2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

Optimizing AODV Routing Protocol to Improve

Quality of Service Performance for V2V
Communication
2024 International Conference on Smart Computing, IoT and Machine Learning (SIML) | 979-8-3503-6410-1/24/$31.00 ©2024 IEEE | DOI: 10.1109/SIML61815.2024.10578106

Ketut Bayu Yogha Bintoro Tri Kuntoro Priyambodo Muhammad Nurul Musthofa
Dept.of Computer Science and Dept.of Computer Science and Dept. of Informatic Engineering,
Electronics, Faculty of Mathematics Electronics, Faculty of Mathematics Faculty of Sciences, Technology and
and Computer Sciences and Computer Sciences Design
Gadjah Mada University Gadjah Mada University Universitas Trilogi
Yogyakarta, Indonesia Yogyakarta, Indonesia Jakarta, Indonesia
[email protected] [email protected] [email protected]

Dept. of Informatic Engineering,

Faculty of Sciences, Technology and
Design
Universitas Trilogi
Jakarta, Indonesia
[email protected]

Abstract—The study focuses on Quality of Service (QoS) We proposed a Learning Automata Ad Hoc On-Demand
issues in learning automata-based AODV (LA-AODV) for V2V (LAAODV) routing protocol to optimize the selection of
communication protocols. It analyzes and compares the LA- relay nodes to improve the efficiency of V2V
AODV protocol with the original AODV protocol in dynamic communication. LAAODV achieves better relay node
traffic situations. The study evaluates critical metrics such as the selection by utilizing real-time information about vehicles'
number of vehicles, average speed, congestion spots, and vehicle
position, acceleration, and speed to predict and select clusters
time loss across various traffic scenarios, including smooth
traffic, steady flow traffic, and heavy traffic. NS3 and SUMO of responsive relay nodes in dynamic traffic scenarios[8].
are used as essential tools for this evaluation. The study finds Previous studies have addressed challenges with the
that the LA-AODV protocol performs better in smooth traffic AODV routing protocol, such as reducing routing overhead
conditions than AODV. It demonstrates a 38.27% higher packet values by forming network routes when necessary[9], an
delivery ratio (PDR), an 85.26% lower packet loss ratio (PLR), optimized and efficient routing protocol for IoV[10],) and
and a remarkable 56.62% increase in average throughput another Alternative approaches, including Node Trends
compared to AODV. Both protocols exhibit similar Prediction[11] and Mobility and Detection Aware AODV
performance in steady flow traffic with high PDRs (97.60% and (MDA-AODV)[12], have also shown promise in improving
96.87%, respectively) and low PLRs (2% and 3%, respectively).
packet delivery ratios and average delays compared to
However, AODV outshines LA-AODV with a higher PDR
(97.60%) than LA-AODV (96.87%) in heavy traffic. LA-AODV, standard AODV. The cluster-based communication
on the other hand, excels in average throughput by 17.24% in strategy[13], integrating prediction supported by learning
heavy traffic. The study offers valuable insights for improving automata[14] and employing various channel reservation and
V2V communication protocols and enhancing communication multipath routing[15] to enhances the learning process in
efficiency to handle dynamic and traffic situations. V2V communication within VANET.
Another study introduces DDSLA-RPL[16], utilizing
Keywords—V2V Communication, Learning Automata, learning automata to adjust parameter weights, enhancing
AODV Routing Protocol, Dynamic traffic, Quality of Sevices. network service quality and extending node lifespan. While
DDSLA-RP exhibits precision and adaptability, it requires
improvement in various situations, and the selection of
I. INTRODUCTION techniques should consider the specific characteristics of the
network, weighing the limitations of fuzzy clustering[17], C-
V2V communication is essential to modern means, and K-means[18]. Another research applying the
transportation systems, particularly with the increasing PSO[19], leap-frog algorithm[20], and basic learning
adoption of autonomous vehicles and V2V technology[1]. automata[21] to ensure channel availability for V2V
Routing protocols play a crucial role in ensuring the communication in VANET. In contrast, LAAODV has
reliability and efficiency of V2V communication in dynamic effectively addressed communication enhancements in
traffic scenarios. AODV protocol is characterized by low data dynamically changing traffic situations. DP-AODV[22] and
transfer rates[2], increased delay times[3], [4], diminished FLOW-AODV[23] serve as routing protocols for vehicular
throughput, and data congestion due to dynamic network communication networks. DP-AODV dynamically adjusts
topologies[5]. These issues significantly impact the overall power, and FLOW-AODV utilizes machine learning to select
efficiency and reliability of V2V communication, making relay nodes intelligently, optimizing QoS parameters and
improvements in these areas critical[6], [7]. communication efficiency.
Several studies have focused on enhancing the quality of
service (QoS) in various contexts using relay nodes

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
979-8-3503-6410-1/24/$31.00 ©2024 IEEE 180
 2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

modification[24], such as in V2V communication within the vehicle's position using speed and acceleration. The
VANET[25]. The AODV routing protocol requires equations used are (2) and (3).
modifications to exploit real-time information on vehicle
= ∑ ,
, ( +( . + !" ($ % ∗ 2(
location, acceleration, and velocity to choose a group of
#
(2)
responsive relay nodes[8], [10]. These modifications are

= ∑ ( +( . !"# ($ % 2(
necessary to implement AODV for V2V communication,
,
) , *
especially in dynamic traffic scenarios[11]. (3)

• $ + = (
This study aims to improve V2V communication's QoS
− − 1 at the beginning of iteration −
in dynamic VANETs using the LA-AODV approach, which Where:

1 = 0,
selects clusters of responsive relay nodes for enhanced
• $ / = ( − − 1 at the beginning of iteration −
reliability and efficiency. Simulations using NS3 show that
1 = 0,
LA-AODV is more effective than standard AODV in traffic
management. Integrating learning automata with the AODV
routing protocol is a promising approach to improve the QoS And:

0 : Maximum Iteration,
parameters of V2V communication and prevent accidents in : Prediction time, where t = 1,2,3, and t < M,

1 : Vehicle k,
dynamic traffic situations. The study section includes the
background of the study, Section 2, discuss about the
proposed method and simulation parameters, Section 3 is : Total number of vehicles within the transmission
result and discussion, then the research conclusion can be range,
found in Section 4. : Vehicle speed at time t.
II. PROPOSED METHOD
Equation (2) predicts vehicle position on the x-axis at a
The research method comprises several vital phases. specific simulation time ('t'). (3) uses additional variables to
Initially, the study identifies critical issues in contemporary forecast the y-axis position accurately. Precise positioning is
vehicle communication systems. Fig.1. depict the LA-AODV essential for efficient communication. The LAAODV
flowchart. protocol utilizes these equations to enhance the efficiency of
the vehicle communication network by predicting vehicle
positions. Vehicles share data to determine their minimum
predicted positions, which updates the routing table, (4) is
then used to determine the vehicle status with the minimum
distance and speed.

234_ ) = 6(|∆ 234_ − ∆ 234 )

| (4)

∆ 234_ = ( 234_ − 234_

Where:
9
∆ 234_ = ( 234_ − 234_
(5)
) )9 ) (6)

Equation (4) predicts vehicle position using changes in

the x and y axes, with values from (5) and (6). (5) calculates
x-axis change by subtracting the predicted vehicle position at
time t+1 from the predicted position at time t. Similarly, (6)
calculates y-axis movement by subtracting two values.
Variable predicts surrounding vehicle position using
expected x and y coordinates from (4). Equation (7) utilizes
the Euclidean distance formula to determine the minimum
value that compares the optimal movement changes of
vehicles along the x and y axes for each vehicle during two
Fig. 1. LA-AODV flowchart diagram prediction time intervals.

(? 234 − 234? 2−
:234 = 0;< =∑ > A (7)
Fig. 1 illustrates the step-by-step process of the
1≤ , ≤0 ++1 +

(? 234 ?2
) 1=1, =1
LAAODV protocol, a communication protocol used by
/+1
− 234
/
vehicles for efficient and reliable V2V communication.
LAAODV is a vehicle communication protocol that uses
parameters such as speed, acceleration, and vehicle positions
to ensure reliable V2V communication. Nodes with a Total Equation (7) predicts vehicle positions to support routing
Weighted Ratio (TWR) below 0.6 are excluded, and nodes decisions. It actively calculates coordinate changes and
with TWR between 0.6 and 1 TWR are rewarded with 1. passively assesses distance for optimal communication. After
LAAODV estimates a vehicle's location and selects the best calculating expected positions, communication reliability is
route using data from the communication network, predicting evaluated before choosing a relay node. The Equation (8)

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
181
 2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

computes the communication stability index between nodes k III. RESEARCH DESIGN
and j. The study addresses challenges in communication systems,

BB_ C; ; /_;<43+ = ?E J?
234
including network instability, data congestion, and
+/ =
D F GHI
(8) communication delays. We created a simulation to replicate
real-world scenarios with accuracy. Fig. 2 shows the research
BB_ C; ; /_;<43+ = KE JT
FLMN,OP
Where:
D
design.
QR FLMN,OPS
(9)

Equation (8) computes the communication stability

index for nodes k and j. We divide the predicted position of
neighboring vehicles by the maximum communication radius
and check if the resulting value is less than or equal to 1. If
so, it indicates a stable communication environment. We
measure distances over time between the "k" node and nearby
Fig. 2. The process structure of research design
vehicles and assign weights based on speed and position
using a formula in (9).
The importance of simulation scenarios in research is
(9
((X ∗ (| R − Y | + (XF ∗ (| R − Y|
UV = W " %
illustrated in Fig. 2. We employ exhaustive simulation
+(XY ∗ (|4R − 4Y | + (XZ ∗ ( BB_[ ;/ models and designs that provide accurate real-world
scenarios to achieve this goal. We used Linux Ubuntu 18.04,

0.6 ≥ UV = 1, _ ;B , <4 UV ≤ 0.59, C ;B .

Where: a reliable operating system, to ensure reliability. The research
team uses NS3[26] for communication modeling and SUMO
[27] for traffic modeling, the primary tools for achieving
LA-AODV employs (9) to calculate the Total Weight accurate real-world scenarios.
Route (TWR), a measure used to evaluate the standard route.
TWR considers several factors such as speed, distance, A. Simulation Environment
acceleration, and communication quality, each assigned an Our research uses V2V communication, evaluated via
equal weighting factor of 1 as specified by (10). SUMO and NS3 v3.35. SUMO creates detailed models of

U = X + XF + XY + XZ = 1
traffic systems, helping us understand vehicle and pedestrian
FM (10) movement. NS3 allows us to model network communications,
simulate configurations, identify issues, and determine
solutions.
Equation (10) assigns speed, distance, acceleration, and
communication quality weights to determine the best route in B. Simulation Environment
the LA-AODV protocol. The FSA engine activates the Our research uses V2V communication, evaluated via
Learning Rate (α) after the final decision state. Source nodes SUMO and NS3 v3.35. SUMO creates detailed models of
share info with relay nodes. The LRI algorithm is used as a traffic systems, helping us understand vehicle and pedestrian
learning rate (α) for reward/punishment in each decision in movement. NS3 allows us to model network communications,
(11). simulate configurations, identify issues, and determine
b( , = 1, 23d 24
= a h
NMNc NY
solutions.
9 b( , +1, OeR fN = 0, <; ℎB3<
(11) C. Simulation Setup
The V2V communication simulation setup parameter
In Equation (11), the LRI algorithm adjusts the learning describe in Table 1.
rate (α) based on past experiences. The reward sets the
learning rate to 1, while the penalty reduces it to 0. The fine- TABLE I. V2V COMMUNICATION SIMULATION PARAMETER
SETUP
tune variable value of the algorithm's learning rate is
associated with its decision-making capability. Equation (12) Parameters Value(s)
explains the addition of the 'a' value to the latest TWR in the Total Numbers of actual Random, based on poisson distribution
Nodes (Vehicles)
predicted iteration (t+1). Simulation Time (s) 200,300,400,500,600 and 700 seconds

UVQ YF N = ∑ ,
,
( UV +
Traffic Scenario Smooth Flow (prob 0.7), Steady Flow
(12) (prob 0.4), and Heavy Traffic (prob 0.1)
*Based on Poisson Distribution
Equation (12) updates TWR values continuously, Route Selection Random Route Selection
Node Speed Random Speed
adapting to changing network conditions and routing
Initial node position Random position
decisions. Adaptability improves communication and routing Node Movement All moving nodes
performance V2V connections during the simulation period Data Packets Real-time traffic data packets from the
(M), with the α value being crucial. Configuration Buluksumur traffic maps.
Type of Protocol AODV and LAAODV
Type of Traffic Passengers cars only, Left-hand drive

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
182
 2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

Performance Matrix PDR, end to end delay, avg.Throughput,

(QOS) Packet Loss ratio, end to end jitter
LAAODV parameter fs: 0.4; fa : 0.3; fd : 0.3; α: 0.2; Reward : 6000 3172

Number of Vehicles (unit)

setup 1; 2366 2696
4000 1919
Table 1 presents the configuration for real-world vehicle 1417
2000 916 1017 1190
simulations using the LAAODV protocol to evaluate its 734
588
effectiveness in dynamic and unpredictable traffic conditions. 262 403 475 490 567
0
The analysis yielded insights into vehicle communication 300 400 500 600 700
efficiency, which depends on three main factors: speed (fs), Simulation Time
acceleration (fa), and distance (fd). To calculate the Smooth Traffic Steady Flow Traffic
Heavy Traffic
probability of certain vehicle numbers in specific time
Fig. 4. The Number of vehicles appearances during all simulation scenarios
simulations, Equation (13) presents the Poisson distribution
formula. Fig. 4 shows the vehicle count in two scenarios: smooth

:(i = =
N jk .ljm
and congested smooth, over time. In the smooth scenario,
O!
(13) traffic increases steadily from 262 to 567. In the congested
smooth scenario, the vehicle count rises from 588 to 1190,
The Poisson distribution (Equation 13) predicts the peaking at 3172 during heavy traffic.
number of vehicles passing through a point based on the
average event rate λ. It can estimate the likelihood of multiple
40
occurrences and involves Euler's number (approx. 2.71). The

Avg Speed (m/s)

Buluksumur area in Yogyakarta is a busy traffic route 9.04 8.92 8.8
30 9.92 8.17
connecting main roads, with four lanes and obstacles, 10.44
11.12 10.22 9.99
including pedestrians, parked vehicles, and motorized 20 9.5
vehicles. Fig. 3 illustrates the need for careful evaluation. 12.14 11.4 10.6
10
10.7 10.55
0
300 400 500 600 700
Simulation Time (s)
Smooth Traffic Steady Flow Traffic

Fig. 5 The average vehicle speed in the SUMO simulation space

Fig. 5 analysis shows the average vehicle speed in

SUMO simulation space at different traffic densities and time
intervals. The smooth scenario had lower speed over time
intervals, while the congested smooth scenario had relatively
higher speed. However, the congested scenario led to slower
traffic and longer travel times due to higher traffic density.
Fig. 1. Visualization of traffic conditions based on the real world in
Bulaksumur using SUMO.

IV. RESULTS AND DISCUSSION 80

Heavy
57
This section compares the AODV and LA-AODV 60 Traffic

protocols in ad hoc networks based on parameters such as 40 33

message overhead, convergence time, throughput, and route 22 Stead
stability. We also explore their practical implications in 20 y
12 7 Flow
different network scenarios. The analysis provides insights 0 3
1 5 2
0 0 0 Traffic
into their relevance in shaping future V2V communication. 300 400 500 600 700
We evaluate QoS parameters in V2V communication by Fig 6. Congestion appearances patterns in traffic simulations
comparing NS3's AODV module with the modified LA-
AODV in three traffic scenarios, each with four Fig. 6 shows how traffic congestion evolves at three
measurements. The results of the number of vehicle traffic density levels: Smooth Traffic, Steady Flow Traffic,
measurements are depicted in Fig. 4. and Heavy Traffic. Congestion is minimal in Smooth Traffic,
gradually increases in Steady Flow Traffic, and experiences
a more pronounced escalation in Heavy Traffic. The analysis
highlights varying degrees of congestion across different
traffic density levels.

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
183
 2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

According to Fig. 9, the increasing vehicle density

300 (s) results in lower average speed and communication efficiency
46.83
83.67 400(s) in steady-flow traffic scenarios, with more congestion spots
500(s) and a significant impact on communication. V2V
58.9 600(s) communication becomes challenging during heavy traffic. A
700(s)
study analyzed three communication protocols under steady
80.7
75.01 traffic conditions: AODV, LA-AODV, and the discrepancy
protocol. AODV and LA-AODV delivered packets with
higher QoS and PDR than the discrepancy protocol. The
Fig. 7 Time loss experienced at different time intervals across three distinct
traffic density levels discrepancy protocol had a negligible PDR and a minimal
PLR of 1% but demonstrated efficiency in handling steady
Fig. 7 analysis shows how traffic density affects time traffic with an Average Throughput of 8.67 Kbps and E2E
loss: Smooth Traffic has a gradual delay increase from 46.83s Delay of 0.28 ns. The QoS discrepancy in Heavy Traffic
to 83.67s, Steady Flow fluctuates from 48.14s to 92.44s, and Scenario is described in Fig. 10.
Heavy Traffic results in a significant delay, escalating from
45.44s to 107.13s. The result helps in traffic management. To
50.00
choose the best communication protocol, compare LA- 40.00
AODV and AODV strengths and weaknesses. Fig. 8 shows 30.00
the QoS Discrepancy in Smooth Traffic. 20.00
10.00
-
(10.00)
50
40
AODV LA-AODV Discrepancy
30
Fig. 10 The QoS Discrepancy in Congested Traffic Situation
20
10 7.45 Fig. 10 tests three communication protocols for V2V
2.67 4 0.7 -6.151.03
0
communication under heavy traffic: AODV, LA-AODV, and
discrepancy protocol. AODV has a high PDR of 97.60% and
-10 a QoS of 11.59 but suffers from congestion. LA-AODV
offers better QoS, PDR, and lower E2E jitter, while the
AODV LA-AODV Discrepancy
discrepancy protocol has almost perfect packet delivery and
Fig. 8 QoS Discrepancy Result in Free-flows Traffic Scenario from 200 to low E2E jitter. Heavy traffic reduces communication
700 simulation seconds efficiency, but LA-AODV and discrepancy protocol can
handle varying traffic dynamics in V2V communication.
Fig. 8 shows that increased congestion reduces average
speed, and addressing congested areas is necessary to V. CONCLUSION
optimize time efficiency. The LA-AODV routing protocol The analysis of V2V communication in different traffic
performs better than the Discrepancy protocol under scenarios showed that LA-AODV outperformed AODV in
smoother traffic conditions, providing valuable insights for smooth traffic with a higher packet delivery ratio (38.27%),
optimizing routing protocol performance. Correlation lower packet loss ratio (85.26%), and increased average
analysis explains the interplay between traffic factors and throughput (56.62%). In steady flow traffic, both protocols
routing protocol performance, which can be used to develop performed similarly. In heavy traffic, AODV showed a higher
optimized traffic management strategies. packet delivery ratio (97.60%), but LA-AODV outperformed
it in average throughput (17.24%). Different traffic scenarios
need different protocols to achieve optimal results. Next, to
optimize V2V communication protocols, research should
50.00 adapt them to various traffic scenarios and investigate the
40.00 impact of emerging technologies. This study can lead to
30.00 adaptable protocols for dynamic traffic conditions by
incorporating vehicular mobility models and network
20.00
8.67 parameters in urban environments
10.00 2.13 (0.01)1.00 0.28 1.08
- REFERENCES
(10.00) [1] A. E. Mezher, A. A. AbdulRazzaq, and R. K. Hassoun, “A
comparison of the performance of the ad hoc on-demand distance
vector protocol in the urban and highway environment,” Indones.
J. Electr. Eng. Comput. Sci., vol. 30, no. 3, pp. 1509–1515, 2023,
AODV LA-AODV Discrepancy doi: 10.11591/ijeecs.v30.i3.pp1509-1515.
[2] K. Ashok et al., “Review on Energy Efficient V2V
Communication Techniques for a Dynamic and Congested Traffic
Fig. 9 QoS Discrepancy Result in Steady-flow Traffic Scenario from 200 to
Environment,” 2022 Int. Conf. Comput. Commun. Informatics,
700 simulation seconds

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
184
 2024 International Conference on Smart Computing, IoT and Machine Learning (SIML)

ICCCI 2022, pp. 0–5, 2022, doi: doi: 10.1109/ICC.2017.7997401.

10.1109/ICCCI54379.2022.9740853. [16] M. H. Homaei, S. S. Band, A. Pescape, and A. Mosavi, “DDSLA-
[3] L. Hota, B. P. Nayak, A. Kumar, B. Sahoo, and G. G. M. N. Ali, RPL: Dynamic Decision System Based on Learning Automata in
“A Performance Analysis of VANETs Propagation Models and the RPL Protocol for Achieving QoS,” IEEE Access, vol. 9, pp.
Routing Protocols,” Sustain., vol. 14, no. 3, pp. 1–20, 2022, doi: 63131–63148, 2021, doi: 10.1109/ACCESS.2021.3075378.
10.3390/su14031379. [17] S. Khadim, F. Riaz, S. Jabbar, S. Khalid, and M. Aloqaily, “A non-
[4] H. A. Ameen, A. K. Mahamad, S. Saon, D. M. Nor, and K. Ghazi, cooperative rear-end collision avoidance scheme for non-
“A review on vehicle to vehicle communication system connected and heterogeneous environment,” Comput. Commun.,
applications,” Indones. J. Electr. Eng. Comput. Sci., vol. 18, no. 1, vol. 150, no. April 2019, pp. 828–840, 2020, doi:
pp. 188–198, 2019, doi: 10.11591/ijeecs.v18.i1.pp188-198. 10.1016/j.comcom.2019.11.002.
[5] T. K. Priyambodo, D. Wijayanto, and M. S. Gitakarma, [18] M. Hasanzadeh-Mofrad and A. Rezvanian, “Learning Automata
“Performance optimization of MANET networks through routing Clustering,” J. Comput. Sci., vol. 24, pp. 379–388, 2018, doi:
protocol analysis,” Computers, vol. 10, no. 1, pp. 1–13, 2021, doi: 10.1016/j.jocs.2017.09.008.
10.3390/computers10010002. [19] M. A. Gawas and S. S. Govekar, “A novel selective cross layer
[6] E. Mustikawati, D. Perdana, and R. M. Negara, “Network Security based routing scheme using ACO method for vehicular networks,”
Analysis in Vanet Against Black Hole and Jellyfish Attack with,” J. Netw. Comput. Appl., vol. 143, no. May, pp. 34–46, 2019, doi:
CommIT (Communication Inf. Technol. J., vol. 11, no. 2, pp. 77– 10.1016/j.jnca.2019.05.010.
83, 2017. [20] B. B. Maaroof et al., “Current Studies and Applications of
[7] F. Belamri, S. Boulfekhar, and D. Aissani, “A survey on QoS Shuffled Frog Leaping Algorithm: A Review,” Arch. Comput.
routing protocols in Vehicular Ad Hoc Network (VANET),” Methods Eng., vol. 29, no. 5, pp. 3459–3474, 2022, doi:
Telecommun. Syst., vol. 78, no. 1, pp. 117–153, 2021, doi: 10.1007/s11831-021-09707-2.
10.1007/s11235-021-00797-8. [21] E. Daknou, N. Tabbane, and M. Thaalbi, “A MAC multi-channel
[8] K. B. Y. Bintoro and T. K. Priyambodo, “Learning Automata- scheme based on learning-automata for clustered VANETs,” Proc.
Based AODV to Improve V2V Communication in A Dynamic - Int. Conf. Adv. Inf. Netw. Appl. AINA, vol. 2018-May, pp. 71–78,
Traffic Simulation,” Int. J. Intell. Eng. Syst., vol. 17, no. 1, pp. 2018, doi: 10.1109/AINA.2018.00023.
666–678, 2024, doi: 10.22266/ijies2024.0229.56. [22] A. M. Bamhdi, “Efficient dynamic-power AODV routing protocol
[9] A. A. Abdullhaj Saif and K. Kumar, “Enhance the performance of based on node density,” Comput. Stand. Interfaces, vol. 70, no.
AODV routing protocol in mobile ad-hoc networks,” J. Phys. November 2019, p. 103406, 2020, doi: 10.1016/j.csi.2019.103406.
Conf. Ser., vol. 2327, no. 1, 2022, doi: 10.1088/1742- [23] M. R. Hasan, Y. Zhao, Y. Luo, G. Wang, and R. M. Winter, “An
6596/2327/1/012057. Effective AODV-based Flooding Detection and Prevention for
[10] K. Afzal, R. Tariq, F. Aadil, Z. Iqbal, N. Ali, and M. Sajid, “An Smart Meter Network,” Procedia Comput. Sci., vol. 129, pp. 454–
Optimized and Efficient Routing Protocol Application for IoV,” 460, 2018, doi: 10.1016/j.procs.2018.03.024.
Math. Probl. Eng., vol. 2021, 2021, doi: 10.1155/2021/9977252. [24] R. Arief, R. Anggoro, and F. X. Arunanto, “Implementation of
[11] G. Arifiyan, S. Djanali, and R. Anggoro, “AODV-PNT Aodv Routing Protocol With Vehicle Movement Prediction in
Performance Study with Added Factor Number of Neighbor Nodes Vanet,” Surabaya Inst. Teknol. Sepuluh Novemb., 2016.
on VANET,” IPTEK J. Proc. Ser., vol. 4, no. 1, p. 69, 2018, doi: [25] S. Haider, G. Abbas, Z. H. Abbas, and T. Baker, “DABFS: A
10.12962/j23546026.y2018i1.3510. robust routing protocol for warning messages dissemination in
[12] K. A. Darabkh, M. S. A. Judeh, H. Bany Salameh, and S. VANETs,” Comput. Commun., vol. 147, no. May, pp. 21–34,
Althunibat, “Mobility aware and dual phase AODV protocol with 2019, doi: 10.1016/j.comcom.2019.08.011.
adaptive hello messages over vehicular ad hoc networks,” AEU - [26] W. Li, X. Ma, J. Wu, K. S. Trivedi, X. L. Huang, and Q. Liu,
Int. J. Electron. Commun., vol. 94, no. July, pp. 277–292, 2018, “Analytical Model and Performance Evaluation of Long-Term
doi: 10.1016/j.aeue.2018.07.020. Evolution for Vehicle Safety Services,” IEEE Trans. Veh.
[13] D. Zhang, C. Gong, T. Zhang, J. Zhang, and M. Piao, “A new Technol., vol. 66, no. 3, pp. 1926–1939, 2017, doi:
algorithm of clustering AODV based on edge computing strategy 10.1109/TVT.2016.2580571.
in IOV,” Wirel. Networks, vol. 27, no. 4, pp. 2891–2908, 2021, [27] M. Malinverno, F. Raviglione, C. Casetti, C. F. Chiasserini, J.
doi: 10.1007/s11276-021-02624-z. Mangues-Bafalluy, and M. Requena-Esteso, “A Multi-stack
[14] G. A. Beletsioti and G. S. Member, “A Learning-Automata-Based Simulation Framework for Vehicular Applications Testing,”
Congestion-Aware Scheme for Energy-Efficient Elastic Optical DIVANet 2020 - Proc. 10th ACM Symp. Des. Anal. Intell. Veh.
Networks,” IEEE Access, vol. 8, 2020. Networks Appl., pp. 17–24, 2020, doi: 10.1145/3416014.34
[15] V. Saritha, P. V. Krishna, S. Misra, and M. S. Obaidat, “Learning
automata based optimized multipath routingusing leapfrog
algorithm for VANETs,” IEEE Int. Conf. Commun., pp. 1–5, 2017,

Authorized licensed use limited to: San Francisco State Univ. Downloaded on September 17,2024 at 00:45:33 UTC from IEEE Xplore. Restrictions apply.
185