A Secure Emergency Framework in an IoT Based Patient Monitoring
System
Neila Mekki
1
, Mohamed Hamdi
2
, Taoufik Aguili
1
and Tai-hoon Kim
3
1
Communications Systems Laboratory ,(SysCom), National Engineering School of Tunis, (ENIT),
University of Tunis El Manar, (UTM), Tunisia
2
Cytekia, Higher School of Communication of Tunis, Sup’Com, University of Carthage, Tunisia
3
School of Information Science, University of Tasmania, Australia
Keywords:
Internet of Things, Healthcare, Monitoring, Patient, Doctor, Diabetic, Security, Authentication.
Abstract:
Today, the Internet of Things (IoT) in healthcare has become more productive to reduce the gap between doc-
tors and patients. If any problem has occurred to the patient, then the doctor approaches the patient and gives
the appropriate treatment. In this context, the main research question is how to identify the methodological
choice (paradigm, approach, and method) in coherent with theoretical foundation of the Internet of things.
Our objective is to involve better management of public IoT healthcare application by following a prevention
methodology. To address this need, we adopt the design science research (DSR) methodology to implement
a smart healthcare application. The ultimate goal is to enable the dynamic prediction and/or detection of the
patient health deterioration, which taking into consideration the patient health evolution. So the prototyping
process suggests some important factors to monitor and assist living diabetic patient at any time. The doctor
have the ability to easily monitor and manage the patient health and can save precious minutes every day.
Without having to manually visit each patient, the doctor can give a remote diagnosis and track the medical
assets. To this purpose, we provides a theoretical contribution which the DSR assists in identifying an intelli-
gent healthcare application based on IoT technology.
1 INTRODUCTION
The Internet of Things (IoT) is an emerging paradigm
that enhance our everyday life through automation
and optimization tasks related to transport health-
care domain, smart-home and smart-enterprise, and
so forth. IoT is a huge network of things, which can
recognize, manage, and control all kinds of objects
around us through sensors and embedded devices.
According to Gartner (Sallabi and Shuaib, 2016), in
2020, there will be 26 billion things for a market ex-
ceeding $300 billion. This new market has been in-
vested by big companies, such as Samsung, Amazon
or Google, that are developed things for the IoT and
try to stay ahead concurrent by proposing emerging
and innovative products.
However, the traffic monitoring system is prone
to some cyberattacks. A malicious user may control
a sensor to upload forged information to the system,
which may cause traffic chaos.
In particular, we address those challenges in
healthcare application context. Healthcare is among
the fastest-growing domains since it affects the whole
world population. A continuous monitoring and real-
time communication of patients have always been the
mean idea of smart healthcare services of cardiac (ren,
2018), Diabetes mellitus (Al-Taee et al., 2017), or res-
piratory problems (Mukhopadhyay et al., 2018).
Despite that security and privacy of client’s data
remain a tremendous challenge to address (Ida et al.,
2016) (Sowmiya et al., 2016), the psychological and
environmental factors interact in decision making and
behavior (Alaiad and Zhou, 2017). Such as, psy-
chosocial problems (Ferrer and Mendes, 2018) are
most common in diabetes patients (Chew et al., 2014)
that often result in serious negative impacts.
Additionally, our previous recherche works were
done on describing the scenario a healthcare applica-
tion (Mekki et al., 2017), solving the noise problem
due to multiple sensor applied WBAN and provid-
ing an authentication protocol to correctly identify the
communication requests from legitimate user (Mekki
et al., 2018),(Mekki et al., 2023) .
For this purpose, our challenge is to allow a com-
844
Mekki, N., Hamdi, M., Aguili, T. and Kim, T.
A Secure Emergency Framework in an IoT Based Patient Monitoring System.
DOI: 10.5220/0012138400003555
In Proceedings of the 20th International Conference on Security and Cryptography (SECRYPT 2023), pages 844-849
ISBN: 978-989-758-666-8; ISSN: 2184-7711
Copyright
c
2023 by SCITEPRESS Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
plete regroup hardware system that gives users control
and check the existence of a threshold event invoked
by a particular sensor (e.g ECG or blood glucose).
However, a design model which gives users control
and lets them check their physical condition by them-
selves was not yet mentioned.
Our contribution consists on how do we construct
a secure protocol to achieve security between users
and sensor nodes? Moreover, how do we transform
data raw into cognitive knowledge information shared
between IoT system components?
To answer those questions, we illustrate our
methodology for a diabetic patient based on De-
sign Science Research Methodology (DSRM)(March
and Smith, 1995). The DSRM supports a research
paradigm that calls the creation of innovative artifacts
to solve real-life problems. We present the empiri-
cal study, which begins with researching the adequate
methodology in coherence with research objectives
and theoretical foundations.
Among others, the following components must be
set:
The key security requirements and challenges in
IoT with a specific focus on healthcare applica-
tions.
The balance between security and functionality by
explaining our system model and methodology.
The ultimate goal of this methodology is provid-
ing a flexible smart IoT-based system able to perceive
the collected data and provide decisions.
This paper is organized as follows. First in Section
2, we begin by some overview of exiting work in IoT
for healthcare application. Second, in section 3, we
propose our adaptive methodology a find to involve
better management of public IoT healthcare applica-
tion. Finally, concluding and future trends are drawn
in Section 4.
2 RELATED WORK
In the past few years, IoT has become most produc-
tive in the area of healthcare, to improve the quality
of care to the patient. Such as diabetes can be actively
treated and monitored if an early diagnosis is avail-
able.
In this context, some contributions have addressed
the focuses on developing an IoT healthcare system,
in which are summarized in table 1 .
Compared to those previous researches, some of a
prior studies did not adopt any sensors (Khan, 2020)
(Rghioui et al., 2021), while some others studies us-
ing these sensor (Islam et al., 2023) technologies for
collecting real-time data.
Different from all the previous researches, our
contribution combine security requirement and DSR
to monitor the diabetic patient.
Such as, our challenge is to provide a secure
emergency framework for healthcare application ,
which include a biomedical sensors, an android inter-
face for monitoring real-time sensor data, and a cloud
infrastructure for processing the real-time data.
3 PROPOSED ADAPTIVE
METHODOLOGY
In this section, we illustrate detailed information in-
cluding the methodology of our system architecture
for diabetic patient-based Design Science Research
Methodology (March and Smith, 1995). Following
the proposed methodology, we have combined emo-
tion and Cognitive IoT for developing a flexible mon-
itoring system. The research aims to create an artifact
to support the decision making for smart healthcare
framework. To build an artifact-based design science
paradigm, we consider two basic activities: build and
evaluate as follows respectively:
Build: is the process of constructing an artifact for
a specific purpose. We build an artifact to perform
a specific task.
Evaluate: is the development of criteria perfor-
mance evaluation.
IoT methodologies can be divided into two categories:
bottom-up approaches and top-down approaches.
The Bottom-Up Approaches: consist of selecting a
priory the residual risk, i.e. the degree of protection
of the system and implementing the countermeasures
that allow reaching it.
The Top-Down Approaches: define scheduled tasks
that are intended to reduce the identified threats.
Both development approaches use the same tool-
chain, and can thus be applied to best suit the needs of
the particular feature development project. It is also
possible to start development with the bottom-up ap-
proach and complete it with the top-down approach.
However, the challenge encountered while imple-
menting our approach, is to achieve and maintain mo-
bility without attacks or vulnerability. Besides, how
the collected data are transformed into insights and
interact with domain experts for better decisions.
We describe as our methodological contribution
in coherence with some elements of qualitative ap-
proaches (case study and narrative analysis (Mekki
et al., 2017)), to design and implement a monitoring
system in IoT-Based on WBAN for a diabetic patient.
A Secure Emergency Framework in an IoT Based Patient Monitoring System
845
Table 1: Benchmarking overview of exiting work in IoT for healthcare.
Ref Embedded platform emergency alerts Sensor
(Khan, 2020) Client computer with IoT sensors No Not done
(Rghioui et al., 2021) IoMT Based cloud infrastructure No Not done
(Islam et al., 2023) Android Device with Iot wearable Device Yes Done
Our main focus is to equip the methodology with hu-
man cognition, which is capable of performing intel-
ligent decision making independently.
These artifacts then become the object of the
study. The methodology followed in our research is
classified into four methods and activities, linked with
the corresponding steps of the DSR approach. DSR
artifact can include: Construct, Models, Method, and
Instantiation.
3.1 Construct
Let’s consider our smart IoT-based healthcare sce-
nario, in which a medical IoT device publishes the
physical condition (such as ECG or blood glucose) of
a diabetic patient to a remote healthcare center peri-
odically.
Despite their large deployment, security issues,
mainly related to data privacy, are often considered
as potential obstacles that might limit the extent of
such solutions. Hence, the constructed artifact itself
presents a challenge to explain how and why it works.
Hence, the attacker blocks critical health information
from being transmitted to physicians, threatening the
life of the patient.
Context: In general, the IoT-base healthcare system
collects sensitive data. The platform must ensure an
adequate level of security to data access and manage-
ment. At this stage, management analysis must clarify
and define functional requirements and design con-
straints. Functional requirements define quality (how
good), environment (know-that), availability (how of-
ten) and the procedural knowledge (know-how) to
make decisions. To ensure such functionalities, the
system should be able to understand the meaning of
the received data and history as well as the disease
management strategies to provide the right decisions.
Problem: If we consider deploying and activating
all the monitoring processes for each patient, this re-
quires increasing the cost. Moreover, hard coding of
all rules causes maintainability problems. Besides,
privacy is also a primary concern in the remote health
monitoring system, as health data are highly relevant
to the patient being monitored.
Solution: To deal with those challenges, our method-
ology design, are used to combine knowledge and se-
curity analyses. We defined within this methodology
a set of (1) delineating the security dynamic coordi-
nation of the management processes to deal with the
system’s context changeability and (2) cognitive abil-
ities to IoT-based systems to interact with the human
through generating new insights and to solve prob-
lems as human do.
3.2 System Model
In this sub-section, we illustrate the system model of
our case study of the diabetic patient. It is depicted
by figure 1, which consists of: (i) a trusted domain
with sensors and IoT devices under a single adminis-
trative domain (e.g. at home, or outside the home),
and (ii) an untrusted domain connected via the public
Internet, including the illegal user.
Note that we consider the application logic to be
separate from its implementation: while the design of
the application logic can be separately verified, vul-
nerabilities can still be introduced due to implemen-
tation bugs.
Proposing a system model for the patient monitor-
ing system should take into account the key character-
istics of IoT. The techniques should be asking enough
to various context. Thus, the IoT system must have
strong security guarantees.
We propose the bottom-up approach for verifying
the security of the software stack in an IoT system, to
provide a guarantee for how the software is secure.
We plan to verify the security of the actual (source
and machine) code that will run on IoT devices. By
bottom-up approach, we mean that the security of
software needs to be established at every level of ab-
straction: at the OS (Operating System), implementa-
tion, and functional design levels.
Specifically, the choice of our methodology refers
to a pluralistic methodological approach (case studies,
architecture, and narratives approach).
As shown the figure 1, the arrow model presents
the precedence relationship between the eight pro-
cesses; denoted by 1- Context; 2-Activity monitoring;
3-Information Capture; 4-Vulnerability; 5-Threats; 6-
Risk; 7-Reduces /Countermeasures; 8- Security re-
quirements.
1-Context: It refers to the vast amount of infor-
mation processing that normally takes place out-
side conscious awareness, even during sleep. Our
methodology should be asking enough two moti-
vating factors.
SECRYPT 2023 - 20th International Conference on Security and Cryptography
846
Figure 1: Proposed Adaptive methodologies Patient Monitoring System.
Attack structure depends on context patient,
time, environment and access network.
The importance of security requirement may
constitute countermeasure selection. For to this
purpose, the key security requirement can be di-
vided into two steps: Information nature secu-
rity requirement, and context security require-
ment (As it has been introduced, in our previ-
ous works (Mekki et al., 2017)).
2-Activity monitoring: This process aims to mon-
itor and to achieve a good decision patient over
time. In this sense, our goal is to counteract his
or her activity to collect human body function. It
is a very important step, to understand what hap-
pened and devise response strategy. Therefore, it
is the association between context and informa-
tion capture. The monitoring of physical activity
facilities the development of adaptive healthcare
research. The real-time decision, based on which
the future action is taken, is made by the cognitive
module. The cognitive healthcare framework is
sufficiently intelligent to make corresponding de-
cisions. Thereafter, the cognitive system makes a
real-time decision on the activities to be provided
to patients based on their state. Such as patients
living with diabetes are not constantly watched,
like in a clinic or a hospital, but are managing
their disease largely by themselves. Therefore,
patients need to make best-individualized care de-
cisions about the daily management of their di-
abetes. The objective is to adapt and personalize
the lifestyle behaviors based on the collected data.
More specifically, positive emotions, such as hap-
piness, excitement, and contentment result in bet-
ter health behaviors and improved adherence to
treatment regimens
3-Information Capture: It is stimulated to mea-
sure and collect a physiological and movement by
using wireless body area networks (WBAN).
4-Vulnerability: Weaknesses and security
breaches of the analyzed system are identified at
this level.
5-Threats: this process aims to identify the attacks
that threaten the analyzed system. Unlike vulner-
abilities, threats are measurable as each of them
can be represented by its frequency and its sever-
ity. Threat rates and impacts depend on the envi-
ronment.
6-Risk: This process aims to identify the risk
that may threaten the asset of the analyzed system
based on the identified vulnerabilities and threats.
This process can be divided into the following
steps:
Identify the global (main) attacks correspond-
ing to the asset.
Build the attack scenario for every main attack.
7-Reduces /Countermeasures: It has been also
demonstrated that security countermeasures can
be viewed as the pseudo inverse of attacks con-
cerning the composition law. These criteria may
include:
The efficiency or the degree of protection
The feasibility of the countermeasure.
A Secure Emergency Framework in an IoT Based Patient Monitoring System
847
8-Security requirements: This process aims to re-
act to security requirements to reduce attacks. The
security requirements become an exigency to re-
duces and adapt the implementation of the coun-
termeasure such as key encryption, encrypted data
transmission, and authenticity.
Our model should implement cognitive capabilities
that allow good decisions at the right time.
3.3 Method
In this section, we study methods to improve an emer-
gency framework in an IoT based patient monitoring
system. To make it possible, we propose to create a
secure authentication in a given temporal and spatial
environment of the subject’s life. By using sensors,
the data will be captured and compared with the pre-
defined threshold. Our study focuses on blood glu-
cose and heartbeat rate, thus in case of emergency no-
tification will be sent to the doctor mobile.
Our challenge is to design the embedded security
framework, by factors good performance, robustness
to attacks and low energy consumption.
3.4 Instantiation
Designing an IoT-health system is a burdensome as-
signment because we should respect some key con-
cerns such as:
A providing authentication protocol to identify
and authorize the communication requests from a
legitimate user.
Designing an authentication protocol to minimize
the power consumption of the smart device (e.g.
blood glucose).
To meet all the requirements and make software
available as soon as possible, the software develop-
ment for embedded the system often uses V-model. It
enables also to perform tests at any time during the de-
velopment of our healthcare application. Here, the de-
velopment method influences the entire development,
from requirements definition to software release. In
the first step, is to meet all functional requirements but
not yet the restrictions of the target hardware. Such as
the functional requirement are checked using a model
methodology based on a bottom-up approach.
In the next step, the function and the model are
optimized for target two use cases of service analytic
capabilities of our application inside or outside home.
However, the capability to perform software tests
faster and more efficiently is of utmost importance to
healthcare application. We plan to introduce a safety
application and front-load more tasks to detect errors
early.
In fact, we develop a Management Application
(MA), is based PHP and Nginx web server eas-
ily configurable and accessible via a web browser.
Specifically, an MA can receive and reply to any re-
quest, coming from the user, in JSON format. A
MySql database, store the information retrieved from
WBAN.
Our solution consists of two parts: the first part is
the web service restful that run in the cloud accessible
by the smart gateway. It was developed by using PHP
and deployed on the Nginx server. The second part
consists of an android application which acts as the
client of our web service restful.
For those reasons, users (patients or doctors) need
to be authenticated before the access to the platform,
accessible via smartphone. such as the proposed
model of a secure authentication protocol based on
Restful approach for monitoring the diabetic patient
has been already introduced in our previous work
(Mekki et al., 2023).
Note That, these interfaces implement restful ser-
vices, which allow the user to communicate with
WBAN through the Smart Health Gateway (SHG). It
offers two main functionalities depending on two pos-
sible users (patient or doctors) as shown in Figure 2.
Patient interface: it allows the patient to register to
SAA. Furthermore, the patient can be equipped with a
smartphone and running application, named MobiDi-
abetic, a customized Android Application with spe-
cific privilege access.
Doctor interface: it allows medical staff to reg-
ister to SAA. Furthermore, doctors can be equipped
with a smartphone and running the same healthcare
Application, to check and interact with patient vital
signs. Since the system collects sensitive and confi-
dential data to ensure the adequate security level.
4 CONCLUSION
In this paper, we have exposed our methodology for
the diabetic patient in IoT healthcare. We discuss
the main process of our adaptive methodology. Our
methodology is mainly guided by knowledge infor-
mation carried by events and objects.
We emphasize describe the systems at multiple
levels and from a variety of perspectives. such as we
plan to validate our MobiDiabetic application through
controlled evaluations taking into account the clinical
environment, to assess its performance, reliability and
safety by the nursing staff.
SECRYPT 2023 - 20th International Conference on Security and Cryptography
848
Figure 2: Users interface visualization.
REFERENCES
(2018). A Novel Cardiac Auscultation Monitoring System
Based on Wireless Sensing for Healthcare. 6:1–12.
Al-Taee, M. A., Al-Nuaimy, W., Muhsin, Z. J., and Al-
Ataby, A. (2017). Robot Assistant in Management of
Diabetes in Children Based on the Internet of Things.
IEEE Internet of Things Journal, 4(2):437–445.
Alaiad, A. and Zhou, L. (2017). Patients’ Adoption
of WSN-Based Smart Home Healthcare Systems:
An Integrated Model of Facilitators and Barriers.
IEEE Transactions on Professional Communication,
60(1):4–23.
Chew, B.-H., Shariff-Ghazali, S., and Fernandez, A. (2014).
Psychological aspects of diabetes care: Effecting be-
havioral change in patients. World Journal of Dia-
betes, 5(6):796–808.
Ferrer, R. A. and Mendes, W. B. (2018). Emotion, health
decision making, and health behaviour. Psychology &
Health, 33(1):1–16.
Ida, I. B., Jemai, A., and Loukil, A. (2016). A survey on
security of IoT in the context of eHealth and clouds.
In 2016 11th International Design Test Symposium
(IDT), pages 25–30.
Islam, M. N., Raiyan, K. R., Mitra, S., Mannan, M., Tasnim,
T., Putul, A. O., and Mandol, A. B. (2023). Predictis:
an IoT and machine learning-based system to predict
risk level of cardio-vascular diseases. 23(1):171.
Khan, M. A. (2020). An IoT framework for heart dis-
ease prediction based on MDCNN classifier. 8:34717–
34727. Conference Name: IEEE Access.
March, S. T. and Smith, G. F. (1995). Design and natural
science research on information technology. Decision
Support Systems, 15(4):251–266.
Mekki, N., Hamdi, M., Aguili, T., and Kim, T.-H. (2017).
Scenario-based Vulnerability Analysis in IoT-based
Patient Monitoring System. pages 554–559.
Mekki, N., Hamdi, M., Aguili, T., and Kim, T.-H. (2018). A
Privacy-Preserving Scheme Using Chaos Theory for
Wireless Body Area Network. In 2018 14th Interna-
tional Wireless Communications Mobile Computing
Conference (IWCMC), pages 774–779.
Mekki, N., Hamdi, M., Aguili, T., and Kim, T.-H. (2023).
An authentication protocol for healthcare application:
A case study of a diabetic patient. pages 434–445.
Mukhopadhyay, B., Sharma, O., and Kar, S. (2018). IoT
Based Wearable Knitted Fabric Respiratory Monitor-
ing System. In 2018 IEEE SENSORS, pages 1–4.
Rghioui, A., Naja, A., Mauri, J. L., and Oumnad, A.
(2021). An IoT based diabetic patient monitor-
ing system using machine learning and node MCU.
1743(1):012035. Publisher: IOP Publishing.
Sallabi, F. and Shuaib, K. (2016). Internet of things network
management system architecture for smart healthcare.
pages 165–170.
Sowmiya, E., Malathi, L., and Selvi, A. T. (2016). A Study
on Security Issues in Healthcare Applications Using
Medical Wireless Sensor Network and Iot. Iioab Jour-
nal, 7(9):575–583. WOS:000397199100004.
A Secure Emergency Framework in an IoT Based Patient Monitoring System
849