Internet of Things with Artificial Intelligence for Health Care Security

Abstract

In recent years, health care facilities are moving towards technological advancements for precise patient monitoring and record management. Though it is technically advanced, the health care information and communication technology network's security is a significant challenge for health care. With the aid of standard algorithms, unstructured data existing outside organized databases (i.e., electronic documents and reports) is difficult to arrange and secure. The existing clustering method has a disadvantage of efficiency issues for recovering data transfer. This paper proposes the Internet of Things with Artificial Intelligence System (IoT-AIS) for health care security. Wireless sensor networks are developed by IoT technology. IoT network is used to bridge the physical and digital world. IoT-AIS is used to monitor the patient’s data and encrypt them. The encrypted data are stored in the cloud to maintain the patient data to access remotely. The IoT-AIS dashboard provides an individualized user interface for individual patients to maintain their records individually with single-user access. The proposed paper's simulation analysis proved that the Patient Record of health care could be encrypted and provide individualized access. The experimental results of IoT-AIS achieve the highest data transmission rate to 98.14% and the highest delivery rate of (98.90%), high period of standard responses (93.79%), less delay estimation (10.76%), improved throughput (98.23%), effective bandwidth monitoring (83.14%) energy usage (8.56%) and highest performance rate (98.4%) when compared to other methods.

Full text

Vol.:(0123456789)
1 3
Arabian Journal for Science and Engineering
https://doi.org/10.1007/s13369-021-06083-8
RESEARCH ARTICLE - SPECIAL ISSUE - FRONTIERS INPARALLEL PROGRAMMING
MODELS FORFOG ANDEDGE COMPUTING INFRASTRUCTURES
Internet ofThings withArtificial Intelligence forHealth Care Security
TaherM.Ghazal1,2
Received: 20 April 2021 / Accepted: 12 August 2021
© King Fahd University of Petroleum & Minerals 2021
Abstract
In recent years, health care facilities are moving towards technological advancements for precise patient monitoring and
record management. Though it is technically advanced, the health care information and communication technology net-
work's security is a significant challenge for health care. With the aid of standard algorithms, unstructured data existing
outside organized databases (i.e., electronic documents and reports) is difficult to arrange and secure. The existing clustering
method has a disadvantage of efficiency issues for recovering data transfer. This paper proposes the Internet of Things with
Artificial Intelligence System (IoT-AIS) for health care security. Wireless sensor networks are developed by IoT technology.
IoT network is used to bridge the physical and digital world. IoT-AIS is used to monitor the patient’s data and encrypt them.
The encrypted data are stored in the cloud to maintain the patient data to access remotely. The IoT-AIS dashboard provides
an individualized user interface for individual patients to maintain their records individually with single-user access. The
proposed paper's simulation analysis proved that the Patient Record of health care could be encrypted and provide indi-
vidualized access. The experimental results of IoT-AIS achieve the highest data transmission rate to 98.14% and the highest
delivery rate of (98.90%), high period of standard responses (93.79%), less delay estimation (10.76%), improved through-
put (98.23%), effective bandwidth monitoring (83.14%) energy usage (8.56%) and highest performance rate (98.4%) when
compared to other methods.
Keywords Technology· Wireless· Dashboard· Encrypted· Patient· Database· Artificial Intelligence
1 Overview ofIoT withArticial Intelligence
forHealth Care Security
The Internet of Things (IoT) is a creative framework that
uses portable devices, sensors, and the cloud to commu-
nicate with a large group of objects and systems without
human interaction [1]. The central point underlying IoT
highlights the relationship between the physical environment
and nature through the Internet [2]. In addition to meeting
everyday life requirements, IoT offers various technologies,
such as travel, farming, smart cities, emergency responders,
and infrastructure [3]. The health care industry for Artificial
intelligence applications is one of the most critical fields [4].
The integration of IoT and medical instruments contrib-
utes to promoting health care and clinical status reporting to
those requiring regular, in-house surveillance and protection
strategies [5]. IoT speeds up early identification and facili-
tates diagnosis and management, such as exercise services,
chronic illnesses, and aged health care [6].
Security in health is the knowledge that one's well-being
is stable; if not, then there are means to receive care to get
back to a healthy state. A basic level of protection against ill-
nesses and bad habits is the goal. Malnourishment and lack
of access to health care, clean water, and other daily essen-
tials make poor rural people more vulnerable to threats to
health. For guaranteeing the health of people, health security
comprises activities and procedures that transcend national
boundaries. It is a developing concept within the areas of
management affairs and strategic communications [7]. They
argue that all states have a responsibility to preserve the
health care and well-being of their citizens. According to its
* Taher M. Ghazal
Taher.ghazal@skylineuniversity.ac.ae
1 Center forCyber Security, Faculty ofInformation Science
andTechnology, Universiti Kebangsaan Malaysia (UKM),
43600Bangi, Selangor, Malaysia
2 School ofInformation Technology, Skyline University
College, University City Sharjah, 1797Sharjah, UAE

Arabian Journal for Science and Engineering
1 3
critics, health security has a detrimental effect on civil free-
doms and the equitable allocation of expenditures. Health
care equipment and facilities must deal with essential patient
details, including specific data on health care applications,
and alter or interact with the protocols adopted. In particular,
such intelligent systems can be integrated into global com-
munication technologies to link individuals anywhere, such
that hackers can target the Web service field [8]. To promote
the complete Internet acceptance of health care technology,
IoT's distinguishing characteristics are efficient [9]. In the
IoT platform, safety criteria and limitations, health care
hazard templates, and detection systems must be defined
and analyzed [10]. Network protection is essential for ensur-
ing access to security and privacy risks following the vast
growth of channels and telecommunications over the recent
century, combined with advanced accessibility and the out-
put of cyber threats [11]. Of course, these attempts nega-
tively impact the channel's efficiency, where the defects and
system interruption causes distress and lag [12].
A fundamental objective of any health care service is the
expense of connectivity facilities via smooth and safeties
between clients, clinics, and specific health care systems
[13]. Chronic diseases, early treatment, actual surveillance,
and critical services are supposed to be assisted by the most
recent wireless connection in health care [14]. The techno-
logical developers, information portals play a significant role
in developing health information to provide approved part-
ners with health care systems on-demand [15]. Services of
computers are linked, exchanged, and measured substantial
data on IoT-based health care organizations.
In the health care sector, IoT applications are vulnerable
to numerous cybercrimes [16]. The health sector is more
likely to experience security challenges than any other sector
and more vulnerable to identity theft [17]. IoT innovations
help detect the automated system and follow-up the clinical
surveillance that helps provide good health care [18]. IoT
leads to cost savings and even decreases the staff's needs
in most hospitals, leading to major problems [19]. Security
and privacy of users are some of the main problems for such
channels. Data protection is of significant significance in
the health care system since the data on such applications
are often the patients' confidential information. Therefore,
IoT is crucial that security risks are mitigated, and large,
stable nets built, taking the financial barriers into account.
A structured evidence template for the confidentiality of
the communication verifies security power. In addition to
precise medical treatment and record-keeping, health care
providers are progressing towards technical change. While
it is technologically sophisticated, the protection of health
information and communication technology network is a
significant health care challenge. The standard algorithms
are difficult to arrange and protect unstructured data from
outside organized databases. The main contribution of IoT-
AIS is described below.
IoT-AIS presents the Internet of Things for the Protection
of Health Care technology in wireless sensor networks. IoT-
AIS is used in the physical and digital world used to bridge,
track and encrypt patient data. The IoT-AIS dashboard offers
a customized user interface to hold records independently
with single-user access.
This paper proposes the Internet of Things with Artifi-
cial Intelligence System (IoT-AIS) for health care security.
Wireless sensor networks are developed by IoT technology.
IoT network is used to bridge the physical and digital world.
IoT-AIS is used to monitor the patient’s data and encrypt
them. The encrypted data are stored in the cloud to maintain
the patient data to access remotely. The IoT-AIS dashboard
provides an individualized user interface for individual
patients to maintain their records individually with single-
user access. The proposed paper's simulation analysis proved
that the Patient Record of health care could be encrypted and
provide individualized access.
The main contributions of the research section are:
• The research develops a system of Internet of Things
with Artificial Intelligence System (IoT-AIS) for the
health care unit.
• IoT network is used to monitor physical and numerical
statistics.
• The encrypted data are deposited in the cloud system to
access patient details distantly.
The remaining article is organized as follows: Sect.2
comprises various background studies concerning the IoT
for health care security. Section3 elaborates the proposed
IoT-AIS model to monitor the patient’s data and encrypt the
information available. Section4 constitutes the results that
validate the performance and predictability with the cor-
responding descriptions. Finally, the conclusion with future
perspectives is discussed in Sect.5.
2 Background Study onIoT forHealth Care
Security
This section discusses several works that various research-
ers have carried out; ÁineMacDermott etal. [20] developed
Securing Things in the Health Care Internet of Things (ST-
HIoT). Based on data from IoT issues, real-time tracking
offers a broader summary of patient treatment, individual
behaviors, and routines. The advantages of implementing
ST-HIoT into health care are apparent; networks and device's
fundamental security weaknesses cannot go unnoticed. ST-
HIoT is set to have a significant effect on society, and with

Arabian Journal for Science and Engineering
1 3
hackers manipulating IoT by various means, the IoT is una-
voidable as the most susceptible region for cybersecurity.
Lanjing Wang etal. [21] proposed Identified Security
Attributes (ISA) framework. ISA introduces a proposal
to test the Internet of Health Things-based devices' safety
features in the health care environment through a Defined
Protection Attribute system. The suggested system employs
hybrid approaches, including the Analytical Hierarchical
Process (AHP). ISA is a two-step framework: the parame-
ters' measurements are extracted by the AHP method during
the first stage. In contrast, the second phase, safety assess-
ments of alternatives by AHP, are carried out regarding
security parameters.
PriyanMalarvizhi Kumar etal. [22] discussed Intelligent
face recognition and navigation system (IFR-NS). IFR-NS
presents a new computer vision and orientation framework
that offers reliable and fast voice signals for visual impair-
ments to communicate directly. Neural learning methods
with mapping techniques and testing components can be
used for face recognition. Relatives and parents' photographs
are saved on the user's mobile phone web page. If an indi-
vidual comes before the blind, the application offers voice
support to the server with the computer program's help.
Relatively, using the Region of Curve study, the output of
the proposed approach is evaluated.
GeethapriyaThamilarasu etal. [23] introduced An Intru-
sion Detection System (IDS). IDS introduces a new threat
detection method focused on mobile nodes to protect inter-
connected health equipment infrastructure. The proposed
framework is specifically centralized, adaptive, and uses
computer vision, regressive techniques to identify interfer-
ences at the device level and sensor data abnormalities. IDS
models the configuration of a medical group and conducts
detailed tests for different Internet sites of medical prod-
ucts, namely wireless communication networks and other
interconnected medical equipment. The simulation results
indicate that IDS can obtain significant tracking precision
with minimal additional resources.
HamedHaddadPajouh etal. [24] proposed an Artificial
Intelligence-powered secure architecture for the IoT's edge
layer (AI4SAFE-IoT). The proposed secure IoT infrastruc-
ture, Artificial Intelligence-powered secure architecture
for the edge layer of the Internet of things Architecture, is
designed into edge layer AI protection modules. Cyber threat
assignment, cyber threat hunting, smart firewall, and Internet
threat intelligence are the main modules suggested by the
architecture. An attack life cycle stage based upon the cyber
kill chain model can be identified, categorized, and further
recognized in the proposed modules. The highest peer IoT
level safety frameworks ranking for the suggested scheme
was 84.7 out of 100.
PriyanMalarvizhi Kumar et al. [25] narrated
CoAP (ChangeCipherSpec and Alert Protocol)-based
authentication scheme. CoAP is developed for an intelli-
gent portal authorized system to avoid and secure the more
important clinical signals from attackers and malicious
behavior to solve the authentication problem. To determine
the efficiency of the improved DTLS, data transfer and inter-
action time are often measured.
Based on the survey, IoT-AIS ensures health care safety in
IoT technology and develops wireless sensor networks. The
IoT-AIS dashboard offers a personalized user experience to
keep the records separately with fixed authentication.
3 The Proposed Internet ofThings
withArticial Intelligence System (IoT‑AIS)
This paper discussed the IoT-based secure patient data trans-
mission and receiving using artificial intelligence. IoT pro-
vides a practical framework for safeguarding information
security, confidentiality, and reliability on the Internet. The
protection, confidentiality, and reliability of medical data
are preserved in different health applications. However, IoT
offers efficient protocols for managing information, attack-
ing, and accessing health information, reducing the whole
Internet health care system's privacy, safety, and reliability.
This research incorporates deep learning to minimize mal-
ware attacks when handling health information and thus fix
these issues. This approach explores medical knowledge in
various layers by the AI concept, which reduces the inter-
mediate attacks to the minimum.
Figure1 demonstrates DNN-based Malware Detection.
Initially, the IoT system is examined using a deep neural
network that analyzes the user's authentication to remove the
unwanted access and attacks in the IoT device. Each request
traffic feature is removed from the database's request to ana-
lyze the malware activity after the authentication process.
The quality value is examined from the extracted features
using the characteristic status and related behavior to assess
information qualitatively. The critical element is to preserve
the protection of the IoT-based DNN method for medical
data transactions.
The protocol analyzes incoming traffic requests for medi-
cal data, which are checked with the above authentication
procedure. Following the authentication mechanism check-
ing, an examination of the process is carried out concern-
ing the IP address request, protocol transmission, file type
sending, frame length, frame number, host post number.
These traffic features eliminate the canal pulse's response,
the signal received, the channel state's details, and the sig-
nal received from the query. Databases are trained by the
specified AI method in networking to detect malware attacks
during IoT health records.
Malware detection details are displayed in IoT health data
in Fig.1. Fig.1 shows that the IoT-Health Data Detection

Arabian Journal for Science and Engineering
1 3
Structure DNN-dependent security analysis helps preserve
IoT health information effectively classifications affected
malware data. The requested traffic data are collected in
response to the abovementioned discussions; listing charac-
teristics are obtained using the IoT-AIS-based data analysis
from the security, privacy, and reliability analysis requests
stored in the database.
Figure2 illustrates the proposed IoT-AIS model-based
data security. Information technology is used to retrieve,
store, restore, analyze, and manage information and unify
electronic health records to disseminate medical infor-
mation. Distant surgeries and health care are often made
possible with the use of smart e-health care tokens and
other innovations. The goal is to improve contact with
patients and improve health care services and the overall
health care industry. Health care services and informa-
tion are being digitized. E-health ensures that clients and
physicians will accomplish everything digitally, ultimately
leading to the lots of paperwork, such as documents and
reports, that take up a big portion of medical centers.
Health care will reap more rewards due to this, as it is
under constant pressure to deliver health care services,
continue to use them, and improve them. E-health systems
are the key to achieving this. E-health data security refers
to data protection measures against unwanted access and
data manipulation during its life cycle. Data protection
e-health requires data encryption, hacking, tokenization,
and key data management activities that secure all apps
and platforms. Medical data are commonly maintained in
electronic health records where patients' ages, symptoms,
procedures, and progress are systemically monitoring. Sur-
vey analysis permits researchers in a comparatively lim-
ited period to gather analytical evidence. According to the
design and scope of surveys, data can be collected from a
representative sample of people, primarily if samples are
used randomized or intentionally unlikely.
All the real, behavioral, and demographic data obtained
from its customers by marketing firms or departments are
referred to in customer data or consumer data. Epidemiol-
ogy is, by concept, the research in particular populations
on the distribution and determinants of health conditions
and events. The information systems in hospitals provide
a shared source of health history information for patients.
The device must hold data in a safe location and controls
that can, in some instances, access the data. Pharmaceutical
data are an essential aspect of the clinical data, providing the
correct drug for the right patient, used at the right dosage;
for direct patient treatment health data are a whole set of
Fig. 1 DNN-based malware detection

Arabian Journal for Science and Engineering
1 3
data for an individual or society related to health conditions,
reproduction outcomes, causes of death, and quality of life.
Health data provided clinical indicators and knowledge
related to health and well-being on the environment, socio-
economic, and behavior. Social media provides health care
with opportunities to exchange knowledge, address health
policy problems and practice, encourage health behavior,
communicate with the public, and inform and connect with
patients, carers, students, and colleagues. Laboratory data
are when a physician takes blood, urine, other bodily fluid,
or body tissue sample for patient health information.
Another type of electronic record is claimed records,
often referred to as administrative data. Claims databases
compile information on millions of visits, bills, insurance
records, and other correspondence from patient providers.
It is a medical system community that enables IoT-AIS to
store, display, update and share its health information. IoT-
AIS center provides protected patient e-Health records for
treating the disorder, testing and other uses, and safe storage
and administration. Implement a centralized framework to
store and update patients' health data within the e-health care
system and track their data completely. Because patient's
health records were stored in the cloud or other parties,
there were large privacy concerns since third-party servers
or unauthorized users might utilize patients' private health
data. To protect patient privacy and improve protection, it is
strongly recommended that patient data be encrypted before
sourcing.
A single repository needs to collect health information
from multiple sources and ensure interoperability with vari-
ous stakeholders; aggregators can use different standards
and protocols. Transit data encryption is the data encryp-
tion, network transmission, and cloud decoding process.
It may have been a vital method since unauthorized eyes
could access the path data, which created problems in data
integrity. Transport layer security (TLS) is used to secure
the interaction between web applications. TLS reserves an
encrypted channel to send the cypher and transmit the key
via a public encrypted file to negotiate between senders and
recipients.
Figure3 shows the sequence diagram for data transmis-
sion and receive. The article does not mention the use of
a group node. Ultrasonic nodes and group nodes authen-
tication failed to implement the IoT-AIS system. Nodes of
ultrasonic are shuffled in multiple radiations to authenticate
security. Group nodes make defined leveling sequence, yet
combining ultrasonic and group cannot authenticate. It
can accelerate data individually without the integration of
channels. The IoT-AIS will send all the data to the server
if sensor nodes directly send it to the Base Station. This
Fig. 2 Proposed IoT-AIS model-
based data security

Arabian Journal for Science and Engineering
1 3
produces more channels, because the base station is con-
nected to every node and attacked by it. In addition, due to
the distance from the sensor nodes to the base station, more
energy is used, and the nodes' lifetime is reduced. The key
authentication process extends; eight sensor nodes are used
by the proposed one; eight authenticate the key for use by
the server. This is addressed in our proposal by group nodes
as interface and base stations (ultra-sensor). These sensor
nodes send the collected data to the group node and the
node to the server. The distance between sensor nodes and
community nodes will then decrease. This decreases energy
usage and increases the node's lifespan.
On the other hand, this improves protection since long
distances do not exist and decrease an attack's chances.
Furthermore, it will minimize the key authentication pro-
cess since it authenticates the group node to obtain the
server key. Our idea is to create and share the group node
through crucial agreements with other ultra-sensor nodes.
The Group node key shares all ultra-sensor nodes the same
key agreement, and not each node needs to communicate
individually to get the authentication key. They have a node
group and eight nodes that scatter across the patient's body
to capture general vital signs, including temperature, blood
pressure, pulsation, and respiration. Collected information
is sent to the group node and the group node to the server
by the sensor nodes. Therefore, the group node would have
limited mobility and a constant, wearable sense node and
patient movement. Suppose the patient does not have the
Fig. 3 Sequence diagram for data transmission and receive

Arabian Journal for Science and Engineering
1 3
status of mobility and body coordinates, then eight sensor
nodes pass their values through the server to the group node
and the base station. Eight sensors have been set in differ-
ent coordination around the patient's body in patients with
wheelchair mobility and are sent into the base station and
server group node.
Until transacting from one place to another, the first step
in the IoT medical health data transmission involves authen-
tication. Due to the importance of confidential medical data,
the established authentication assesses IoT networks and pre-
vents intermedia attacks and unauthorized access. Restricted
memory resources, batteries, and computations have been
built in the IoT devices used to build Sybil attacks in the
network. The physical layer uses different characteristics,
such as channel pulses, signal strength indicators, channel
status information, signal strength, and data privacy. How-
ever, this network functionality provides efficient security
since creating IoT devices is resource-based and leads to
less protection when transmitting health data. In this paper,
deep learning neural networks (DNN) are implemented to
maintain authentication, reducing data leakage, because they
efficiently learn IoT features. Before using this system, the
DNN system is implemented on the IoT device for medical
data transactions. The IoT system must initially check its
control range under the evaluation. A particular collection
of authentication requests from the IoT system must be sub-
mitted to the IoT testing area due to the privacy verifica-
tion of the health data transaction. Different signal features,
including channel pulse response standard, battery status
indicator, channel recorded data, and transmission power,
are disabled when the authentication is needed. Depending
on the extracted functionality, the parcel request and envi-
ronmental radio signals are analyzed via DNN. First, the
functionality extracted is trained to achieve the successful
outcome of IoT-AIS authentication. It effectively trains the
function even if there are few errors or noise in the extracted
features. The training phase is carried out by the IoT system
feature in Eq.(1):
Figure4 shows the training feature of the IoT-AIS system
that includes the pooling layer, feature extraction, training
feature, and high recognition rate. Once the pooling layer
parameter connects with the summation process, it adds to
the training feature. The extraction feature filters the coor-
dinate data and merges it to the summation value. These
data are loaded into a training feature for recognition ideas.
Therefore, the system classifies into three different methods
as classification 1, classification 2, and classification L state.
Figure4 and Eq.1 show the Training Feature.
𝜎i
are the
characteristics of the layer of pooling,
gi
is a better-extracted
feature. For authentication, training features are stored in the
database.
i
is the neuron. The corresponding extracted signal
functions are processed via a deep learning network, con-
sisting of three layers: input, hidden, and output layer when
the new authentication request enters the IoT system. These
determined layers use the basic weights and biases in the
measurement of the authentication-related result estimated
in Eq.(2) as follows:
Figure5 and Eq.(2) deliberate the neural network output.
Yj
is the authentication process input
Zj
signifies the specific
weights
a
is the bias value the network is further trained in
a deep learning system that uses the weights and bias value
defined as updating weights and the authentication perfor-
mance estimation process in Eq.(3):
As initialized in Eq.(3) authentication performance
estimation process has been evaluated.
YL+1
denotes the
authentication performance with layer,
YL−
expresses
(1)
E
(y)=
i
∑
i=1
𝜎igi(y
)
(2)
Netoutput
=
M
∑
j=1
Yj∗Zj+
a
(3)
YL+1
=Y
L
−
[
ITI−𝜇J
]−1
ITf
Fig. 4 Training feature

Arabian Journal for Science and Engineering
1 3
the previous authentication with layer,
I
is explored the
authentication request,
f
demonstrated the training func-
tion,
T
is a time,
IT
denotes the authentication request with
time transition,
𝜇
is updating weights,
J
denotes incom-
ing authentication. Authentication on the authentication
request is based on the above method to see if IoT device
is authenticated, contrasted with the training feature. This
authentication process is performed at a particular time
to ensure that the transmission of health data is fast. This
method of authentication removes intermediate attacks
when IoT is used to transmit health information. An IoT
access control is identified in the health data transaction
and approved users effectively reach the IoT system using
the authentication procedure. The health data transac-
tion protection is further investigated using the IoT-AIS
method based on learning after examining an IoT system
authentication. To assess the data security, privacy, and
durability of the data, the IoT-AIS approach implements
the functionality suggested in the above discussions and
the queries in a database. The medical transit informa-
tion requested is retrieved. IoT-AIS is a powerful training
tool that does not require an appropriate model to identify
health information that is safe and malware identified. The
Network uses
Q
values or quality functions at the time of
this detection process to decide each state's action for the
successful choice of particular data.
According to clustering, data or things with similar quali-
ties are grouped into homogeneous subgroups, while vari-
ability is maximized within each cluster. Consequently, data
or objects of interest are clustered together based on features
that make them related, with the ultimate goal of separat-
ing them from other groupings. Cluster arrangements that
are generally homogenous within each grouping, resulting
in high intra-class similarity, are sought, while diversity
between groups is maximized, resulting in low inter-class
resemblance.
Furthermore, IoT-AIS often includes the set of
T
states;
every state belongs to a particular
B
action, for which each
action gives a special reward. In addition to the state, behav-
ior and networks have special weights to determine the price
and value of discounts. The reduction factor calculated is
between 1 and 0. Then, every state's quality is calculated
accordingly (4):
As found in Eq.(4), every state's quality has been
obtained. The
Q
value is defined as a fixed value chosen by
the procedure before the method
P
is every state's quality.
The new value is defined with the aid of the arbitrary value
by action
bj
state
Ts+1
, at a time
s
, that provides the reward
value
qs
. The value changes the current weighted average
value in Eq.(5):
In Eq.(5) current weighted average value has been evalu-
ated.
𝜎
is a learning rate,
qs
expresses the reward value,
𝛽
denotes the discount factor,
𝛽
.max
b
P
(
T
s+1
,b
)
explores the
optimal upcoming value,
(
q
s
+𝛽.max
b
P
(
T
s+1
,b
))
described
the quality of learned value.
P(
T
s
,b
s)
is interpreted by each state as an old value; the
process is continuously repeated until the quality values
of each state and related activities are calculated, and the
Te
,P
(
T
s
,b
)
is final, not modified, the rewards
q
and observed
state
Te
and
P(
T
s
,b
)
is considered to be 0. The characteristics
state is checked, and data protection is checked effectively
using qualitative metrics. In addition, the process of malware
sensing is calculated by an extensive neural learning network
that effectively classifies safe and malware-detected health
data. For revised purpose using the weight and bias value
, the classification process is further optimized in Eq.(6):
As obtained in Eq.(6), classification process optimization
has been determined.
E(y)
is the trained function classifica-
tion process.
(
E
1
(y),E
2
(y),…E
L
(y)
)S
denotes the all-state
trained function classification process with layer and overall
time. The neural network's weight and bias value are modi-
fied based on that protocol to its previous value. In addition,
the sigmoid function is used to train the extracted features
to detect malware and highly recognized IoT health informa-
tion. In the interests of security, safety, and data reliability,
this method is constantly repeated.
The average response time is the time the Edge Server
will send to the patients the information processed. The
data rate, processing, and communications speed, number,
and types of jobs submitted are factors that deteriorate the
(4)
P
∶T∗B→Q+
[
ITI−𝜇J
]−1
ITf
(5)
Pnew (
T
s
,b
s)
←(1−𝜎).P
(
T
s
,b
s)
+𝜎
(
q
s
+𝛽.max
b
P
(
T
s+1
,b
))
(6)
E
(y)=
(
E
1
(y),E
2
(y),…E
L
(y)
)S
Fig. 5 Neural network output

Arabian Journal for Science and Engineering
1 3
answer time. The delivery percentage of packets (DPR) for
packets is based on the number of packets sent, and the num-
ber of packets received successfully. The ratio of sent pack-
ets to the packet received is defined as calculated in Eq.(7):
As discussed in Eq.(7), the delivery percentage of pack-
ets has been determined, where
Tj
is the number of packets
sent and
Qj
is the number of packets received. The delay
𝛿
is
the total time necessary to receive a packet successfully at
the destination. In contrast, the average delay is the number
of all delay samples, as measured in Eq.(8).
As shown in Eq.(8), a delay function has been found,
where
𝜏
is when a packet is being transmitted, and
𝜇
the
time packet reached its destination successfully. The packet
transmission is traversed in the form of a bucket list to lev-
eled destination. Thus, the packet with less delay confirms
to perform more data transfer at a time.
F(𝛿)
) is given the
average delay in Eq.(9):
As described in Eq.(9), the average delay has been calcu-
lated. The network output is typically calculated per second
in a bit (bps) or per second in packets (PPS). Network output
is the sum of the data rates given to all network nodes. As in
Eq, it is measured [4].
As deliberated in Eq.(10), throughput has been com-
puted.
𝜌
is throughput. Total Control Overhead is expressed.
j
is the IoT node
M
number of messages. The total number
of controlling messages created by each network node is
calculated according to the total number of packets received
successfully. The first of these is in Eq.(11) as follows:
As initialized in Eq.(11), Total Control Overhead has
been derived, where
Dj
is the number of control messages.
The proposed IoT-AIS method provides reliable data trans-
mission and data security which are achieved based on the
five metrics, such as high period of standard responses,
enhanced delivery percentage of packets, less delay
(7)
DeliveryPercentageofPackets
=
∑
M
j=1Tj
∑
M
j=1
Qj
×
100
(8)
𝛿=𝜏−𝜇
(9)
F
(𝛿)=
∑
M
j=1𝛿j
M
+M−T
j
(10)
𝜌
=
∑
M
j=1Qj
∑
M
j=1
Tj
−E1(y)
S+1+
M
(11)
𝜐
=
∑
M
j=1Dj
∑
M
j=1
Pj
+𝜌
𝜏−
𝜇
estimation, improved throughput, and effective bandwidth
monitoring.
The proposed system of IoT-AIS is created to conse-
quence data transfer with the protection of less traffic for-
mation. Therefore, the development system creates a pro-
cessed formation of IoT sensor networks. The encryption
and decryption of patient data are secured into a summing
unit for admin access. Thus, the output access of patient
security data to individual access is guaranteed.
4 Results andDiscussion
The proposed IoT-AIS method provides reliable data security
and receives data transmission securely which is evaluated
based on the five metrics (1) Period of Standard Responses,
(2) Delivery Percentage of Packets, (3) Delay estimation, (4)
throughput, (5) Bandwidth Monitoring.
4.1 Period ofStandard Responses
The estimated response duration is when the Edge Processor
can transfer the data and transmit it back to clinicians. Data
transmission rate and interaction rate, amount of jobs, and
working experience presented are all factors which affect
the reaction time. The device's database upload/download
period for clinicians' data generation calculates the period of
standard responses. The remaining period is the time needed
for a job, and the duration to upload/download is the period
to register. Likewise, the processing period is much less than
the Central server for work on the Network devices. The
period of standard responses is shown in Fig.6.
Fig. 6 The period of standard responses

Arabian Journal for Science and Engineering
1 3
4.2 Delivery Percentage ofPackets
The delivery percentage of packets is determined by the
number of packages delivered and the number of data pack-
ets obtained effectively. The percentage between sent and
obtained packets is calculated as the delivery percentage of
packets. The first possibilities represent packets' delivery in
health care services with a conventional network, while the
second or third illustrates the devices' data transfer. When
the amount of data transmission people rises, the delivery
percentage of packets increases. The delivery rate of packets
is shown in Fig.7.
4.3 Delay Estimation
The delay estimation is the maximum time needed to get the
packet to its destination with completion, and the data rate
amounts to a maximum of all defects separated by several
delays are measured. The machine delay toward the percent-
age of participants is seen in these statistics. The late deliv-
ery, contact, uploading/downloading of patient data is worth
notice. The delay is slight, and the IoT devices are high in
processing. The processing on devices works together to
support network management, load balance, and practical
resource usage. The delay estimation is shown in Fig.8.
4.4 Throughput
The system output is typically calculated in bytes. Network
output reflects the number of transfer rates transmitted to
all network devices. Edge cooperation, and effective use
of Internet services, IoT devices have a higher throughput
based on smart load balance decision, Edge cooperation.
The Region detection network provides higher efficiency
than conventional networks, although large amounts of data
cannot be processed efficiently on standard IoT connected
devices. The throughput of IoT-AIS is shown in Fig.9.
4.5 Bandwidth Monitoring
Bandwidth monitoring represents each device's ratio in
the total range of performance communications produced
from the packet's output. The conventional system situa-
tion reduces overlap control because of few control packets.
Simultaneously, the Edge nodes share additional control
packets for node coordination and upload/download data,
leading to greater overlap management. The bandwidth
monitoring of IoT-AIS is shown in Fig.10.
Fig. 7 The delivery rate of packets
Fig. 8 The delay estimation of IoT-AIS
Fig. 9 The throughput rate of IoT-AIS

Arabian Journal for Science and Engineering
1 3
In line with the mathematical expression, access and
transmission of data through IoT-health are accomplished
using minimal network resources and effectively removing
the intermediate malware attack. Despite the appropriate
security of these networks, verification is less important
during health data transmission than IoT devices' energy
development. The transmission rate of IoT-AIS is shown in
Table1.
The network consistency is measured using IoT, system
life, efficiency, threat identification performance, and threat
error rate identification. There are many datasets taken for
predicting performance on affecting the dataset size. This
section imports the actions for evaluating the presentation
part by increasing the output level. Whenever the dataset
processing develops, the significance of measuring results
will be influenced. The devices support a range of benefits
for IoT implementations for effective interaction between IoT
nodes, IoT applications in-vehicle networks, detector, and
sensor networks. The energy usage between the IoT devices
is evaluated based on the delay rate. The energy usage of
IoT-AIS is shown in Table2.
The proposed IoT-AIS method provides reliable data
transmission and data security which is achieved based on
the five metrics, such as high period of standard responses,
enhanced delivery percentage of packets, less delay estima-
tion, improved throughput, effective bandwidth monitoring
when compared to other existing intrusion Detection Sys-
tems (IDS), Intelligent face recognition and navigation sys-
tem (IFR-NS), Securing Things in the Health Care Internet
of Things (ST-HIoT).
Thus, the approach of IoT-related patient data is trave-
led with security based on fundamental metrics. The data
on period standard responses, delivery percentage packets,
delay estimation, throughput, and bandwidth rate are identi-
fied. The tabulations over transmission rate and energy usage
are significantly created for requiring outputs.
5 Conclusion andFuture Works
This paper presents IoT-AIS for health care protection of
data in the IoT platform. IoT technology is used to develop
wireless sensor networks. The physical and digital worlds
are linked with the IoT network. In an attempt to track and
encrypt patient data, IoT-AIS is used. Encrypted data are
recorded and stored for remote sharing of patient data.
Besides individual patients keeping their records separately
with a single access, the IoT-AIS dashboard offers a cus-
tomized user interface. The mathematical expression proved
that the health care medical record could be encrypted, and
individual access can be provided. Health care providers are
moving to technological development for accurate patient
tracking and registration. The health care system's future
aspects enable mobile application development for users of
every standard people. The experimental results of IoT-AIS
achieve the highest data transmission rate to 98.14% and
Fig. 10 The bandwidth rate of IoT-AIS
Table 1 The transmission rate
of IoT-AIS Number of
devices
Transmis-
sion rate
(%)
10 86.34
20 87.02
30 88.13
40 89.33
50 90.27
60 91.45
70 92.18
80 93.67
90 94.56
100 95.11
Table 2 The energy usage of
IoT-AIS Number of
devices
Energy usage (%)
10 22.11
20 21.56
30 20.32
40 15.78
50 11.52
60 10.89
70 9.45
80 9.02
90 8.44
100 8.56

Arabian Journal for Science and Engineering
1 3
the highest delivery rate of (98.90%), high period of stand-
ard responses (93.79%), less delay estimation (10.76%),
improved throughput (98.23%), effective bandwidth moni-
toring (83.14%) energy usage (8.56%), and highest perfor-
mance rate (98.4%) when compared to other methods.
References
1. Thota, C.; Sundarasekar, R.; Manogaran, G.; Varatharajan, R.;
Priyan, M.K.: Centralized fog computing security platform for
IoT and cloud in the healthcare system. In Fog Computing: Break-
throughs in Research and Practice. IGI global, pp. 365–378 (2018)
2. Amin, R.; Kumar, N.; Biswas, G.P.; Iqbal, R.; Chang, V.: A light-
weight authentication protocol for IoT-enabled devices in distrib-
uted Cloud Computing environment. Future Gen. Comput. Syst.
78, 1005–1019 (2018)
3. Amudha, G.; Jayasri, T.; Saipriya, K.; Shivani, A.; Praneetha,
C.H.: Behavioural Based Online Comment Spammers in Social
Media
4. Muthu, B.; Sivaparthipan, C.B.; Manogaran, G.; Sundarasekar, R.;
Kadry, S.; Shanthini, A.; Dasel, A.:. IoT-based wearable sensor for
diseases prediction and symptom analysis in the healthcare sector.
Peer-to-peer networking and applications, 1–12 (2020)
5. Gao, J.; Wang, H.; Shen, H.: Smartly handling renewable energy
instability in supporting a cloud datacenter. In: 2020 IEEE Inter-
national Parallel and Distributed Processing Symposium (IPDPS).
IEEE, pp. 769–778 (2020)
6. Ogudo, K.A.; Muwawa Jean Nestor, D.; Ibrahim Khalaf, O.; Daei-
Kasmaei, H.: A device performance and data analytics concept for
smartphones’ IoT services and machine-type communication in
cellular networks. Symmetry 11(4), 593 (2019)
7. Liu, B.H.; Nguyen, N.T.: An efficient method for sweep coverage
with the minimum mobile sensor. In: 2014 Tenth International
Conference on Intelligent Information Hiding and Multimedia
Signal Processing. IEEE, pp. 289–292 (2014)
8. Gheisari, M.; Najafabadi, H.E.; Alzubi, J.A.; Gao, J.; Wang, G.;
Abbasi, A.A.; Castiglione, A.: OBPP: an ontology-based frame-
work for privacy-preserving in IoT-based smart city. Future Gen.
Comput. Syst. 123, 1–13 (2021)
9. Lakshmanaprabu, S.K.; Shankar, K.; Ilayaraja, M.; Nasir, A.W.;
Vijayakumar, V.; Chilamkurti, N.: Random forest for big data
classification in the Internet of things using optimal features. Int.
J. Mach. Learn. Cybern. 10(10), 2609–2618 (2019)
10. Amudha, G.; Narayanasamy, P.: Distributed location and trust-
based replica detection in wireless sensor networks. Wirel. Pers.
Commun. 102(4), 3303–3321 (2018)
11. Kimani, K.; Oduol, V.; Langat, K.: Cybersecurity challenges for
IoT-based smart grid networks. Int. J. Crit. Infrastruct. Prot. 25,
36–49 (2019)
12. Nguyen, V.C.; Kostarakis, P.: The impact of green systems and
signals on the health of green residences’ habitants. Ann. Gen.
Psychiatry 17(1), A12 (2018)
13. Yang, P.; Yang, Y.; Wang, Y.; Gao, J.; Sui, N.; Chi, X.; Zou, L.;
Zhang, H.Z.: Spontaneous emission of semiconductor quantum
dots in inverse opal SiO2 photonic crystals at different tempera-
tures. Luminescence 31(1), 4–7 (2016)
14. Alabdulatif, A.; Khalil, I.; Yi, X.; Guizani, M.: Secure edge of
things for smart healthcare surveillance framework. IEEE Access
7, 31010–31021 (2019)
15. Janarthanan, R.; Doss, S.; Baskar, S.: Optimized unsupervised
Deep learning assisted reconstructed coder in the on-nodule
wearable sensor for Human Activity Recognition. Measurement
(2020). https:// doi. org/ 10. 1016/j. measu rement. 2020. 108050
16. Tao, H.; Bhuiyan, M.Z.A.; Rahman, M.A.; Wang, G.; Wang, T.;
Ahmed, M.M.; Li, J.: Economic perspective analysis of protect-
ing big data security and privacy. Future Gen. Comput. Syst. 98,
660–671 (2019)
17. Liu, B.H.; Pham, V.T.; Nguyen, N.T.: A virtual backbone con-
struction heuristic for maximizing the lifetime of dual-radio
wireless sensor networks. In: 2015 International Conference on
Intelligent Information Hiding and Multimedia Signal Processing
(IIH-MSP). IEEE, pp. 64–67 (2015)
18. Abou-Nassar, E.M.; Iliyasu, A.M.; El-Kafrawy, P.M.; Song, O.Y.;
Bashir, A.K.; Abd El-Latif, A.A.: DITrust chain: towards block-
chain-based trust models for sustainable healthcare IoT systems.
IEEE Access 8, 111223–111238 (2020)
19. Kuthadi, V.M.; Selvaraj, R.; Baskar, S.; Shakeel, P.M.; Ranjan,
A.: Optimized energy management model on data distributing
framework of wireless sensor network in IoT system. Wirel. Pers.
Commun. (2021). https:// doi. org/ 10. 1007/ s11277- 021- 08583-0
20. MacDermott, Á.; Kendrick, P.; Idowu, I.; Ashall, M.; Shi, Q.:
Securing things in the healthcare internet of things. In 2019
Global IoT Summit (GIoTS). IEEE, pp. 1–6 (2019)
21. Wang, L.; Ali, Y.; Nazir, S.; Niazi, M.: ISA evaluation framework
for security of Internet of health things system using AHP-TOP-
SIS methods. IEEE Access 8, 152316–152332 (2020)
22. Kumar, P.M.; Gandhi, U.; Varatharajan, R.; Manogaran, G.; Jid-
hesh, R.; Vadivel, T.: Intelligent face recognition and navigation
system using neural learning for smart security in the Internet of
Things. Clust. Comput. 22(4), 7733–7744 (2019)
23. Thamilarasu, G.; Odesile, A.; Hoang, A.: An intrusion detec-
tion system for the Internet of medical things. IEEE Access 8,
181560–181576 (2020)
24. HaddadPajouh, H.; Khayami, R.; Dehghantanha, A.; Choo,
K.K.R.; Parizi, R.M.: AI4SAFE-IoT: an AI-powered secure archi-
tecture for edge layer of the Internet of things. Neural Comput.
Appl. 32(20), 16119–16133 (2020)
25. Kumar, P.M.; Gandhi, U.D.: Enhanced DTLS with CoAP-based
authentication scheme for the Internet of things in healthcare
application. J. Supercomput. 76(6), 3963–3983 (2020)