in one night. This screening software can The Figure shows that the signals obtained from the PSG
provide information that requires analysis and interpretation were compared to a smart sleep system in sleep-laboratory,
of results from a specialist in sleep disorders. It can record Fig. 8(a) including pulse oximetry was measures oxygen
indicators for monitoring sleep, including sleep monitoring saturation (SpO2) level in the blood and nasal airflow
start time until waking up, the amount of wakefulness during respiratory signal was measured by a pressure transducer as
sleep, sleep efficiency, all night respiration rate. In addition, shown in Fig. 8(b). The RIP belts measure respiratory effort
an asleep monitor is an efficient and convenient way to signals, the placement of the RIP belts over the abdomen.
detect sleep disorders while sleeping. The software can The experiment results of the respiratory signal (Fig. 8(c))
automatically detect sleep apnea events that are equipped compared with the raw respiratory signal of the OSA Pt-
with an alarm when the person loudly snoring in a supine prototype (Fig. 8(d)) and the raw respiratory signal from the
position. bed-sheet device (Fig. 8(e)). The signal of all devices
represents a temporary breathing stop. From the graph, it can
V. RESULTS AND DISCUSSION be seen that the sleep breathing disorder has the crescendo-
decrescendo pattern in the absence of breathing effort. This
The selection of volunteers to participate in a clinical pattern is typically associated with an imbalance of the
sleep screening to observe respiratory disorders during sleep. oxygen and carbon dioxide levels in the blood. This
The sleep test covers patients at risk of sleep apnea to assess experiment proves that a pressure sensor can measure and
and diagnose the severity of the disease. The beginning and reference a breathing pattern signal with a non-contact
end of the in-bed period in the electronic record were marked method. The amplitude of the respiration signal close to the
as “Lights Out” and “Lights On,” respectively. PSG signal from the PSG device using the Airflow and the RIP-
recordings were performed overnight. The sleep test by Abdo sensors.
smart sleep system was done in parallel with
polysomnography at the sleep lab. The experiment was In the case of sleep quality, we have demonstrated the
certified by documentary proof of ethical clearance system with evaluating sleep efficiency based on the
committee on human rights related to research involving detection of the frequency of head movement during sleep.
human subjects’ faculty of medicine Ramathibodi hospital, To compare the sleep efficiency of standard commercial
Mahidol University. The protocol number ID 02-56-59 (No. measuring devices between polysomnography with the smart
MURA2013/140). Respiratory scoring rules are according to sleep system. The calculation of sleep efficiency can be
the American Academy of Sleep Medicine adult rules 2012, referenced by the equation below:
Version 2.0. The results of all PSG tests were analyzed by a %SE = [(TST)/(TRT)] x 100 ….………………… (5)
technician while the results of our smart sleep tests were
automatically analyzed by software. Sleep efficiency (%SE) was defined as the percentage of
total sleep time divided by total recording time. Total
In a preliminary case-control study, we performed the recording time (TRT) was defined as the total time of study
test on 50 subjects that had already been diagnosed with from lights off to lights on. Total sleep time (TST) was
sleep apnea. In this case, we separated to test on two defined as the total time spent sleeping. Sleep latency (SL)
categories; firstly, we evaluated the sleep efficiency of all 27 was defined as the recording time in bed to time of sleep
patients by comparing their accuracy between the smart onset. Using a sleep sensor to calculate waking events based
pillowcase and the PSG device. After that, we used the OSA on the frequency of motion and body movement. When the
PT-device monitoring of the 24 participants to record the sleep position changes, the sum of the sensors during that
sleep position, respiratory signal, and microphone amplifier time exceeds the threshold level. Usually, we define the
to record the snoring signal. Only one volunteer used the duration of the large gesture movement as a reference to the
pillowcase and the OSA PT-device at the same time. awakening. Small posture movements below the threshold
level are referred to as the sleeping states.

TABLE I. SLEEP CHARACTERISTICS BASED ON POLYSOMNOGRAPHY.

Sleep Polysomnography (PSG) Pillowcase-device
Parameter
(n = 27) %SE SL %SE SL
Mean ±SD 85.83±9.20 8.73±8.93 85.41±8.26 7.54±8.26

In statistics, reporting mean and SD of %SE and SL from

patients measured by PSG was 85.83% ± 9.20 and
8.73±8.93, respectively. Pillowcase devices showed a high
correlation of Mean ±SD for %SE (85.41±8.26) and SL
(7.54±8.26), respectively.

Fig. 8. Example of apnea signal from the experimental records: (a)-(c)

output of spO2(%), airflow (thermistor), and Abdo-movement (the rip
belts) monitor from PSG, versus (d)-(e) output from OSA PT-prototype and
bedsheet systems based time-domain response.

212
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.
Fig. 9. Linear regression for (a) sleep efficiency and (b) sleep latency from classify sleep poses for constructing sleep positional therapy
polysomnography (PSG) and Pillowcase device. devices to treat patients with positional sleep apnea. The
From Fig. 9, A scatterplot is used to graphically represent advantages of this methodology, we found that our sleep
the relationship between two variables, showing a scatter system can take benefit from the biokinetics movement to
diagram of the sleep efficiency (Fig.9(a)) and sleep latency estimate sleep performance. The validation accuracy of sleep
(Fig.9(b)) between the pillowcase device and the PSG pose is greater than 90 percent. In addition, the smart sleep
device. Linear regression of SE and SL reveals strong system can be analyzed in real-time that easy to use, and
correlations between the Pillowcase and the PSG device. The faster than a polysomnography device. Furthermore, the
shape of the scatterplot from Fig. 9(a, b) indicates that there device can provide sleep duration estimates close to
is a positive linear correlation. commercial devices that alternative requiring unobtrusive
sensors. Moreover, we embedded a microphone sensor
TABLE II. SLEEP CHARACTERISTICS BASED ON POLYSOMNOGRAPHY. module on the circuit for detected snoring sound which is
Supine Left Right Prone equipped with an alarm when the person loudly snoring. The
Sleep vibrating alarm can help to change the sleep position when
Parameter Mean ±SD Mean ±SD Mean ±SD Mean ±SD
(n = 24) the patient sleeps in the supine position and snores loudly
than the threshold level.
The RIP Belts 59.23±31.64 15.27±21.17 24.67±29.61 0.29±1.43
a ACKNOWLEDGMENT
OSA PT-device 62.05±31.85 13.82±18.39 23.77±34.56 0.0±00.0
a
5.83x10-4 ± 2.08x10-3
This research was supported by the Faculty of Science,
The purpose of this study is to compare sleep positional Mahidol University, and National Nanotechnology Center
change recorded by the commercial RIP-belts with our OSA (NANOTEC) through the Research Network of NANOTEC
positional Therapy in adults. Overall, we found specific (RNN). The Commission on Higher Education (CHE) has
categories (the MEAN(SD) supine and lift/right lateral) of partial supports through its National Defense Research
device data that did not differ from PSG measures as shown Program.
in Table II.
REFERENCES

[1] C. Mermigkis, J. Chapman, J. Golish, D. Mermigkis, K. Budur, A.

Kopanakis, V. Polychronopoulos, R. Burgess, and N. Foldvary-
Schaefer, “Sleep-related breathing disorders in patients with
idiopathic pulmonary fibrosis,” Lung, Vol. 185, pp. 173–178. April
2007.
[2] P.E. Peppard, T. Young, M. Palta, J. Dempsey, and J. Skatrud
“Longitudinal Study of Moderate Weight Change and Sleep-
Disordered Breathing,” JAMA, Vol. 284, pp.3015-3021, Dec 2000.
[3] D. Burman, “Sleep Disorders: Sleep-Related Breathing
Disorders,” FP essentials, Vol. 460, pp. 11–21, Sep 2017.
[4] E. Shahar, C. W. Whitney, S. Redline, E.T. Lee, A.B. Newman, F.J.
Nieto, G.T. O'Connor, L.L. Boland, J.E. Schwartz, and J.M. Samet,
“Sleep-disordered breathing and cardiovascular disease: cross-
sectional results of the Sleep Heart Health Study,” Am J Respir Crit
Care Med, Vol. 163, pp. 19-25, Jan 2001.
[5] R.J. Schwab, “Merck Manual Professional Version: Snoring,”
Fig. 10. Linear regression plots the relationship between the OSA Retrieved, July 24, 2020.
positional therapy and the commercial RIP-belt evaluations (a) supine, (b) [6] G. Marchesini, A. Pontiroli, G. Salvioli, R.F Novi, R.E. Vitacolonna,
left position, and (c) right position. C. Taboga, A.M. Ciccarone, E. Grossi, and QUOVADIS Study
Group, “Snoring, hypertension and Type 2 diabetes in obesity.
Protection by physical activity,” Journal of endocrinological
The OSA positional therapy aims to prevent patients from investigation, Vol. 27, pp. 150–157. Feb 2004
the effect of loud snoring on supine sleep position. The [7] R.B. Berry, R. Budhiraja, D.J Gottlieb, et al., “Rules for scoring
device records the behavioral signals of body position and respiratory events in sleep: update of the 2007 AASM Manual for the
vibrated every ten seconds when the patient snored in the Scoring of Sleep and Associated Events. Deliberations of the Sleep
supine position. However, if the patient snoring on the side Apnea Definitions Task Force of the American Academy of Sleep
Medicine,” J Clin Sleep Med, Vol. 8, pp. 597-619, Oct 2012.
position, the device did not vibrate because not all snored
[8] S. Lokavee, V. Tantrakul, W. Suwansathit, J. Pengjiam, and T.
pose has OSA. Kerdcharoen, “Detection of Sleep-Disordered Breathing and Sleep
Efficiency with Wireless Pillow-Sheet Sensor Network,” IEEE
VI. CONCLUSION AND FUTURE WORK International Conference on Electrical Engineering/Electronics,
Computer, Telecommunications and Information Technology
In summary, we have developed a sleep monitoring (ECTICON), pp. 838-841, July 2019.
technology for patients as well as everyday people based on [9] S. Lokavee, V. Tantrakul, W. Suwansathit, J. Pengjiam, and T.
multi-sensor devices. All the device records the behavioral Kerdcharoen, “Efficacy of Positional Therapy Prototype for Patients
signals from biokinetics movement. The asleep monitoring with Positional Obstructive Sleep Apnea,” 2020 12th International
system was used to monitor many of our sleep-related Conference on Knowledge and Smart Technology (KST), pp. 159-
features in one night. The pillowcase and Bedsheet system 163, Jan-Feb 2020.
can detect sleep efficiency value with effective autonomous
scoring capability. In addition to measuring the quality of
sleep, our prototype can work as a sleep apnea screening
device. The goal of this study is to identify sleep apnea and

213
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on November 23,2023 at 04:15:41 UTC from IEEE Xplore. Restrictions apply.