SVM-Based System for Detecting

Blackhole Attacks

Abstract— Mobile Ad hoc Networks (MANETs) are now and others. When it comes to routing challenges, nodes
essential in the modern world. They are crucially required in a require a routing function in order to communicate with one
variety of circumstances where it is necessary to promptly put another efficiently. The fundamental purpose of the routing
together a network without having the luxury of time or function is to make sure that the message from the sender
resources to set up infrastructure or even have human takes the shortest path possible to get to its target location.
involvement. There are numerous uses for ad hoc networks. For a variety of reasons, the traditional infrastructure routing
They can be utilized, among many other things, in education, protocols are ineffective for MANETs. One of them is that
on the battlefield, during rescue operations, etc. These
MANETs lack infrastructure, hence routers are not present to
networks are characterized by high mobility and limited
computing power, storage, and limited energy. They don't
execute the routing task in these networks. Rather, practically
employ infrastructure-related communication tools because every network node performs this role. For MANETs, new,
they lack infrastructure. As an alternative, these networks rely improved routing protocols were created for this reason.
on one another for communication and routing. In MANETs, The two primary types of MANETs routing protocols are
each node locates another node within its communication range proactive (table driven) and reactive (on demand) [4] [5].
and uses it as a hop to pass the message through another node, When using a table-driven protocol, routing data is routinely
and so on. Routers, servers, firewalls, and other specialized
updated anytime a change occurs. In contrast, the on-demand
equipment are present in traditional networks. While in ad hoc
networks, each node serves several purposes. As an example,
routing protocol only collects routing data when it is required
nodes handle the routing task. They are therefore more [6]. Ad hoc On Demand Distance Vector (AODV) is one of
susceptible to attacks than traditional networks. The primary the well-known on-demand routing protocols. Compared to
objective of this research is to provide a method for blackhole other on-demand routing protocols, it performs better [7][8].
attack detection utilizing anomaly detection based on Support However, MANETs are vulnerable to several attacks because
Vector Machine (SVM). This detection system seeks to examine of their limitations. The blackhole attack is one of these. The
network traffic and spot anomalies by observing node activity. Blackhole is a significant attack that has a major impact on
Attacking nodes in blackhole situations exhibit specific network performance. The attacker's node acts as the shortest
behavioral traits that set them apart from other nodes. Our path to the target in this attack. The packets it receives are
SVM-based detection system is capable of efficiently detecting then dropped. As a result, it significantly influences the
these properties. A traffic under blackhole attack is generated network delivery ratio.
using the OMNET++ simulator to test the efficacy of this
method. The malicious node is then identified based on the The key contribution of this paper is as below:
classification of the traffic into malicious and non-malicious.
• Analyzing and classifying the various methods and
The proposed solution's results shown high accuracy in
contrasting the various mitigation strategies for blackhole
identifying blackhole attacks.
Keyword: Adhoc Networks, MANET, Routing Protocols, attacks in MANETs.
Blackhole Attacks. • Creating a dataset for analyzing blackhole attacks
with OMNET++ in order to carefully examine the traffic and
I. INTRODUCTION research nodes behavior under attack.
T Mobile Ad hoc Networks (MANETs) can be created • Developing a detection system to identify malicious
without the aid of administrators or traditional base nodes.
infrastructure; instead, they are pre-configured to operate on
their own. In general, there are two primary types of wireless The rest of this paper is structured as follows:
networks: infrastructure-based and infrastructure-less [1]. In Background information on mobile ad hoc networks,
infrastructure networks, wireless devices are configured by including their uses, traits, and difficulties, is provided in
administrators so they may connect to fixed base equipment Section 2. The related study and initiatives made by other
to be provided with certain services such as routing, storage, researchers on this subject are reviewed in Section 3. The
and security. Nodes in infrastructure-less networks, like technique of our suggested solution is described in section 4.
MANETs, are self-configured and rely on one another rather The paper is concluded in Section 5, which also offers ideas
than fixed base infrastructure. In other words, each node for additional research.
carries out a variety of tasks [2].
II. BACKGROUND
Due to these restrictions, MANETs have several
particular challenges that make them different from Mobile ad hoc networks are distinguished from
traditional infrastructure networks. Two vital factors taken traditional networks in a variety of ways, particularly in
into consideration when building MANETs are the terms of security. This section covers the uses of MANETs,
challenges with security and routing [3]. Regarding security issues, and typical attacks.
challenges with security, MANETs generally lack the 1. MANETs Characteristics
infrastructure that enables conduct perimeter security
functions using devices such as firewalls, routers, IDS, IPS,

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

 MANETs have a variety of unique characteristics that routing protocol [16]. Routing protocols can often be divided
support their distinctive uses. These characteristics are listed into three categories: proactive, reactive, and hybrid [17].
below:
In proactive routing protocols, each node maintains a
• Lack of infrastructure: MANETs are described as table of all possible routes, and the table is updated at
networks with no infrastructure. As a result, they are predetermined intervals whenever there is a change in the
effective in terms of both time and money. They can be network (for example, when a node enters or exits the
created quickly and at relatively low cost [9]. However, it network). It is called a table-driven protocol for this reason.
also makes them more exposed to attacks than traditional The main goal of proactive protocol is effectiveness. The
networks. nodes respond to changes well and are constantly prepared to
transfer data effectively. The main problem is the network
• Distributed management: Due to the lack of
overhead associated with updating modifications at regular
centralized control, these functions are divided up among the
intervals.
nodes. Thus, network structure, node authentication, and data
security are all impacted [10]. 4. Blackhole Attack
• Cooperativeness: MANETs use peer-to-peer The AODV routing protocol is exploited by the blackhole
architecture rather than client-server architecture which is attack. Every node in the network has a routing table, where
usually used in traditional networks. Hence, nodes should it stores all the data on the most effective routes to certain
work together to provide the functions that are not performed destinations, as part of the AODV routing protocol. A node
by the MANETs' lack of infrastructure security and initially determines if it has the necessary information in its
centralized management. The goal of this collaboration is to own routing table before attempting to send a packet to
increase nodes confidence. another node. It starts a discovery procedure if it is not found
or if the necessary route is no longer operational. The node
• Multi-hop routing: Routing is one of the tasks
broadcasts a route request (RReq) to all of its neighbors in
carried out by nodes themselves. A node uses nearby nodes
this situation. The route reply (RRep), which includes the
as hops to get to the destination in order to transmit a
most recent sequence number, broadcast ID, and hop count,
message to another node [11]. This process is known as
is subsequently sent back if the receiving node is the
multi-hop routing.
destination node [18]. If not, it compares the routing
• Dynamic topology: Because MANETs have no destination sequence and determines whether its own routing
perimeter borders, nodes can enter and leave the network at database has any routing entries to the destination. It sends
any time without warning. Furthermore, networks can form out an RReq to its neighbors if the sequence number is less
independently at any time because there is no centralized than or equal to the one it currently holds. A fresh route
management [12] [13]. indicates a higher sequence number, which it uses to update
its own table and send an RRep back to the node it sent the
• Decentralized architecture: Each network node is RReq to [19]. A malicious node injecting itself into a
autonomous. Since they are self-configured, they can join or network and then claim to have the fastest route to the target
leave the network without any assistance. They are free to causes a black-hole attack [20]. An illustration of this
make these choices on their own. Additionally, even if they procedure is shown in Fig. 2. Node "S" is trying to go to
are meant to, they are free to forward or drop data packets node "D." Therefore, it broadcasts "RREQ" to the nearby
[13]. nodes. Node "M" quickly responds by injecting itself and
• Limited resources: Because they rely on batteries announcing that it has the optimal path. As soon as
and have weaker processing units, nodes in MANETs have communication begins, node "M" discards all data
minimal power and processing capabilities. The main issue transmitted through it.
with a limited power supply is that it increases the 4. Blackhole Attack
vulnerability of MANET nodes to denial-of-service attacks
[14]. In this instance, the attacker sends extra packets to The AODV routing protocol is exploited by the blackhole
nodes to drain their batteries. attack. Every node in the network has a routing table, where
it stores all the data on the most effective routes to certain
2. MANETs Security Challenges destinations, as part of the AODV routing protocol. A node
Due to their constrained resources, lack of perimeter initially determines if it has the necessary information in its
protection, no physical security, and unpredictable topology, own routing table before attempting to send a packet to
MANETs are more vulnerable than traditional wired another node. It starts a discovery procedure if it is not found
networks. They are more vulnerable to both internal and or if the necessary route is no longer operational. The node
external network threats. Passive attacks and Active attacks broadcasts a route request (RReq) to all of its neighbors in
are the two primary types of MANETs attacks [15]. this situation. The route reply (RRep), which includes the
most recent sequence number, broadcast ID, and hop count,
Active attacks are those where the attacker tries to change is subsequently sent back if the receiving node is the
or distort the data being sent into the network. Active attacks destination node [18]. If not, it compares the routing
include the blackhole attack, which is the key focus of this destination sequence and determines whether its own routing
work, routing table overflows, impersonations, rush attacks, database has any routing entries to the destination. It sends
denial-of-service attacks, Byzantine attacks, packet out an RReq to its neighbors if the sequence number is less
replication attacks, and distributed denial-of-service attacks. than or equal to the one it currently holds. A fresh route
3. MANETs Routing Protocols indicates a higher sequence number, which it uses to update
its own table and send an RRep back to the node it sent the
The procedures or guidelines that specify how nodes will RReq to [19]. A malicious node injecting itself into a
communicate with one another are known as routing network and then claim to have the fastest route to the target
protocols. Finding the most efficient path for a message to go causes a black-hole attack [20]. An illustration of this
from the originator to the destination is the main goal of the
procedure is shown in Fig. 2. Node "S" is trying to go to between (S) and (D) is not the same, (S) sends a control
node "D." Therefore, it broadcasts "RREQ" to the nearby packet to its neighbors and waits for responses. The
nodes. Node "M" quickly responds by injecting itself and communication ratio is then calculated using the requests
announcing that it has the optimal path. As soon as submitted and responses received.
communication begins, node "M" discards all data
To monitor the node's actions, determine if they are
transmitted through it.
normal (N) or malicious (M), and estimate trust and
III. RELATED WORK reputation, the authors of [24] suggested a Node Activity-
based Trust and Reputation Estimation (NA-TRE) solution.
Since ad hoc networks are becoming more and more The author claims that there are three different node states.
common and are being used in a variety of fields, blackhole Nodes make their best efforts to cooperate and adhere to
attacks have attracted the interest of numerous academics. routing requirements in the Normal State (NS). Resource
According to Fig. 3, the most often suggested solutions can Limitation State (RS), in which nodes cooperate little
be divided into four categories. because they use little power heavy traffic, out of
A. Enhanced Routing Protocol Based communication range, etc. Malicious State (MS), in which
nodes cause network disruption by starting denial of service
According to this mitigation strategy, improving the attacks, creating new paths, delaying packets, or engaging in
current protocols will make them better able to detect and other malcious behaviours that have an effect on the network.
stop blackhole attacks, as suggested by several researchers. A "Semi-Markov probability decision procedure" is used for
The MBDP-AODV advanced routing protocol was put prediction in order to proactively separate several situations.
forth in [20]. This improved methodology employs several A reputation and trust system against blackhole attacks
statistical elements, such as the mean and standard deviation. was suggested by the authors in [25]. This is based on nodes'
Figures ought to increase reasonably under typical mutual trust. If node (A) trusts node (B) in this scenario, then
circumstances. The numbers, however, suspiciously increase (B) can also trust (A). In the same way, if (B) trusts (C) and
quickly when there is an attack. Three phases make up the (C) trusts (A). So (A) can trust (B). Every node in the
suggested remedy. Dynamic threshold calculation and network is outfitted with a reputation table in order to
suspicion come first. The source node determines a threshold implement this solution. Information regarding the neighbor
value for the destination's sequence number at this phase. node's behavior is kept in the table. The behavior is measured
The second phase is detection, during which the suspicious and kept up to date. Therefore, after a message is transmitted
packet is found and the network's nodes are all informed of from the source to the destination, the destination should
the malicious node ID. The malicious node is stopped from acknowledge receipt of the message. Each node receives an
participating in the network during the third step, which is acknowledgement, which is then transmitted back. The
prevention. trusted table is modified negatively in the same way if a
By setting bait timers in each node, the authors of [7] message is not received.
suggested a method for finding blackhole nodes. The timer C. Acknowledgment Based
for the baiting is set to a random number. When the baiting
timer reaches the predetermined time, fake ID broadcasts The acknowledgement-based solution uses a mechanism
start to run. All requests, regardless of their nature, will to generate acknowledgment packets across source or
receive a response from the blackhole. They respond to those intermediate nodes in order to mitigate blackhole attacks.
false queries that serve as bait. As a result, the sending node Prior to route determination, acknowledgement packets are
locates and maintains the blackhole node in a specific table. sent. The nodes that choose not to respond are either self-
Malicious nodes are ignored when the true requests are sent centered, low-energy, or malicious nodes.
out in accordance with the data kept in the malicious nodes Ad hoc On-demand Multipath Secure Routing
table. (AOMSR), an improved routing system based on
The authors in [22] suggested to modify the existing acknowledgement, was suggested by the authors in [27].
AODV protocol by adding Neighbor Credit Table to each Based on the maximum delay in receiving the data, this
node in the network. Whenever a data packet is sent from a routing protocol states that a source node must maintain
neighbor or forwarded by a neighbor, the neighbor is several pathways from source to destination.
assumed as a genuine node and its credit value is increased in The authors of [28] proposed an expansion of the
the table. Even the genuine node, when not participating, get acknowledgment-based approach that takes into account the
poor credit values. Then when a node wants to use the choice of the intermediate nodes that are energy-efficient and
neighbor node for transmitting a message, it checks first the uncongested for communication, as well as session key
value of the table. If the neighbor node doesn’t have enough agreements, a counter base end to end the cycle of
credit, then it is not trusted and another hop should be used. acknowledgement, and an authenticating of Ack packets by
B. Reputation and Trust Based message digest.
A reputation system is a system that collects, examines, D. Intrusion Detection System Based
and disseminates data about nodes' actions based on their An alarm system is used by an intrusion detection system.
prior interactions. It sends a warning to the system when it finds an attack [29].
According to [23], a selfish node can be eliminated from The IDS system has an audit register where all the data is
the network using the Selfish Node Removal using stored for analysis and output that is used to make
Reputation Model (SNRRM). The selfish node is identified judgements.
by the node's current energy level and communication rate, IDS was the remedy suggested by the authors in [30].
according to the author. Only the sender's reputation value is (DPAA-AODV). The two modes of operation for this
examined if both the sender (S) and the distention (D) are protocol are online and offline. To identify a trustworthy
inside the communication range. If the communication range
feature of the blackhole detection dataset (BDD) data set, the 6 can deceive all of its neighbors into considering that it was
ReliefF Model's offline phase one is applied. This is done to the one closest to them. As a result, it gets as many requests
improve selection accuracy. The learned features from the as it can. To put it another way, when nodes look for the best
previous mode are chosen for the online mode. If the findings routes and send their RReq, they will initially appear as a
indicated that the threshold had been surpassed, a malicious neighboring node and will send their RRep as soon as they
node might be responsible. can.
Host-based IDS was employed in [31], and data on nodes' C. Data Collection
typical activity was gathered there. For the purpose of
Results were gathered from two scenarios: one in which
simulating typical malicious node activity, they used the
it was considered that all nodes were cooperative and acting
GloMoSim simulator. The feature selection was then applied
normally, and another in which it was assumed that one node
using a machine learning technique (Weka 3.7.11), which
was acting maliciously. The detection system was then used
helped to identify the rogue node. This was accomplished by
to analyze the outcomes of the two scenarios. The most
applying six features: the quantity of RREQ delivered,
important data from the simulation ran in OMNET++ were
quantity of RREP forwarded, quantity of high destination
included in the dataset. It included the AODV request, the
sequence number, quantity of low count of hops to
nodes' transmission power when sending packets, and the
destination, quantity of source nodes, and quantity of
kind of data transfer (Broadcast vs. Unicast).
destination nodes.
D. Feature Selection
IV. METHODOLOGY
There are three elements that contribute to blackholes'
Four steps make up the methodology utilized in this
attack conduct in this research, as indicated in section 5. The
paper: Making the data necessary for machine learning
first characteristic is that the malicious node strengthens its
analysis is the initial step. This is accomplished using the
ability to deceive the other nodes into considering it is the
OMNET++ simulator to produce traffic data that closely
one closest to them. The second is that rogue nodes hardly
resembles real traffic while conducting a blackhole attack.
ever send any RReq requests. They would rather respond to
The generated data is then gathered in a certain manner so
as many inquiries as they can. Last but not least, they almost
that it can be examined further at a later time. The traffic
never transmit and always use unicast.
records that are gathered share a few basic traits or
characteristics. Support Vector Machine (SVM) analysis is E. Machine Learning Using (SVM)
used to classify the traffic into malicious and normal traffic
based on these behaviors. This analysis allows for the Support Vector Machine (SVM), the well-known
identification and blocking of malicious nodes. machine learning technique is frequently utilized for
activities involving pattern classification. Figure 10 shows
A. Proposed Solution the model, which is made up of three lines. The best
categorization line is the line in the center. The two
All nodes in MANETs should cooperate. In other words,
additional lines are known as the margin lines. Patterns are
they must rely on one another to fill in for the functions that
divided into two classes using these lines [34]. Based on the
the absence of infrastructure prevents. Otherwise, the
traffic analysis, this model is employed in our situation to
network as a whole won't function effectively. One attack
distinguish between malicious and normally behaving nodes.
that seeks to corrupt MANETs is called Blackhole. By
observing the behavior of such hostile nodes on a network,
such attacks can be discovered. These common behavioral
traits of theirs can be summed up as follows:
They nearly never send any RREQ, they boost their
transmission power such that they can respond to the
majority of RREQ, and they almost never broadcast and
almost always unicast. Our solution is to create a detection
system based on SVM algorithm, which can identify any
rogue node based on the above-mentioned behavioral traits.
Label those nodes after which they should be isolated from
the network.
B. Data Generation
It is possible to simulate both the actions of good and bad Fig. 10. Support Vector Machine (SVM)
nodes using OMNET++ 5.7. On 7 nodes, the simulation was
performed. One of them served as a transmitter and is V. RESULTS
stationary (node1). Since it is irrelevant to illustrating the The simulator was configured to examine seven nodes
behavior of the mobile nodes, which form the simulation's interacting with one another. A rogue node that imitated a
central nodes, this node is excluded from the graphs. This blackhole attack was set up as one of the nodes. 10 minutes
simulation involves two scenarios. In scenario one, all nodes were given for running the simulation. 21,337 records
are cooperative and there are no malicious nodes. produced by the simulation were among the records the
Accordingly, all of the nodes are acting normally. In the system examined. The algorithm eventually succeeded in
second scenario, node 6 was configured to act maliciously categorizing the records into two groups: legitimate records
while the other nodes carry on with their regular operations. and harmful records. 16,609 of the 21,337 records were
All nods have a 1 mW radio transmission power setting. classified as normal, while 4,726 were classified as
In this instance, node 6's radio transmission power was malicious. This was supported by the following three main
set at 5 mW. By boosting its radio transmission power, node aspects:
 whether a change is made to the transmission power. The Fig. 8. Results of the simulator in presence of a
rogue node modifies its transmission power as previously blackhole attack
mentioned in order to look close to the Rreq sender. . Conclusion and Future Work
• Exceptionally increased response to as many requests as MANETs are networks without any physical
feasible. infrastructure; instead, they rely on nodes cooperation by
• Almost never transmit broadcast messages; only ever enabling both client and router functionality. These networks
send unicast. are underresourced and lack numerous security elements.
They are therefore more vulnerable than networks using
Based on the attributes, the machine learning algorithm traditional infrastructure. In this paper, we covered the
clearly detected the malicious documents. Through analysis various MANET applications, security issues, and one of the
of its behaviors based on the features listed above, the system most widespread attacks, the blackhole attack. We reviewed,
has demonstrated a high degree of accuracy in identifying organized, and contrasted the remedies suggested in the
malicious nodes. literature to decrease blackhole attacks. Then, using machine
learning, a suggested method for identifying and preventing
such attacks was put out.
Normal Behaviour
180 640
We used OMNET++ to simulate a malicious node in a
160 MANET network in order to thoroughly explore blackhole
620 attacks, and we produced a data set that we used for analysis
140
120 600 and examining the behavior of the malicious node serving as
100 the blackhole attack. Transmission power, the amount of
580 responses in comparison to the other nodes, and the type of
80
60 560 communication—broadcast or unicast—were our three main
40 areas of attention for spotting blackhole attacks. Machine
540
20 learning was used to thoroughly investigate these three
0 520 features. This research's limitation is that just seven nodes
Node 2 Node 3 Node 4 Node 5 Node 6 Node 7 were used for the simulation, and only one node served as the
attacker. A larger network may be put up in the future for
AODV Rrep AODV Rreq greater analysis, and the attackers may consist of more than
one node. This will make it possible to analyze blackhole
Fig. 7. Results of the simulator in absence of a attacks in greater networks in greater depth and to analyze
blackhole attack network data when there are multiple attacker nodes present.

In our scenario, node 6 has its radio transmission power References

boosted to 5 mW while the other nodes are left with their
default settings of 1 mW. The second characteristic of
blackhole attackers is that they respond to as many requests 1. M. H. a. M. S. A. a. M. H. a. M. A. a. R. A. A. a. J. M. A.
as they can while remaining nearly silent in terms of sending Hassan, "Mobile ad-hoc network routing protocols of
route requests (RReq). The blackhole attacker hardly rarely time-critical events for search and rescue missions,"
broadcasts, and all of its communication takes the form of Bulletin of Electrical Engineering and Informatics, vol.
unicast, which is the final but not the least aspect. As seen in 10, no. 1, pp. 192--199, 2021.
Fig. 7, the graph's six nodes behave normally. They all send 2. P. a. V. S. a. R. D. B. a. D. S. a. o. Rani, "Mitigation of
RReps and RReqs in a typical manner. In the words, each black hole attacks using firefly and artificial neural
node has a relative proportion between the number of RReq network," Neural Computing and Applications, pp. 1--
and the number of RRep. 11, 2022.
The numbers of RRep sent by node 6 to the other nodes 3. S. K. a. S. T. Prasad, "Performance comparison of
are vastly different, as seen in Fig. 8. The transmission power multipath routing protocols for mobile ad hoc
of node 6 has been increased to 5 mW, whereas the network," International Journal of Systems, Control and
transmission power of the other nodes is 1 mW, explaining Communications, vol. 13, no. 1, pp. 82--98, 2022.
why this is the case. Additionally, compared to the rest of the 4. Shrivastava, Prashant Kumar, and L. K. Vishwamitra.
nodes and nodes that are close toBehaviour
Malcious one another, node 6 sends "Comparative analysis of proactive and reactive routing
almost no RReqs. protocols in VANET environment." Measurement:
6000 1400
Sensors 16 (2021): 100051.
5000 1200
1000
5. Mukti, Fransiska Sisilia, et al. "A Comprehensive
4000 Performance Evaluation of Proactive, Reactive and
800
3000 Hybrid Routing in Wireless Sensor Network for Real
600
2000 Time Monitoring System." 2021 International
400
Conference on Computer Science and Engineering
1000 200
(IC2SE). Vol. 1. IEEE, 2021.
0 0
Node 2 Node 3 Node 4 Node 5 Node 6 Node 7 6. A. M. a. K. S. a. M. A. H. Shantaf, "Performance
evaluation of three mobile ad-hoc network routing
AODV Rrep AODV Rreq protocols in different environments," in 2020
International Congress on Human-Computer
Interaction, Optimization and Robotic Applications
(HORA), 2020.
7. A. a. A. Z. M. Yasin, "Detecting and isolating black-hole 23. A. M. a. D. H. El-Semary, "BP-AODV: Blackhole
attacks in MANET using timer based baited technique," protected AODV routing protocol for MANETs based
Wireless Communications and Mobile Computing, on chaotic map," IEEE Access, vol. 7, pp. 95197--95211,
2018. 2019.
8. D. a. M. A. a. A. S. a. P. S. Ramphull, "A review of 24. M. a. o. ponnusamy, "Detection of selfish nodes
mobile ad hoc NETwork (MANET) Protocols and their through reputation model in mobile adhoc network-
Applications," in 2021 5th international conference on MANET," Turkish Journal of Computer and
intelligent computing and control systems (ICICCS), Mathematics Education (TURCOMAT), vol. 12, no. 9,
2021. pp. 2404--2410, 2021.
9. D. Kanellopoulos, "Congestion control for MANETs: 25. A. a. O. M. a. D. N. a. T. A. Hammamouche,
An overview," ICT Express, vol. 5, no. 2, pp. 77--83, "Lightweight reputation-based approach against simple
2019. and cooperative black-hole attacks for MANET,"
10. D. a. S. V. K. Kanellopoulos, "Survey on power-aware Journal of information security and applications, vol.
optimization solutions for manets," Electronics, vol. 9, 43, pp. 12--20, 2018.
no. 7, p. 1129, 2020. 26. R. L. a. R. C. Raghavendar, "Node activity based trust
11. B. S. K. a. A. M. a. K. A. R. a. G. P. Anibrika, "A Survey and reputation estimation approach for secure and QoS
of Modern Ant Colony Optimization Algorithms for routing in MANET," International Journal of Electrical
MANET: Routing Challenges, Perpectives and and Computer Engineering, vol. 6, no. 9, p. 5340, 2019.
Paradigms," International Journal of Engineering 27. D. a. D. P. Dave, "An effective Black hole attack
Research and Technology (IJERT), 2020. detection mechanism using Permutation Based
12. N. a. C. U. Yadav, "Secure Routing in MANET:A Acknowledgement in MANET," 2014.
Review," in 2019 International Conference on Machine 28. M. A. a. D. B. Hussain, "Preventing Malicious Packet
Learning, Big Data, Cloud and Parallel Computing Drops in MANETs by Counter Based Authenticated
(COMITCon), 2019. Acknowledgement.," Ingénierie des Systèmes
13. S. J. J. a. R. A. a. S. S. Thangaraj, "Comprehensive d’Information, vol. 25, no. 2, pp. 173--181, 2020.
Learning on Characteristics, Applications, Issues and 29. M. P. a. M. R. a. A. S. Preet, "RESEARCH
Limitations of Manets," International Journal of TECHNOLOGY INTRUSION DETECTION SYSTEM
Innovative Technology and Exploring Engineering FOR MANET," INTERNATIONAL JOURNAL OF
(IJITEE) ISSN, pp. 2278--3075, 2019. ENGINEERING SCIENCES, 2020.
14. K. A. a. A. S. Alshaker, "Availability in IOT for MANET 30. F. a. Y. M. B. a. N. A. Albalas, "Detecting black hole
network," Materials Today: Proceedings, 2021. attacks in MANET using relieff classification
15. N. a. M. R. Sivapriya, "Analysis on Essential Challenges algorithm," in Proceedings of the 5th International
and Attacks on MANET Security Appraisal," Conference on Engineering and MIS, 2019.
JOURNAL OF ALGEBRAIC STATISTICS, vol. 13, no. 3, 31. M. B. a. K. Y. M. a. A. M. Yasin, "Feature Selection for
pp. 2578--2589, 2022. Black Hole Attacks," J. Univers. Comput. Sci., vol. 22,
16. M. a. A. L. a. A. R. S. a. H. M. A. a. A. S. a. S. M. A. no. 4, pp. 521--536, 2016.
Maad Hamdi, "A Review of Applications, 32. M. Y. Ms Katakam and Adilakshmi, "Black hole Attack
Characteristics and Challenges in Vehicular Ad Hoc Detection Using Machine Learning Algorithms in
Networks (VANETs)," in 2020 International Congress MANET--Performance Comparision," 2020.
on Human-Computer Interaction, Optimization and 33. J. a. F. S. a. L. W. a. K. J. a. M. X. a. Z. R. a. S. D. Xu, "An
Robotic Applications (HORA), 2020. algorithm for determining data forwarding strategy
17. S. Yogarayan, "Wireless Ad Hoc Network of MANET, based on recommended trust value in MANET,"
VANET, FANET and SANET: A Review," Journal of International Journal of Embedded Systems, vol. 12, no.
Telecommunication, Electronic and Computer 4, pp. 544--553, 2020.
Engineering (JTEC), vol. 13, no. 4, pp. 13--18, 2021. 34. Ibrahim, Ibrahim, and Adnan Abdulazeez. "The role of
18. H. Al-Refai, "An Enhanced AODV Protocol Against machine learning algorithms for diagnosing diseases."
Black Hole Attack Based on Classification Algorithm," Journal of Applied Science and Technology Trends 2.01
Int. J. Open Problems Compt. Math, vol. 13, no. 2, 2020. (2021): 10-19.
19. F.-H. a. C. H.-P. a. C. H.-C. Tseng, "Black hole along
with other attacks in MANETs: a survey," Journal of
Information Processing Systems, vol. 14, no. 1, pp. 56--
78, 2018.
20. S. a. C. S. Gurung, "A dynamic threshold based
approach for mitigating black-hole attack in MANET,"
Wireless Networks, vol. 24, no. 8, pp. 2957--2971, 2018.
21. P. Sarao, "Performance Analysis of MANET under
Security Attacks," Journal of Communications, vol. 17,
no. 3, 2022.
22. K. a. S. M. Rama Abirami, "Preventing the impact of
selfish behavior under MANET using Neighbor Credit
Value based AODV routing algorithm," vol. 43, no. 4,
pp. 1--7, 2018.