Content uploaded by Hoe Tung Yew
Author content
All content in this area was uploaded by Hoe Tung Yew on Mar 31, 2021
Content may be subject to copyright.
IoT Based Real-Time Remote Patient Monitoring
System
Hoe Tung Yew
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
htyew@ums.edu.my
Seng Kheau Chung
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
kheau@ums.edu.my
Ming Fung Ng
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
mingfung93@gmail.com
Ali Chekima
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
chekima@ums.edu.my
Soh Zhi Ping
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
sohzhiping95@hotmail.com
Jamal A. Dargham
Universiti Malaysia Sabah
Kota Kinabalu, Sabah
jamalad@ums.edu.my
Abstract—Healthcare technology is one of the most popular
studies nowadays. With the development of healthcare
technology, the lifespan of people has successfully extended.
However, people in the rural area are still having a hard time to
obtain professional healthcare services due to the barrier of
distance and lack of doctors. A remote patient monitoring
system is one of the best solutions to overcome this issue. This
paper proposes an Internet of Things (IoT) based real-time
remote patient monitoring system that is able to guarantee the
integrity of the real-time electrocardiogram (ECG). Message
Queuing Telemetry Transport (MQTT) protocol is used for
transmitting the real-time ECG from the proposed system to the
webserver. The doctor can access the webserver via smartphone
or computer to monitor the real-time or previously recorded
ECG data. The proposed system has been tested in both Local
Area Network and Wide Area Network environments. The
results show that the proposed system has no package loss and
packet error in both Local Area Network and Wide Area
Network.
Keywords—: electrocardiogram, real-time monitoring,
MQTT protocol, remote patient monitoring system.
I. I
NTRODUCTION
The ratio of doctors to the population in Malaysia is
improving from year to year. However, it is still lacking doctor
in rural areas. The report from the Ministry of Rural
Development Malaysia showed that 24.5% of the population
are staying in rural areas. The majority of them are from East
Malaysia and the East Coast of Peninsular Malaysia [1].
According to the statistic from the Ministry of Health
Malaysia, the ratio of doctor to the population in west coast of
Peninsular Malaysia, east coast of Peninsular Malaysia and
East Malaysia is 1:482, 1:730 and 1:786, respectively [1].
Obviously, the number of doctors in East Malaysia and the
east coast of Peninsular is much lower compared to the west
coast of Peninsular Malaysia.
A remote patient monitoring system is one of the solutions
to overcome the barriers of health service. It provides
professional health care services to the rural population
through information and communication technologies (ICT).
Patel et al. stated that the continuation of health and wellness
monitoring can help in the treatment of the patient [2]. Remote
patient monitoring system allows the doctor to observe the
patient’s health condition remotely at any time and any place
[3]. The prevention can be done during the early detection of
the disorder.
This paper proposes an IoT based real-time
electrocardiogram (ECG) monitoring system that is able to
support both real-time and store-and-forward modes. The
store-and-forward mode has no data integrity issue and it also
does not require high network quality. However, the real-time
mode is challenging because transmission delay and packet
loss during real-time transmission will affect the data
integrity, especially for real-time ECG signal transmission.
Inaccurate health data might lead to misdiagnosis.
II. R
ELATED
W
ORKS
A WLAN based remote health monitoring system with
end to end flow had been presented in [4]. The system collects
the patient vital sign such as heart rate, blood pressure, pulse
oximetry and body temperature by using medical sensors.
Real-time ECG signal was not included in this work. Arduino
Yun is used as the main controller to process and analyze the
data. The data is then uploaded to the cloud using the
Application Programming Interface (API) through WLAN.
Doctors can log in to the online server for checking the patient
medical data and send comments to the patient through the
server. If the device lost the WLAN connection, the data will
be temporarily stored in the memory card. The system will
synchronize the data with the server once the connection with
the server is recovered.
Lee et al. proposed a smartphone based remote patient
monitoring system [5]. In this system, a smartphone
application (app) is implemented. Patient can have a video
conference with the doctor by using a smartphone to discuss
the treatment process. This system is collaborating with
Skype’s AES-style encryption to secure the patient’s video
data. This system does not support real-time mode. Patient is
required to upload the collected vital sign to the server
manually.
Roy et al. presented an RF based remote patient
monitoring system which enables the user to use medical care
information through web and mobile application platforms
[6]. The system is built up with sensor nodes, coordinator
node, web and database server and graphical user interface
(GUI). The sensor node is used for collecting the data. The
acquired data is uploaded to the central server. Users can
observe the data and analyse result through the GUI. Due to
the delay in transmission and slow server response, this
system is unable to support real-time mode.
An IoT based health care system for doctors remotely
monitoring patient’s health conditions via the internet had
been presented by Ghosh et al. [7]. E-Health Sensor Shield Kit
This research is supported by Ministry of Education Malaysia and
Universiti Malaysia Sabah, Geran Penyelidikan UMS (SBK0364-2017).
2020 16th IEEE International Colloquium on Signal Processing & its Applications (CSPA 2020), 28-29 Feb. 2020, Langkawi, Malaysia
176
was used in this module for collecting the biometric
information and then transmitting to the online server through
wireless technology. The data collected by the sensor was
temporarily stored in the microcontroller and uploaded to the
online database after some time. The user is required to log in
with unique ID for accessing the data. The total time required
by the system from collecting health data, transmitting data to
the webserver and visualizing the data on the webpage is
around 15 seconds.
Mustapha et al. proposed a patient remote monitoring
system using Web Real-Time Communication (WebRTC)
and Edge Cloud [8]. Edge Cloud is not only the storage but it
is also used as a communication platform between the patients
and remote patient monitoring system. The system has two
modes: push–mode and pull-mode. In push-mode remote
monitoring, the system sends alert to the user if abnormal
health data is detected. While pull-mode remote monitoring is
for the user to review back the previously collected data.
WebRTC supports multiple ways of communication between
the cloud server and the main system. The collected data and
video will be analyzed by the analytic engine of the system.
This system is mainly focusing on optimizing and transmitting
video data instead of vital sign data.
Pap et al. implemented an IoT based e-health system by
using a Raspberry Pi [9]. Node.js server-side application was
used for recording the data in a specific chart and providing a
web interface for the user to study the chart. The e-health
system is equipped with blood pressure, pulse oximeter,
airflow, galvanic skin and temperature sensors. This system
allows the user to select the data recording mode either a live
session or a recording session. In the live session mode , the
data will only be temporarily stored. For the recording session,
the data will be permanently stored.
III.
METHODOLOGY
The proposed patient monitoring system consists of four
modules: a sensor, a main controller, a communication
module and a webserver as shown in Figure 1. An ECG sensor
AD8232 is used to collect the ECG data from the cardiac
patient. It measures and amplifies the ECG signal based on the
electrical activity of the heart muscle. In the proposed system,
Arduino ESP 32 which equipped with WiFi is used as the
main controller. The collected real-time ECG data will be
processed by the main controller and then transmitted to the
MQTT broker via WiFi module. The ECG data is published
to the webserver through Mosquito Message Queuing
Telemetry Transport (MQTT) broker which is hosted by the
Raspberry Pi 3.
An MQTT-DB relay is applied in this work. It subscribes
the same Topic with the webserver in the MQTT broker,
buffering the ECG JSON data and then storing the ECG data
in the database. This allows the proposed system to display the
real-time ECG signal on the webserver and stores it in a
database simultaneously. Besides webserver, the doctor also
can view the real-time ECG signal via Android mobile
application (App). The App can display both heartbeat rate
(BPM) and a real-time ECG chart. The BPM is extracted from
the received real-time ECG data, it updates every 3 seconds.
Doctors or nurses can access the database server to view the
recorded history ECG data.
This work uses MQTT network protocol instead of
Hypertext Transfer Protocol (HTTP) because MQTT requires
low bandwidth, low power consumption and suitable for IoT
development [10]. The ECG signals data are packed into a
JSON Format string before sending it to the webserver and
database. This greatly improves the efficiency of data
converting at the client side because both webserver and
database have a built-in function that can process the JSON
format data.
The webserver was implemented by using JavaScript
programming language. The web sockets are used to display
a real-time ECG signal based on the ECG JSON data that is
transmitted from the proposed patient monitoring system. The
webserver also performs package loss and jitter delay
calculation. The equation for calculating the packet loss is
given as follow:
=
−
R
(1)
where
is package loss,
is the number of packages sent
by the main controller and
R
is the number of packages
received by the webserver. Equation (2) is used for calculating
the jitter delay.
=
T
R
–
T
S
(2)
where
is jitter delay, T
S
is the time interval between the
packages sent by the main controller and T
R
is the time interval
between the package received by the webserver. The average
jitter delay
can be computed by using equation (3).
∑
(3)
where S
is the total jitter delay and
is the total number of
packages.
Fig. 1. Architecture of the proposed real-time patient monitoring system
Mobile Apps
Webpage
Main Controller
ECG Signal
Cardiac
Patient
WiFi Database
Publish
MQTT
Broker
Proposed System
2020 16th IEEE International Colloquium on Signal Processing & its Applications (CSPA 2020), 28-29 Feb. 2020, Langkawi, Malaysia
177
IV. E
XPERIMENT
S
ETUP
In order to test the performance of the proposed system in
different environments, two experiments had been set up in
real networking environments. The first experiment is set up
in a private network or local area network (LAN). The
proposed system, database system and computers as shown
in Figure 1, are connected to the same access point. Assuming
the patients in the hospital are equipped with the proposed
system. The collected ECG data are stored in the in-house
server. Doctors monitor their patients’ real-time or previously
recorded ECG data from their office. The second experiment
is set up in the public network or Wide Area Network
(WAN), as shown in Figure 2. Assuming the proposed system
is connected to the home WiFi. The collected real-time ECG
data is transmitted to the cloud server via internet. Doctors
access the cloud server for observing the patient real-time
ECG through their own mobile cellular network.
The performance of the proposed system is measured
based on the number of packet loss and average jitter delay.
The real-time ECG signal was observed at Arduino Serial
Monitor and webserver. The ECG signal displayed on the
Arduino Serial Monitor is the real-time ECG signal collected
by the AD8232 sensor. This ECG signal will be published by
the MQTT broker to store in the database system and also
display on the webserver.
Fig. 2. Experiment setup in the public network
V. R
ESULTS AND
D
ISCUSSION
The experiment results show that there are no packet loss
and packet errors in both private and public networks as
shown in Table I and II, respectively. This is because MQTT
protocol has a checksum function. It will resend the packet if
subscribers failed to receive the packet. However, this leads
to delay in sending the next packet.
To further verify there is no packet error, we compared
the total value of ECG data of the selected range of 5000
packets between sender (main controller) and subscriber
(database system or webserver). The results show that both
the sender and subscriber have an identical sum of ECG
values, as shown in Table I and II. In term of jilter delay, the
performance in public network is about ten times greater than
private network. This is because the public network has
greater transmission distance compared to the private
network. The delay of 50.08 ms in public network is
acceptable because the maximum tolerance can be up to 1
second [11]. Figure 3 and Figure 4 show the ECG signal that
displayed at Arduino Serial Monitor and webserver in private
network and public network, respectively.
TABLE I. P
ERFORMANCE OF THE
P
ROPOSED
S
YSTEM
I
N
P
RIVATE
N
ETWORK
TABLE II. P
ERFORMANCE OF THE
P
ROPOSED
S
YSTEM
I
N
P
UBLIC
N
ETWORK
Module
Number
of ECG
packages
Sum of
values
Average
Jitter Delay,
A
JD
(ms)
ECG packets sent by main
controller 5000 9477828
50.0832
ECG packets received by
Webserver/database 5000 9477828
(a)
(b)
Fig. 3. ECG signal displayed on the (a) Arduino serial monitor and (b)
webserver in private network
Module
Number of
ECG
packages
Sum of
ECG
values
Average Jitter
Delay, A
JD
(ms)
ECG packets sent by
main controller 5000 9511560
5.714
ECG packets received
by Webserver 5000 9511560
Smartphone
Main Controller
ECG Signal
Cardiac Patient
WIFI
Public
Network
Cloud Server + MQTT Broker
Internet
2020 16th IEEE International Colloquium on Signal Processing & its Applications (CSPA 2020), 28-29 Feb. 2020, Langkawi, Malaysia
178
(a)
Fig. 4. Section of ECG signal displayed on the (a) Arduino serial monitor
and (b) webserver in public network
Apart from the webserver, Android mobile application
(App) is developed for the doctor to view the real-time
heartbeat rate (RPM) and ECG signal through the
smartphone. Figure 5 and Figure 6 show the interface of real-
time BPM and ECG signal on the Android mobile App.
Fig. 5. Real Time heart rate meter in Android App
Fig. 6. Section of Real-Time ECG signal displayed in Android App
VI. C
ONCLUSION
The developed real-time patient monitoring system has
been tested in both private and public networks. The real-time
ECG signal was displayed on the web and storing in the cloud
server successfully. The experiment results show that there
are no package loss and package errors during transmission
of ECG packages in both private and public networks. The
proposed system can be applied in the hospital. The doctors
or nurses can observe the ECG signal of any patients at any
time any place through the computer or smartphone, without
going to the wards. This system is able to reduce the patients'
traveling time and cost especially for those patients who are
from suburban or rural areas. Doctors can monitor the ECG
of the chronic cardiac patient in distance via the internet. The
proposed system has the potential of improving the quality of
healthcare services all over the nation.
As future work, ECG self-interpretation algorithm
can be implemented into the system so that the system can
detect the abnormal ECG signal and generate an alert. This
system also can be expanded by adding more e-health sensors
to collect various health parameters. Research on reducing
the jitter delay and eliminating the noise signal are also
required to improve the performance of the proposed system.
R
EFERENCES
[1] Ministry of Health Malaysia, “Health indicators,” 2018.
[2] S. Patel, H. Park, P. Bonato, L. Chan, and M. Rodgers, “A review
of wearable sensors and systems with application in
rehabilitation,” J. Neuroeng. Rehabil., vol. 9, no. 1, p. 21, 2012.
[3] H. T. Yew, Y. Aditya, H. Satrial, E. Supriyanto, and Y. W. Hau,
“Telecardiology system for fourth generation heterogeneous
wireless networks,” ARPN J. Eng. Appl. Sci., vol. 10, no. 2, 2015.
[4] N. B. Krishnan, S. S. S. Sai, and S. B. Mohanthy, “Real Time
Internet Application with distributed flow environment for
medical IoT,” Proc. 2015 Int. Conf. Green Comput. Internet
Things, ICGCIoT 2015, pp. 832–837, 2016.
[5] E. Lee, Y. Wang, R. Davis, and B. Egan, Designing a low-cost
adaptable and personalized remote patient monitoring system.
2017.
[6] S. Roy, A. Rahman, M. Helal, M. S. Kaiser, and Z. I. Chowdhury,
“Low cost RF based online patient monitoring using web and
mobile applications,” 2016 5th Int. Conf. Informatics, Electron.
Vision, ICIEV 2016, pp. 869–874, 2016.
[7] A. M. Ghosh, D. Halder, and S. K. A. Hossain, “Remote health
monitoring system through IoT,” 2016 5th Int. Conf. Informatics,
Electron. Vision, ICIEV 2016, pp. 921–926, 2016.
[8] H. Moustafa, E. M. Schooler, G. Shen, and S. Kamath, “Remote
monitoring and medical devices control in eHealth,” in 2016 IEEE
12th International Conference on Wireless and Mobile
Computing, Networking and Communications (WiMob), 2016, pp.
1–8.
[9] I. A. Pap, S. Oniga, I. Orha, and A. Alexan, “IoT-based eHealth
data acquisition system,” 2018 IEEE Int. Conf. Autom. Qual.
Testing, Robot. AQTR 2018 - THETA 21st Ed. Proc., pp. 1–5,
2018.
[10] T. Yokotani and Y. Sasaki, “Transfer protocols of tiny data blocks
in IoT and their performance evaluation,” 2016 IEEE 3rd World
Forum Internet Things, WF-IoT 2016, pp. 54–57, 2017.
[11] H. T. Yew, E. Supriyanto, M. H. Satria, and Y. W. Hau, “Adaptive
network selection mechanism for telecardiology system in
developing countries,” in 3rd IEEE EMBS International
Conference on Biomedical and Health Informatics, BHI 2016,
2016.
2020 16th IEEE International Colloquium on Signal Processing & its Applications (CSPA 2020), 28-29 Feb. 2020, Langkawi, Malaysia
179
