Wireless Personal Communications (2022) 124:189–205

https://doi.org/10.1007/s11277-021-09335-w

IoT for Health Monitoring System Based on Machine Learning

Algorithm

S. Balakrishnan1 · K. Suresh Kumar2 · L. Ramanathan3 · S. K. Muthusundar4

Accepted: 4 November 2021 / Published online: 17 November 2021

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2021

Abstract
Internet of things (IoT) is one of the futuristic and upcoming technologies which have led
to the concept of automation in many daily chores. IoT has placed its fame nearly in all
possible platforms and the main idea is to enhance this technology in the current health-
care system, for the purpose of automatic monitoring of hospitals and patient’s health by
spreading positive vibes to the vision of IoT. With the support of certain technologies such
as Radio Frequency Identification (RFID), Brainsense headband, Wireless Sensor Network
and smart mobile, by keeping the IoT as its connecting platform, a Smart E-Healthcare
System is framed. A low power wireless personal area connection helps in incorporating
these technologies together through a Constrained Application Protocol/IPv6. To track the
state of the patient, Smart Healthcare Sensor (SHS) and RFID are used that are attached
to the person’s wrist bands automatically by using electromagnetic fields. Each tag held
by the person has a unique ID and it contains information that is electronically stored in
a separate cloud environment. Based on the various information sensed by the SHS sen-
sors, required prescription is given to the victim in the absence of doctors. This technology
represents a significant step in the development of health care sectors. In this paper, we are
proposing pro prediction algorithm for detecting the prescription for the patient.

Keywords E-Health · Internet of things · Pro prediction algorithm · Radio frequency

identification · Brainsense headband

* S. Balakrishnan
[email protected]
1
Department of Computer Science and Business Systems, Sri Krishna College of Engineering
and Technology, Coimbatore, Tamil Nadu, India
2
Department of Information Technology, Saveetha Engineering College, Thandalam, Chennai, India
3
School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, India
4
Sharad Institute of Technology College of Engineering, Yadrav, Ichalkaranji, India

13
Vol.:(0123456789)
190 S. Balakrishnan et al.

1 Introduction

Improving the efficiency of healthcare infrastructure and the bio medical system is one
of the major challenging goals of many people in the current scenario. Normally people
expect quality healthcare system from the healthcare units. But providing quality health-
care to the patients by reducing the costs and tackling the nursing staff shortage problem is
hard. It is highly expected that this type of smart health care units helps both the villagers
and the sick people. Designing a health care system which would be beneficial to all peo-
ple is highly mandatory. After the arrival of numerous technologies in the field of medi-
cine and healthcare, many changes were brought in improving the standards of healthcare
systems.
Currently the endorsement of IOT is playing a vital role in framing and designing the
smart healthcare units. These advanced technologies have to be introduced in the field
of medicine in order to “support and improve the standard of healthcare and biomedical
related processes”. The Smart Healthcare Sensor (SHS) is used in tracking the body state
of the patient. The SHS is placed in the wrist bands of each patient. This SHS designed
efficiently and it is capable of identifying some of the person’s health factor such as pres-
sure, temperature, ECG, motion, heartbeat etc. automatically. After detecting the body
state of the victim, the collected details will be sent to a health monitoring server. The
RFID bands that are placed in the patient’s wrist holds a unique ID which is similar to the
patient ID used in the hospitals. Based on the details sensed, exact prescription is given to
the patients by the proposed Pro Prediction algorithm, which uses deep learning as its plat-
form. A huge cloud database acts as a repository for all the solution to various symptoms.
This database holds the solution to various symptoms by referring to the health records and
treatments given for various patients. Here, the database does not holds the name or bio
data of any patient in his/her health record. The database only holds the health record of a
patient and the treatment given to them by the doctors.

1.1 Objectives

The main objective of this system is that enable a real time system of monitoring the health
of the patients, and it also enables early means of detection of clinical deterioration. This
idea of smart health care system is enabled by three promising technologies such as Radio
Frequency Identification (RFID) which holds the patient’s details, Wireless Sensor Net-
work (WSN) for enabling connection between database and patient, finally smart mobiles
for sending and receiving diagnostic solutions to various symptoms. It is given in the
Fig. 1.
Radio Frequency Identification (RFID) is made of antenna and a small chip. RFID
reader device gives signal for the antenna. These reader devices later transmit the ID to
the database, generally let us consider cloud database. The available tags are of active and
passive type. RFID is also used for transmission information for a range of few kilometres.
Thus, RFID is efficient and they can be easily embedded in any real life devices. These
devices lead to the development of RFID sensor networks and in development of IoT. This
RFID tags lead to new a concept of “Anytime, anywhere, any media” and has also achieved
its goals successfully. One of the major defects and problems in the sensor network is
“management of data and communication, increasing the traffic overhead and requirements
for bandwidth of the sensor networks”. In order to handle emergency and important cases
the doctors must be near the patients in hospital for generating the needed prescriptions.

13
IoT for Health Monitoring System Based on Machine Learning… 191

Fig. 1  Promising technologies

Micro Electro Mechanical Systems (MEMS) Walsh et al. [1] plays a major role in the
implementation of smart environments and recent advancement in this technology has
made it possible. Also, the advancement in sensor technologies [2] extends it support-
ive hands in analysing the “patient’s heartbeat, body pressure and temperature, ECG and
motion rate”. Thus, the advancement of these technologies has helped in providing health-
care to the citizens’. Further in this field, UHF and RFID technologies plays a major role in
tracking and monitoring the health state of patients in hospitals and other nursing institutes.
The idea is to combine the functionality of two tags. The wearable tags and the ambient
tags are combined together and this combined technology helps in developing a complete
passive RFID system. This system can be designed based on the patient’s caring require-
ment. One of the parameter in sensing patient’s health state is NIGHT-Care. NIGHT-Care
is entirely based on futuristic intelligence platform and this helps in analysing and estimat-
ing the person’s sleeping parameters, activity of the human, and it also helps in predicting
the abnormal events in patient’s body that requires some immediate assistance in advance.
For the purpose of localising the hospital, passive RFID is used. Under the coverage region
of the readers these RFID tags works perfectly and efficiently. But rather than the useful
features of RFID technology, they also have its own limitations, monitoring and tracking
the patient/devices is quite impossible in this technology. Hence as a solution to this case,
WSN is used in localizing the patient and in a health care centre, they helps in tracking and
monitoring various services. Balakrishnan et al. [3] propose a “modified Multi-manifold
Deep Metric Learning (MMDML) approach used for image set classification of any cancer
diseases”.
The engine used for localizing and tracking is held mainly by two components. The
major components used for this purpose is Received Signal Strength Indicator and in case
of filtering particle filters are used, while the movements of the patients are classified by
using the biaxial accelerometers. Also in order to track and monitor the state and condition
of the patients admitted in the hospital or in the indoor environments, a wireless localiza-
tion network is used for tracking. This wireless localization network helps in monitoring
both the health and the physical status of the person. A three tier architecture system is

13
192 S. Balakrishnan et al.

specifically used in “denoting the WSN4QoL [4], the lower tier is composed of a Bluetooth
enabled Wireless Body Area Network (WBAN) [5–8])”. This WBAN enables a connec-
tion between the nodes of the sensor and to a collector that is managed locally; as a result
it sends the details regarding the measurement report towards the gateway via a Zig Bee
network which is of IEEE standard. This gateway is designed for performing local compu-
tations and further it helps in forwarding data to the public IP networks Page et al. [9]. For
the purpose of real time analysis, it sends the data towards professional caregivers. Further,
in order to get details about the health state of the patients frequently and in providing
effective healthcare services to the patients, 6LoWPAN [10] standards are used and imple-
mented. Furthermore in detail, for the purpose of measuring the Photo Plethysmo Gram
(PPG) signals, WSN devices are used as a measuring tool and then it delivers then to a
server mentioned via internet. By means of creating a customised application [11], in order
to extend this healthcare service to the people, an android device is used.
More prominently in a health care centre, the doctors are mostly not available to handle
all cases [12–14]. In these cases a long queue is added in the hospitals and it causes lots
of annoyance to the patients. Environment integration and the methodology of monitoring
patients do not exist in many rural areas. The report of patient is normally framed only
by the doctor. The doctor must be present in the hospital for the purpose of giving pre-
scriptions and in handling many emergency cases. In a worst case if the doctor is in other
location apart from the hospital, the patients report is sent to the doctor via messages or
by calls, and sometimes this may lead to conflicts. Thus by creating a mobile based appli-
cation for health care units is mandatory in monitoring the status and health reports of
the patients. For the purpose of avoiding the above mentioned limitations, the proposed
model aims at addressing the patients without having any actual or physical contact with
the health care unit, also the prescriptions for the patients can be generated even in the
absence of the doctor.
Data storage security involves “protecting storage resources and the data stored on
them—both on-premises and in external data centers and the cloud—from accidental or
deliberate damage or destruction and from unauthorized users and uses”. It’s an area that
is of critical importance to enterprises because the majority of data breaches are ultimately
caused by a failure in data storage security.
In addition to storing data effectively, organizations need to take proactive steps to pre-
vent bad actors from using the Internet of things to breach their networks. A few best prac-
tices when it comes to IoT data protection and security are: Limit connectivity and access,
Keep software and passwords updated, Implement firewalls and Encrypt IoT data.

2 Proposed System

The idea of this proposed system is to implement an “IOT based Smart Healthcare System”
which uses Pro-prediction algorithm for predicting the exact prescription for the patients.
Sensors are used in sensing the environmental factors such as temperature, humidity, ambi-
ent light etc. Using IOT and other enabling technologies, the available health care centres
are good in managing both common and serious diseases. Also these health care centres
enables connections with other health care centres via network and they can communicate
and share resources among each others and all the details of patients can be stored in the
cloud database. A villager or the patient can complete his registration process when he vis-
its the health care centre for the first time and they are guided my other medical assistants

13
IoT for Health Monitoring System Based on Machine Learning… 193

in completing this process. A unique RFID band will be issued to the patient which holds
all the medical details of the patient along with his medical history. These RFID is fitted
in the wrist bands for the convenient use of patients and these tags will always remain
online. This system functions by, if a person is sick, then he or she need not visit the hos-
pital. Instead the can press the button in the RFID which in turn transmits a signal to the
reader at the health centre. When this function is triggered, the patient is identified using
his registration number. “Later the health record of the patient is analysed from the health-
care server”. The patient symptoms that are sensed is submitted to an expert technological
medical system. This expert medical system is capable of analysing the current condition
of the patient and prescribes needed medicines using pro prediction algorithm. This expert
system works by sensing the present state of the patient and comparing it with the patient’s
medical history. The detail of the patient along with their health records is stored in a cloud
database and so it can be accessed from any part of the country. The “graphical chart and
the prescription will be created as a PDF File and send to that patient”. Figure 2 describes
about the proposed system architecture.

3 PRO Prediction Algorithm

The proposed pro prediction algorithm is used in detecting the prescription that has to
be given for the patient. This algorithm has some flavours of “ensemble learning tech-
nique in the prediction of the exact prescription that could be given to the patient”.
Initially RFID wrist bands are read to initiate the functioning of pro prediction algo-
rithm. By this process, we can archive the details of the patient. The entire bio data
of the patient is stored in a medical database. This medical database holds the history
of the patient’s health records. It also holds the detail of what type of treatment suits/
does not suits for a patient; for e.g., based on the health conditions few treatments does
not supports for a patient body. By analysing these data records, the time spent for test-
ing various new medicines on the patients can be eliminated and the needed treatment
can be given to the patient directly. The entire functioning of the algorithm is based
on the inputs sensed by the Smart Health Sensor (SHS). The Smart Health Sensor is

Fig. 2  Proposed system architecture

13
194 S. Balakrishnan et al.

designed to sense parameters of the patient such as blood pressure, temperature, heart
beat rate and other necessary parameters. Once the RFID tags of the patients are read,
the SHS sensors senses and based on the data sensed by SHS, the algorithm looks for
the patient’s health record in the database. The algorithm looks for the symptoms sensed
in the patient’s health record. If the patient has already experienced these detected
symptoms, then the resulted prescription is delivered directly to the patient. In case, if
the patient has not experienced these detected symptoms, then the algorithm looks for
the solutions in the medical database. Using the one of the machine learning algorithm
[15] that is dimensionality reduction algorithm, the records with similar health records
and solutions are listed. Dimensionality reduction algorithm is used for the purpose of
reduction of unmatched health records and it eliminates the manual work of eliminating
unmatched health records. The needed prescription is given at the output layer of the
algorithm and finally the record of the patient is updated for future reference. Pro pre-
diction algorithm is given in the Fig. 3.
In the Pro Prediction algorithm X1, X2, X3, and X4 denotes various input parameters
sensed by the Smart Health Sensor (SHS). The input parameters includes various envi-
ronmental factors, pressure, heart beat rate, temperature etc.
Here H represents the health record of the patient that is archived from the medical
database when the RFID tag is read. The formula needed for calculating H is given below,
n
∑ D
H = Xi + log , (1)
i−0
Xi

where H—health record of the patient; D—Data taken from dataset; X—various param-
eters sensed by SHS; n—total number of paramenters sensed.
If the solution for the symptoms exists in the H, the needed prescription is prescribed
directly. Or else, using dimesionality reduction, prescription is given based on similar
health records of other patient’s. The concept of machine is used inorder to eleminate
the time needed for manual/other technical search that has to be done in searching for
similar health records.

Fig. 3  Pro prediction algorithm

13
IoT for Health Monitoring System Based on Machine Learning… 195

3.1 P Pseudo Code

4 Experiment

We did our experiment in “four modules which includes Environment Sensing, Regis-
tration and Ward Allocation, Patient and Environment Monitoring and Report Analy-
sis”. Here the modules are explained as follows:

4.1 Environment Sensing

The environmental conditions of the hospital wards are sensed by placing the sensors in
the needed wards and these sensors helps in sensing some of the environmental condi-
tions such as temperature, humidity, ambient light etc. of that particular ward. A unique
tag ID is provided to each sensors. The sensors starts sensing the data and it upload
the data about each ward to the cloud server. The patient details form must be filled
out by the patient who is admitted. The patient or his/her care taker must fill the form
that holds the symptoms he feels or various issues he has. Based on the description
given by patients, the ID of the patients will be updated automatically to the ward. The

13
196 S. Balakrishnan et al.

localization problem is also resolved by using WSN nodes combined with UHF RFID
readers.

4.2 Registration and Ward Allocation

Normally in healthcare units, each ward has a two or three doctors and few nurses for their
help. The doctors and nurses working in the health care units also have their unique ID’s
and they have to use these ID’s to manage the hospital web application. In absence of the
doctors, the nurses are responsible in monitoring the health state of the person. They are
primarily responsible for tracking the environmental conditions and in updating the health
records of their wards to the database server.

4.3 Patient and Environment Monitoring

The nurses keeps on verifies the health state of the person and a dynamic chart is gener-
ated that gives a diagrammatic representation of the patient’s health state. This dynamic
graph is automatically updated based on the changes sensed by the sensor. The nurses or
the care taker can use these dynamic charts for analysing a person’s health state. In case
if the patient is going to face a critical situation, the dynamic graph chart gives a warning
based on the assumption made by the sensor.

4.4 Report Analysis

Finally based on the loaded data sensed by the sensors, the pro prediction algorithm is
implemented in predicting the exact solution to the detected symptoms. Further if needed,
these reports can be consulted with the doctors for future analysis. An image chart is gener-
ated and converted to PDF and it can be maintained as the patient’s health status. The doc-
tor can use these records and it helps the doctors in handling the patients in a better way.
The health records of the patients held in the database will be automatically updated. Thus
this system helps the doctors in providing prescription and referring to the report of the
patient’s from remote location.

5 Results and Discussion

This paper “holds an IOT based smart healthcare system that plays a major role in auto-
matic monitoring and tracking of patients when the doctors are away from the patients”. In
fact, providing treatment to a person in the absence of a doctor is really hard and this could
be made possible by the implementation of SHS technology. The idea is not to eliminate
doctors completely in giving treatment to a patient. This idea is proposed to handle the
tough cases a patient deals with during the absence of doctors. This does not support emer-
gency cases where the patients have to be admitted in the hospitals but it helps in giving
appropriate first aid techniques that has to be given in handling the emergency cases. This
idea acts as a helping hand to many victims who suffer from normal symptoms of com-
mon diseases. The patient’s Normal Blood Pressure Values are given in the Table 1. The
heart beat rate of a normal human is listed in Table 2. This data collected from hospitals in
Chennai.

13
IoT for Health Monitoring System Based on Machine Learning… 197

Table 1  Blood pressure values of Age Gender Min/Max (mmHg)

normal human
< 18 Male 80/120
18 to 20 Male 80/125
21 to 40 Male 85/135
40 and above Male 85/135
< 20 Female 80/123
21 to 40 Female 85/133
40 and above Female 85/133

Table 2  Rate of normal heart State BPM

beat
Relax 60–100
Sleep 40–50
Tachycardia > 100

Fig. 4  Brainsense headband

And also, this section analyzes the experimental results of the proposed algorithm. To
test the “500 performance of the proposed method the live brain waves is collected from
the 500 users with the help of brainsense headband”. For visualization, a brainsense head-
band is shown in the Fig. 4.
The smart health care system proposed by Hemalatha et al. [16] is composed of RFID
tag, local server, internet, servers for medical purpose and service provider. The physio-
logical condition of the patient is viewed via RFID tags and with the help of LAN connec-
tions, doctors give treatments to the patients. The main objective and theme of this paper
is to monitor and handle the patients regularly and carefully. A battery-less RFID tags
with has additional features of sensor capabilities were proposed by Fellini et al. [17]. This
paper holds RFID tags that have some additional sensors attached to it. In order to trace the
parameters like temperature, blood pressure, weight sensors etc. bio sensors were used. A
low power wearable ECG monitoring system was proposed by Elisa Spanò et al. [18], for
Multiple-Patient Remote Monitoring. Hoda Ramin Hossein [19] proposed a concept that
monitors the status of the patients admitted in the hospital with the help of smart phones.
Moeen Hassanalieragh et al. [20] highlighted the goals and functions of IOT in handling
the future vision of healthcare units. Vivek Pardeshi et al. [21] proposed “Raspberry-Pi
based health monitoring system using IoT”. Any abnormalities in the health conditions can

13
198 S. Balakrishnan et al.

be known “directly and are informed to the particular person through GSM technology or
via internet”.

5.1 Data Analysis

This model is built by using the available training and meta data. The dataset descrip-
tion consists of 500 users with 150 number of trails. The total records held in a training
data ranges above 500 based on two distinct vital signs and the health state of the victim
or patient. The data used for testing comes under the category of the machine learning
approach which is based on supervised learning method and the solution to this model is
composed using an algorithm proposed called pro prediction algorithm. The health condi-
tion of a patient varies from one person to another. The information regarding the health
state of the patient is stored in the form of Meta data in the cloud database. These data
holds various details regarding the conditions of the patients.
Our proposed system was tested using IMedBox simulator compared to the work of
Vivek Pardeshi et al. [21] and Hemalatha et al. [16] and simulation parameters are given in
the Table 3.
Table 4 shows “the values of room temperature, body temperature and heart rate” (i.e.
the values are measured at different times in a day).
From this, we know that “the body temperature is directly proportional to the heart rate
and the body temperature slightly depends on room temperature”. This is shown in the
Fig. 5.
Figure 6 shows the classification of instance of vital signs that are correctly and incor-
rectly classified by applying and using different classification algorithms. The algorithms
used for classification are J48, SVM (Support Vector Machine) and Naive Bayes algo-
rithms. J48 and SVM are used among these algorithms as they give better results.
To measure the performance of the system, precision is used, Recall and F-Measure.
The percentage of messages is denoted by the precision in the test data classified which
truly are.
|{Relevant doc} ∩ {Retrieved doc}|
Precision = (2)
|{Retrieved doc}|

The proportion of actual messages is denoted by recall in the set for test that are catego-
rized by the classifier.

Table 3  Simulation parameters Parameter Value

Number of nodes 30 wireless nodes

Routing protocol CoAP/IPv6
No. of gateway 5
No.of internet server 1
Propagation model Two ray ground
Terrain dimensions 800 × 800
Channel Wireless channel
Queue type Priqueue
Queue length 50

13
IoT for Health Monitoring System Based on Machine Learning… 199

Table 4  Values of room Time Room temperature Body temperature Heart

temperature, body temperature (°F) (°F) rate
and heart rate (bpm)

6–7 A.M 32.22 95.72 60

32.75 96.64 63
7–8 A.M 32.93 96.76 63
33.49 97.7 65
10–11 A.M 33.77 97.15 65
34.55 97.22 65
12–1 P.M 38.78 97.07 67
39.49 98.03 70
1–2 P.M 39.68 97.18 68
39.26 97.35 65
4–5 P.M 39.54 96.28 65
37.63 96.15 62
5–6 P.M 31.76 95.45 63
31.21 96.05 63
6–8 P.M 39.68 97.18 68
39.26 97.35 65
8–9 P.M 30.73 97.12 66
30.25 96.5 64
9–10 P.M 30.6 96.58 63
30.57 96.41 63

Fig. 5  Relationship between sensors values

|{Relevant doc} ∩ {Retrieved doc}|

Recall = (3)
|{Relevant doc}|
For combining both precision and recall F-measure is used.

13
200 S. Balakrishnan et al.

Fig. 6  Classification of vital signs

Table 5  Different classification algorithms’ accuracy measure

Vital signs Algorithms of classification Precision Recall F-measure

Heart rate Support vector machine 0.942 0.96 0.951

Naïve Bayesian 0.803 0.659 0.624
J48 0.959 0.956 0.956
Temperature Support vector machine 0.918 0.954 0.933
Naïve Bayesian 0.816 0.752 0.744
J48 0.918 0.924 0.953

Fig. 7  Performance analysis of body temperature

13
IoT for Health Monitoring System Based on Machine Learning… 201

Fig. 8  Performance analysis of heart rate

Fig. 9  The accuracy based on the experience of patient’s admitted in health care units under each test case

2 × Precision × Recall
F − measure = (4)
Precision + Recall
Table 5 shows the measure of accuracy of various algorithms of classifications for
vital signs. The performance analysis of body temperature is shown in the Fig. 7.
The performance analysis of heart rate is shown in the Fig. 8.
In this two different use cases with different results has been included to validate the
purpose of the proposed Smart Healthcare System. The results achieved by this method
demonstrates the success of this proposed technique, it not only performs the identifi-
cation and patient tracking system, but also provides a means of immediate handling
of emergency cases and power effective remote patient monitoring. Figure 9 shows the
experience of patient’s admitted in health care units under each test case. As for as we
concern, based on the various observations that is made in all the cases, the patient
learns to interact with the platform is almost same and in particularly emergency cases
it was high.

13
202 S. Balakrishnan et al.

6 Conclusion

6.1 Future Enhancements

In this paper, we tried to emphasize on machine learning based IoT device for monitor-
ing the patients using cloud environment must have its aim “to improve the standard of
not only realizing the illustration and traceability of healthcare user but also guarantee
the improved healthcare services”. And we observed that in all of the cases, the patient
learns to interact with the platform is almost same and in particularly emergency cases
it was high. So, our system provided more accuracy in each usability test according to
the patient experience in particularly emergency situation.
The proposed system makes use of technology to face many challenges in the health
care units such as increasing the cost of the healthcare delivery, sharing of information
and shortage of healthcare professionals who can take better care of the patients and
enhanced various services for the patients for the betterment in their health. However,
due the trust, privacy and security issues the benefits are gained but it also faces some
of the issues that are technical that must be addressed before healthcare providers and
actors can fully adopt the trust the proposed system. However the model can have some
additional features that make it a user friendly model.

6.2 Recommendation

Healthcare sensors will produce enormous quantity of data and so the concept of Big
Data was introduced. To find the health status of the patient Map reduce algorithm can
be used. Security and privacy of patients’ are the main factors in constructing a healthy
health care framework.

Funding There is no funding provided to prepare the manuscript.

Data Availability Data sharing not applicable to this article as no datasets were generated or analysed during
the current study.

Code Availability No code is available for this manuscript.

Declarations
Conflict of interest There is no conflict of Interest between the authors regarding the manuscript preparation
and submission.

Ethical Approval This article does not contain any studies with human participants or animals performed by
any of the authors.

Informal Consent Informed consent was obtained from all individual participants included in the study.

Consent to Participate There is no consent to participate or any concerns in the manuscript.

Consent to Publish There is no consent or any copyright needed to get concerns in the manuscript.

13
IoT for Health Monitoring System Based on Machine Learning… 203

References
1. Walsh, S., Linton, J., Grace, R., Marshall, & Knutti, S. (2000). MEMS and MOEMS technology and
applications. SPIE—The International Society for Optical Engineering.
2. Kartsakli, E., Antonopoulos, A., Alonso, L., & Verikoukis, C. (2014). A cloud-assisted random linear
network coding medium access control protocol for healthcare applications. Sensors, 14, 4806–4830.
3. Balakrishnan, S., Aravind, K., & Jebaraj, R. A. (2018). A novel approach for tumor image set clas-
sification based on multi-manifold deep metric learning. International Journal of Pure and Applied
Mathematics, 119(10c), 553–562.
4. Tennina, S., Santos, M., Mesodiakaki, A., Mekikis, P.-V., Kartsakli, E., Antonopoulos, A., Di Renzo,
M., Stavridis, A., Graziosi, F., Alonso, L., & Verikoukis, C. (2016). WSN4QoL: WSNs for remote
patient monitoring in e-Health applications. 2016 IEEE International Conference on Communications
(ICC). IEEE. https://​doi.​org/​10.​1109/​ICC.​2016.​75115​97
5. Bazzani, M., Conzon, D., Scalera, A., Spirito, M., & Trainito, C. (2012). Enabling the IoT paradigm
in e-health solutions through the VIRTUS middleware. IEEE 11th international conference on trust,
security and privacy in computing and communications (TrustCom) (pp. 1954–1959). IEEE.
6. Benharref, A., & Serhani, M. (2014). Novel cloud and SOA-based framework for e-health monitoring
using wireless biosensors. IEEE Journal of Biomedical and Health Informatics, 18(1), 46–55.
7. Milenkovi, A., Otto, C., & Jovanov, E. (2006). Wireless sensor networks for personal health monitor-
ing: Issues and an implementation. Computer Communications, 29(1314), 2521–2533.
8. Paradiso, R., Loriga, G., & Taccini, N. (2005). A wearable health care system based on knitted inte-
grated sensors. IEEE Transactions on Information Technology in Biomedicine, 9(3), 337–344.
9. Page, A., Kocabas, O., Soyata, T., Aktas, M., & Couderc, J.-P. (2014). Cloud-based privacy-preserving
remote ECG monitoring and surveillance. Annals of Noninvasive Electrocardiology (ANEC). https://​
doi.​org/​10.​1111/​anec.​12204
10. Bui, N., & Zorzi, M. (2011). Health care applications: A solution based on the internet of things. Pro-
ceedings of the 4th international symposium on applied sciences in biomedical and communication
technologies. ISABEL’11.
11. Olorode, O., & Nourani, M. (2014). Reducing leakage power in wearable medical devices using mem-
ory nap controller. 2014 IEEE dallas circuits and systems conference (DCAS) (pp. 1–4). IEEE.
12. Sheela, K. G., & Varghese, A. R. (2020). Machine learning based health monitoring system. Materials
Today: Proceedings, 24, 1788–1794. https://​doi.​org/​10.​1016/j.​matpr.​2020.​03.​603
13. Pandey, H., & Prabha, S. (2020). Smart health monitoring system using IOT and machine learning
techniques. 2020 sixth international conference on bio signals, images, and instrumentation (ICBSII)
(pp. 1–4). IEEE. https://​doi.​org/​10.​1109/​ICBSI​I49132.​2020.​91676​60
14. Aldahiri, A., Alrashed, B., & Hussain, W. (2021). Trends in using IoT with machine learning in health
prediction system. Forecasting, 3, 181–206. https://​doi.​org/​10.​3390/​forec​ast30​10012
15. Jebaraj Ratnakumar, A., & Balakrishnan, S. (2018). Machine learning based grape leaf disease detec-
tion. Journal of Advanced Research in Dynamic & Control Systems, 10, 775–780.
16. Rallapalli, H., & Bethelli, P. (2017). IOT based patient monitoring system. International Journal of
Computing, Communications & Instrumentation Engineering (IJCCIE), 4(1), 115–118.
17. Fellini, C., Merenda, M., & Della Corte, F. G. (2012). Battery-less smart RFID tag with sensor capa-
bilities, RFID-technologies and applications (RFID-TA). 2012 IEEE international conference on
RFID-technologies and applications (RFID-TA) (pp. 160–164). IEEE.
18. Spanò, E., Di Pascoli, S., & Iannaccone, G. (2016). Low-power wearable ECG monitoring system for
multiple-patient remote monitoring. IEEE Sensors Journal, 16(13), 5452–5462.
19. Hossein, H. R. (2015). SPHMS: Smart patient m-healthcare monitoring system with NFC and IOT.
International Journal of Computer Applications Technology and Research, 4(12), 956–959.
20. Hassanalieragh, M., Page, A., Soyata, T., Sharma, G., Aktas, M., Mateos, G., Kantarci, B., &
Andreescu, S. (2015). Health monitoring and management using internet-of-things (IoT) sensing with
cloud-based processing: Opportunities and challenges. 2015 IEEE international conference on ser-
vices computing (pp. 285–292). IEEE. https://​doi.​org/​10.​1109/​SCC.​2015.​47
21. Pardeshi, V., Sagar, S., Murmurwar, S., & Hage, P. (2017). Health monitoring systems using IoT and
raspberry Pi—A review. 2017 international conference on innovative mechanisms for industry appli-
cations (ICIMIA 2017) (pp. 134–137). IEEE.

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

13
204 S. Balakrishnan et al.

S. Balakrishnan completed his UG and PG in the domain of Computer

Science and Engineering. He completed PhD in the domain of Com-
puter Science and Engineering in the Bharath Institute of Higher Edu-
cation & Research Institute, Chennai. Currently he is a Professor in Sri
Krishna College of Engineering and Technology, Coimbatore. His area
of interests are Artificial Intelligence, Cloud Computing and IoT.

K. Suresh Kumar completed his UG and PG in the domain of Informa-

tion Technology. He completed PhD in the domain of Computer Sci-
ence & Engineering majoring in Web Security from Anna University,
Chennai. Currently he is an Associate Professor in Saveetha Engineer-
ing College. His area of interest are web security, Cloud Computing,
and Image Processing.

L. Ramanathan completed his UG and PG in the domain of Computer

Science Engineering. He completed his PhD in the domain of Com-
puter Science Engineering in the Vellore Institute of Technology, Vel-
lore. He is currently working as an Associate Professor in VIT, Vellore,
His area of interests are networking, Machine learning and Artificial
Intelligence.

13
IoT for Health Monitoring System Based on Machine Learning… 205

S. K. Muthusundar completed his UG and PG in the domain of Com-

puter Science and Engineering. He completed PhD in the domain of
Computer Science and Engineering in the Bharath Institute of Higher
Education & Research Institute, Chennai. Currently he is a Professor of
Computer Science Engineering in Sharad Institute of Technology Col-
lege of Engineering. His area of interests are Artificial Intelligence,
Image processing and IoT.

13