A Robust Data Security Model Development to

Maintain the Integrity of Information using

Artificial Intelligence and Cyber Laws
G.Muthu Kumar, D. Hemanand,
Department of Artificial Intelligence and Data Science, Department of Computer Science and Engineering,
S.A. Engineering College(Autonomous), S.A. Engineering College (Autonomous),
Chennai, Tamil Nadu, India Chennai, Tamil Nadu, India
[email protected] [email protected]

Abstract—The field of artificial intelligence (AI) has seen the capabilities of artificial intelligence (AI) both today as
significant advancements in recent years. These days, artificial well as in the near future [2, 3].
intelligence (AI) tools are being utilized by organizations in
both the public and commercial sectors all over the world. On the other hand, these very same technical
2024 International Conference on Intelligent Systems for Cybersecurity (ISCS) | 979-8-3503-7523-7/24/$31.00 ©2024 IEEE | DOI: 10.1109/ISCS61804.2024.10581024

Individuals, organizations, and society as a whole will reap advancements bring up significant concerns, such as those
broad and significant advantages as a result of the capabilities
regarding the conflict that exists between artificial
of artificial intelligence (AI) both today and in the near future.
Nevertheless, these very same technical advancements give rise intelligence and data protection legislation [4]. Furthermore,
to significant concerns, such as the question of how to ensure in light of the technical realities of the 21st century, we are
that artificial intelligence technology is built and implemented in a position to review the efficacy of the data protection
in a manner that is in accordance with the applicable data rules that are now in place, which are both an opportunity as
privacy laws and standards. The fast development of artificial well as an obligation. We require data protection rules and
intelligence presents substantial hurdles in terms of protecting policies that not only successfully protect privacy in an era
customers' privacy and the confidentiality of their data. The of artificial intelligence and the vast data on which it
purpose of this essay is to suggest an all-encompassing strategy frequently depends, but also do not impose additional
for the development of a framework to solve these concerns.
hurdles for the development of these breakthrough
First, an overview of prior research on security and privacy in
artificial intelligence is presented, with an emphasis on both technologies in the future. We cannot make a decision
the progress that has been made and the limits that still among the already routine advantages of artificial
remain. In the same vein, open research topics and gaps that intelligence and the safety of personal data; rather, we must
need to be addressed in order to improve existing frameworks develop practical solutions to ensure both [5]. This is
are recognized. Regarding the development of the framework, something that has been emphasized in multiple reports by
the topic of data protection in artificial intelligence is the government and regulators. There has been a significant
discussed. This includes elaborating on the significance of increase in the analytical capacity of modern computers,
protecting the data that is utilized in artificial intelligence which is posing an increasing challenge to the laws and
models, as well as elaborating on the policies and practices that
conventions that govern data protection [6]. The term
are in place to ensure the data's safety and the methods that
are utilized to maintain the data's integrity. Additionally, the "artificial intelligence" is frequently used to refer to these
security of artificial intelligence is investigated, which includes advancements. This word is used to define the overarching
an analysis of the vulnerabilities and dangers that are present objective of enabling "computer systems to perform tasks
in artificial intelligence systems, as well as the presentation of which normally require human intelligence, including visual
instances of potential assaults and malevolent manipulations, perception, speech recognition, decision-making, and
as well as security frameworks that are designed to prevent translation among languages."1. This one term incorporates
these risks. The model that has been suggested is known as the a vast range of technical advancements, each of which may
Artificial Intelligence based Secured Data Handler (AISDH), provide a unique set of difficulties to the data protection
and in order to test the effectiveness of the proposed scheme, it
regulations that are already in place [7].
is cross-validated with the traditional Blowfish Algorithm
(BA).
This is owing to advancements in processing power
Index Terms—Data Security, Data Integrity, Artificial as well as the massive amounts of digital data that are
Intelligence, AI, Cyber Laws, Secured Data Handler, AISDH, available for study [8]. The foundation of other tools, a few
BA, Blowfish Algorithm of which are described below, is machine learning, which is
widely used today to perform a wide variety of tasks. These
I. INTRODUCTION tasks include the detection of fraud, the filtering of email,
the detection of cyber-threats including network intruders or
In the past few years, there has been a rapid malicious insiders, the recommendation of books or films,
development in artificial intelligence (AI) [1]. These days, and the provision of other services based on previous or
artificial intelligence (AI) tools are being utilized by unusual behavior. Deep learning frequently makes use of
organizations in both the public and private sectors all over larger data sets in order to generate larger models and train
the world. Individuals, organizations, and society as a whole those models in the most effective manner [9].
will reap widespread and significant benefits as a result of

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 II. RELATED STUDY computing over building their own IT infrastructure due to
the significantly reduced cost of the services offered.
With the rise of new technologies like distributed Computing and data storage are outsourced to third-party
computing, utility computing, and grid computing, the cloud companies under the cloud computing model. Apps and
computing movement is picking up speed [11]. One of the customer data are not entirely under the control of the
most important trends in modern computing, cloud customer. Customers' principal concern when contemplating
computing has the ability to give accessible, elastic, and cloud services is the security of their data, which carries
powerful resources over the internet at a reasonable price. with it the additional burden of security concerns. This study
Virtual computers access data through a network resource delves into different cloud computing service deployment
and the cloud provides the ability to store and access data patterns and the Internet of Things (IoT), identifies the data
remotely. In addition, the fourth industrial revolution is security challenges that come along with each, and proposes
spearheaded by cloud computing. When using Dropbox, a metric-based approach to evaluate the security of cloud
other Google services, or Microsoft Office 365, everyone is services [13].
always interacting with the cloud. Although there are
numerous benefits to such an environment, there are also A critical issue for the network community is
security concerns, including data availability, privacy, and managing the massive IoT infrastructure [14]. Many believe
security; access control; cyber-attacks; and performance and that Software Defined Networking (SDN) is the best way to
reliability challenges. Providers of cloud services have an oversee the Internet of Things (IoT) because of its
obligation to their customers to take reasonable precautions centralized network management features. In order to
to protect their data in a way that is both private and localize cloud demands, edge computing provides cloud
accessible at all times. However, consumers have not been resources close to the IoT. Thus, a framework that combines
well served by cloud service providers in terms of SDN, the Internet of Things (IoT), and edge computing may
dependable and secure services. Cloud computing is being be built to establish a powerful SDIoT-Edge architecture.
enhanced by a technology called blockchain. In order to This architecture can effectively orchestrate cloud services
address security issues, this groundbreaking technique and make flexible use of IoT devices with constrained
provides data integrity features that are convincing. The resources. In addition to the widespread use of IoT,
difficulties associated with cloud privacy and security is attackers can exploit the OpenFlow channel through
examined in depth in this study. In this case study, we focus Distributed Denial of Service (DDoS) attacks due to the
on smart campus security to highlight the significance of vulnerabilities in this less secure architecture. A major
security challenges; hopefully, this will inspire future challenge with SDIoT-Edge is providing security for the
studies to investigate cloud computing security [11]. OpenFlow channel, since DDoS attacks on this channel
might impair the entire network. In order to address the
Everyone can benefit from enhanced physical security issues with this design, we offer a framework that
health by getting enough quality sleep [12]. In order to we term SDIoT-Edge Security (SIESec). For effective
maintain a healthy body and mind, it is essential to evaluate defence against DDoS attacks, the SIESec prototype uses a
the sleep cycle, since irregular sleep patterns may be a sign categorization technique based on machine learning,
of the illness leading to persistent sadness. The provision of integrates blacklists, and filters contextual network flows.
high-quality health care facilities through the use of cost- Using the Mininet network simulator and the Floodlight
effective instruments and technology has been one of many controller, we run comprehensive simulations. Our findings
issues investigated in the context of globalization, alongside demonstrate that SIESec offers robust protection against
the expansion of facilities. Creating a reliable and affordable distributed denial of service (DDoS) attacks over OpenFlow
system to track patients' sleep quality is one of the stated channels while imposing minimal network burden [14].
goals of IoT technology development. There are a number
of alternative systems that provide this function, but they are Physical characteristics of wounds, such as their
prohibitively expensive and complex to set up. An width, form, color, etc., have been the primary focus of
innovative approach to track and analyze sleep patterns clinical wound assessment studies [15]. However, the most
based on environmental factors is proposed as a solution to important component influencing the healing process is the
this problem in this study. The suggested method is robust wound's appearance. But there are other aspects outside the
enough to accurately track patients' sleep using COS sensors wound's appearance that play a role in the healing process.
and make predictions based on the data provided by the An example of a factor that might have an effect on healing
random forest model. The study's findings demonstrate that is the wound's internal and exterior environment. The
the recommended method has a 95% success rate. Through widespread use of the Internet of Things (IoT) in nearly
the validation of manual results, which yield the lowest error every industry and household over the past decade has
rate, this method is tested using the patient's sleep data. The contributed significantly to its meteoric rise in popularity.
use of the Internet of Things (IoT) to create a smart system Thus, in this study, we present an Internet of Things (IoT)
to monitor sleep quality on a variable number of people at a based intelligent wound assessment system that uses
low cost is highlighted in this work [12]. decision tree entropy and information gain statistics to
classify assessment results as good, satisfactory, or
The term "cloud computing" refers to an emerging alarming, reflecting the status of the wound assessment.
paradigm in the IT industry that uses the Internet to provide Using MATLAB, we built a decision tree and used the ID3
computer resources to users on demand [13]. Small and algorithm to choose the optimum feature to split the tree
medium-sized enterprises (SMEs) often choose cloud based on entropy and information gain. Decision tree

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 performance and training accuracy were both enhanced by
an effective feature split [15].

III. METHODOLOGY

When it comes to physical protection, the

application of artificial intelligence in information defense
creates new obstacles. Even if it is essential to make use of
artificial intelligence technologies in order to identify and
combat malware threats, it is also possible for cyber
attackers to use AI tools in order to carry out progressive
behavior attacks. In part, this is due to the fact that access to
advanced artificial intelligence technology, in addition to
machine learning methodologies, is increasing at a time
when the costs of creating and deploying these
breakthroughs are decreasing. As a result, malicious
software can be developed by computer attackers in a more
expedient and cost-effective manner, allowing them to
create increasingly sophisticated and effective applications.
A definition of hostile artificial intelligence provided by
Accenture describes it as a situation that "causes machine
learning systems to misunderstand inputs into the structure
and react in a way that is beneficial to the intruder." In
essence, this happens when neural networks in an artificial
Fig.1 Flow Diagram for the Construction of a Structure that Assures the
intelligence programme are tricked into incorrectly Privacy and Security of AI
identifying or erroneously depicting artifacts as a result of
modified inputs that have been purposefully altered. When it Artificial intelligence (AI) comes with both
comes to building privacy remedies for big data, one of the advantages and disadvantages. Despite the fact that
challenges that arise is the sheer volume of data sources. technology is now helping workers in a variety of
When there are several data sources, it is possible for there professions, it is conceivable that AI will lessen the
to be a variety of settings in which different stakeholders necessity for workers in other professions. If the data used
can have differing degrees of possession of the processed in the development of artificial intelligence only represents
data. some segments of the population or reflects current societal
bias, it may add bias and new kinds of discrimination. This
A single data owner has the ability to encrypt their is especially true if the data is used to develop AI. The
data using their own keying material and to apply data conventional ideas of urban and residential design, which
analytics to the encrypted data neither locally or offloaded often involve substantial areas that are devoted to parking
to an external platform. This means that the data owner has lots and garages, are expected to be challenged by artificial
the ability to encrypt their data. On the other hand, in today's intelligence. Additionally, artificial intelligence may give
world, data are gathered by a wide variety of apps and rise to significant antitrust concerns, particularly in the event
services, which are utilized by a variety of businesses. The that the data required for its development is centralized in
data in question are frequently subjected to in-depth analysis the hands of a small number of firms. When compared to the
in order to derive knowledge that is beneficial to the way that is now in use, which is known as Blowfish
companies in question. The following figure Fig.1 shows Algorithm (BA), the proposed method in question, which is
the flow diagram for the construction of a structure that dubbed Artificial Intelligence based Secured Data Handler
assures the privacy and security of AI. (AISDH), considerably improves the capabilities of systems
in terms of cyber-security. The following figure Fig.2
describes the process of implementing the framework for AI
security and privacy.

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 our horizons, develop our skills, and find better ways to
organize what we already know. This calls for computer
learning. From the most fundamental kind of learning-
parametric learning, or understanding the meaning of such
quantities to the most advanced forms of abstract teaching
concept, grammar, usability, and behavioral learning, there
is a vast spectrum of difficulties.

IV. RESULTS AND DISCUSSIONS

Machine learning is an AI technique that eliminates

the need for explicit programming and allows computers to
learn from their own experiences and the data given to them.
Many areas, including data security, can benefit from this
rapidly developing technology. These days, data security is
a hot subject because of all the dangers that come with
dealing with massive amounts of data. When it comes to
data security, machine learning technologies can help
identify potential risks. This paper's objective, then, is to
investigate how machine learning contributes to data
protection. Machine learning will make it feasible to secure
massive volumes of data. We are living in the data-rich era
Fig.2 Process of Implementing the Framework for AI Security and Privacy of the internet and rapidly expanding technology across all
industries. The correct protection of data is of the utmost
Deep learning has become increasingly important importance; hence, machine learning is one area that might
in computer security due to the rapid mobility of both its be merged with data security to facilitate data management.
implementation in hardware and its use in graphical Additionally, data security will be achieved by appropriate
chipsets. New neural net innovations include third- data handling. However, data security using machine
generation cognitive nets, which speed up machine learning learning is still in its infancy and is only beginning to
by better simulating artificial neurons and opening up more develop in tandem with related research and best practices.
potential uses.
Data security can benefit from some of the machine
Agents with intelligence are programmes that run learning methods that have recently been developed. Data
on computers. They could be capable of planning, security may be achieved by the application of machine
organizing, and assessing. In fact, there is a concept of learning techniques such as supervised learning,
software agents within the software development unsupervised learning, and reinforcement learning. It is also
community. These are objects that actively employ the possible to utilize other algorithms. When it comes to
agent-like language used for networking. Subjects, in protecting sensitive information, machine learning has
contrast to agents, can be non-active and unable to several potential uses. Secure data storage will benefit from
communicate in order to be understood. Effective protection all of these. Data security is aided by machine learning, a
of cooperating agents from DDoS attacks was demonstrated prominent technology. Artificial Intelligence based Secured
in simulations using intelligent agents. A "cyber police Data Handler (AISDH) is a machine learning algorithm and
force" made up of mobile smart officers could be a reality if approach that can aid in data security; to evaluate the
all contractual and legal hurdles have been cleared. Included suggested scheme's efficacy, it is cross-validated with the
in this would be gear that allows cyber professionals to be standard Blowfish Algorithm (BA). Appropriate data
mobile and connected, but which adversaries cannot access. classification is facilitated by the suggested classification
Works well with Internet service providers (ISPs). methods. It aids data security by classifying fraudulent and
Nevertheless, the search can be somewhat more effective if accurate or secure data. As an example, we may learn about
additional experience can be utilized to guide it. A smart fraudulent emails since machine learning models categorize
system's total success is frequently dependent on the quality spam emails differently. Therefore, we shall take
of its quest, which is present in nearly all of them. In order precautions to protect our data by avoiding providing any
to address specific search challenges, a wide range of search information in response to any unsolicited communication,
strategies are developed. There are a lot of known AI search especially if it appears to be spam. Classification of user
strategies that are utilized in a lot of applications, but they behaviour patterns allows for user authentication, which in
aren't employed as AI. One issue is that search isn't turn helps in identifying both authorized and unauthorized
considered an AI function because it's part of the application users of data or systems. Malware detection, questionable
stack. This is the main way that optimal security issues are data access, and how to prevent it can all be taught to
addressed by dynamic analysis programming. Learning classification algorithms. Data and security issues, such as
fortifies the data structure by extending, rearranging, or fraudulent users, may be better anticipated using this
improving the existing body of knowledge. There has been a information. Face recognition is one of the most well-known
lot of research into it because it is one of the most crucial uses of machine learning. To prevent unauthorized
areas of AI. With the use of computation, we may expand individuals or entities from gaining access to systems or

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 data, facial recognition technology may be utilized as a
login credential for any system. System administrators can
determine who should have access to sensitive information
by implementing biometric authentication, which compares
a user's face to a database of previously stored images.
Using facial recognition technologies, it is possible to detect
and track people who are unauthorized accessing data or
systems.

This approach may be utilized to verify the

identification of individuals in many settings, such as
business enterprises, government offices, educational
institutions, and more. It can authorize users to make
transactions and prevent fraudulent activities in the financial
industry. Data security is achieved by the widespread usage
of facial recognition technology, which aids in
authenticating users. One use of machine learning is human
identity identification, which helps identify people by seeing
Fig.3 Classification Accuracy
patterns in their individual characteristics. Users can be
granted access to data and systems using this method.
The prediction accuracy of the proposed model,
Online marketplaces are becoming increasingly popular
AISDH, is shown in Fig-4. To evaluate the accuracy of the
these days. In order to safeguard the system and its data, we
model, it is cross-validated with the traditional BA. What
must first identify the specific data security issue that needs
follows is a descriptive table of the same information,
fixing, and then think about the best machine learning
Table-2.
approach to take. Overall, data security may benefit from
machine learning if its many ideas are adequately TABLE-2: PREDICTION ACCURACY
understood.
Data Size (bits) BA (%) AISDH (%)
Fig. 3 shows the results of a cross-validation test
between the suggested model, AI-based Secured Data 1500 86.54 96.38
Handler (AISDH), and the traditional Blowfish Algorithm
2000 86.39 96.37
(BA), which was used to determine the model's
classification accuracy. What follows is a descriptive table 2500 86.52 96.42
of the same information, Table-1.
3000 86.47 96.49
TABLE-1: CLASSIFICATION ACCURACY
3500 86.39 96.51

Data Size (bits) BA (%) AISDH (%) 4000 86.40 96.48

1500 91.26 98.35 4500 86.37 96.82

2000 91.32 98.39 5000 86.35 96.73

2500 91.36 98.36 5500 86.33 96.79

3000 91.39 98.64 6000 86.31 96.85

3500 91.42 98.29 6500 86.29 96.90

4000 91.47 98.64 7000 86.26 96.96

4500 91.51 98.59

5000 91.55 98.63

5500 91.58 98.67

6000 91.62 98.71

6500 91.66 98.75

7000 91.70 98.79

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 Figure 6 shows the data integrity ratio of the
AISDH model, which is cross-validated with the traditional
BA to assess the data integrity ratio of the plan. What
follows is a descriptive table of the same information, Table
4.
TABLE-4: DATA INTEGRITY ANALYSIS

Data Size (bits) BA (%) AISDH (%)

1500 90.26 95.64

2000 89.45 95.39

2500 90.01 95.48

3000 89.63 95.62

Fig.4 Prediction Accuracy 3500 89.51 95.54

Fig. 6 shows the data security ratio of the AISDH 4000 89.37 95.59
model, which is cross-validated with the standard BA to 4500 89.24 95.58
measure the degree to which the proposed system improves
data security. Table 3 provides a descriptive representation 5000 89.11 95.59
of the same. 5500 88.97 95.60

TABLE-3: DATA SECURITY ANALYSIS 6000 88.84 95.61

6500 88.71 95.62

Data Size (bits) BA (%) AISDH (%)
7000 88.57 95.63
1500 91.83 98.76

2000 91.76 98.64

2500 91.82 98.52

3000 91.76 98.79

3500 91.83 98.67

4000 91.79 98.67

4500 90.64 98.66

5000 90.39 98.65

5500 90.52 98.73

6000 90.40 98.82

Fig.6 Data Integrity Analysis
6500 90.20 98.78

7000 90.01 98.76

V. CONCLUSION

For the purpose of ensuring the safety of data, we

may make use of machine learning, which is one of the
technologies that are experiencing rapid growth. Data may
be protected through the use of machine learning by
recognizing and anticipating potential dangers to the data.
For the purpose of ensuring the safety of data in a variety of
domains, a variety of machine learning algorithms and
methods can be employed. Anomalies in data are discovered
by machine learning through the analysis of patterns in data.
Additionally, machine learning generates security warnings,
which contributes to the prevention of data loss. Because
machine learning models are able to process massive
volumes of data and learn from their previous experiences
and data, they are able to provide a variety of various ways
Fig.5 Data Security Analysis to safeguard the data. Through these means, we are able to

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.
 utilize the ideas of machine learning in the context of data
security if we have a thorough grasp of them.

REFERENCES

[1] Wang Shaohui, Pan Xiaoxiao, Wang Zhiwei, Xiao Fu and Wang
Ruchuan, "Identity-based cloud storage data integrity verification scheme
supporting forward security [J]", Journal of Nanjing University of Posts
and Telecommunications (Natural Science Edition), vol. 39, no. 01, pp. 79-
86, 2019.
[2] V. M et al, "Deep Reinforcement Learning for Energy Efficient Routing
and Throughput Maximization in Various Networks," (I-SMAC), Dharan,
Nepal, 2022, pp. 204-210, doi: 10.1109/I-SMAC55078.2022.9987395.
[3] Huang Xuerong and Guo Rongzuo, "Data integrity audit protocol based
on identity proxy off-line signature [J]", COMPUTER ENGINEERING
AND DESIGN, vol. 41, no. 06, pp. 1553-1561, 2020.
[4] T. M. et al, “Effective control of non linear parameters using artificial
intelligence,” 2016 IEEE International Conference on Computational
Intelligence and Computing Research (ICCIC), Dec. 2016, Published, doi:
10.1109/iccic.2016.7919659.
[5] Wang Yanling, "Cloud data integrity verification algorithm for
accounting informationization in sharing mode [J]", Modern Electronics
Technique, vol. 42, no. 05, pp. 87-89, 2019.
[6] S. S. Abdul-Jabbar, A. Aldujaili, S. G. Mohammed and H. S. Saeed,
"Integrity and security in cloud computing environment: a review", Journal
of Southwest Jiaotong University, vol. 55, no. 1, 2020.
[7] R. Umamaheswari, D. Lakshmi, V. S. Pandi, B. Geetha, S. Sumithra
and R. P. Y, "An Advanced Deep Learning Approach for Primary
Osteoporosis Prediction Using Radiographs with Clinical Covariates," 2023
7th International Conference on Electronics, Communication and
Aerospace Technology (ICECA), Coimbatore, India, 2023, pp. 788-793, doi:
10.1109/ICECA58529.2023.10395285.
[8] Liu Xueyan, Lu Tingting and Yang Xiaotao, "Verifiable Attribute-
based Keyword Search Scheme with Privacy Preservation [J]", Journal of
Electronics & Infonnation Technology, vol. 43, no. 01, pp. 218-225, 2021.
[9] R. K. Sadavarte, G. D. Kurundkar and D. Parbhani, "Data security and
integrity in cloud computing: Threats and Solutions", 2020.
[10] B. Sarwar, I. S. Bajwa, N. Jamil, S. Ramzan and N. Sarwar, "An
intelligent fire warning application using IoT and an adaptive neurofuzzy
inference system", Sensors, vol. 19, no. 14, pp. 3150, 2019.
[11] J. J. J, B. V. Prabha, B. Yasotha, J. J, C. Senthilkumar and V. S. Pandi,
"Enhancing Residential Security with AI-Powered Intrusion Detection
Systems," 2023 International Conference on Sustainable Communication
Networks and Application (ICSCNA), Theni, India, 2023, pp. 1510-1515,
doi: 10.1109/ICSCNA58489.2023.10370042.
[12] K. Saleem, I. S. Bajwa, N. Sarwar, W. Anwar and A. Ashraf, "IoT
healthcare: design of smart and cost-effective sleep quality monitoring
system", Journal of Sensors, 2020.
[13] D. K. Saini, K. Kumar and P. Gupta, "Security Issues in IoT and CC
Service Models with Suggested Solutions", Security and Communication
Networks, vol. 2022, 2022.
[14] W. Rafique, M. Khan, N. Sarwar and W. Dou, "A security framework
to protect edge-supported software-defined Internet of Things
infrastructure", In International Conference on Collaborative Computing:
Networking Applications and Worksharing, pp. 71-88, 2019.
[15] H. Sattar, I. S. Bajwa, R. U. Amin, N. Sarwar, N. Jamil, M. A. Malik,
et al., "An IoT-based intelligent wound monitoring system", IEEE Access,
vol. 7, pp. 144500-144515, 2019.

Authorized licensed use limited to: Pontificia Universidad Catolica Del Peru. Downloaded on November 10,2024 at 16:29:26 UTC from IEEE Xplore. Restrictions apply.