Journal of Network and Computer Applications 213 (2023) 103607

Contents lists available at ScienceDirect

Journal of Network and Computer Applications

journal homepage: www.elsevier.com/locate/jnca

Review

A comprehensive survey on security, privacy issues and emerging defence

technologies for UAVs
Hassan Jalil Hadi a , Yue Cao a,b ,∗, Khaleeq Un Nisa a , Abdul Majid Jamil a , Qiang Ni c
a
School of Cyber Science and Engineering, Wuhan University, China
b
Shenzhen Research Institute of Wuhan University, China
c
School of Computing and Communications, Lancaster University, UK

ARTICLE INFO ABSTRACT

Keywords: In the past two decades, there has been a rapid development in the drone industry known as Unmanned
Cyber security and privacy Aerial Vehicles (UAVs). Currently, the use of commercial UAVs has increased a lot due to their affordability,
Vulnerabilities but lack of security implementations has introduced many threats and vulnerabilities in UAVs. In addition,
Adversarial machine learning
software, and hardware complexity in UAVs also triggers privacy and security issues as well as causes critical
Intrusion detection system
challenges for government, industry and academia. Firstly, in this research review, we broadly survey privacy
Digital forensic
Blockchain
and security issues of UAVs by dividing them into three classes: Software, Hardware and Communication.
Particularly, for each class, we systematically survey the common vulnerabilities causing potential attacks
to UAVs. Secondly, a review of prevailing threats that are threatening civilian UAVs’ applications is also
a part of this survey. Thirdly, a comprehensive discussion of passive and active attacks from adversaries,
for compromising privacy and security of UAVs is given as well. Fourthly, we provide detail description
of existing mitigation techniques and countermeasures, to protect UAVs. Fifthly, the solution architecture
part includes discussion about emerging technologies such as, blockchain usage, machine learning, intrusion
detection systems and secure communication protocols. To store all data in transit, blockchain can be used
cryptographically and protect it from eavesdropping and tampering. In this research review, key points that
highlighted the lessons learned about security and privacy of UAVs are also summarized. Lastly, the survey
is concluded by discussing important pitfalls as well as suggestions for future research directions, concerning
privacy and security of UAVs.

1. Introduction surveys and damage assessments. Fig. 1 shows the applications of

UAVs.
The use of mobility and locality aware vehicles has increased in The applications of UAVs are dramatically increasing as a result
nearly every domain, this has led to the significant increase in the of their mobility flexibility, low upkeep cost and capacity, to monitor
number of these vehicles and associated applications. A popular class of real-time environment. The following are the key applications of UAVs:
such vehicles is Unmanned Aerial Vehicles (UAVs), capable to fly with
or without any pilot. UAVs can be controlled remotely through radio 1.1.1. Monitoring, security and surveillance
or Wi-Fi and their advance features have capability of carrying heavy UAVs are a crucial component of military surveillance missions.
payloads, mine’s detection and unethical activities’ scanning in specific UAVs have been included in the defence strategies of several na-
areas (Abbaspour et al., 2016a). tions. These flying robot devices are used by nations for border con-
trol, maritime monitoring of important sea lanes, enemy identification,
1.1. Applications of UAVs anti-poaching, and enemy detection.

In recent years, the usage of UAVs by military for crucial operations 1.1.2. Disaster management
has intensified. The other applications of UAVs include rescue missions, Information gathering, the need for quick responses, and debris
search missions, courier services, fire-fighting surveillance, ecological navigation can all be assisted by UAVs. UAVs equipped with radars,

∗ Corresponding author.
E-mail addresses: [email protected] (H.J. Hadi), [email protected] (Y. Cao), [email protected] (K.U. Nisa),
[email protected] (A.M. Jamil), [email protected] (Q. Ni).

https://doi.org/10.1016/j.jnca.2023.103607
Received 3 November 2022; Received in revised form 23 January 2023; Accepted 20 February 2023
Available online 24 February 2023
1084-8045/© 2023 Elsevier Ltd. All rights reserved.
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 1. Applications of UAVs.

sensors, and high-quality cameras can help rescue crews locate damage, 1.2. Security concerns raised by UAVs applications
launch recovery efforts right away, and deliver supplies like manned
helicopters and first aid packages. UAVs can aid with rapid disaster However, majority of such application domains has caused great
estimation, disaster alarms, and the discovery of effective countermea- threat to human lives, because of unidentified weather and other en-
sures. vironmental factors. These UAVs are being attacked in different ways,
because they are susceptible to numerous security risks (Nassi et al.,
1.1.3. Aerial photography
2019). Therefore, results of many attacks might be overwhelming,
One of the wonderful uses for UAVs is aerial photography. This
where some of these attacks are intended to steal information as well as
technology is outfitted with heavy camera equipment that will greatly
affect confidentiality, integrity, and availability (AbdAllah et al., 2015).
help enthusiasts give aerial pictures of the targeted areas. UAV aerial
photography have the ability to give clear images. Additionally, with the help of communication channels, UAVs carry
a lot of information, while that must be secured. The information
1.1.4. Packet delivery carried by UAVs is of multiple forms like text, images, videos, and
The use of UAVs for transportation and delivery is another signifi- audios. Although, many encryption algorithms have been proposed for
cant evolutionary sector. It speeds up the distribution process, enhances security of sensitive data gathered by UAVs, yet UAVs are insecure (Li
system efficiency, and uses less manual labor overall. et al., 2018) because these encryption algorithms have vulnerabilities
that can be exploited. This has raised many security concerns about the
1.1.5. Rescue and search operation communication protocols of UAVs. Currently, the main security con-
UAVs can be deployed through flood zones for the purpose of rescue cern related to these communication protocols is to secure information
operations and care. Before the assistance of a rescue team arrives, exchange using connections like Wi-Fi.
these are more useful for providing food and medications to those Usually, to send data to base stations, UAVs use a wireless link
human resources who are outside of these institutions. Additionally, which is vulnerable for attackers. Hence, it is necessary to prevent data
UAVs can significantly reduce the amount of the labor, materials and interception by attackers. For this purpose, one widely used mechanism
time required for SAR by providing real time imaging data of tar- is encryption like Advanced Encryption Standard (AES) being used
get sites. As a result, a SAR (search and rescue) team may precisely
commonly at present. However, its use is inefficient when it comes to
determine where immediate aid is needed and identify the location
real-time applications due to communication overhead, particularly in
in real-time. In emergency scenarios including missing individuals,
case of extremely high data transfer rate (Li et al., 2018).
avalanches, wildfires, and gas penetration, UAVs can expedite SAR
activities. Another important security concern is the attackers interfering with
UAVs, with intention to take control of UAV for disabling communi-
1.1.6. Wildlife monitoring cation between Ground Control Station (GCS) and the UAV. For this
UAVs are quite useful for observing animals because they are fitted purpose, numerous attacks such as spoofing (Arthur) jamming and false
with laser lights, cameras altimeters, GPS tracking, altimeters and other data injection are being used. In this paper, we will analyze the security
devices. In addition to using night and day cameras, effective conser- and privacy challenges faced by UAV networks, but it is necessary to
vation UAVs have used infrared and long-range focal-point cameras to know first the UAV system’s components and information flow among
observe wildlife. these components, in order to effectively analyze these challenges.

2
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 2. Block diagram of a typical UAV.

1.3. UAV system This review paper is arranged in this way: Section 2 presents related
work, vulnerabilities in UAVs are discussed in Section 3. Defence and
The base system of a UAV is manufactured only to link together all emerging technologies to secure UAVs are described in Section 4. In
components. It is utilized for communication among these components, Section 5, application of Blockchain in UAVs network are explored,
to control sensors and navigation systems, as shown in Fig. 2. The base while challenges and possible future research directions are discussed
system of UAVs is also used to integrate the optional components like in Section 7. In addition to above, to give better understanding of paper
weapon systems etc. Besides, the sensors system in UAVs is made up of
flow, the taxonomy of the paper is presented graphically in Fig. 3.
sensory equipment having integrated functionalities for pre-processing,
e.g., sensors having GPS, sensors along with cameras and sensors with
radars. 1.4. Motivation
In addition, avionics systems have the responsibility to execute
the control commands issued by controllers like spoilers, engine com-
mands, stabilizers and flaps. The UAVs mainly depend on a wireless The estimated value of the commercial UAVs’ market was $20.8
communication channel, a direct communication line or indirect com- billion in 2021 and it is expected to increase to $ 501.4 billion by the
munication line, by using satellites. For attacking a UAV, the attacker end of 2026 (Anon, 2021). UAVs are becoming more popular due to
must affect the UAV externally to access UAV system physically. The their application in a variety of fields and their use will further increase
communication system of UAVs relies mainly on the inputs from exter- in future. This has also posed serious privacy and security threats to
nal environment, due to wireless condition in these system, this offers UAVs also there exist many vulnerabilities in UAVs. In last few decades,
numerous input channels to an attacker. many research surveys have been conducted about security and privacy
The information in UAVs streams between the UAVs and external
concerns of UAVs but none of these has thoroughly expressed security
environment over different channels. Therefore, the two-way com-
and privacy issues related to UAVs. Most of the existing surveys have
munication between the GPS and UAV communication system is the
just discussed common security risks and concerns related to UAVs.
highly open channel, which can be easily utilized for attacking UAVs.
Another most critical and sensitive component of the UAVs is the flow Hence, there is a need to conduct a survey that provides high cover-
of information from the external environment to UAVs sensors, where age of this issue. Thus, we are motivated to thoroughly explore the
such links are highly prone to manipulations. However, the reliability vulnerabilities and threats present in UAVs, as well as related security
of these sensors is not trustworthy. In such case, the knowledge of the and privacy concerns. Moreover, we are determined to discuss solution
host, about receptiveness of components to commands, is a key for architecture (like blockchain, IDS, MLIDS) to mitigate these threats and
controlling UAVs during cyber-attacks. vulnerabilities.

3
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 3. Flow of the survey paper.

1.5. UAV cyber security threat model hijacking, identity spoofing, eavesdropping, and multiprotocol han-
dover. Such attacks can be caused by integrity loss like increased Signal
The UAV threat model is based on fundamental security services to Noise Ratio (SNR), tapping, subroutine exploits, re-transmitting, and
including availability, confidentiality, and integrity. Under these three capturing feed.
services, there are natural events and malicious attacks. These, ma- Similarly, attacks on availability can be performed by an intruder
licious attacks on the confidentiality of UAV communication include using DoS. Table 1 shows threats and projected control for UAVs, while
malware, viruses, key-loggers, hacking, trogons, hijacking, cross-layer, natural events causing confidentiality loss are greed, social engineering,

4
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 4. UAV cyber security threat model.

and life threatening in Fig. 4. All related cyber security threats are 1.5.6. Man-in-the-middle (MitM) attack
explained below: In this attack, attackers secretly relay and alter communication
among two or more parties, so that parties trust they are communi-
1.5.1. Eavesdropping cating to each other.
This is one type of man-in-the-middle attacks. Here, attackers would
create an individualistic connection with the victim and transmit mes-
1.5.7. Trojan
sage, in such a way that victim cannot identify that the communication
party is illegitimate. It is a way through which the user can bypass normal security and
get access to the system. The backdoor can have a legal use and pre-
1.5.2. Distributed and denial of service (DDoS, DoS) attack exist for enabling system administrator’s access. Like other attacks, this
In the DoS attack, attacker try to make a system unavailable to attack might be used for evil purposes, such as allowing malicious users
intended purpose. Specifically they flood the system with requests continuous, secret access to the system.
more than the processing capacity of that system. This overloads the
system and causes interruptions of service. The Dos attack is simple 1.5.8. Brute force password attack
that unskilled hackers can conduct it, while DDoS attack is conducted In this attack, the password is cracked systematically by attempting
by originating the attack source from numerous infected hosts. numerous passwords, in order to eventually guess a correct password.

1.5.3. Identity spoofing

In this attack, attackers act like legitimate users in Internet of 1.5.9. Compromised URL
Drones (IoD) network. They use spoofing ID of legitimate users, and This involves changing parameters in a URL, so that the web server
access to the network as well as communication links. To prevent such would send information without authorization.
attacks, pseudo-ID or encrypted IDs might be useful.
1.5.10. Fabrication
1.5.4. Malicious code The authenticity of UAV is targeted by this attack. It permits access
This is the code inserted in the software or web script aimed to to UAVs components by giving incorrect information.
cause unwanted effects, damage to systems or security breaches. The
self-protection of software can be used to prevent from vulnerabilities
1.5.11. Natural issues
and attacks.
Due to the lightweight nature, UAVs cannot stand wind and extreme
1.5.5. Malware infection temperature, causing failure of engine, bringing the UAV down. More-
Communication protocols in UAVs permit users to pilot UAVs over, the battery may explode causing serious harm and damage. In
through remote control like tablets, mobile phones, and laptops. Un- fog, the visibility becomes limited which narrow downs from meters
fortunately, this is an insecure method because it permits hackers to to a meter, causing communication disruption between UAVs and GPS.
make reverse-shell payload of TCP, injecting into the UAVs memory Therefore, this can send the UAVs outside the range of control area
for installation of malwares. until it crashes.

5
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 1
Threats and countermeasures for UAVs.

Types of attacks Entity under attack System limitations Possible consequences Countermeasures

Malware, Key loggers, Malicious code GCS, UAV Less computational and Loss of Critical Data & UAV Antivirus, IDS, IPS
limited power at UAV,
Processing overhead

Jamming UAV Processing overhead and Loss of UAV Use of high gain directional
Power constraints antennas

Hijacking UAV Processing overhead and Loss of Critical Data & UAV Encryption
Power constraints

Eavesdropping UAV Processing overhead and Loss of Critical Data& UAV Encrypt communication fully
Power constraints

Multiprotocol attacks UAV Processing overhead and Loss of Critical Data Use tamper resistant protocols
Power constraints across communication links

Compromised link UAV Processing overhead and Loss of Critical Data Strong authentication,
Power constraints encryption

Data retransmitting Fabrication Replay UAV Processing overhead and Loss of Critical Data Time Stump, Packet signing
Power constraints

Dos/DDoS UAV Processing overhead and Loss of data & control Use resource and Bandwidth
Power constraints throttling techniques

Distortion UAV Processing overhead and Loss of UAV & control Anti-jamming techniques,
Power constraints ECCM Features, Frequency
Hopping.

Subroutine exploit UAV Processing overhead Loss of Critical Data & UAV Encryption

Cross-Layer Attack UAV Weak network performance Loss of Critical Data & UAV MAC binding

Interception of Transmission UAV Processing overhead and Loss of Critical Data & UAV Encryption
power constraints

GPS Spoofing GPS Processing overhead and Loss of Critical Data & UAV Filter packets, Verify identity
power constraints

Physical Security UAV Increase operational cost Loss of Critical Data & UAV Deployment of physical
security assets

Masquerading UAV Processing overhead Loss of Critical Data & UAV Encryption

1.5.12. Keyloggers 1.5.18. Signal-to-noise ratio

keyloggers are thought of as an internal threat because IoD is It is a method used in engineering and science, making a comparison
developed and deployed, keyloggers are integrated with software of of the level of a required signal with level of noise. It can be defined
IoD. Hence, the secret information is transferred to attackers through as the ratio of signal power to the background noise power.
attached keyloggers.
1.5.19. Signal integrity
1.5.13. GPS jamming and spoofing UAVs Signal Integrity (SI) includes the measurement of quality of electric
Hackers can hack UAVs as well, for example, through GPS jamming signals. Critical issues regarding SI are crosstalk, ringing, distortion,
and spoofing. In such a situation, UAVs cannot get legitimate GPS ground bounce, power supply noise, and signal loss.
signals and the GPS Jamming occurs when attackers successfully launch
signals that can interfere with GPS signals. While GPS spoofing is more 1.5.20. Subroutine exploit
sophisticated, thus attackers create incorrect signals causing UAV to It is also known as function, method, procedure, and subprogram.
change its position. It is a code called and executed somewhere in program, e.g., a routine
can be used for saving files or to show time.
1.5.14. Physical attacks
These attacks are conducted on hardware’s and a have key motiva-
1.5.21. Retransmission
tion to destroy devices.
An Round Trip timeout (RTO) has a significant role in TCP, pri-
marily for achieving reliable transmission. In the TCP, if a sender
1.5.15. UAV hijacking
GPS spoofing or Jamming is used for UAV hijacking, such as terror- sends information and there is no acknowledgment, also the RTO
ist attacks. timer is expired, then sender will consider that the information is lost.
Commonly used attacks against UAV are shown in Fig. 5.
1.5.16. Keystroke logging
It is the practice to record keystrokes of a keyboard, usually with- 1.5.22. Malfunction
out knowledge of users, for capturing instructions as well as data The malfunction of a sensor may lead to loss of assets and human
transferred between UAV and operator. life as well as abandoning missions. The communication between UAV
and GCS includes an uplink transferring command. Further, it involves
1.5.17. Distortion control to down link and UAV from the controller that gives back data
Once the trust in data integrity is lost, the modifications in sensitive displaying UAV status on the GCS. The GCS may be connected directly
data as well as access monitoring become critical, like the development to UAV through ground relay or satellite in case of Line of Sight (LOS)
of complicated management procedures in case of incident. range. There are numerous threats to UAV communication that are

6
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

attack risk either form physical or cyber domain. Similarly, security

issues of UAVs communication are also reviewed in Hentati and Fourati
(2020). Shakhatreh et al. in Shakhatreh et al. (2019) surveyed civil
applications of UAVs and their key challenges. Likewise, Krishna et al.
in Krishna and Murphy (2017) surveyed security vulnerabilities present
in UAVs, recent research by Shafique et al. (2021) discussed security
protocols and vulnerabilities in UAVs.
Another research work (Altawy and Youssef, 2016) have highlighted
safety, privacy and security aspects of commercial UAVs. Particularly,
authors have discussed physical and cyber threats, major vulnerabilities
and possible attacks that may cause crashing during a flight mission.
Authors in Yaacoub et al. (2020) identified emerging challenges as
well as cyber-attacks in commercial UAVs. Furthermore, malicious use
of UAVs and present threats were reviewed in Yaacoub et al. (2020).
Nassi et al. (2021) conducted a systematic review of privacy and
security challenged on commercial UAVs. Researchers in Zhi et al.
(2020) investigated challenges of wireless communication in UAVs like
Wi-Fi based communication in UAVs. They have also discussed the
privacy leak, caused by a UAV because of aerial photos.
Fig. 5. Common types of attacks on UAV.
2.2. Security and privacy issues of UAV communications

necessary to be tackled, and vulnerabilities in communication protocols Fotouhi et al. (2019) reviewed the critical security challenges caused
should be protected. by UAV wireless communications. Mishra and Natalizio (2020) argued
that UAVs integration with wireless networks causes security challenges
1.6. Contributions of this survey in UAVs. They further pointed that such issues a thorough investigation
by research community. Hayat et al. (2016) explored the security,
In this survey, we provide an extensive survey addressing privacy privacy and safety challenges of UAV networks from a communication
and security issues in UAVs. Our main contributions are given below: perspective. Also, they have discussed communication requirements for
a secure and safe UAVs deployment. Another research work (Sharma
• In this survey, we present a systematic division of privacy and et al., 2020), presented a comprehensive literature review of emerging
security issues by categorizing them into three classes: Software, UAV communication technologies, and discussed the need for securing
Hardware and Communication. Particularly, for each class, we the data transmitted to GCS. Hassija et al. (2021) have surveyed key
systematically survey the common vulnerabilities causing poten- security challenges and vulnerabilities in UAV communications.
tial attacks to UAV.
• Applications vulnerabilities in security protocols and prevailing 2.3. Security and privacy issues of UAV networks
threats concerning civil applications of UAVs are also reviewed.
• We thoroughly discuss passive and active attacks from adversaries Boccadoro et al. (2021) presented a survey of IoD. The authors
for compromising privacy and security of UAVs. A detailed de- have discussed privacy and security issues of drone-to-drone commu-
scription of existing mitigation techniques and countermeasures nications. They have also highlighted security concerns for specific
for protecting UAVs is also given. application scenarios in IoD architecture like smart framing and public
• Moreover, the solution architecture part includes discussion about safety. Likewise, privacy and security challenges related to design of
emerging technologies. These comprise of Blockchain usage, Ma- UAV networks are discussed in Noor et al. (2020). In this research
chine Learning, Intrusion detection systems and secure communi- study authors have argued that the key challenge is an Ad-hoc fashion
cation protocols. communication among multiple UAVs (Rondon et al., 2022). This
• In this research review, takeaways that highlighted the lessons communication is called Flying Ad hoc Network (FANET). Chriki et al.
learned about security and privacy of UAVs are also summarized. (2019) have also reviewed FANETs security challenges. They have
Lastly, we conclude the survey by discussing important pitfalls highlighted the need for developing robust security mechanisms before
as well as suggestions for future research directions, concerning deployment of FANET in real scenarios. Maxa et al. (2017) discussed
privacy and security of UAVs. key security issues of UAV routing protocols. Sharma et al. (2020) have
presented security threats and vulnerabilities of UAV communications
2. Related work protocols.

UAVs are thought of as an emerging category of flying IoT de- 2.4. Edge AI implementation challenges for UAVs
vices. UAVs include many applications such as immediate help for
patients, like medical supplies and delivering blood etc. Many privacy Edge AI, which runs AI close to consumers on devices or edge
and security issues have been addressed in the past decade, due to servers, may be appropriate for enhancing UAV-based IoT applications.
introduction of UAV technology. In this regard, several research works Several important UAV technological elements, including formation
have been published covering numerous aspects of UAV privacy and control, autonomous navigation, power management, computer vision,
security issues. security and privacy, communication, and applications, have been ex-
amined in a study on the effects of edge AI (McEnroe et al., 2022).
2.1. Security and privacy challenges of UAVs Although there are many difficulties in implementation, the develop-
ment of security and privacy, distributed training algorithms, resource
The privacy and security challenges in UAV networks are presented allocation, and real-time needs pose the most problems. There are two
in Wang et al. (2019) from a Cyber–Physical System perspective. Wang key components of application delivery system that require additional
at el. in this study, discussed important UAV components that have high study. The first component is reaction of system to errors, the key is

7
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 6. Vulnerability in UAV.

reaction of the. Second component is coordination of numerous UAVs which made UAVs more vulnerable to several certain cyber attacks
flying simultaneously in the same airspace while being controlled by (Tlili et al., 2022). Therefore, this section will describe how attackers
various operators. This can call for identical resources and operating use different strategies to breach security protocols for UAVs. Different
frequencies (Huda and Moh, 2022). The effective handling of image vulnerabilities and countermeasures are summarized in Tables 3, 4 and
processing, the development and creation of precise and independent 5, with schematic diagram shown in Fig. 6.
real-time UAV power line detection methods, and multi-UAV collabo-
ration for connectivity inspection are all areas of the implementation 3.1. Hardware threats and vulnerabilities
infrastructure inspection that require additional research.
Hardware threats and vulnerabilities consists of physical collision of
UAV, trojans, issues in flying skills and hardware failures (see Table 3).
2.5. Differences from existing surveys
3.1.1. Hardware Trojan
Unlike other reviews, this review paper is intended to broadly sur- These trojans include tampering with electronic hardware like mod-
vey privacy and security challenges of UAVs by categorizing them into ifying logic gate and hardware circuit modifications, etc. Bhunia and
different levels. Most of the existing surveys have categorized privacy Tehranipoor (2018). Hardware trojans are specially indented to tar-
and security challenges, in terms of attack vectors or based on basic get Flight Controllers (FC) and make UAV vulnerable to numerous
information security principles. Yet, this categorization cannot explain attacks. A non-trusted party maliciously embeds hardware trojans in
completely the threats, attacks, vulnerabilities, and countermeasures supply chain of FC (Rahman et al., 2020a). The enemy leverages such
etc. On contrary, our survey considers the privacy and security issues modifications for compromising the functions and security features of
of a complete UAV system e.g., end to end components like hardware, integrated circuits in FC, e.g., decrease in propellers’ rotation speed,
communication, and software. access to the secret keys of FC, etc. A similar trojan was discovered
Besides, privacy and security issues in commercial UAVs, are also in Boeing 787 jet (Casals et al., 2013). This trojan was found in the
reviewed in this survey. Particularly, dissecting from a security perspec- chip called Actel ProASIC (Casals et al., 2013). The backdoor in the
tive, threats, vulnerabilities, attacks, and countermeasures into three jet helped the attacker in monitoring the avionics system and taking
different categories. The first category is based on hardware, the second control of the aircraft, by putting the security of passengers at risk.
is based on software and the third is communication based. These
categories are significant classes of UAVs functionality. Moreover, this 3.1.2. Physical UAV collision
survey also presents attacks of UAV privacy and security aspect along Physical collision might happen during a mission that needs col-
laboration and cooperation among multiple UAVs. This will result in a
with techniques for their mitigation. The survey has also categorized
crash of UAVs and to avoid such a collision, UAVs depend on Collision
the emerging defence technologies for security and privacy risks in all
Avoidance System (CAS) (Yasin et al., 2020). Unfortunately, there are
categories mentioned earlier. The categories of these defence technolo-
no built-in security features in these systems, so these systems cannot
gies, for securing a UAV, include I) Secure Communication Protocols
fulfill the requirement of collision prevention threat posed by malicious
II) Machine Learning II) Intrusion detection Systems Iv) Blockchain
entities (Hannah et al., 2020).
Technologies.
To show differences among this survey and other existing works, a 3.1.3. Hardware failures
comparison table for existing surveys is presented in Table 2. UAVs can have defects of hardware components like motor issues
or battery life. Such defects pose a threat to the mission of flight
3. Potential vulnerabilities in UAVs and could cause unsafe landing in unexpected area (Alwateer et al.,
2019). In such a situation, if UAVs have unencrypted information, then
Numerous methods have been introduced to strengthen security of enemy can disclose secret mission-information and can also challenge
UAVs. Yet, all the existing methodologies have many shortcomings, the confidentiality of flight mission.

8
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 2
Comparison of our survey and existing surveys on security, privacy issues and emerging defence technologies for UAVs.
Ref & year Security Issues in Privacy Emerging Defence
UAV Issues in Technologies
UAV
Hardware Software Communication SC PY IDS MLIDS BCT
Vulnerability Vulnerability Vulnerability
A C T V A C T V A C T V A C
Wang et al.
(2019)
Hentati and
Fourati (2020)
Shakhatreh et al.
(2019)
Krishna and Mur-
phy (2017)
Shafique et al.
(2021)
Altawy and
Youssef (2016)
Yaacoub et al.
(2020)
Nassi et al.
(2021)
Zhi et al. (2020)

Mishra and Na-

talizio (2020)
Hayat et al.
(2016)
Sharma et al.
(2020)
Hassija et al.
(2021)
Noor et al.
(2020)
Chriki et al.
(2019)
Maxa et al.
(2017)
Newaz et al.
(2021)
Rondon et al.
(2022)
Boccadoro et al.
(2021)
Nassi et al.
(2019)
Khan et al.
(2020)
Syed et al.
(2021)
Tsao et al.
(2022)
Tlili et al. (2022)

Alladi et al.
(2020a)
Huda and Moh
(2022)
This work
A=Attack, C= Countermeasure, T=Threat, V= Vulnerability, SC= Secure communication, Py= Physical layer, IDS= Intrusion Detection system, MLIDS =Machine Learning Based
intrusion detection system, BCT= Blockchain Technology.

3.1.4. Flying skills issues UAVs may crash as well as operational failures may occur. As a result,
UAVs might become an easy target for physical theft.
Such issues occur when semi-autonomous or non-autonomous UAV
is operated by human remotely. Particularly those UAVs that are ex-
3.2. Software vulnerabilities and threats
tremely sensitive to wind disturbance, due to their size and complex
dynamics (Lee et al., 2020). That is why, flying skills like orientation Software vulnerabilities and threats for UAVs include zero-day vul-
of UAVs, height, and remote control of speed are needed, otherwise nerabilities and malicious software (see Table 4).

9
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 3
Hardware level vulnerabilities, existing countermeasures, and their limitations.

Ref Hardware-Based Countermeasure Limitation

Bhunia and Tehranipoor (2018) Hardware Trojans Designing machine learning based IDS solution and Hardware based complicated approaches
fine-grained circuit analysis perceive Trojan in hardware. might evade existing detection
approaches.

He et al. (2020) Physical collisions Design and implement collision avoidance systems for Collision avoidance system mostly avoid
physical collision protection. from security measures.

Alwateer et al. (2019) Hardware failure Implement strong encryption techniques to secure store data Strong encryption on data might be
in case of hardware failure. hurdle in forensic analysis.

Hodgkins (2015) Hijacking Validated encryption secure against unauthorized access of Attackers can hijack authentic UAVs by
UAV and GCS. Flight pattern can be avoided by constant counter UAVs technology.
changing flying path.

Table 4
Software level vulnerabilities, existing countermeasures, and their limitations.

Ref Software-Based Countermeasure Limitation

Paganini (2015) Malicious Software Firewall would be an excellent measure to restrict malicious Real-time detection and counter action
attacks. Along with mentioned measure IDS and antivirus enhance computational cost.
solutions would be beneficial protective measure.

Hooper et al. (2016) Zero day vulnerabilities Zero days vulnerabilities reduce attacks by periodic system Manufactures release patches after
update. zero-day exposure.

3.2.1. Malicious software 3.3.2. Network layer vulnerabilities and threats

The FC and GCS are vulnerable to malicious software and threats The UAV network works in the ad hoc fashion, known as FANETs.
caused by UAV malware. These threats may lead to control of UAV as FANETs have unique characteristics like computational power for rout-
well as loss of sensitive information. Also, an attacker’s access to a UAV ing data and latency. These UAV networks consist of dynamic topology
flight stack can lead to shut down of UAV, resulting a DoS attack and posing critical threats. Research work in He et al. (2017) has in-
disruption of flight mission. Implantation of such malware in UAV can vestigated common security threats to UAVs aided public security
compromise the privacy and security of UAV, e.g., Maldrone is a virus networks. The authors found if complexity of UAV networks increases,
that infects FC and enables the attacker to get control of UAV (Paganini, the vulnerabilities and attacks also increase, targeting at communi-
2015). Currently, there is an emerging kind of malware that comprises cation modules and sensor inputs. Moreover, vulnerabilities in UAV
of locking a system or encrypting the data of a user. It holds the data communication like blocking or intercepting the communication link
encrypted until an user is willing to pay a ransom to the adversary, such among FCs and GCS, can cause a DoS attack. To avoid this, it is required
malware is called ransomware (Oz et al., 2021). However, ransomware to develop cryptographic algorithms for FANTEs, taking such charac-
has not attacked UAVs yet. Therefore, it is necessary to consider that it teristics into consideration (Chriki et al., 2019). Next, the adversary
may target UAVs in future, because of their civilian applications. can disorder UAVs network directly or indirectly. By sending malicious
traffic directly through GCS or thorough UAV indirectly. Adversaries
3.2.2. Zero-day vulnerabilities continuously threaten the GCS with either centralized or decentralized
Hidden vulnerabilities can exist in GCS software or flight stack, architecture. In these architectures, GCS shows a single point of failure
for example, overflow, DoS, and buffer, etc. These vulnerabilities are and security of complete UAV network relies on the security of GCS.
known to the manufacturers of UAVs and can pose critical threats to Although, security frameworks have been implemented for GCS,
operators. The attackers can exploit such vulnerabilities until appropri- the adversary can still disrupt the flight mission by compromising
ate patches are released by UAVs manufacturers. However, operators flying UAVs (Tsao et al., 2022). In some cases, the mission can be
must update the system when a patch is released. considered as success even one or more UAVs get compromised. In
civilian application, operators need minimum uncompromised UAVs
3.3. Communication vulnerabilities and threats for accomplishing missions, as shown in Fig. 7.
In the centralized architecture, the attacker has to target the UAVs.
Communication threats and vulnerabilities can be classified based For this the attacker sends malicious traffic to a particular number
on the communication layers as follows (see Table 5): of UAVs. The attacker does this in a way that the minimum number
of uncompromised UAVs necessary for successful completion of the
3.3.1. Physical and MAC layer vulnerabilities mission is not satisfied. As a result, the attacker causes failure to flight
Many potential vulnerabilities are due to the complex nature of mission through disruption of whole UAVs network.
UAV to GCS communication network, e.g., the authors in Hooper et al. Moreover, in case of decentralized UAVs network architecture, the
(2016) have presented three attacks that effect commercial UAVs based attackers need to compromise only a specific UAV for failure of flight
on Wi-Fi. These include (1) DoS attack, (2) Buffer overflow (3) ARP mission, as shown in Fig. 8. Besides, in case of single backbone UAVs
cache poisoning. Yet, their results indicate a lot of security issues to network, attackers just send malicious traffic to backbone UAV, because
GCS network. Besides, selection of adequate wireless communication backbone UAV serves like a gateway among GCS and other UAVs. The
technologies depends on specific mission requirements, such as oper- whole network is disturbed when the backbone UAV is compromised.
ating frequency, category, and transmission range, etc. Still, this does Hence, the weakest link is constituted by single backbone UAV in the
not ensure success of flight due to consideration for security issues UAVs network. In this architecture, the security of the whole network
in these technologies. Thus, it remains undecided, regarding which depends on single backbone UAV, security level of GCS, and commu-
wireless communication technology offers the highest level of security nication link between them. Yet, if UAVs are same in size, shape and
for UAV. color, it becomes a bit for difficult attackers to identify the backbone

10
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 5
Communication level vulnerabilities, existing countermeasures, and their limitations.
Ref Network and Transport Countermeasure Limitation
Layer
Deligne (2012) Attacks on Communication Developing a high-level architecture protective against Need to implement and maintain a balance
Protocols attack, with the ability of pliability and reliability. between security and performance of trade-offs.
Security services can be provided by embedding into a
hardware module. Encryption and IDS
methods utilization’s are important tools for protection.
Emerging technologies like Blockchain provide security
against attacks.

He et al. (2017) Eavesdropping Implementing highly secure encryption approaches and IDS Time delays due to computation overheads.
DoS/DDoS solutions can be a fruitful tool to guard against these attacks. Limited routing protocols provide security feathers.
Blackhole Secure routing protocols also enhance security. Performance dent of GCS to UAV communication.
Flooding Attackers change the pattern to compromise
Wormhole signature-based-IDS. False positives and false
Forwarding attacks negatives defeat anomaly-based IDS.

Fig. 8. Single backbone in a decentralized architecture.

Fig. 7. Centralized architecture.

UAV. In more than one backbone UAV architecture, backbone UAVs

and GCS of every swarm are especially critical to the success of mission.
Although, the backbone UAV gets compromised but mission might
be completed. It can be observed from Fig. 9, that attackers must
compromise the GCS or four backbone UAVs for disrupting complete
UAVs network. Further, advanced architecture of UAVs network is by
mixing UAVs, as shown in Fig. 10. In this architecture, securing the
network of backbone UAV is equally important as the whole network
security, because the threats are prone to increase of complexity and
number of UAVs.
Additionally, the UAV protocols for routing are vulnerable because
of limited resources, dynamic topology, and lack of encryption in
wireless links (Maxa et al., 2017). In such situations, attackers take
advantage of such constraints for performing numerous routing attacks
at network level. The attacker may disclose sensitive data in UAV
network which does not have security mechanisms.
Adversaries can leak routing and topology information as well as
UAV location by using eavesdropping techniques (Maxa et al., 2017).
Fig. 9. Swarm of UAVs in a decentralized architecture.
Therefore, without considering authentication and integrity, routing
protocols of UAVs are prone to attacks, like route-cache poisoning (Wu
et al., 2007), where the attackers add incorrect routing information into
UAV cache.

11
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

protocols were proposed. The nature of UAVs’ application decides the

use of these security protocols, while there are three main categories of
UAVs security protocols. The first is secure communication. The second
is security of the physical layer and the third is intrusion detection
system. Fig. 11 shows the emerging defence technologies for UAVs.

4.1. Secure communication in UAVs

UAVs can observe a wider area without taking any aid from a
network. While in flight, UAVs continuously exchange sensitive infor-
mation by communicating with GCS, many new challenges are created
by the exchange of information due to dynamic topology. Since UAVs
are often used to transmit data from any node to GCS, attackers can
attack the transmitted data in different ways. In maximum military
application, important information is transmitted between authorized
users via wireless communication channels (Khan et al., 2020). As these
communication channels are insecure mediums, so it is easy to steal
information by launching cyber-attacks, like confidentiality attacks,
integrity attacks and availability attacks. To protect the information
from attackers, numerous security protocols are being used.
These security protocols help in securing the transmission and in
authentication of users, for example, symmetric protocols for UAV
security and asymmetric protocols for security. These two types of
Fig. 10. Mixed UAVs in a decentralized architecture.
protocols are used for securing communication among UAVs and GCS.
Whereas, a private key is used for encrypting and decrypting data
in symmetric cryptography security protocols while asymmetric cryp-
4. Emerging defence technologies for UAVs tography protocols use pair keys (one public and one private). In
these protocols, data encryption is done through a public key while
No doubt, UAVs have brought different benefits including commer- decryption is done through private key. Asymmetric and symmetric
cial and personal benefits. Yet, there are many security, privacy and protocols are scribed in the following sections and after that lightweight
safety drawbacks (Kafi et al., 2013) through which cyber criminals authentication protocols that are also discussed.
can use to invade the safety and privacy of individuals as well as
the general public. Moreover, many attributes of UAVs are used in
4.1.1. Symmetric cryptography protocols
attacks including unauthorized inspections and high-level operations.
These days, cryptographic protocols are used for ensuring confi-
Similarly, images taken by these UAVs might be used to carry out
dentiality, integrity, and availability. Particularly, symmetric crypto-
illegal activities like scamming. Therefore, UAVs should not be used for
graphic protocols are being used for protecting sensitive data like text,
capturing images and for recording the videos of individuals without
audio, video and images. While in these protocols, sender and receiver
their consent (Peterson, 2010). Currently, majority of UAVs are Wi-
both share the same key for encryption and decryption. For example,
Fi enabled for broadcasting the captured videos to smart devices,
One Time Pad (OTP) is used frequently for securing transmission. For
attackers easily access Wi-Fi with weak password and interfere with
this purpose, OTP needs to have the same size of data. When an
the communication (Jones and Kovacich, 2015). Thus, UAVs also pose
image containing pixels of 𝑁 rows and M columns is transmitted, key
serious threats to public.
should be same as the size of the image (Atoev et al., 2019). The OTP
Similarly, attackers may use unauthorized UAVs for destroying
encryption is also applied to increase security of Micro Air Vehicle link
authorized UAVs with the help of physical collisions. The unauthorized
(MAV link) (Atoev et al., 2019).
and authorized UAVs run into each other frequently, so it is important
to prevent any collision among them. Further, decryption and encryption methods are used for securing
Furthermore, different algorithms are proposed in literature for the data transmission. Here many commands are used to control UAVs,
preventing collision among UAVs, e.g., using Sense and Avoid (SAA) like start, and takeoff commands. Whereas all commands are given
algorithm (Zeitlin et al., 2006) to sense UAV and referring to the in a bit format which can be shown by 0 to 1. In addition, a long
hurdles placed by an attacker. Barfield has presented another model text, secured by using encryption, is created with the combination
for sense and avoid. In this method, he proposed an independent of these bits. With OTP encryption, this mechanism also has some
collision prevention system that could protect UAVs from accidents. disadvantages, for example, the size of key and data should be equal.
Also, many algorithms aiming for collision avoidance have been de- In order to send large data, it is necessary to share a key at the receiver
signed. These algorithms have helped in accomplishing some crucial side. Therefore, key sharing is a problem because it consumes most
challenges including Individual Collision Avoidance (ACA) and Group of the bandwidth. Also, since the key might be used for once due to
Collision Avoidance (GCA) (Sharma et al., 2020). Yang et al. (2014) authorization risk, a new key for each transmission is required (Atoev
have presented another method in 2014, this is based on a 3D path et al., 2019).
organization for UAV that comprises of tracking down a collision- Other than this, an algorithm in Atoev et al. (2019) can be improved
free path. Moreover, Ueno and Higuchi (2011) have given a novel by implementing robust transformation mechanisms like discrete cosine
algorithm that is capable of correctly locating the objects in UAVs transformations and discrete wavelet. This type of transformation is
vicinity. Authors in Brandt and Colton (2010), have argued that UAVs faster than transforms performed on actual message directly (Ma et al.,
are more appropriate for operating indoors. This is because of their 2016). Likewise, a Chaotic Lorenz system presented in Kirichenko
flexibility as well as well controlled operations in confined and small (2015) can encrypt, decrypt those actual and transformed messages
areas. In Israelsen et al. (2014), presented an algorithm to manually with long term unpredictability. This Chaotic Lorenz may generate
control UAVs by applying automated obstacle collision avoidance. higher randomness with minimum differences in seed values.
Avoiding UAVs’ collisions is also crucial to secure communication In the Chaotic Lorenz system, UAVs collect data from cameras,
among UAVs and GCS. To secure this communication, different security sensors and later pass data to the Lorenz encoder. However, no direct

12
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Fig. 11. Emerging defence technologies for UAVs.

encryption is applied to plain messages, Instead all information is Furthermore, application of asymmetric cryptography through ADS-
changed into bits and later encrypted. Following encryption, UAVs B is costly as well as time consuming, because different keys are
send encrypted information towards a receiver. While the receiver provided for encrypting and decrypting the actual text. The commu-
then decrypts information by reversing the Lorenz system. This en- nication security protocols may also be used for ensuring data integrity
cryption algorithm lies in the category of symmetric protocol. This being transmitted. Diffie as well as Hellman presented a protocol for
protocol (Kirichenko, 2015) also has some drawbacks, e.g., no data key exchange between the sender and receiver (Bresson et al., 2007) a
scrambling mechanism is present in this system, whereas the security of few decades ago. However, both parties have no prior information that
encryption algorithms relies on both diffusion and confusion (Shannon, keys sent over the communication channel are shared by authorized
1949). person or not.
Additionally, in Sahingoz (2013) authors have proposed a protocol
4.1.2. Asymmetric cryptography protocols for exchanging public keys. In this protocol, keys are exchanged by
Asymmetric protocols for security use two separate keys including UAVs sensor nodes, and communication between nodes starts after
one public and one private key. These keys are used for encryption authenticating the surrounding nodes. In this framework, sensor nodes
and decryption of data by sender and receiver, respectively. While are actual communication parties. Here, one sensor sends a message
the confidentiality of a public key is not very important since the encrypted with a public key, and another sensor decrypts message
information encrypted through a public key cannot be decrypted using using a private key and random numbers generated by the sensor itself.
the same key. So, there is always a need for a private key to decrypt While, the other party transmits an encrypted message using public
data. For this, the author has proposed a mechanism in Wesson et al. key. The cipher message is then decrypted by the first party through
(2014), for checking the originality of the data. This protocol verifies a private key as well as random number. If both parties have the same
whether data is sent by the original GCS or eavesdroppers. This protocol messages, then the sensor is considered as an original entity and will
falls into the category of asymmetric protocols, where asymmetric be allowed to continue communication.
protocols are used for ensuring the integrity of data being transmit- developed trust-based security protocol to secure UAVs (Valentin-
ted among different devices or sensors. With the help of X.509, the Alexandru et al., 2019). In this protocol, sensors of UAVs are assigned
length of generated signatures has become shorter. Thus, this made the with trust values to check accuracy and correctness of data. This
authentication process faster. Another scheme presented in Pan et al. methodology has three modules: (1) direct determination of trust value,
(2012), verifies signatures once the data is received by receiver. Here, (2) indirect determination of trust value and (3) final determination of
the verification process is executed by UAV (after receiving the data trust value calculated by UAVs. Here, each sensor determines its own
from GCS) to confirm confidentiality before executing the last action. trust value. If an attacker places a sensor, that sensor can also produce
For verification, a 164 bits hash is created through SHA-1 Algorithm on trust value, but UAV will determine the ultimate trust value. For this,
sender sides. This hash is encrypted through a public key before it is trust values saved in the UAV log file will be compared with those
sent to the other side. While, the private key is utilized for decrypting received by a UAV. If the result of comparison is a negative trust value,
the hash on receiving side. Later, the hash is calculated by receiver for then sensors are not trustworthy, thus UAV will not take more data
the original message. Finally, both hashes are compared by the receiver from sensors and attacker will not have any access to UAV. Contrary to
to check the originality of message. this, a positive trust value indicates that sensors are trustworthy, UAV
Moreover, security of Automatic Dependent Surveillance Broadcast can take further communications with sensors. While zero trust value
(ADS-B) can be enhanced by asymmetric protocols (Wesson et al., shows that more information is required to determine whether sensors
2014). However, ADS-B, being a protocol for air traffic surveillance, is can be trusted.
insecure. As it is used for detecting the other surrounding UAVs. ADS-B Yoon et al. (2017), have implemented a protocol that can verify
has some disadvantages like it lacks built-in mechanisms for security. whether information received by UAVs, is transferred from an attacker
Next, Wesson et al. (2014) have investigated either an asymmetric or original GCS. In this scheme, a UAV sends randomly encrypted mes-
protocol can increase security in ADS-B or not. For this purpose, they sage to GCS and the data received by GCS is decrypted. For decryption
have evaluated the existing security protocols based on their ability to GCS use a private key and then encrypts again using a public key.
ensure security of ADS-B. Then after evaluation, they have contended Finally, the data is sent to UAV where it is compared to values stored
that key elliptic curve computerized signature is more practical. in UAV log file. If the verification is successful, the UAV can take off. If

13
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

verification fails, this tells that UAV attackers are trying to get control to make the system lightweight, selective encryption is performed.
over UAV. Hence, no further communication between sender (GCS) and In this algorithm, other than cryptographic method, watermarking is
receiver (UAV) will take place. In this mechanism, only random text also used for increasing data confidentiality and integrity. The rea-
encryption and comparison are used for checking if the UAV is hijacked son behind using specific encryption is to offer stabilization among
or secure. This scheme provides authentication of data. Yet, in case UAVs within limited resources. However, selective encryption has an
of larger data transmission for UAV authentication, high bandwidth, advantage, particularly in the case of real-time application where quick
processing time and cost are required. Therefore, instead of whole text processing is necessary. To point out authentication problems and data
encryption, a small hash can be used to overcome these issues. insecurity, a two-step lightweight correlative authentication protocol
Steinmann et al. (2016) have developed a significant negotiation is implemented in Alladi et al. (2020b,a). In this method, a Software
framework for providing security and confidentiality to data stored Defined Network (SDN) is assisted with the UAV network in necessary
in UAV chip. The major theme is to develop a scheme that changes spying areas.
random keys continuously. For example, One-Time PAD technique can Besides, evidence for security of the protocol is also provided
be used to encrypt data because in this technique, the size of original to highlight security features. For UAVs, Smart IoD assisted frame-
message and key are the same. In this method, if one key is explored, work is proposed in Deebak and Al-Turjman (2020), which collects
then whole message can be revealed easily. Therefore, random keys required information independently, to reduced computational cost
generation can increase the security of original message and keys. In through Lightweight Privacy-Preserving Scheme (L-PPS) (Deebak and
this mechanism, the sender sends a message to the receiver along with Al-Turjman, 2020). This scheme provides robustness among IoD devices
a hash code after encrypting with a public key. The receiver uses a along with a proper authentication time. Due to of limited resources
private key for data decryption, calculates value of hash, and later and substantial risk around UAV, attackers can perform numerous at-
compares both received as well as calculated hashes. If both hash codes tacks including confidentiality, wireless and Man-in-the-Middle (MitM)
match, then it ensures the integrity of the message and authentication attacks. To protect UAVs from these attacks, authentication is necessary
is implemented through this message, but it is not good for keeping to implement prior to start of UAV communication. It is also crucial
keys a secret. that implemented authentication mechanism guarantees that original
UAVs is prime concern of UAVs network security. Table 6 presents a
4.2. Lightweight authentication protocols for UAV comparison among security protocols for intrusion detection and secure
communication in UAVs.
Lightweight encryption protocols, belong to another class of secu- Furthermore, standardized authentication mechanism have an user-
rity protocols used to hide confidential information from intruders. name and a password which makes private key less secure. Hence, the
Also, these lightweight protocols can help to reduce the information RSA certification needs long-lasting keys that cannot fulfill lightweight
encoding time. Lightweight protocols do not take much memory which demand in UAVs infrastructure. Another, lightweight recognition mode
permits UAVs to perform faster. A lightweight protocol is implemented for authentication based on Elliptic Curve Cryptography (ECC) is de-
in Driscoll (2018), which is more suitable for switching context fre- veloped in Teng et al. (2019). It has three phases: (1) initiation of
quently, in a highly multi-task environment. Next, in Wang et al. ECC certification, (2) identity authentication and (3) verification of key
(2021) developed a blockchain-based solid routing scheme for Un- compatibility. The authors have mentioned that phase one and phase
armed Arial System (UAS) networking. In this algorithm, lightweight two completely match the two-way authentication. While the third
blockchain is used like a bargaining chip for strengthening routing phase validates consistency of verification keys. Contrary to traditional
of UAS, using 5G cellular technology. This algorithm is not the same modes of authentication in a UAV network, the method proposed
as traditional algorithms for routing, because it can easily identify in Teng et al. (2019) depends upon short keys as well as reduced
spiteful UAS, reduce attack intensity and avoid the attackers. Moreover, computing workload. In Barka et al. (2018), Barka et al. proposed
the proposed algorithm is intended to secure UAS and expand the lightweight communication technique for aerial Named Data Network-
deployment of swarm UAS network over a wider range. For minimizing ing (NDN). This technique can assist NDN security, also has 80%
effects of cyber-attacks sessions keys known by both participating prediction accuracy. This technique has reduced end-to-end pause to
parties should be used for data encoding. lower than a second in the worst conditions.
Conversely, it is difficult to get required abilities for both encoding However, the lightweight security protocols suit UAVs well, because
and session keys with low-cost IoT integration, due to limited perfor- UAVs have small batteries and limited memory. A lightweight security
mance. Demeri et al. (2020) have implemented an integrated secure protocol is presented by Srinivas et al. (2019), known as tempo-
data transfer mechanism by using low-cost arial policy that integrates ral Credential-Based Anonymous Lightweight Authentication Scheme
numerous cryptographic accelerators. In this method, the components (TCALAS) for IoD networks. The scheme proposed in Srinivas et al.
are combined through moldable Application Programming Interference (2019) can work in condition only where there is a single flying zone.
(API), hardware and software design yielded costless UAVs. yet, with Further, the exploration provided in Ali et al. (2020) indicates, the
recent advancements in wireless communication systems as well as scheme proposed by Srinivas et al. (2019) does not prove to be efficient
electronic devices, great relaxation is being offered to the public by for availability and traceability attacks.
UAVs. In Ali et al. (2020), an upgraded version of TCALAS scheme, uT-
The UAV security is getting higher attention due to emerging secu- CALAS is provided by Ali et al. (2020). This method provides security
rity issues, financial as well as strategic information, and significance against different attacks including availability and traceability attacks.
associated with UAV applications. For providing privacy and security It increases extendibility and can work in numerous areas, where
to communicating parties, lightweight protocol for authentication is more different flying zones are available. Besides it has successfully
proposed in Pu and Li (2020). This protocol offers safe communication achieved fast computation for authentication by only taking 2.29 ms
between the GCS and UAVs. This framework has also used Packet cap- to accomplish the whole authentication process.
ture (PCAP) for ensuring secure communication between both parties.
The prime idea behind PCAP is, UAV as well as a GCS using seed values 4.3. Security of UAV physical layer
from chaotic maps. These value can shuffle actual message randomly
based on chaotic sequence (Pu and Li, 2020). One frequently adopted performance measure in security of physical
Additionally, the concept of lightweight along with data security layer is rate of secrecy at which data is transmitted securely (Li et al.,
is discussed by literature (Haque and Chowdhury, 2017). Another, a 2014), but conventional encryption protocols are vulnerable due to
novel framework was developed for achieving the desired task. Here, key distribution as well as longer processing time. The examination

14
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 6
Comparison of various protocols used in UAVs.
Classification Method Feature Limitation
Symmetric Security protocols One time PAD communication links Secure information due to large key size Extra bandwidth is necessary for sharing
the secret key
Information security Protocols Randomness Lack of confusion part makes the system
insecure
Asymmetric security protocol Information authentication Computationally faster due to short hash encrypted form of Hash, attacker may
length try to perform
eavesdropping attack to retrieve the
original hash
Inspection of cryptography in Automatic Elliptic curve digital signature algorithm Costly and time inefficient
Dependent Surveillance-Broadcast is feasible
Public key sharing Authentication of the received Select only one random number for
commands from the GS public key, An attacker can
easily judge the key
Trust-based security protocols Verify whether sensors placed in UAV Packet can be Fabricated
network are trusted or not
Authentication scheme Protects data even after hijacking the Costly and time inefficient
UAV by the hacker
Convex optimization method Decreases eavesdropping effects. Costly and time inefficient
Lightweight authentication protocols Blockchain-based Maintains 5G cellular network Different than traditional algorithms
Authentication using public key Use of moldable and extensible API Slower than traditional authentication
technique
Packet capture Physical unclonable function less dimensional chaotic map is used
SDN-based authentication Combine of security and SDN Costly and time inefficient

of physical layer properties can make transmission secure, besides the least square (Ra et al., 2008). Authors in Bejiga et al. (2017) have
security of physical layer is useful to get maximum secrecy rate for proposed a method, It is used to calculate jamming signal’s using
transmitted data over UAVs network. The physical layer security is convolutional neural network (CNN). In this algorithm, UAV receives
crucial for all communication devices and security controls of UAV. values of weights and that from GCS. After that, Bit Error Rate (BER) is
Unlike traditional approaches of cryptographic security, Physical used to select a relay power element. If learning parameter has a greater
layer security (PLS) benefits from the attributes of cellular channels value as compared to relay power, then a random relay is selected
like interference, fading, noise for boosting signals reception at re- by the device. If it is greater than zero, UAV uses reinforcement and
ceiver end and reducing the quality of signals for eavesdroppers (Yang randomly selected values to send a message. The error rate can be
et al., 2015). Moreover, PLS can be ensured with incorporation of increased due to randomly selected relay power with above features,
cryptographic protocols. There are numerous cryptographic protocols this algorithms may prevent communication and UAV from jamming
presented in literature. These protocols provide a significant security attacks, but it might be expensive when the error rate is high.
level, but unfortunately there is no protocol that can provide ideal Besides, a technique for attack detection is presented in Hoang
security. Hence, PLS is getting more attention. et al. (2019). In this technique, K-mean clustering and Support Vector
Additionally, to maximize secrecy rate and increase security of Machine (SVM) are used. Such algorithms have the ability to learn
UAV communication, many works have been proposed on PLS (Choi from available data and then make a decision for future samples.
et al., 2014; Li et al., 2019a; Zeng et al., 2016). In recent years, The proposed technique has two phases. The first phase involves both
static relay-based systems for communication have gained importance parties to send signals to a UAV and in the second phase, UAVs transmit
for improving existing PLS algorithms. A novel framework relying on the same signals to third party to detect deviation. When the third
a technique, called UAV empowered mobile relay, is considered as party takes the signals, creation of a dataset is required to classify the
valuable technology. Authors have developed an upgraded version of output using ML algorithms. Here, signals are received by a third party
PLS algorithm based on UAV enabled mobile relying (Wang et al., continuously and mean as well as standard deviation is found for each
2017). For improving communication system security, buffer aided signal (see Table 7).
mobile based relay is implemented that permits quick and independent Consequently, standard deviation as well as mean become two
data arrival. separate feature points in dataset. Common types of attacks detected
by UAV MLIDS are shown in Fig. 12.
4.4. Intrusion detection system Once the model training is completed the third party gets the new
sample, it is given to ML algorithms, where the labels (as Attacked
4.4.1. Machine learning based IDS and Not attacked) are assigned according to the calculated values. For
Digital machines can perform multiple tasks depending on user increasing the accuracy of model, there should be more features in ML
instructions. In these machines, for the automation of tasks, different data set and this ensures that there is no event declassified. Lastly, the
mechanisms and techniques are used, like, Machine Learning (ML), accuracy of the model is compromised because only two features are
Neural Networks, and deep learning are used commonly. There are two used .
phases in ML algorithms including training and testing. In training, the On the contrary, maximum approaches are focused on ML algo-
model is trained with the help of data to predict future events depend- rithms for addressing vulnerabilities, and the most secure as well as
ing on training. The testing phase is used to measure the accuracy of the trusted process among them is data analytic. Yet, it is not true in each
model with the help of different strategies. Also, these ML techniques case (Chen and Chen, 2019) because such techniques have not received
can also be applied to UAVs to detect intrusion. much attention from research communities. The security in big data
Moreover, a novel algorithm was presented in Lu et al. (2020). analytic is necessary, because it must deal with sample amounts of sen-
This algorithm is based on deep reinforcement as well as weighted sitive data. This data requires guaranteed confidentiality, availability,

15
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 7
Machine learning scheme for UAV intrusion detection system.
Ref Attacks on UAV Machine Learning approaches Algorithm
Hoang et al. (2019) Eavesdropping Man in the middle attack Unsupervised One-class SVM K-means clustering
Xiao et al. (2018) Jamming Supervised Hill climbing
Xiao et al. (2017) Jamming, Spoofing, Eavesdropping Reinforcement learning Deep q-learning
Liu et al. (2019a) GPS spoofing Supervised learning K-nearest neighbors
Panice et al. (2017) GPS spoofing Supervised learning SVM
Arthur (2019) Jamming and Spoofing Reinforcement learning SVM, Deep Q-learning
Liu et al. (2019b) Eavesdropping attack Reinforcement learning Dyna-Q-based
Sedjelmaci et al. (2017) Black hole, Jamming, Gray hole Supervised learning SVM
Vanitha and Ganapathi (2020) Multiple type of attacks Hybrid approach SVM,KNN, K-mean clustering
Lu et al. (2017) Jamming and Spoofing Reinforcement learning Deep q-networks
Manesh et al. (2019) GPS spoofing Semi-supervised KD-trees
Mitchell and Chen (2013) Different attacks Supervised learning Feed forward neural network
Eavesdropping Reinforcement learning Deep Q-networks
Min et al. (2018) Interception of malicious UAVs. Reinforcement Learning Deep Q Learning
Jamming Supervised learning SVM, K-NN
Feng et al. (2020) GPS spoofing Supervised learning Genetic algorithm, SVM
Li et al. (2019b) Jamming Reinforcement learning Reinforcement learning
Lin et al. (2019) Jamming Reinforcement learning Deep Q learning
Challita et al. (2019) Cyber–physical attacks Ai neural network Artificial neural networks
Abbaspour et al. (2016b) Fault data injection Neural network Neural network
Sheng et al. (2018) Multiple type of attacks Unsupervised Ant-colony clustering
Lu et al. (2020) Jamming Neural network Deep conventional neural network

data analytics (Darwish and Bakar, 2018). Some studies have been con-
ducted for approximation algorithms’ integration based on ML (Bangui
et al., 2020). For example, data approximation for detecting continuous
jamming attacks.

4.4.4. Machine learning based IDS datasets

These datasets have a significant role in validation of ML based IDS.
Currently, numerous intrusion detection datasets are available over the
internet (Anon, 2018). However, there is a need for creating many
other IDS datasets in UAV research for exploring recent attacks over
IDS. Mostly, research work used KDD CUP dataset (Anon, 1999) for
validating IDS models. In these works, other realistic datasets like CIC-
IDS-2018 (Anon, 2018) are not considered. While, it is necessary to
investigate and learn the features of these attacks for building a reliable
IDS model. It is significant to use multiple datasets for evaluating ML
based IDS, this will be useful for reinforcement learning models and
estimating learning impact for numerous attacks in real-time scenarios.
Thus, new dataset is required that might be paramount to train and
validate credibility of IDS for UAV environments.

4.4.5. Training data security

Fig. 12. Types of attacks detected by UAV MLIDS. In the future, there will be potential but unknown vulnerabilities
due to an increase in smart applications, therefore IDS and Intrusion
Prevention System (IPS) must be more reliable for mitigating attacks.
Yet, single IDS cannot detect all types of attacks because of limited
and integrity. Hence, data-driven security needs particular attention to
information related to unknown attacks. For this, a collaborative IDS
support advances of AI based transport systems in a better way.
can be a solution to detect all attacks efficiently and effectively. On
4.4.2. Challenges of machine learning based intrusion detection the contrary, this collaboration also has certain risks, because expertise
An important asset for IDS is big data, because it is effective to shared by each IDS is based on results of classification as reliable
protect against known as well as unknown attacks, so as to help feedback. Other collaborators may use this to choose whether to give
identify abnormal nodes. Still, there are some critical data issues that an alarm or not, still an attacker can hack and alter the training data
need consideration to address, e.g., challenges relevant to UAVs data as well as modify the classifier’s function for launching other threats.
affecting security in UAV communication are identified. Besides, these threats can immediately impact the accuracy of detection
in IDS as well as collaborators. In Chen and Chen (2019), authors have
4.4.3. Intensive data computation particularly highlighted the significance of data security, they have also
No doubt, the advances in ML techniques have helped in intrusion discussed how attackers might use ML classifiers for decrypting UAV
detection and prevention. However, challenges of UAV data augmen- messages. There is still need to develop new techniques for ensuring
tation are still not well addressed. As stated by Intel, each day, 4 confidentiality of the shared attacks in collaborative IDS.
terabytes of data will be generated by UAV and more time will be
required for big data analytics. For protecting communication in UAV 4.4.6. Fairness and trust
networks, lightweight ML based IDS are highly recommended for re- A classical IDS is deployed in a cloud as it has sample computation
ducing computation time, the data approximation might be a solution. resources. Yet, lightweight IDS may need deployment in a distributed
Moreover, this can be done by choosing highly valuable data for big environment. There is need to timely minimize network problem and

16
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

protect against malicious attacks. For this distributed models like feder- by authors in Choudhary et al. (2018). According to information-
ated learning, deep learning, and reinforcement learning have attracted gathering sources, deployment tactics, detection strategies, detection
researchers’ attention. For instance, deep learning models can achieve stages, IDS acknowledgment, and intrusion kinds, they categorized
prediction with high accuracy. Unfortunately, these models mostly do the existing IDS mechanisms. The potential software and hardware-
not display how internal features are extracted because these models based assaults on the UAV power devices were briefly examined. The
assume that all their predictions are reliable, without any explanation researchers have only suggested developing IDSs to protect the flight
regarding the reliability of outcomes. Therefore, it is still a question that control system from such onboard attacks; however, there has not been
how these models can explain transparently the prediction results. Like- any implementation or discussion of such a system or the necessary
wise, what might be a typical form of such explanations for ensuring parameters (Rahman et al., 2020b) .
fairness and trust of IDS in UAVs? Moreover, the misbehavior of node might be temporary due to some
environmental condition like rain, fog and extreme temperature etc.
4.5. Signature based intrusion detection There is risk for higher false positive rate due to immediate ejection
of node. Table 8 presents a signature based schemes for UAV’s IDS.
For making devices intelligent, instructions must be given to these There are four significant factors that were used for analysis of these
devices and rules should be defined in rule-based tasks, so devices act techniques. These are True Positive (TP), False Positive (FP), True
depending on these rules. While in UAVs, different rules are stored in Negative (TN) and False Negative (FN). These factors can be used
the chips and levels for rule acceptance are created. A novel IDS is pro- for finding statistical parameters like precision, accuracy, F-Score, and
posed based on rules of behavior for minimizing false-negative predic- recall.
tions (Sedjelmaci et al., 2017). Seven attacks related to integrity, con- The accuracy tells the numbers of predictions made correctly by
fidentiality, and availability were discussed in this detection method. a model and correct predictions mean higher accuracy. The following
When any of these attacks is detected, the UAV takes defensive action method can be used to calculate accuracy:
as below: 𝑁𝑜. 𝑜𝑓 𝑐𝑜𝑟𝑟𝑒𝑐𝑡 𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑖𝑜𝑛𝑠
𝐴𝑐𝑐𝑢𝑟𝑎𝑐𝑦 = . (1)
• Firstly, when a UAV successfully reaches a secure location, it turns 𝑇 𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑃 𝑟𝑒𝑑𝑖𝑐𝑡𝑖𝑜𝑛𝑠
on weapons for defending against attacks. The next parameter Precision can be calculated by ratio between
• Secondly, an action is taken when there is difference in sensor sum of true positive, true negative and true positive. A better model
reading taken from trusted nodes. shows high precision. Precision is calculated as below:
• Thirdly, the proper action is taken when good recommendations 𝑇𝑃
are received for misbehaving UAVs. Similarly, it is required when 𝑃 𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 = . (2)
𝑇𝑃 + 𝐹𝑃
unrighteous recommendations are received for trusted nodes.
Another parameter is Recall. It is the sensitivity of a system. It is
• Fourthly, when landing gear is deployed by a UAV in an inappro- calculated as below:
priate location, the indicator is used for handling it.
𝑇𝑃
• Fifthly, when a UAV begins to send data to unauthorized party, 𝑅𝑒𝑐𝑎𝑙𝑙 = . (3)
𝑇𝑃 + 𝐹𝑁
the indicator is activated.
The final parameter is the F-Score, where Recall and Precision are
• Sixthly, when a UAV uses countermeasures without investigating
used in calculation. It is weighted average of Precision and Recall. The
an attack the attack indicator is activated.
F-Score lies between 0 and 1. 1 tells that Recall and Precision are
• The seventh attack occurs when more trust is used by UAV to
perfect. While 0 shows that either Recall or Precision is 0. F1-Score
cross the restricted altitude.
is calculated as:
The Last two attacks correspond to availability attacks, and a de- ( )
(𝑅𝑒𝑐𝑎𝑙𝑙)−1 + (𝑃 𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛)−1 𝑅𝑒𝑐𝑎𝑙𝑙 × 𝑃 𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛
fence phase is initiated by UAV defence for protection against these 𝐹 1 − 𝑆𝑐𝑜𝑟𝑒 = =2 ×
attacks. No doubt, IDS can be used for the detection of deviation in 2 𝑅𝑒𝑐𝑎𝑙𝑙 + 𝑃 𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛
a network. Thus, IDS undoes the impact of attacks to prevent UAV (4)
from hazards, therefore, in UAV network IDS is the main mechanism
for detection of malicious attacks as well as for protection from these 5. Blockchain application in UAV networks
attacks.
Intrusion detection along with ejection of malicious node was ad- At present, blockchain technology is emerging to solve increasing
dressed in Sedjelmaci et al. (2016). For more accurate detection of problems related to UAVs network. This section reviews the state-of-
intrusion, a technique based on Bayesian model was given in Xu the-art in UAVs having deployment of numerous versions of blockchain
et al. (2020). It focuses on the detection of internal intruders and the as shown in Fig. 13. Blockchain in UAVs is deployed for exploiting
ejection of nodes which might be harmful for the network. Initially, many benefits like transparency, immutability, auditability, and trace-
different nodes activate the system intrusion detection mode. For this, ability. Latest applications of blockchain enabled UAVs for networks
misbehave rate (MR) is calculated by IDS for neighbor UAV which are presented in Table 9.
is in the network. If the threshold value is lower than MR, then IDS Cybersecurity works as an inclusive aim to deploy blockchain tech-
begins monitoring of neighbor nodes and detection system is activated, nology in UAVs networks. Certainly, the feature of immutability in
otherwise no task is performed by IDS. The MR of a node is computed this technology makes it perfect to ensure accountability of attackers.
by intrusion ejection system, when the threshold value is lower than Many existing works related to deployment of blockchain in UAVs are
MR, node is declared as malicious and rejected from network. motivated by cybersecurity needs. These needs include availability,
Further, hybrid approach for anomaly estimation inside UAV net- integrity, non-repudiation, authentication, and confidentiality. Though,
works based on both spectral traffic analysis and a reliable con- most of the existing solutions address all these aspects simultaneously,
troller/observer (Condomines et al., 2019). The suggested hybrid so- but we group them based on a dominant one. Since, confidentiality
lution considers a quantitative signature of the data exchanged in the is the protection of information from unauthorized users. Therefore,
network. The discrepancies are utilized to choose the most accurate UAVs networks are also prone to confidentiality attacks. In Wu et al.
model for a precise assessment of that irregular traffic,by looking at the (2021), Wu et al. have described different scenarios where blockchain
resulting signatures. The second step is to robust observation based on helps to ensure UAV network privacy. They have proposed a tamper-
the observer or controller. Modern IDS systems that deal with vulner- proof and blockchain oriented ID management system developed for
abilities and assaults in networked UAV situations have been reviewed authentication and authorization of UAVs, according to requirements

17
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 8
Signature based schemes for UAV’s intrusion detection system.

IDS type Used method Benefits Drawback

Signature Based IDS Intrusion detection method (Lu et al., Prediction and Detection of reduced More rules mean higher process time.
2020) false-negative defends UAVs from false this will result in the delay of the
information being injected decision taken by the UAV.

Signature Based IDS Intrusion Detection and evil node If nodes are rejected immediately based Instead of immediate ejection of nodes,
(Sedjelmaci et al., 2016) on misbehavior there is a risk of more rounds are required. Once a
increased false positive rate number of rounds are complete average
will be taken for the final decision of
ejection regarding malicious nodes.

Signature Based IDS Hybrid IDS (Condomines et al., 2019) Hybrid experimental system to take Statistical signature.
advantage of the low cost of a
simulation while still obtaining the
accuracy of a real protocol stack

Signature Based IDS Hybrid IDS (Rahman et al., 2020b) Machine learning and Signature based More rules mean higher process time.
intrusion detection systems implement, it
can deal with malicious data/control
commands initiated due to hardware
attacks

Fig. 13. Blockchain based UAVs network.

of Federal Aviation authority (FIA). They grasped the potentiality of Another research (Khan et al., 2022) work has presented a coop-
Distributed Ledger Technology (DLT) for preserving privacy of tra- erative strategy based on the blockchain hyperledger fabric and the
jectory information related to UAVs. They have also stated that ho- B-Drone genetic algorithm for fog node management. This method
momorphic obfuscations strategies and asymmetric encryption devised secures the gathering, planning, optimization, processing, management,
in blockchain might be used for improving confidentiality of UAV and preservation of drone-based data in the fog node. The use of
networks. the hash-encryption (SHA-256) algorithm ensures the privacy of each
Next, data integrity is significant for ensuring that data transferred transaction between a drone and a fog node before exchange.
is not modified or altered by an attacker. Islam and Shin (2019) Federated Learning (FL) for UAV networks supporting B5G com-
have developed the architecture for preserving integrity of information munications with BC (blockchain) support is discussed in this sur-
transferred among Mobile Edge Computing (MEC) servers, devices, and vey (Saraswat et al., 2022). The technical advancement specifics con-
IoT (Internet of Things) devices. They have used UAVs as reliable relays cerning architectures, protocols, and concepts are offered. When com-
for ensuring integrity of information before transferring it to MEC bined with BC, it guarantees a reliable FL ecology. The core principle
servers. This framework has an encryption mechanism, to refine general of upcoming 6G or B5G networks, as well as their ability to support
integrity of transmitted data and lessens direct requests to MEC servers. massive UAV networks, are also discussed in this survey. Also, it
On contrary, in Singh et al. (2020) used blockchain for improving is concluded that the unification of FL and BC would be a better
the integrity of the data broadcasted among IoD environment and and more reliable solution. However, in order to meet the objective
UAVs. This mechanism efficiently chooses miner node by applying a of a complete integration for UAV enabled applications, 6G com-
Boltzmann machine. One more application requiring higher integrity of munication channels must still be the energy-efficient and reliable
data is power plant surveillance. In Kanade (2021), Kanade et al. pre- answers for UAV networks. To offer communication services over the
sented a safe power plant surveillance system using UAV swarms. They UAV network with dynamic, flexible, and on-the fly decision capabili-
have employed Distributed Ledger Technology (DLT) for preserving ties, blockchain-enabled UAV software may be used. For safe network
integrity of data sensed by a UAV. To decrease the cost of computation, administration and communication among UAVs, authors suggest a
transactions are allowed just in case of higher-risk signals. blockchain-enabled softwarization architecture (Kumari et al., 2020).

18
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Table 9
Latest application on blockchain enabled UAVs networks.
Ref CIA Triad Blockchain application
Wu et al. (2021) Confidentiality ID management system
Qian et al. (2020) Confidentiality Privacy of collected content
Xiao et al. (2021) Confidentiality Privacy of collected content
Alkadi et al. (2021) Confidentiality Privacy of collected content
Lv et al. (2021) Confidentiality Privacy of collected content
Khan et al. (2022) Confidentiality Privacy of collected content
Islam and Shin (2019) Integrity Overall integrity is improved by used encryption mechanism
Singh et al. (2020) Integrity Deep Boltzmann machine
Gupta et al. (2021) Integrity Integrity is improved by used encryption mechanism
Saraswat et al. (2022) Integrity Using federated learning for providing security between communicating nodes.
Kanade (2021) Integrity used DLT for preserving integrity of plant information sensed by a UAV
Yazdinejad et al. (2020) Availability Proposed decentralized zone-based system
Kapitonov et al. (2019) Availability Implemented rhobonomic protocol for providing security between communicating
nodes.
Kumari et al. (2020) Availability Implemented blockchain-based protocol for providing security between
communicating nodes.
Barka et al. (2019) Non-repudiation Trusted blockchain-based
Ossamah (2020) Non-repudiation Digital fingerprint

They provided a softwarization architecture for the 5G-TI UAV network work (Allouch et al., 2021), authors have implemented Unmanned
based on the blockchain. The authors have claimed that for its effective Traffic Management (UTM). The UMT Chain is a flimsy hyperledger
functioning, the 5G-TI permits a reduction in communication delay. fabric-based blockchain safety solution for low-altitude UTM. UAVs
However, they did not provide any experimental to prove the claim. are made to operate within the limitations of available computational
Hence, there is a need to test this system in a real-time scenario. Other- and storage power. UAVs may be targeted by jamming, hijacking, and
wise, the security and privacy issues in UAVs still remain a challenge. GPS spoofing. To determine the performance of the suggested solution
The study suggested a blockchain-based drone communication archi-
two factors were used. The first is the delay in transaction, calculated
tecture that would support 5G and AI to reduce security and privacy
by multiplying system threshold with confirmation Transection time
concerns (Gupta et al., 2021). Though this technique helps to minimize
and subtracting it from submission transaction time. While the second
security issues, but it cannot solve the problem completely due to
element (delay for read) is calculated by subtracting response time from
the following reasons: The suggested system 26 uses the Ethereum
blockchain, open to the public and available to all users. Therefore, submission transaction time. The experimental evaluation of the system
gathering and storing data in the blockchain offers security and en- indicated that the peer’s average and highest CPU usage, respectively,
hances UAV communication performance, but the sensitive information are 7.32% and 11.34%. Additionally, the peer’s average and extreme
of the users is also accessible. Thus, the threat of privacy and data memory use are separately 81.4 MB and 95.5 MB. The assembly node’s
integrity is still present. Using public-key cryptography, the information typical CPU and memory consumption were recorded as 75.2 MB and
on a blockchain network is securely encrypted (digital signatures). 7.53%, respectively.
However, public key encryption can be broken using the emerging As technology has advanced, the data control and aggravation
technology known as quantum computing. Therefore, it is necessary difficulties in UAVs have also increased. Another solution based on
to create a blockchain network that is quantum secure. blockchain technology (BCT) for improving security and privacy of de-
In addition, validation of block is given to humans for reducing vice data is presented in Ch et al. (2020). In this solution, fundamental
computational costs. An important concern in airspace is to maintain
data aboutvalidity, dependability, and reactions of the UAV is saved in
availability of service for UAVs, because blockchain is decentralized in
a cloud platform. This platform uses SHA and Pentatope-based Elliptic
nature and resistant to a single-point-of failure (Agrawal et al., 2018).
Curve Cryptography to safeguard the privacy of data. The information
All presented blockchain-oriented UAV Traffic Management (UTM)
is stored on a public blockchain built on Ethereum to support seamless
architectures are decentralized without having exceptionally low or no
central authority. A decentralized system is proposed for authenticating BCT transactions. An IoT-based application is used to test the proposed
and registering UAVs in Yazdinejad et al. (2020). In this architecture, methodology in a fictitious automotive monitoring system. The results
a reliable ground-based UAV controller agent is assigned for managing reveal that blockchain enhances the latency and processing time, which
authentication in a predefined perimeter. In this work, the availability increases overall time consumption. This is because there is now more
of authentication is maintained by permitting neighbor UAV-controller latency and processing time owing to involvement of the blockchain.
to replace a failing one. In context of UTM, researchers of Kapitonov This has led to an increase in data control and aggravation concerns in
et al. (2019) gave a blockchain architecture by implementing Robo- UAVs.
nomic protocol for providing security among communication nodes. A hybrid ML blockchain paradigm has been proposed in one of
They have argued that this architecture solves issues of latency found in the most recent research works (Abualsauod, 2022). A built-in UAV-
de-centralized systems. This mechanism is assisted with smart contrasts Network Simulator with Python as the implementation language is
that give immutability and transparency. Therefore, this architecture
chosen for the suggested hybrid solution for instilling security for IoT
is implemented through decentralized interplanetary file system and
services, FlyNetSim. The suggested methodology is evaluated against
ethereum.
many criteria, including latency, system utility and processing speed,
with the overall accuracy of attack detection. On a centralized server,
5.1. Existing and proposed reference architecture of blockchain application
in UAV network UAVs typically process the data using sophisticated machine learning
techniques. Data can be transferred to and stored in UAVs using all
In this section we have presents existing architecture based on types of conservative cyber attacks. UAVs rely on intelligent systems
blockchain application in UAV network. After that we have analyzed using machine learning techniques to make judgements in the ab-
these proposed architecture and then identify a hybrid solution ar- sence of people, despite their considerable effect. The system improves
chitecture based on ML and blockchain better. In a recent research reliability with the use of a hybrid technique.

19
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

5.2. Solution taxonomy 5.3.2. Absence of third party management

Despite it is decentralized nature, the P2P network offers huge ben-
efits and mitigates considerable security risks, but are still vulnerable
In contrast to merely protecting data integrity alone, the dynamics
to nearly 51% attacks (Saad et al., 2019). This is particularly terri-
of new services and technologies requirements in 5G present significant
fying for UAVs, where UAVs may leave and join without centralized
difficulties related to security and privacy. Recently, blockchain-based management (Su et al., 2020). To solve this problem, authors in Dasu
solutions for UAV security have drawn greater attention to address et al. (2018), proposed a hybrid model of traffic management. In this
these new 5G challenges (Puthal et al., 2018). In addition to main- model, a centralized authority manages parts of airspace, while others
taining secure UAV networks, blockchain-based technologies can be are tackled by block-chain based decentralized management. They have
used to assure the effective launch of 5G services. Immutability, de- also given a unique method for preventing DoS attacks by taking some
centralization, and transparency are essential security elements that fees from airspace users, on the local demand. This can also limit up
satisfy the increased service requirements in the 5G and B5G future. to 51% attacks by restricting attackers to join UAV with multiple UAVs
Blockchain is quickly replacing other security measures as the most simultaneously. One more way for improving security of the blockchain
popular technology for addressing new security issues. Moreover, it was proposed by Kim (Kim et al., 2019). Here, a blockchain governance
has the potential to change the 5G communication landscape (Nguyen game was developed for dealing with attackers, trying to get control
et al., 2020). The blockchain network uses public key cryptography over blockchain with addition of more illegal nodes. However, one
for secure functioning of blockchain. The integrity of data kept on should focus on trust issues and associated communication overheads.
the blockchain is less vulnerable thanks to blockchain technology. The
5.3.3. Limited storage and computational resources
execution of threats to integrity and availability can be significantly
It is a key hurdle when it comes to using blockchain in UAV network
impacted when blockchain technology is applied to every UAV in the
as stated in Mehta et al. (2020), Wu et al. (2021). Especially, the
network. Since each UAV maintains a copy of the blockchain, it will
most recent mechanisms are power consuming due to their design.
choose its own route independently of other UAVs across the network.
A large memory is required to keep the copy of blockchain UAV
Further, different ML and blockchain based solutions available to board, therefore focus has shifted to enhance. computational power and
secure UAVs have been analyzed in Khan et al. (2021). Numerous tests capacity of UAVs with minimum of power consumption. Still, a game-
and experiments have been conducted to determine the potential ap- changing proposal is required, to benefit from lightweight consensus
plications of machine learning and blockchain technologies to enhance mechanism (Su et al., 2020), because the trust and communication
UAVs security. Although the blockchain is a relatively new technology overheads are still concerned.
for decentralized UAVs and security, however, MLtechniques play a
crucial role in intelligently boosting UAV security. The overall reliabil- 5.3.4. Compatibility of blockchain ledgers
ity towards the security and quality of information can be increased by After review of literature, it is noticed the blockchain from each
adopting the hybrid blockchain model with ML. The obtained results research works are incompatible. The work in Rahman et al. (2021)
in the study show that ML in conjunction with blockchain can serve emphasized on permissioned blockchain for ensuring authentication of
miners. However, there can be security vulnerabilities if permissionless
the purpose of secure and reliable data communication for a number
architecture is used by other ledgers in cross-blockchain. This can de-
of UAV application areas. Such models can support a variety of appli-
crease the security level provided by permissioned blockchain. Hence,
cations, especially in healthcare and other sectors where security and
certain blockchain can be used either as private or public, permission-
quality cannot be compromised (Newaz et al., 2021).
less or permissioned. If combined, levels of security of a whole system
would be based on the least-protected component. One more issue that
needs research is either integrating numerous blockchain ledgers will
5.3. Open issues and challenges cause by blockchain application in UAVS
create new vulnerabilities in a system or not.

Both UAV network and blockchain interoperability are under devel- 6. Challenges and future direction for research
opment with great efforts. This section describes challenges as well as
research opportunities to guide future work. The following are major There are numerous research areas that lack proper investigation
challenges: and solutions. As there are many entities such as UAV, gateway, plane
etc, therefore, key management and authentication have high complex-
ity. This can help to generate more secure and trustworthy UAVs in
5.3.1. Selection of cross-blockchain technology future. The innovative solutions should be restricted with specific needs
Peer-to-peer (P2P) networks are decentralized in nature, offer huge and limitations, for example, unwavering quality and intricacy etc.
benefits and mitigate considerable security risks, but it is still vulner- The mobility of UAV is with certain regulation and under hierarchical
able to nearly 51% attacks (Saad et al., 2019). This is particularly management, therefore, the threaten can be created by prediction of
terrifying in UAVs, where UAVs may leave and join freely (Wu et al., mobility or investigate the backdoor of weak network entity in man-
2021). The more availability of inexpensive UAVs can facilitate these agement chain. This section provides some probable future solutions
for security of UAVs.
attacks. To solve this problem, Dasu et al. (2018) proposed a hybrid
model of traffic management. In this model, a centralized authority
6.1. Connectivity insecurity
manages parts of airspace, while others are tackled by block-chain
based decentralized management. They have given a unique method
There are two frequently used methods for connecting UAVs to GCS.
for preventing DoS attacks, by taking some fees from airspace users de- The first is connecting through Wi-Fi and the second is through radio
pending on the local demand. This can also restrict almost 51% attacks signals. No doubt, wired connections are more secure as compared to
because it makes it difficult for attackers to join UAV with multiple wireless, but former methods cannot be used due to range restrictions,
UAVs at a time. One more way for improving security of blockchain for this reason wireless connections are mostly used. Security of these
was proposed in Kim (2019, 2021). Here, blockchain governance game wireless connections can be increased through encryption (Kriz and
was developed for dealing with attackers trying to get control over Gabrlik, 2015) and watermarking methods (Marcinak and Mobasseri).
blockchain with addition of more illegal nodes (Kim et al., 2019). The novel solution can be proposed by combining watermarking and

20
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

cryptography in a single protocol, because watermarked text can re- Table 10

tain itself. Thus, it can be assumed that if password is hacked, the UAV forensic process steps.

cracked password might be used by attackers before applying reverse Steps Details
watermarking.
Preparation • Chain of command identification
6.2. Authentication drawback in UAVs • Identification of any fingerprint or identity of UAV
• Use of offensive method to track down the owner with the
help of date, time, location, and identity recorded on the
Authentication is the most important, in UAV networks, because
UAVs
most of the time eavesdroppers use false information for hijacking
UAVs. It is also crucial for identifying right signals but still a big Examination • Assessment and inspection of data available on a UAV like
challenge and researchers are trying to tackle this problem. Although, video, audio and pictures etc.
• Identifying methods as well as ports used for
there are numerous authentication protocol, still there is a necessity to
communication
improve, e.g., one way is to use single random numbers but these can
be predicted with a bruit force attack. Instead of using these numbers
Analysis and After extraction from a UAV all the information is compiled
a chaotic map can be used with a key to generate multiple random
reporting and analyzed to find the culprit
numbers (Marcinak and Mobasseri; Cheng et al., 2012; Rosalie et al.,
2018, 2016; Sun et al., 2020). Yet, low dimensional chaotic is not good,
but a high dimensional map can solve this problem (Gohari et al.,
2019). 6.5. Secure UAVs communication

6.3. Intrusion detection systems (IDS) in UAVs Our study shows that there is a need for developing more realistic
communication protocols, secure and reliable communication among
Real-time network traffic analysis is needed to find intrusions numerous components and UAVs systems. Security of UAVs communi-
against UAVs while they are in flight. To achieve this, deploying
cation channels with achievement of high network throughput is still
an IDS for UAVs enables the detection of many intrusion classes,
challenging. Also, authentication of UAV can secure communication
including signal alteration, routing attacks, malware, and message
links, reply to attacks, and prevent impersonation (Rodrigues et al.,
forging attacks (Choudhary et al., 2018). Additionally, the creation of
2019). The Development of UAVs access policies like authentication
frameworks for detecting anomalies that indicate malicious behavior
schemes and authorization is still a challenging topic (Yaacoub et al.,
is crucial for identifying attack patterns (Moosbrugger et al., 2017).
2020). In fact, authenticated UAVs should not be a part of flight or
Also, using honeypots and honeynets in addition to IDS can assist to
protect the flying mission from hostile actors (Franco et al., 2021). On collect transferred data from other UAVs in a network. On the contrary,
the other hand, ML have been applied to predict known/unknown event in case of multi-UAVs, use of a particular network model for UAVs like
with already learnt data. To increase the accuracy of these algorithms, FANETs (Bekmezci et al., 2013) permits multi-UAVs operations. Even
features must be relevant, e.g., prediction results cannot be correct so, FANETs are vulnerable to numerous attacks, therefor implemen-
if the features in the dataset are irrelevant. For improving ML based tation of secure communications in multi-UAVs remain a challenging
IDS techniques, there is a need to include deep learning. This will research topic. Numerous FANETs protocols for routing are presented
help to make predictions more accurately. More number of features in literature (Arafat and Moh, 2019), but they cannot meet the privacy
for intrusion detection can increase model accuracy. and security requirements completely and demand more research in
this domain (Maxa et al., 2017).
6.4. UAVs digital forensic
6.6. Realistic implementations
If a security incident happens in a flight, there is need of forensic
analysts for compromised UAVs, but they cannot collect evidence from The development and deployment of UAVs need an emphasis on
UAVs that lack logging capabilities (Altawy and Youssef, 2016). More exchange between performance and security. From a security perspec-
precisely, sensitive data like onboard flight information and flight tive, we must consider privacy and security requirements of UAVs
trajectories are saved in volatile Random-Access Memory (RAM) of systems. Besides, we should consider computation costs, storage over-
Flight Controller (FC). Consequently, making the process of recovery head, and energy from a performance point of view. The development
more challenging. Additionally, model creation is most the significant of lightweight security protocols or authentication mechanism for UAVs
for providing UAVs with forensic analysis (Al-Dhaqm et al., 2021). includes use of cryptographic primordial. Still, these implementations
Still, even with stronger forensic techniques, surprisingly current anti- may consume high energy and more computation cost. Hence, bal-
forensic strategies could thwart forensic analysis process (Atkinson
ancing these two aspects as well as finding strategic solutions is a
et al., 2021). On the other hand, mitigation strategies normally consider
major research topic. Present security solutions operate under specified
adopting a forensic-by-design technique, which combines forensics re-
software and hardware settings, so when proposing a real-time imple-
quirements with design of UAVs systems (Ab Rahman et al., 2016).
mentation, deployment challenges in numerous UAVs systems must be
Forensic analysis of UAVs is inadequately investigated in the area of
considered. It must be emphasized that imitating UAVs’ cyber-attack
UAV security. At present forensics models do not have proper stan-
scenarios could demonstrate flexibility of current security solutions
dardization and unification to enclose a wider domain of commercial
UAVs, here is key problem with UAVs forensics where attackers could against such attacks prior to their deployment. Likewise, present sim-
compromise UAVs. ulation environments, to analyze UAVs security, are limited (Javaid
Likewise, a number of other researchers have put forth various et al., 2013), so this topic needs further research.
strategies for using forensic techniques to defend against sophisticated
and complicated attacks. The forensic approach is being highlighted 6.7. Privacy preservation
because, with time, attack types and goals become more sophisticated
and challenging to recognize (Salamh et al., 2021). The attacker and Combining UAVs with national airspace has increased privacy is-
the attack’s methodology can be detected through forensics. The type sues, causing leakage of secret information gathered by UAVs, and
of attack may be identified to prevent any further incidents, and the transfer it to third party. In such situation, it is necessary to safeguard
relevant countermeasures can be applied to avoid such incidents in the privacy of exchanged data. Currently, numerous privacy preser-
future. Moreover, the process of UAVs digital forensic can be divided vation techniques have been proposed by researchers. For example,
into three main steps (see Table 10). the mitigation privacy attacks comprises of using privacy increasing

21
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

technologies for preserving data of consumers. Specifically, numerous In addition, there is still a need for UAV dataset to train models,
privacy techniques and secure computation maintain privacy of users, like malware dataset, network traffic dataset, etc. Some models may
and allow coordination of data among UAVs with privacy ensured. fail when it comes to detection of UAVs’ cyber-attacks (Hassija et al.,
More examples include using homo-morphic encryption for performing 2021). Next, SDN enables UAV security communications and provides
computational operations on encrypted data (Acar et al., 2018) as well a programmable network as well as dynamic flow control for numerous
as Zero Knowledge Proof (ZPF) for validating data without disclosure. security functions, thus can secure UAVs network from cyber-attacks.
Further, Privacy issues are prevalent, due to the simplicity with The main disadvantage of this technology is high-end-to-end delay in
which anyone can obtain UAVs with high-definition cameras and other the case of non-delay-tolerant applications. However, the link in control
devices. Within their private property, people can readily be seen or plane and data plane can be attacked, and integrating UAVs in smart
recorded without their knowledge. UAVs have given rise to numerous
cities requires processing as well as large amount of stored data, for
safety, security, and privacy concerns, according to Canadian Public
this fog computing may help to store and process data.
Safety (CPS) (Abro et al., 2022). Without their permission, persons have
Fog computing provides secure communication among fog layer and
been photographed or recorded resulting in being blackmailed and
other illegal actions. To help control UAV operations, legal regulations UAVs with low latency and high stability, but existing fog architecture
related to flying through buildings, hovering at window level, or taking is not modified for UAVs model. Thus, adopting this architecture may
private images or films without permission should be implemented. enhance time for data processing, particularly in Multi UAV networks.
More diverse technologies may emerge to be incorporated in next
6.8. Secure data aggregation generation of UAVs (Nayyar et al., 2020). Hence, further research is
needed in the domain of emerging technologies for securing UAVs in
The wide use of UAVs in numerous applications domains have civil applications.
increased quantity of gathered and shared data. The gathered data
is normally aggregated for efficient use of resources but its data ag-
7. Conclusion
gregation process requires protection against malicious entities. The
deployment of these aggregation mechanisms should take into account
encryption schemes for providing confidentiality, to ensure a secure In this paper, we have broadly surveyed privacy and security issues
exchange of data among GCS and UAVs. Additionally, providing secure of UAVs by dividing them into three classes: software, hardware and
and efficient data aggregation techniques for UAVs will minimize the communication. A systematic review of the common vulnerabilities,
communication cost and energy with grantee of confidentiality. Still, specifically linked to each class was also provided. We have thoroughly
developing such techniques remains a continuous challenge. discussed passive and active attacks from adversaries for compromising
privacy and security of UAVs. A detailed description of mitigation tech-
6.9. AI for UAVs niques and countermeasures for protecting UAVs was presented. The
solution architecture part also includes a discussion about emerging
Future research direction further involves the exploration and im- technologies like blockchain etc. In several applications, blockchain
plementation of Edge AI for securing UAVs. The overlap between edge technology appears to provide intriguing advantages. It has been sug-
computing and UAV systems allows UAVs (fitted with edge servers) to gested to use a variety of cross-blockchain interoperability techniques
either function as users themselves and offload duties to edge servers,
to increase scalability while preserving decentralization, transparency,
or to offer edge computing services for terrestrial user devices (McEnroe
and immutability. We outlined the difficulties involved in setting up a
et al., 2022). The other concept, which uses the conventional cloud
network of blockchains to simultaneously enable various UAV network
computing approach, includes distributing the jobs between UAVs as
operations. Cost, availability of data centers, and interoperability be-
well as a distant central computer. Contrarily, edge computing makes
tween various blockchain ledgers are a few of these difficulties. Briefly,
these computing services accessible to end users (UAVs) at the network
due to increased use of UAVs in civil airspace, privacy and security
edge, preventing the requirement for data to travel far distances to
distant centralized servers (Ansari et al., 2020). issues have become a very crucial stuff of national security. Thus,
collaboration among academia, industry and law enforcement is needed
6.10. Emerging technologies to develop new frameworks, regulations and standards of security. In
short, this survey provides a worthy reference for researchers to learn
Recently, for securing UAVs, emerging technologies are being used about designing and developing secure UAVs architectures.
extensively. These technologies include Blockchain technology, Artifi-
cial Intelligence, Fog Computing and SDN (Hassija et al., 2021; Syed
Declaration of competing interest
et al., 2021). Blockchain technology has distributed architecture that
adds extra security layer for communication. It uses smart contracts
with hash functions for storing data as chains of blocks. Therefore, it The authors declare the following financial interests/personal rela-
becomes difficult for attackers to tamper with the communication of tionships which may be considered as potential competing interests:
UAVs. The main applications of blockchain in UAV communication suf- Yue Cao reports financial support was provided by Wuhan University.
fer real-time deployment when it comes to highly mobile UAV (Kumari
et al., 2020).
Data availability
Furthermore, the real-time application of Blockchain for securing
UAVs is yet ongoing research area. The Introduction of AI technology
like ML algorithms has shown great benefits for security-oriented appli- The data that has been used is confidential.
cations, like privacy leakage and protection of UAVs from attacks. Nu-
merous ML security mechanisms have been presented in literature for
Acknowledgments
addressing security issues, e.g., DoS attacks and UAV detection (Yang
et al., 2019). Currently, federated learning methods are replacing ML
algorithms, as these techniques have shown best results, e.g., using The work is supported in part by the Guangdong Basic and Applied
UAVs authentication model relying on Radio Frequency features of Basic Research Foundation (2022B1515120067); Wuhan Knowledge
UAVs in IoT networks (Yazdinejad et al., 2021). Innovation Program (2022010801010117).

22
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

References Bhunia, S., Tehranipoor, M., 2018. The Hardware Trojan War: Attacks, Myths, and
Defenses. Springer.
Ab Rahman, N.H., Glisson, W.B., Yang, Y., Choo, K.-K.R., 2016. Forensic-by-design Boccadoro, P., Striccoli, D., Grieco, L.A., 2021. An extensive survey on the Internet of
framework for cyber-physical cloud systems. IEEE Cloud Comput. 3 (1), 50–59. Drones. Ad Hoc Netw. 122, 102600.
Abbaspour, A., Yen, K.K., Noei, S., Sargolzaei, P.c.s., 2016a. Detection of fault data Brandt, A.M., Colton, M.B., 2010. Haptic collision avoidance for a remotely operated
injection attack on uav using adaptive neural network. 95, pp. 193–200. quadrotor uav in indoor environments. In: 2010 IEEE International Conference on
Abbaspour, A., Yen, K.K., Noei, S., Sargolzaei, A., 2016b. Detection of fault data Systems, Man and Cybernetics. IEEE, pp. 2724–2731.
injection attack on uav using adaptive neural network. Procedia Comput. Sci. 95, Bresson, E., Chevassut, O., Pointcheval, D., 2007. Provably secure authenticated group
193–200. Diffie-Hellman key exchange. ACM Trans. Inf. Syst. Secur. 10 (3), 10–es.
AbdAllah, E.G., Hassanein, Surveys, I.C., Tutorials, 2015. A survey of security attacks Casals, S.G., Owezarski, P., Descargues, G., 2013. Generic and autonomous system
in information-centric networking. 17, (3), pp. 1441–1454. for airborne networks cyber-threat detection. In: 2013 IEEE/AIAA 32nd Digital
Abro, G.E.M., Zulkifli, S.A.B., Masood, R.J., Asirvadam, V.S., Laouti, A., 2022. Com- Avionics Systems Conference. DASC, IEEE, pp. 4A4–1.
prehensive review of UAV detection, security, and communication advancements Ch, R., Srivastava, G., Gadekallu, T.R., Maddikunta, P.K.R., Bhattacharya, S., 2020.
to prevent threats. Drones 6 (10), 284. Security and privacy of UAV data using blockchain technology. J. Inf. Secur. Appl.
Abualsauod, E.H., 2022. A hybrid blockchain method in internet of things for privacy 55, 102670.
and security in unmanned aerial vehicles network. Comput. Electr. Eng. 99, 107847. Challita, U., Ferdowsi, A., Chen, M., Saad, W., 2019. Machine learning for wireless
Acar, A., Hidayet, A., Selcuk, U.A., Conti, M., 2018. A survey on homomorphic connectivity and security of cellular-connected UAVs. IEEE Wirel. Commun. 26
encryption schemes: Theory and implementation. ACM Comput. Surv. (Csur) 51 (1), 28–35.
(4), 1–35. Chen, X.-C., Chen, Y.-J., 2019. A machine learning based attack in UAV communication
Agrawal, R., Verma, P., Sonanis, R., Goel, U., de, A., kondaveeti, s.a., shekhar, s., 2018. networks. In: 2019 IEEE 90th Vehicular Technology Conference. VTC2019-Fall,
Continuous security in IoT using Blockchain. In: IEEE International Conference on IEEE, pp. 1–2.
Acoustics, Speech and Signal Processing. ICASSP, pp. 6423–6427. Cheng, Z., Tang, Y.X., Liu, Y.L., 2012. 3-D path planning for UAV based on chaos
Al-Dhaqm, A., Ikuesan, R.A., Kebande, V.R., Razak, S., Ghabban, F.M., 2021. Research particle swarm optimization. In: Applied Mechanics and Materials. Vol. 232, Trans
challenges and opportunities in drone forensics models. Electronics 10 (13), 1519. Tech Publ, pp. 625–630.
Ali, Z., Chaudhry, S.A., Ramzan, M.S., Al-Turjman, F., 2020. Securing smart city Choi, D.H., Kim, S.H., Sung, D.K., 2014. Energy-efficient maneuvering and commu-
surveillance: A lightweight authentication mechanism for unmanned vehicles. IEEE nication of a single UAV-based relay. IEEE Trans. Aerosp. Electron. Syst. 50 (3),
Access 8, 43711–43724. 2320–2327.
Alkadi, R., Alnuaimi, N., Shoufan, A., Yeun, C., 2021. Blockchain interoperability in Choudhary, G., Sharma, V., You, I., Yim, K., Chen, R., Cho, J.-H., 2018. Intrusion
UAV networks: state-of-the-art and open issues. arXiv preprint arXiv:2111.09529. detection systems for networked unmanned aerial vehicles: a survey. In: 2018 14th
Alladi, T., Chamola, V., Kumar, N., et al., 2020b. PARTH: A two-stage lightweight International Wireless Communications & Mobile Computing Conference. IWCMC,
mutual authentication protocol for UAV surveillance networks. Comput. Commun. IEEE, pp. 560–565.
160, 81–90. Chriki, A., Touati, H., Snoussi, H., Kamoun, F., 2019. FANET: Communication, mobility
Alladi, T., Chamola, V., Sahu, N., Guizani, M., 2020a. Applications of blockchain in models and security issues. Comput. Netw. 163, 106877.
unmanned aerial vehicles: A review. Veh. Commun. 23, 100249. Condomines, J.-P., Zhang, R., Larrieu, N., 2019. Network intrusion detection system
Allouch, A., Cheikhrouhou, O., Koubâa, A., Toumi, K., Khalgui, M., Nguyen Gia, T., for UAV ad-hoc communication: From methodology design to real test validation.
2021. Utm-chain: blockchain-based secure unmanned traffic management for Ad Hoc Netw. 90, 101759.
internet of drones. Sensors 21 (9), 3049. Darwish, T.S., Bakar, K.A., 2018. Fog based intelligent transportation big data analytics
Altawy, R., Youssef, A.M., 2016. Security, privacy, and safety aspects of civilian drones: in the internet of vehicles environment: motivations, architecture, challenges, and
A survey. ACM Trans. Cyber-Phys. Syst. 1 (2), 1–25. critical issues. IEEE Access 6, 15679–15701.
Alwateer, M., Loke, S.W., Zuchowicz, A., 2019. Drone services: issues in drones for Dasu, T., Kanza, Y., Srivastava, D., 2018. Geofences in the sky: herding drones with
location-based services from human-drone interaction to information processing. J. blockchains and 5G. In: Proceedings of the 26th ACM SIGSPATIAL International
Locat. Based Serv. 13 (2), 94–127. Conference on Advances in Geographic Information Systems. pp. 73–76.
Anon, 1999. KDD cup 1999 data. [Online] http://kdd.ics.uci.edu/databases/kddcup99/ Deebak, B.D., Al-Turjman, F., 2020. A smart lightweight privacy preservation scheme
kddcup99.html. (Accessed 16 December 2021). for IoT-based UAV communication systems. Comput. Commun. 162, 102–117.
Anon, 2018. CIC-IDS-2018. [Online] http://www.unb.ca/cic/datasets/ids-2018.html. Deligne, E., 2012. ARDrone corruption. J. Comput. Virol. 8 (1), 15–27.
(Accessed 15 December 2021). Demeri, A., Diehl, W., Salman, A., 2020. Saddle: Secure aerial data delivery with
Anon, 2021. Global anti-drone markets 2021–2026 - increasing terrorism and illicit lightweight encryption. In: Science and Information Conference. Springer, pp.
activities through Drones & amp; rising investment by governments. 204–223.
Ansari, M.S., Alsamhi, S.H., Qiao, Y., Ye, Y., Lee, B., 2020. Security of distributed in- Driscoll, K., 2018. Lightweight crypto for lightweight unmanned arial systems. In: 2018
telligence in edge computing: Threats and countermeasures. In: The Cloud-to-Thing Integrated Communications, Navigation, Surveillance Conference. ICNS, IEEE, pp.
Continuum. Palgrave Macmillan, Cham, pp. 95–122. 1–15.
Arafat, M.Y., Moh, S., 2019. Routing protocols for unmanned aerial vehicle networks: Feng, Z., Guan, N., Lv, M., Liu, W., Deng, Q., Liu, X., Yi, W., 2020. Efficient drone
A survey. IEEE Access 7, 99694–99720. hijacking detection using two-step GA-XGBoost. J. Syst. Archit. 103, 101694.
Arthur, M.P., Detecting signal spoofing and jamming attacks in UAV networks using a Fotouhi, A., Qiang, H., Ding, M., Hassan, M., Giordano, L.G., Garcia-Rodriguez, A.,
lightweight IDS. In: 2019 International Conference on Computer, Information and Yuan, J., 2019. Survey on UAV cellular communications: Practical aspects, stan-
Telecommunication Systems. CITS, IEEE, pp. 1–5. dardization advancements, regulation, and security challenges. IEEE Commun. Surv.
Arthur, M.P., 2019. Detecting signal spoofing and jamming attacks in UAV networks us- Tutor. 21 (4), 3417–3442.
ing a lightweight IDS. In: 2019 International Conference on Computer, Information Franco, J., Aris, A., Canberk, B., Uluagac, A.S., 2021. A survey of honeypots and
and Telecommunication Systems. CITS, IEEE, pp. 1–5. honeynets for internet of things, industrial internet of things, and cyber-physical
Atkinson, S., Carr, G., Shaw, C., Zargari, S., 2021. Drone forensics: The impact and systems. IEEE Commun. Surv. Tutor. 23 (4), 2351–2383.
challenges. In: Digital Forensic Investigation of Internet of Things (IoT) Devices. Gohari, P.S., Mohammadi, H., Taghvaei, S., 2019. Using chaotic maps for 3D boundary
Springer, pp. 65–124. surveillance by quadrotor robot. Appl. Soft Comput. 76, 68–77.
Atoev, S., Kwon, O.-J., Kim, C.-Y., Lee, S.-H., Choi, Y.-R., Kwon, K.-R., 2019. The secure Gupta, R., Kumari, A., Tanwar, S., 2021. Fusion of blockchain and artificial intelli-
UAV communication link based on OTP encryption technique. In: 2019 Eleventh gence for secure drone networking underlying 5G communications. Trans. Emerg.
International Conference on Ubiquitous and Future Networks. ICUFN, IEEE, pp. Telecommun. Technol. 32 (1), e4176.
1–3. Hannah, J., Mills, R., Dill, R., 2020. Traffic collision avoidance system: threat actor
Bangui, H., Ge, M., Buhnova, B., 2020. Improving big data clustering for jamming model and attack taxonomy. In: 2020 New Trends in Civil Aviation. NTCA, IEEE,
detection in smart mobility. In: IFIP International Conference on ICT Systems pp. 17–26.
Security and Privacy Protection. Springer, pp. 78–91. Haque, M.S., Chowdhury, M.U., 2017. A new cyber security framework towards
Barka, E., Kerrache, C.A., Benkraouda, H., Shuaib, K., Ahmad, F., Kurugollu, F., 2019. secure data communication for unmanned aerial vehicle (uav). In: International
Towards a trusted unmanned aerial system using blockchain for the protection of Conference on Security and Privacy in Communication Systems. Springer, pp.
critical infrastructure. Trans. Emerg. Telecommun. Technol. e3706. 113–122.
Barka, E., Kerrache, C.A., Hussain, R., Lagraa, N., Lakas, A., Bouk, S.H., 2018. A trusted Hassija, V., Chamola, V., Agrawal, A., Goyal, A., Luong, N.C., Niyato, D.,
lightweight communication strategy for flying named data networking. Sensors 18 Yu, F.R., Guizani, M., 2021. Fast, reliable, and secure drone communication: A
(8), 2683. comprehensive survey. IEEE Commun. Surv. Tutor.
Bejiga, M.B., Zeggada, A., Nouffidj, A., Melgani, F., 2017. A convolutional neural Hayat, S., Yanmaz, E., Muzaffar, R., 2016. Survey on unmanned aerial vehicle networks
network approach for assisting avalanche search and rescue operations with UAV for civil applications: A communications viewpoint. IEEE Commun. Surv. Tutor. 18
imagery. Remote Sens. 9 (2), 100. (4), 2624–2661.
Bekmezci, I., Sahingoz, O.K., Temel, Ş., 2013. Flying ad-hoc networks (FANETs): A He, D., Chan, S., Guizani, M., 2017. Drone-assisted public safety networks: The security
survey. Ad Hoc Netw. 11 (3), 1254–1270. aspect. IEEE Commun. Mag. 55 (8), 218–223.

23
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

He, D., Yang, G., Li, H., Chan, S., Cheng, Y., Guizani, N., 2020. An effective Liu, J., Yang, W., Xu, S., Liu, J., Zhang, Q., 2019b. Q-learning based UAV secure
countermeasure against UAV swarm attack. IEEE Netw. 35 (1), 380–385. communication in presence of multiple UAV active eavesdroppers. In: 2019
Hentati, A.I., Fourati, L.C., 2020. Comprehensive survey of UAVs communication 11th International Conference on Wireless Communications and Signal Processing.
networks. Comput. Stand. Interfaces 72, 103451. WCSP, IEEE, pp. 1–6.
Hoang, T.M., Nguyen, N.M., Duong, T.Q., 2019. Detection of eavesdropping attack Liu, G., Zhang, R., Wang, C., Liu, L., 2019a. Synchronization-free gps spoofing detection
in UAV-aided wireless systems: Unsupervised learning with one-class SVM and with crowdsourced air traffic control data. In: 2019 20th IEEE International
k-means clustering. IEEE Wirel. Commun. Lett. 9 (2), 139–142. Conference on Mobile Data Management. MDM, IEEE, pp. 260–268.
Hodgkins, K., 2015. Anti-drone shoulder rifle lets police take control of UAVs with Lu, X., Xiao, L., Dai, C., Dai, H., 2020. UAV-aided cellular communications with deep
radio pulses.(2015). reinforcement learning against jamming. IEEE Wirel. Commun. 27 (4), 48–53.
Hooper, M., Tian, Y., Zhou, R., Cao, B., Lauf, A.P., Watkins, L., Robinson, W.H., Lu, X., Xu, D., Xiao, L., Wang, L., Zhuang, W., 2017. Anti-jamming communication game
Alexis, W., 2016. Securing commercial wifi-based uavs from common security for UAV-aided VANETs. In: GLOBECOM 2017-2017 IEEE Global Communications
attacks. In: MILCOM 2016–2016 IEEE Military Communications Conference. IEEE, Conference. IEEE, pp. 1–6.
pp. 1213–1218. Lv, Z., Qiao, L., Hossain, M.S., Choi, B.J., 2021. Analysis of using blockchain to protect
Huda, S.A., Moh, S., 2022. Survey on computation offloading in UAV-enabled mobile the privacy of drone big data. IEEE Netw. 35 (1), 44–49.
edge computing. J. Netw. Comput. Appl. 103341. Ma, Y., Wu, X., Yu, G., Xu, Y., Wang, Y., 2016. Pedestrian detection and tracking from
Islam, A., Shin, S.Y., 2019. BUAV: A blockchain based secure UAV-assisted data low-resolution unmanned aerial vehicle thermal imagery. Sensors 16 (4), 446.
acquisition scheme in Internet of Things. J. Commun. Netw. 21 (5), 491–502. Manesh, M.R., Kenney, J., Hu, W.C., Devabhaktuni, V.K., Kaabouch, N., 2019. Detection
Israelsen, J., Beall, M., Bareiss, D., Stuart, D., Keeney, E., van den Berg, J., 2014. of GPS spoofing attacks on unmanned aerial systems. In: 2019 16th IEEE Annual
Automatic collision avoidance for manually tele-operated unmanned aerial vehicles. Consumer Communications & Networking Conference. CCNC, IEEE, pp. 1–6.
In: 2014 IEEE International Conference on Robotics and Automation. ICRA, IEEE, Marcinak, M.P., Mobasseri, B.G., Digital video watermarking for metadata embedding
pp. 6638–6643. in UAV video. In: MILCOM 2005-2005 IEEE Military Communications Conference,
Javaid, A.Y., Sun, W., Alam, M., 2013. UAVSim: A simulation testbed for unmanned Vols. 3–5. IEEE, p. 1637.
aerial vehicle network cyber security analysis. In: 2013 IEEE Globecom Workshops. Maxa, J.-A., Mahmoud, M.S.B., Larrieu, N., 2017. Survey on UAANET routing protocols
GC Wkshps, IEEE, pp. 1432–1436. and network security challenges. Adhoc Sensor Wirel. Netw. 37.
Jones, A., Kovacich, G.L., 2015. Global Information Warfare: The New Digital McEnroe, P., Wang, S., Liyanage, M., 2022. A survey on the convergence of edge
Battlefield. CRC Press. computing and AI for UAVs: Opportunities and challenges. IEEE Internet Things
Kafi, M.A., Challal, Y., Djenouri, D., Doudou, M., Bouabdallah, A., Badache, N., 2013. J.
A study of wireless sensor networks for urban traffic monitoring: applications and Mehta, P., Gupta, R., Tanwar, S., 2020. Blockchain envisioned UAV networks:
architectures. Procedia Comput. Sci. 19, 617–626. Challenges, solutions, and comparisons. Comput. Commun. 151, 518–538.
Kanade, V.A., 2021. Securing drone-based ad hoc network using blockchain. In: 2021 Min, M., Xiao, L., Xu, D., Huang, L., Peng, M., 2018. Learning-based defense against
International Conference on Artificial Intelligence and Smart Systems. ICAIS, IEEE, malicious unmanned aerial vehicles. In: 2018 IEEE 87th Vehicular Technology
pp. 1314–1318. Conference (VTC Spring). IEEE, pp. 1–5.
Kapitonov, A., Berman, I., Manaenko, V., Rzhevskiy, V., Bulatov, V., Zenkin, A., 2019. Mishra, D., Natalizio, E., 2020. A survey on cellular-connected UAVs: Design challenges,
Robonomics as a blockchain-based platform for unmanned traffic management of enabling 5G/B5G innovations, and experimental advancements. Comput. Netw.
mobile vehicles. In: 2019 Workshop on Research, Education and Development of 182, 107451.
Unmanned Aerial Systems. RED UAS, IEEE, pp. 9–17. Mitchell, R., Chen, R., 2013. Adaptive intrusion detection of malicious unmanned air
Khan, N.A., Brohi, S.N., Jhanjhi, N., 2020. UAV’s applications, architecture, security vehicles using behavior rule specifications. IEEE Trans. Syst. Man Cybern. A 44
issues and attack scenarios: a survey. In: Intelligent Computing and Innovation on (5), 593–604.
Data Science. Springer, pp. 753–760. Moosbrugger, P., Rozier, K.Y., Schumann, J., 2017. R2U2: monitoring and diagnosis
Khan, A.A., Khan, M.M., Khan, K.M., Arshad, J., Ahmad, F., 2021. A blockchain-based of security threats for unmanned aerial systems. Form. Methods Syst. Des. 51 (1),
decentralized machine learning framework for collaborative intrusion detection 31–61.
within UAVs. Comput. Netw. 196, 108217. Nassi, B., Bitton, R., Masuoka, R., Shabtai, A., Elovici, Y., 2021. SoK: Security and
Khan, A.A., Laghari, A.A., Gadekallu, T.R., Shaikh, Z.A., Javed, A.R., Rashid, M., Es- privacy in the age of commercial drones. In: 2021 IEEE Symposium on Security
trela, V.V., Mikhaylov, A., 2022. A drone-based data management and optimization and Privacy. SP, IEEE, pp. 1434–1451.
using metaheuristic algorithms and blockchain smart contracts in a secure fog Nassi, B., Shabtai, A., Masuoka, R., Elovici, Y., 2019. Sok-security and privacy in the
environment. Comput. Electr. Eng. 102, 108234. age of drones: threats, challenges, solution mechanisms, and scientific gaps. arXiv
Kim, S.-K., 2019. The trailer of strategic alliance for blockchain governance game. arXiv preprint arXiv:1903.05155.
preprint arXiv:1903.11172. Nayyar, A., Nguyen, B.-L., Nguyen, N.G., 2020. The internet of drone things (IoDT):
Kim, S.-K.A., 2021. Enhanced IoV security network by using blockchain governance future envision of smart drones. In: First International Conference on Sustainable
game. Mathematics 9 (2), 109. Technologies for Computational Intelligence. Springer, pp. 563–580.
Kim, S.-K., Yeun, C.Y., Damiani, E., Al-Hammadi, Y., Lo, N.-W., 2019. New blockchain Newaz, A.I., Sikder, A.K., Rahman, M.A., Uluagac, A.S., 2021. A survey on security and
adoptation for automotive security by using systematic innovation. In: 2019 IEEE privacy issues in modern healthcare systems: Attacks and defenses. ACM Trans.
Transportation Electrification Conference and Expo, Asia-Pacific. ITEC Asia-Pacific, Comput. Healthc. 2 (3), 1–44.
IEEE, pp. 1–4. Nguyen, D.C., Pathirana, P.N., Ding, M., Seneviratne, A., 2020. Blockchain for 5G and
Kirichenko, V., 2015. Information security of communication channel with UAV. beyond networks: A state of the art survey. J. Netw. Comput. Appl. 166, 102693.
Electron. Control Syst. (3), 23–27. Noor, F., Khan, M.A., Al-Zahrani, A., Ullah, I., Al-Dhlan, K.A., 2020. A review on
Krishna, C.L., Murphy, R.R., 2017. A review on cybersecurity vulnerabilities for communications perspective of flying ad-hoc networks: key enabling wireless
unmanned aerial vehicles. In: 2017 IEEE International Symposium on Safety, technologies, applications, challenges and open research topics. Drones 4 (4), 65.
Security and Rescue Robotics. SSRR, IEEE, pp. 194–199. Ossamah, A., 2020. Blockchain as a solution to drone cybersecurity. In: 2020 IEEE 6th
Kriz, V., Gabrlik, P., 2015. Uranuslink-communication protocol for uav with small World Forum on Internet of Things. WF-IoT, IEEE, pp. 1–9.
overhead and encryption ability. IFAC-PapersOnLine 48 (4), 474–479. Oz, H., Aris, A., Levi, A., Uluagac, A.S., 2021. A survey on ransomware: Evolution,
Kumari, A., Gupta, R., Tanwar, S., Kumar, N., 2020. A taxonomy of blockchain-enabled taxonomy, and defense solutions. ACM Comput. Surv.
softwarization for secure UAV network. Comput. Commun. 161, 304–323. Paganini, P., 2015. A hacker developed maldrone, the first malware for drones. https:
Lee, J., Ryu, S., Kim, H.J., 2020. Stable flight of a flapping-wing micro air vehicle //securityaffairs.co/wordpress/32767/hacking/maldrone-malware-fordrones.html.
under wind disturbance. IEEE Robot. Autom. Lett. 5 (4), 5685–5692. Pan, W.-J., Feng, Z.-L., Wang, Y., 2012. ADS-B data authentication based on ECC and
Li, J., Kamin, S., Zheng, G., Neubrech, F., Zhang, S., Liu, N., 2018. Addressable X. 509 certificate. J. Electron. Sci. Technol. 10 (1), 51–55.
metasurfaces for dynamic holography and optical information encryption. Sci. Adv. Panice, G., Luongo, S., Gigante, G., Pascarella, D., Di Benedetto, C., Vozella, A.,
4 (6), eaar6768. Pescapè, A., 2017. A SVM-based detection approach for GPS spoofing attacks to
Li, Z., Lu, Y., Shi, Y., Wang, Z., Qiao, W., Liu, Y., 2019b. A dyna-q-based solution for UAV. In: 2017 23rd International Conference on Automation and Computing. ICAC,
uav networks against smart jamming attacks. Symmetry 11 (5), 617. IEEE, pp. 1–11.
Li, K., Voicu, R.C., Kanhere, S.S., Ni, W., Tovar, E., 2019a. Energy efficient legitimate Peterson, P.A., 2010. Cryptkeeper: Improving security with encrypted RAM. In: 2010
wireless surveillance of UAV communications. IEEE Trans. Veh. Technol. 68 (3), IEEE International Conference on Technologies for Homeland Security. HST, IEEE,
2283–2293. pp. 120–126.
Li, Q., Yang, Y., Ma, W.-K., Lin, M., Ge, J., Lin, J., 2014. Robust cooperative beam- Pu, C., Li, Y., 2020. Lightweight authentication protocol for unmanned aerial vehicles
forming and artificial noise design for physical-layer secrecy in AF multi-antenna using physical unclonable function and chaotic system. In: 2020 IEEE International
multi-relay networks. IEEE Trans. Signal Process. 63 (1), 206–220. Symposium on Local and Metropolitan Area Networks. LANMAN, IEEE, pp. 1–6.
Lin, Z., Lu, X., Dai, C., Sheng, G., Xiao, L., 2019. Reinforcement learning based UAV Puthal, D., Malik, N., Mohanty, S.P., Kougianos, E., Das, G., 2018. Everything you
trajectory and power control against jamming. In: International Conference on wanted to know about the blockchain: Its promise, components, processes, and
Machine Learning for Cyber Security. Springer, pp. 336–347. problems. IEEE Consum. Electron. Mag. 7 (4), 6–14.

24
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Qian, Y., Jiang, Y., Hu, L., Hossain, M.S., Alrashoud, M., Al-Hammadi, M., 2020. Tsao, K.-Y., Girdler, T., Vassilakis, V.G., 2022. A survey of cyber security threats and
Blockchain-based privacy-aware content caching in cognitive internet of vehicles. solutions for UAV communications and flying ad-hoc networks. Ad Hoc Netw.
IEEE Netw. 34 (2), 46–51. 102894.
Ra, W.-S., Whang, I.-H., Park, J.B., 2008. Robust weighted least squares range estimator Ueno, S., Higuchi, T., 2011. Collision avoidance law using information amount. In:
for UAV applications. In: 2008 SICE Annual Conference. IEEE, pp. 251–255. Numerical Analysis-Theory and Application. IntechOpen.
Rahman, M.S., Khalil, I., Atiquzzaman, M., 2021. Blockchain-powered policy enforce- Valentin-Alexandru, V., Ion, B., Victor-Valeriu, P., 2019. Energy efficient trust-based
ment for ensuring flight compliance in drone-based service systems. IEEE Netw. 35 security mechanism for wireless sensors and unmanned aerial vehicles. In: 2019
(1), 116–123. 11th International Conference on Electronics, Computers and Artificial Intelligence.
Rahman, M.A., Rahman, M.T., Kisacikoglu, M., Akkaya, K., 2020a. Intrusion detection ECAI, IEEE, pp. 1–6.
systems-enabled power electronics for unmanned aerial vehicles. In: 2020 IEEE Vanitha, N., Ganapathi, P., 2020. Traffic analysis of UAV networks using enhanced deep
CyberPELS (CyberPELS). IEEE, pp. 1–5. feed forward neural networks (EDFFNN). In: Handbook of Research on Machine and
Rahman, M.A., Rahman, M.T., Kisacikoglu, M., Akkaya, K., 2020b. Intrusion detection Deep Learning Applications for Cyber Security. IGI Global, pp. 219–244.
systems-enabled power electronics for unmanned aerial vehicles. In: 2020 IEEE Wang, Q., Chen, Z., Mei, W., Fang, J., 2017. Improving physical layer security using
CyberPELS (CyberPELS). IEEE, pp. 1–5. UAV-enabled mobile relaying. IEEE Wirel. Commun. Lett. 6 (3), 310–313.
Rodrigues, M., Amaro, J., Osório, F.S., RLJC, B.K., 2019. Authentication methods for Wang, J., Liu, Y., Niu, S., Song, H., 2021. Lightweight blockchain assisted secure routing
uav communication. In: 2019 IEEE Symposium on Computers and Communications. of swarm UAS networking. Comput. Commun. 165, 131–140.
ISCC, IEEE, pp. 1210–1215. Wang, H., Zhao, H., Zhang, J., Ma, D., Li, J., Wei, J., 2019. Survey on unmanned aerial
Rondon, L.P., Babun, L., Aris, A., Akkaya, K., Uluagac, A.S., 2022. Survey on enterprise vehicle networks: A cyber physical system perspective. IEEE Commun. Surv. Tutor.
Internet-of-Things systems (E-IoT): A security perspective. Ad Hoc Netw. 125, 22 (2), 1027–1070.
102728. Wesson, K.D., Humphreys, T.E., Evans, B.L., 2014. Can cryptography secure next
Rosalie, M., Danoy, G., Chaumette, S., Bouvry, P., 2016. From random process to generation air traffic surveillance? IEEE Secur. Priv. Mag.
Wu, B., Chen, J., Wu, J., Cardei, M., 2007. A survey of attacks and countermeasures
chaotic behavior in swarms of UAVs. In: Proceedings of the 6th ACM Symposium
in mobile ad hoc networks. In: Wireless Network Security. Springer, pp. 103–135.
on Development and Analysis of Intelligent Vehicular Networks and Applications.
Wu, Y., Dai, H.-N., Wang, H., Choo, K.-K.R., 2021. Blockchain-based privacy
pp. 9–15.
preservation for 5g-enabled drone communications. IEEE Netw. 35 (1), 50–56.
Rosalie, M., Danoy, G., Chaumette, S., Bouvry, P., 2018. Chaos-enhanced mobility
Xiao, W., Li, M., Alzahrani, B., Alotaibi, R., Barnawi, A., Ai, Q., 2021. A blockchain-
models for multilevel swarms of UAVs. Swarm Evol. Comput. 41, 36–48.
based secure crowd monitoring system using UAV swarm. IEEE Netw. 35 (1),
Saad, M., Spaulding, J., Njilla, L., Kamhoua, C.A., Nyang, D., Mohaisen, A., 2019.
108–115.
Overview of attack surfaces in blockchain. Blockchain Distrib. Syst. Secur. 51–66.
Xiao, L., Lu, X., Xu, D., Tang, Y., Wang, L., Zhuang, W., 2018. UAV relay in VANETs
Sahingoz, O.K., 2013. Multi-level dynamic key management for scalable wireless sensor
against smart jamming with reinforcement learning. IEEE Trans. Veh. Technol. 67
networks with UAV. In: Ubiquitous Information Technologies and Applications.
(5), 4087–4097.
Springer, pp. 11–19.
Xiao, L., Xie, C., Min, M., Zhuang, W., 2017. User-centric view of unmanned aerial
Salamh, F.E., Karabiyik, U., Rogers, M.K., Matson, E.T., 2021. A comparative uav
vehicle transmission against smart attacks. IEEE Trans. Veh. Technol. 67 (4),
forensic analysis: Static and live digital evidence traceability challenges. Drones
3420–3430.
5 (2), 42.
Xu, J., Deng, Z., Song, Q., Chi, Q., Wu, T., Huang, Y., Liu, D., Gao, M., 2020. Multi-
Saraswat, D., Verma, A., Bhattacharya, P., Tanwar, S., Sharma, G., Bokoro, P.N., UAV counter-game model based on uncertain information. Appl. Math. Comput.
Sharma, R., 2022. Blockchain-based federated learning in UAVs beyond 5G 366, 124684.
networks: a solution taxonomy and future directions. IEEE Access 10, 33154–33182. Yaacoub, J.-P., Noura, H., Salman, O., Chehab, A., 2020. Security analysis of drones
Sedjelmaci, H., Senouci, S.M., Ansari, N., 2016. Intrusion detection and ejection systems: Attacks, limitations, and recommendations. Internet Things 11, 100218.
framework against lethal attacks in UAV-aided networks: A Bayesian game-theoretic Yang, B., Matson, E.T., Smith, A.H., Dietz, J.E., Gallagher, J.C., 2019. UAV detection
methodology. IEEE Trans. Intell. Transp. Syst. 18 (5), 1143–1153. system with multiple acoustic nodes using machine learning models. In: 2019 Third
Sedjelmaci, H., Senouci, S.M., Ansari, N., 2017. A hierarchical detection and response IEEE International Conference on Robotic Computing. IRC, IEEE, pp. 493–498.
system to enhance security against lethal cyber-attacks in UAV networks. IEEE Yang, L., Qi, J., Xiao, J., Yong, X., 2014. A literature review of UAV 3D path planning.
Trans. Syst. Man Cybern. A 48 (9), 1594–1606. In: Proceeding of the 11th World Congress on Intelligent Control and Automation.
Shafique, A., Mehmood, A., Elhadef, M., 2021. Survey of security protocols and IEEE, pp. 2376–2381.
vulnerabilities in unmanned aerial vehicles. IEEE Access 9, 46927–46948. Yang, N., Wang, L., Geraci, G., Elkashlan, M., Yuan, J., Di Renzo, M., 2015. Safe-
Shakhatreh, H., Sawalmeh, A.H., Al-Fuqaha, A., Dou, Z., Almaita, E., Khalil, I., guarding 5G wireless communication networks using physical layer security. IEEE
Othman, N.S., Khreishah, A., Guizani, M., 2019. Unmanned aerial vehicles (UAVs): Commun. Mag. 53 (4), 20–27.
A survey on civil applications and key research challenges. IEEE Access 7, Yasin, J.N., Mohamed, S.A., Haghbayan, M.-H., Heikkonen, J., Tenhunen, H.,
48572–48634. Plosila, J., 2020. Unmanned aerial vehicles (uavs): Collision avoidance systems
Shannon, C.E., 1949. Communication theory of secrecy systems. Bell Syst. Tech. J. 28 and approaches. IEEE Access 8, 105139–105155.
(4), 656–715. Yazdinejad, A., Parizi, R.M., Dehghantanha, A., Karimipour, H., 2021. Federated
Sharma, A., Vanjani, P., Paliwal, N., Basnayaka, C.M.W., Jayakody, D.N.K., Wang, H.-C., learning for drone authentication. Ad Hoc Netw. 120, 102574.
Muthuchidambaranathan, P., 2020. Communication and networking technologies Yazdinejad, A., Parizi, R.M., Dehghantanha, A., Karimipour, H., Srivastava, G., Aled-
for UAVs: A survey. J. Netw. Comput. Appl. 102739. hari, M., 2020. Enabling drones in the internet of things with decentralized
Sheng, G., Min, M., Xiao, L., Liu, S., 2018. Reinforcement learning-based control for blockchain-based security. IEEE Internet Things J. 8 (8), 6406–6415.
unmanned aerial vehicles. J. Commun. Inf. Netw. 3 (3), 39–48. Yoon, K., Park, D., Yim, Y., Kim, K., Yang, S.K., Robinson, M., 2017. Security
Singh, M., Aujla, G.S., Bali, R.S., 2020. A deep learning-based blockchain mechanism authentication system using encrypted channel on uav network. In: 2017 First IEEE
for secure Internet of drones environment. IEEE Trans. Intell. Transp. Syst. International Conference on Robotic Computing. IRC, IEEE, pp. 393–398.
Srinivas, J., Das, A.K., Kumar, N., Rodrigues, J.J., 2019. TCALAS: Temporal credential- Zeitlin, A., Lacher, A., Kuchar, J., Drumm, A., 2006. Collision Avoidance for Unmanned
based anonymous lightweight authentication scheme for Internet of drones Aircraft: Proving the Safety Case. Report, MITRE CORP MCLEAN VA.
environment. IEEE Trans. Veh. Technol. 68 (7), 6903–6916. Zeng, Y., Zhang, R., Lim, T.J., 2016. Throughput maximization for UAV-enabled mobile
Steinmann, J.A., Babiceanu, R.F., Seker, R., 2016. Uas security: Encryption key relaying systems. IEEE Trans. Commun. 64 (12), 4983–4996.
negotiation for partitioned data. In: 2016 Integrated Communications Navigation Zhi, Y., Fu, Z., Sun, X., Yu, J., 2020. Security and privacy issues of UAV: a survey.
and Surveillance. ICNS, IEEE, pp. 1E4–1. Mob. Netw. Appl. 25 (1), 95–101.
Su, Z., Wang, Y., Xu, Q., Zhang, N., 2020. LVBS: Lightweight vehicular blockchain for
secure data sharing in disaster rescue. IEEE Trans. Dependable Secure Comput.
Sun, J., Wang, W., Kou, L., Lin, Y., Zhang, L., Da, Q., Chen, L., 2020. A data
Hassan Jalil Hadi received M.S Information Security degree
authentication scheme for UAV ad hoc network communication. J. Supercomput.
from Riphah International University, Islamabad, Pakistan.
76 (6), 4041–4056.
He is currently pursuing the Ph.D. degree in Cyberspace
Syed, F., Gupta, S.K., Hamood Alsamhi, S., Rashid, M., Liu, X., 2021. A survey
Security at the School of Cyber Science and Engineering,
on recent optimal techniques for securing unmanned aerial vehicles applications.
Wuhan University in China. He is a cyber-security analyst
Trans. Emerg. Telecommun. Technol. 32 (7), e4133.
performing threat detection, monitoring, mitigation and au-
Teng, L., Jianfeng, M., Pengbin, F., Yue, M., Xindi, M., Jiawei, Z., Gao, C., Di, L., dit on industry proven standards for CS challenges at Cyber
2019. Lightweight security authentication mechanism towards uav networks. In: Reconnaissance and combat (CRC) Lab, National Center for
2019 International Conference on Networking and Network Applications. NaNA, Cyber Security, Islamabad Pakistan. His research interest
IEEE, pp. 379–384. includes, IoT security, Digital Forensic, Edge Computing and
Tlili, F., Fourati, L.C., Ayed, S., Ouni, B., 2022. Investigation on vulnerabilities, threats Botnet.
and attacks prohibiting UAVs charging and depleting UAVs batteries: Assessments
& countermeasures. Ad Hoc Netw. 129, 102805.

25
H.J. Hadi et al. Journal of Network and Computer Applications 213 (2023) 103607

Yue Cao received his Ph.D. degree from the Institute for Abdul Majid Jami is currently a Master Student of Cy-
Communication Systems (ICS) formerly known as Centre for berSpace Security at the School of Cyber Science and
Communication Systems Research, at University of Surrey, Engineering, Wuhan University in China. He obtained his BS
Guildford, UK in 2013. Further to his Ph.D. study, he in Information Technology from Arid University, Pakistan
had conducted research fellow at University of Surrey, and his research interest in Malware Analysis, UAV Security,
and academic faculty at Northumbria University, Lancaster Intrusion Detection and Digital Forensics.
University, UK and Beihang University, China; and he is
currently the Professor at School of Cyber Science and En-
gineering, Wuhan University, China. His research interests
focus on Intelligent Transport Systems, including E-Mobility,
V2X, Edge Computing.
Qiang Ni received the B.Sc., M.Sc., and Ph.D. degrees in
engineering from the Huazhong University of Science and
Khaleeq Un Nisa received the B.S Software Engineering Technology, Wuhan, China. He is currently a Professor
degree from Fatima Jinnah Women University, Rawalpindi, and the Head of the InfoLab21, Communication Systems
Pakistan. She has received MS Software Engineering degree Research Group, School of Computing and Communications,
from Riphah International University, Islamabad, Pakistan. Lancaster University, Lancaster, U.K. His research interests
She is Currently working in Cyber Reconnaissance Combat include future generation communications and networking
Lab, Pakistan. Her research interest includes, IoT security, systems, including green communications, cloud systems,
Network security and quality testing. cognitive radio network systems, heterogeneous networks,
5G and SDN, the IoT, and big data analytics.

26