Software defined network-based controller system in intelligent transportation system

IAES International Journal of Artificial Intelligence (IJ-AI)
Vol. 13, No. 3, September 2024, pp. 2645~2655
ISSN: 2252-8938, DOI: 10.11591/ijai.v13.i3.pp2645-2655  2645
Journal homepage: http://ijai.iaescore.com
Software defined network-based controller system in intelligent
transportation system
Arjunan Mari Kirthima1
, Pushpa Sothenahalli Krishnaraju2
1
Department of Computer Science and Engineering, BMS Institute of Technology and Management, Bengaluru, India
2
Department of Information Science and Engineering, BMS Institute of Technology and Management, Bengaluru, India
Article Info ABSTRACT
Article history:
Received Dec 2, 2023
Revised Feb 6, 2024
Accepted Mar 5, 2024
Intelligent transportation system (ITS) is meant for redefining the
conventional transport by incorporating various analytical features that not
only offers safety but also enriches traffic data quality extensively. Review
of existing literature shows that there is a significant gap towards utilizing
vehicular adhoc network (VANET) for optimal performance in ITS
environment. Therefore, this paper contributes towards a simplified and yet
intelligent controller system harnessing potential of software defined
network (SDN) towards effective directional management of complex
transportation system. The novelty of this model is dual-fold. The first
novelty is about the usage of locally and globally processed traffic
information for undertaking decision towards clearing the waiting vehicles
in observation point in specific route segment. The second novelty is
associated with relaying of distinct traffic clearance signal to the distinct
vehicles unlike any of teh existing transportation management scheme.
Keywords:
Decision making
Intelligent transportation system
Route
Software defined network
Vehicular adhoc network
This is an open access article under the CC BY-SA license.
Corresponding Author:
Arjunan Mari Kirthima
Department of Computer Science and Engineering, BMS Institute of Technology and Management
Doddaballapur Main Road, Avalahalli, Yelahanka, Bengaluru 560064, India
Email: krithi.a@bmsit.in
1. INTRODUCTION
With the changing dynamics of transportation system from conventional to intelligent transportation
system (ITS), vehicular adhoc network (VANET) has been still proving to be highly essential technology that
facilitates various beneficial perspective towards transportation [1]. Some of the prime reason for choosing
VANET in ITS system is because of its enhanced road safety [2], traffic management [3], supportability
towards emergency services [4], improved navigation system [5], and analytically enhanced information
services. Although, it sounds quite good theoretically, but it is quite challenging to implement it on practical
ground. Some of the identified challenges in deploying VANET system in ITS are associated with lack of
reliability and connectivity issues, privacy concerns, inferior scalability, poor network management,
fluctuating quality-of-service, lack of standardization and interoperability [6]–[9]. Apart from this, the
dynamic state of vehicular network with vehicular nodes consistently leaving and entering to a network
further makes it more challenging to establish a stabilized link [10]. It was also noticed that the power
constraints associated with the on-board units in vehicle can also influence the energy efficiency of a
VANET. This is essential in order to maintain an overall resources in order to support the networking
operation [11]. However, it is also required to highlights the challenges associated with ITS which makes the
integration of VANET in ITS slightly non-incompatible as per present state of solutions [12]. An ITS
environment encounters a significant challenge in providing the seamless communication as well as
interoperability among different types of components. Further, controlling and integrating a massive quantity

 ISSN: 2252-8938
Int J Artif Intell, Vol. 13, No. 3, September 2024: 2645-2655
2646
of traffic data originating from vehicles and different infrastructure units of ITS poses a significant challenge.
Although, there are studies related to data storage, sharing, and processing in ITS, still they are yet in its
beginning stages [13]. Even if a proper ITS environment has been formed than another hurdle will be to carry
out an infrastructure upgrading process. It is because not all the regions will have an updated infrastructure of
transportation. Further, integrating ITS into urban planning will potentially involve coordination between
technology providers, city planner, and transportation agencies while it is computationally complex process
to align all these heterogeneous entities to meet a different level of goal in ITS. Finally, there is a need to
offer robust planning and coordination for an ITS system in order to efficiently handle the emergency
situations. Hence, both ITS as well as VANET system has their own futuristic scope as well as inherent
loopholes, which are required to be addressed.
All these significant issues and gap between VANET and ITS can be solved using software defined
network (SDN) owing to its potential capabilities towards an effective traffic management [14]. An
administrator can programmatically control as well as adjust the network resources in ITS using specific event,
incidents and traffic patterns thereby facilitation more dynamic and real-time traffic management in ITS. This
property of SDN is highly contributory towards enhancing overall efficiency of transportation, minimizing
congestion, and optimizing traffic flow. Further, SDN can offer higher degree of flexibility, scalability, and
interoperation towards deploying VANET system in ITS with reduced operational cost. SDN allows for
dynamic and real-time adjustments to network traffic routing. In ITS, this can be applied to dynamically reroute
traffic based on current conditions, incidents, or emergencies to optimize traffic flow and minimize congestion.
SDN facilitates the dynamic allocation of resources based on demand. In ITS, this is valuable for handling
varying traffic loads, ensuring that the network can scale up or down as needed. Therefore, the prime aim of the
proposed study is to develop a novel SDN-based controller system which can deploy VANET in ITS with more
effectiveness towards controlling and managing dynamic traffic condition. Further, value-added contribution of
proposed scheme are presented with respect to research objectives as follows:
− To develop a model for performing processing using both local and global information for SDN controller
for better decision making,
− To design an ITS environment with VANET considering on-board units, roadside units, gateway node,
and cloud in order to construct a highly collaborative SDN-based framework,
− To construct a novel progressive and non-iterative scheme towards offering a navigational services by the
SDN-controller is designed to generate a distinct route to be undertaken considering the distinct
requirement of a vehicle to reach their destination.
This section of introduction will further illustrate about the background of study, identified research
problem, and proposed solution. This section of introduction discusses about all the related study
methodologies deployed toward improving ITS system with respect to communication perspective as a
continuation of our prior study [15]. Ali et al. [16] have developed traffic model considering conventional
adhoc routing protocol for unmanned vehicle. Rai et al. [17] have constructed a signalling-based framework
where a programmable micro-controller and inductive loops were used for assess the traffic environment.
Chowdhury et al. [18] have emphasized about addressing issues pertaining to emergency vehicles guidance
system considering priority-based management with respect to congestion. Study towards congestion control
in VANET is carried out by Nazar et al. [19] where a pre-caching methodology is adopted using machine
lerarning approach. Ahmed et al. [20] have used reinforcement learning scheme in order to structurize the
traffic system that supports exchange of heavier multimedia data in VANET. Adoption of reinforcement
learning is also witnessed in work of Guo et al. [21] that is meant for performing a cooperative control over
the connected vehicles.
Beenish et al. [22] have presented a Markov modelling scheme towards computing the traffic density
in ITS. Kannadhasan et al. [23] have presented a systematic framework that is claimed to optimize the data
transmission in VANET in order to meet the quality-of-service demands. Makridis et al. [24] have presented a
control strategy of traffic that can handle varies conflicts among vehicle trajectories. Adoption of k-nearest
neighboring algorithm along with dynamic time warping is witnessed to be considered in model presented by
Zibner et al. [25] emphasizing on encapsuting traffic data for classifying various critical events in ITS.
Adoption of deep learning is witnessed in work of Gao et al. [26] that is meant for performing predictive
analysis of ITS environment. A unique study model is presented by Ma and Zou [27] where traffic noise is
used for annalyzing the state of traffic where particle swarm optimization and support vector machine is
further used for performing classification of traffic states. Ma et al. [28] have constructed a traffic guidance
system using game theory with a target to address the issues of clogged traffic. Rafter et al. [29] have
presented an adaptive controller system which is meant to reduce the delay of vehicles in urban transporation.
Adoption of federated learning towards performing predictive analysis is carried out by Sepasgozar and Pierre
[30] where long short-term memory (LSTM) can be used to make it decentralized operation. Apart from this,
existing system has also witnessed utilizing SDN towards ITS. Huang et al. [31] have used reinforcement

Int J Artif Intell ISSN: 2252-8938 
Software defined network-based controller system in intelligent transportation … (Arjunan Mari Kirthima)
2647
learning approach for controllong routing operation. Kurungadan and Abdrabou [32] have also presented an
SDN-based framework for effective data control in smart city. Masood et al. [33] have developed a
prioritization scheme in traffic controlling system over SDN-based large scale environment by integrating
machine learning with reinforcement learning. Hence, it can be noted that there are various schemes towards
improving the performance of ITS considering various case studies with proven beneficial outcomes.
However, the above-mentioned studies are not found to combinely address the problems of ITS and VANET
using SDN based approach much. The next section of methodoutlines the identified research issues extracted
from existing studies.The potential set of identified issues in existing research work are:
− Adoption of learning-based approaches are frequently witnessed with better predictive accuracy but at the
cost of higher computational burden associated with it.
− There are lack of frameworks or problem solution where VANET is deployed in typical ITS environment
to offer a sustainable communication.
− Majority of the existing approaches towards traffic management is carriedout considering the local data
perspective which is not practical approach for large scale global scenario in ITS.
− There is a lack of inclusion of the dynamic and stochastic state of vehicles in any exising models using
SDN based traffic management scheme.
Hence, the statement of the research problem is, “it is computationally challenging task to relay a reliable and
appropriately calculated information for the vehicular nodes in VANET seeking assistance in large and
dynamic environment of ITS.”
2. METHOD
The overall research objective is to develop a novel computational model which can harness the
potential of SDN and intelligence learning stategy for assisting in effective traffic management in VANET.
The core problem addressed by proposed system is to develop highly distributed information propagation
mechanism in large scale environment of vehicular network. The development of this part of implementation
will be carried out in analytical research methodology. An architecture will be designed exclusively meant
for VANET system considering route segments with multiple observation points along with traffic signal in
the junction point to manage the traffic. The proposed study considers that complete algorithm is meant to
run in two different positions viz. i) on board unit of a vehicle and ii) smart hotspots present in junction point
which runs SDN equipped for information propagation and routing in VANET. The Figure 1 highlights the
plan of implementation:
Figure 1. Methodology of proposed solution
The discussion of core operational blocks involved in proposed architecture are as follows:
− On-board units: these are the embedded units deployed within the vehicles and is meant as a
communication bridge between the driver/vehicle and external communication environment in vehicular
network. All the on-board units (or vehicular nodes) forward its characteristics information of vehicle to
road side units installed at the terminals of each lane system approach junction.
− Local/global processing: the collected information is subjected to local processing (about individual
vehicles) and to global processing (for all the vehicles present in lane).

 ISSN: 2252-8938
2648
− Road-side units: these are the infrastructural units in vehicular network which acts as a bridge of
communication between vehicle to another vehicle or other extended infrastructural unit (gateway node).
− SDN controller: this is the core part of the implementation where this controller is assumed to reside in
the centre of junction point in urband transportation system. All this aggregated information by road side
unit is forwarded to the SDN controller, where the aggregated information is maintained in a matrix.
− Matrix for repositioning traffic map: this is basically a buffer system present within the core VANET
system that is meant to retain updated topological information.
− Indexing for emergency vehicles: this is basically module that maintains an indexes of the vehicles in order to
differentiate normal and emergency vehicles. This module allocates new indexes for every incoming and
existing vehicles are respectively updates the associated shared information among all the infrastructural units.
− Algorithm for intelligent traffic management: this is the prime contribution of proposed system where a
novel mechanism is constructed to clear the awaiting vehicles and thereby maintain a better traffic flow.
The elements of this matrix for repositioning traffic map further act as an input for proposed algorithm for
intelligent traffic management for obtaining best path to be given as an outcome for each vehicle for
better communication and navigational system.
− Selection of auxilary node: in order to offer better coverage, the proposed system also identify possible
form of an auxiliary nodes, which are normal vehicles, meant for acting as a temporary mobile node to
transport adhoc information from one to another vehicular node located in different sensing area but
within common sensing area of auxiliary node.
This system assists in better vehicle-to-vehicle communication system. Further, the SDN controller
also considers inclusion of indexed emergency vehicles in order to formulate route and offer different and
discrete information to every traffic lanes. This is to ensure both emergency vehicles and regular vehicles
communication demands are met.
3. ALGORITHM IMPLEMENTATION
This section of algorithm implementationdiscusses about the algorithmic operation carried out
towards deploying an SDN-based framework in ITS. A novel SDN-based controller system is designed
which is introduced in traffic environment for controlling the flow of transportation system in highly
distributed manner. The core idea of the proposed model is to ensure a proper identification of the actual state
of traffic in urban transportation system followed by minimizing the queue for waiting vehicles at the
junction point. The prime logic for the algorithm development is also to assign a priority-tags for the vehicles
with longer waiting time on specific point of routes. Some of the essential attributes considered for this
purpose are position of vehicle, its velocity, and its direction. Further node density for each specific route
segment connecting the junctions is computed on the basis of number of vehicles traversing on it. Each route
segment is initialized with a threshold value of maximum number of vehicles to be accommodated for
traversing on that specific route segment. The condition of saturated and unsaturated traffic on each route
segment is carried out by comparing the current number of vehicles with this threshold value. Such threshold
value can be altered when implemented over real-time application and hence the modelling is carried out in
such a way that it can be mapped with any practical environment. The initiation of the algorithm design starts
from selecting a primary route out of all the possible routes by the vehicle traversing on a specific route
segment, as shown in Algorithm 1.
Algorithm 1: Selection of primary route
Input: α
Output: β
Start
1. Forα= 1:m
2. If (j=nstate)
3. β=γθ
4. End
5. End
End
The discussion of the lines of the algorithm is as follows: the algorithm takes the input of α (route
segments) that after processing yields an outcome of β (primary route). This computation is carried out
within SDN controller system which is assumed to be located at the center of junction point connecting
multiple routes in ITS. The algorithm considers m number of joining routes at the junction whose values are

2649
typically ranging from 2-5 conventionally in urban area (line-1). The SDN controller system evaluates the
state of vehicle nodes (line-2) by considering all the information associated with node density compared with
the threshold T to yield four types of node states i.e., very high (nden>>T), high (nden>T), low (nden<<T), and
medium (nden=T) score, where nden is node density for specific routes. The next line of operation consists of
computing number of possible routes θ from the junction point followed by computing the primary route
considering γ i.e., number of constraints to yield β primary route (line-3). This can be illustrated in Figure 2
where it can be seen that a matrix is formulated considering γ i.e., number of constraints to maximum level
max. Further, each node states will carry an index value assigned by the controller within four different
specific ranges with respect to rows and columns i.e., v1,1, v2,1, …, vmax,1.
Figure 2. Internal constraint management in SDN controller
It should be noted that Algorithm 1 assists in choosing an appropriate condition of the traffic in
particular route segment on the basis of node density in it. The output generated i.e., β (primary route) is
further considered as an input for the next part of algorithmic operation towards choosing the optimal route
ρ. Hence, Algorithm 2 is responsible to generate a distinct traffic decision by the SDN-controller system
considering the local information of current route segment. The steps of the Algorithm 2 are as follows:
Algorithm 2: Route selection by SDN controller
Input: η, β
Output: ρ
Start:
1. For i=1: (η, β)
2. Jnc→φ1(μ, σ)
3. If Jnc<T
4. drs→ φ2(Jnc, σ)
5. For l=1:argmax(λ)
6. If drs≤ λ
7. ρ=U
8. End
9. End
10. End
End
The discussion of Algorithm 2 is as follow: the algorithm takes the input arguments of η (population
of nodes) and β (primary route) that after processing yields an outcome of ρ (chosen route). The initiation of
the algorithm is carried out by considering the input arguments (line-1) where the evaluation of the traffic
state is carried out in the junction point Jnc (line-2). For this purpose, a function φ1 is developed which
considers its input argument of μ current route and σ controller decision (line-2). It is to be noted that prime
logic of constructing a function φ1 is to develop a novel structure of junction point in urban environment with
more monitoring control over the information acquired from μ current route and σ controller decision. The
matrix Jnc stores all the global information from other junction points Jnc1, Jnc2, … and store them on cloud
storage unit that can be accessed at a point of time via the gateway node that are synced with the SDN
controller nodes. The elements within the matric Jnc are then compared with the threshold T (line-3) while
the condition Jnc<T will represent an ideal condition ranging between low and medium traffic scenario. This
condition also calls for the SDN controller system to clear the vehicle by relaying a clearance decision-based
signal. Further, another function φ2 is constructed which takes the input of junction Jnc and σ controller

 ISSN: 2252-8938
2650
decision in order to generate decision ruleset of SDN controller drs (line-4). It should be noted that
Algorithm 2 is executed in line with continuous updating process by Algorithm 1 where the primary routes
are often updated prior to generate a final decision of route selection by Algorithm 2. The study also
considers that the SDN controller is capable of generating λ which are number of traffic decision. It will
mean that based on local and global information, the SDN controller system will generate a number of
decision pertaining to clearance, waiting, halting with respect to routes to be undertaken by the vehicles. As
this information of decision by SDN controller is specific to different vehicles waiting to be cleared from the
junction, hence, the numbers of such decision-based information vary depending upon current number of
vehicles in the current route. Hence, the algorithm considers multiple conditions ranging from 1 to maximum
value of traffic decision argmax(λ) (line-5) followed by further assessing the decision ruleset of SDN
controller drs with number of traffic decision λ (line-6). If the number of decision ruleset of SDN controller
system is found to be less than number of traffic decision λ, it will mean that proposed scheme has
successfully generated valid number of decisions by SDN controller (line-6). Finally, the controller selects all
individual (or decision for local route information) l and stores them in global decision for selected routes
i.e., ρ (line-7) in form of a matrix. The Relationship between Algorithms 1 and 2 is illustrated in Figure 3.
Algorithm-1 Algorithm-2
β (primary route)
η (population of nodes)
ρ (Chosen Route)
α (route segments)
Figure 3. Relationship between Algorithms 1 and 2
The final algorithm is responsible for further optimizing the decision made by the SDN controller
system in order to ensure a proper distribution of decision making towards effective traffic management
system in urban environment. This was a necessary step as proposed scheme relays a distinct traffic decision
uniquely for every vehicle considering the collaborative network and auxiliary nodes. The prime role played
by the auxiliary nodes is towards assisting in message propagation between the requesting vehicular node
towards the destination without using much infrastructural units. The steps of the Algorithm 3 are as follows:
Algorithm 3: Optimized relaying of traffic decision
Input: cs, Jnc, ρ
Output: σopt
Start
1. Fori=1:(Jnc, ρ)
2. Ifcinit≥cs
3. param(i)→χ(Jnc)
4. Ifρ(iter) ≠ ρ(Jnc-1)
5. Ifiter≠f
6. σ(ρ(iter-1), ρ(Jnc))=(τ1, τ2)
7. h=ψ
8. σopt(ρ(i-1), ρ(i))= (τ1, τ2)
9 End
10. End
11. End
12. End
End
The discussion of the Algorithm 3 is as follows: the algorithm takes an input of cs (clearance signal),
Jnc (Junction point), and ρ (chosen route) that after processing yields an outcome of σopt (optimized traffic
decision). The algorithm first initialized Junction Jnc and chosen route ρ (line-1) obtained from Algorithm 2
while analysis is carried out by comparing initialized clearance decision cinit with clearance signal cs (cs∈cds)
(line-2). If the value of cinit is found to be more than cs than it will mean that there are a greater number of
vehicles waiting to be cleared. In such case, the algorithm constructs a parameter param which is an array
consisting of local controller decision σ, initialized clearance decision cnit, and chosen route ρ. A matrix χ is
constructed further to access all the global information of junction point Jnc and they are stored back in an
array of param (line-3). A logical condition is evaluated to assess if the score of the chosen route ρ is not at

2651
par with the iteration iter is found similar to chosen route ρ corresponding to prior junction (Jnc-1) (line-4).
Further, it also assesses if the number of iterations iter is not similar to lowest possible number of adjacent
routes from junction point f. For simplicity the value of f can be considered as 2 (line-5). If the condition
stated in line-4 and line-5 is found to be true than it will signify multiple number of vehicles in different
(global) junctions are also waiting to be cleared. It will eventually mean that the algorithm will require to
undertake a decision towards relaying a decision which justifies towards clearning the vehicle not only in
present junction but also in all the connected junctions too. Further, the algorithm set up two variables
τ1 and τ2 which corresponds to two states of chosen routes viz. the first state ρ(iter-1) is towards local
processing while the second state ρ(Jnc) is for global processing with respect to controller decision σ. The
value of τ1 can initialized to medium statistical value of 0.5 while the value of τ2 can be initialized to lowest
score of 0. The system then computes current queue time ψ of vehicles in specific lanes and assign it to
global matrix of h (line-7). The information is further updated with respect to chosen routes ρ in order to
undertake a final decision of optimized controller-based clearance signal σopt (line-8). The relationship among
Algorithms 1 to 3 is shown in Figure 4.
Algorithm-1 Algorithm-2
β (primary route)
η (population of nodes)
ρ (Chosen Route)
α (route segments)
Algorithm-3
cs (clearance signal)
Jnc (Junction point)
σopt (Optimized traffic decision)
Local processing Global processing
Local processing
Global processing
Figure 4. Relationship among Algorithms 1 to 3
However, there are also known challenges associated with implying SDN controller system in
vehicular network. Vehicles in a vehicular network move at high speeds, leading to rapid changes in network
topology. SDN controllers need to handle fast handovers seamlessly, ensuring continuous connectivity and
efficient routing. Vehicular networks can have a large number of vehicles, and scaling SDN controllers to
handle the increasing volume of devices and data traffic can be challenging. However, proposed scheme
addresses this challenge by effectively introducing a local and global traffic data processing as well as novel
SDN-based controller system that is capable of relaying discrete clearance signal to vehicles differing from
any conventional scheme.
4. RESULTS AND DISCUSSION
Prior to illustration of the result discussion, it is essential to understand the operation of the
proposed study model in practical environment. A sample use-case is shown in Figure 5 where a global
scenario of a vehicular network is shown to consists of two regions that are considered as local attributes.
SDN controller resides in center of junction and has its own coverage where each SDN controllers are synced
among each other via gateway node. Each local regions consist of one gateway node that consists of
extensive traffic information that are accesisble by cloud-based environment. Each gateway nodes are further
linked to cloud network using streaming server where the traffic information can be disseminated as well as
can be stored within the cloud storage units in distributed manner. The prime benefit of this representative
use-case is that SDN controller has an accessibility of both local and global traffic information which
faciliates quite a reliable as well as practical traffic command for clearing the queues in traffic.
The scripting of the proposed study model is carried out in MATLAB considering 500×500 m of
network size with 200 m of route segment. Each vehicle is considered to posses 200 m of transmission range
while the vehicle moves with a velocity of 60-80 km/h. Further, the transmission range of roadside unit is
also considered as 200 m. The study implementation is assessed with respective to consistency and duration
associated with the relaying of the clearance signal by the SDN controller system. Figure 6 showcases all the
accomplished outcome of the simulation study. All the above-mentioned evaluation are carried out
considering 150 number of iteration where each iteration represents increment in traffic size randomly.
Figure 6(a) showcases that performance outcome of consistency in relaying clearance signal for all the four
observations point. The observation point is considered as location of junction where the SDN controller is
installed. The performance parameter consistency in relaying clearance signal represents total number of the
primary routes generated as a consequence of Algorithm 1 for all the observation points. A closer look into

 ISSN: 2252-8938
2652
Figure 6(a) showcases that there is a good amount of consistency for almost all the respective observation
points in every round of iteration.
This analysis is carried out by dynamic presence of emergency vehicles by suddenly relaying a
request to SDN controller system for optimal route towards destination while the controller had no
predefined knowledge of that node as an emergency node. This represents that proposed scheme can offer a
better form of equilibrium between normal and non-predefined emergency vehicle. The outcome of
Figure 6(b) represents duration of route selection by controller which is basically a consequence of
Algorithm 2. Although, the duration is found to be slightly increased for higher iteration rounds, but majority
of the iteration round is found to maintain a nearly-similar duration values with least significance differences
with increasing iteration. The prime reason behind this is an effective syncing among all the roadside units
with the SDN controller system, while SDN controller system will only have local information access, their
coverage towards accessing global information is boosted by their synced connectivity with roadside units
that are linked with local gateway, while all the gateways are lined to cloud. Hence, decision-making doesn’t
undergo much time towards route generation by an SDN controller.
Figure 6(c) showcases the duration required for relaying the optimal signal considering all the global
information by the SDN controller. A closer look into the trend of the bars shows that there is an
approximately reduction of 60% of time period for relaying the optimal signal in contrast to Figure 6(b) for
all the observation points. The prime justification behind this significantly reduced duration is because of the
broader accessibility of navigational routes accessed by the SDN controller using Algorithm 3.
Figures 6(d) and 6(e) exhibits the practical proof-of-concept for the effectiveness of proposed
scheme towards clearing the vehicles thereby reducing the waiting time in the observation point. Further,
reduction in waiting time can also be proven by the outcome exhibited in Figure 6(c). Figure 6(d) exhibts that
different observation points to receive different values of relayed clearance signal by the SDN controller
while Figure 6(e) exhibits that overall clearance of vehicles from all the observation point from global
perspective. This analysis is carried out by calculating the total number of vehicles being cleared from all the
individual respective observation points in complete simulation area. The outcomes showcases that in least
duration, each observation points are capable to clear minimal 400-800 number of vehicles considering the
complete simulation area. A closer look into the analytical outcomes also shows that the performance of
existing literatures [16]-[33] are must restricted to one segment considering local variables of respective parto
of the road, whereas proposed scheme is implemented considering global attributes of traffic environment.
This consideration in implementation plan of proposed system makes it more applicable towards practical
traffic management in vehicular network.
Table 1 highlights the comparison of the proposed system with existing literatures with respect to
adopted methodology, network type supportability, and possible limitation. It is explored that proposed
system is capable of addressing concurrent request deploying its SDN controller in order to make definite
decision of clearning the vehicles from junction points. Apart from this, it also supports highly distributed
and decentralized network which is not reported in existing reviewed methodologies towards traffic
management. This is the prominent novelty and significance of proposed study.
SDN Controller Gateway node Streaming Server
Distributed
Cloud
Storage
Units
-
-
Region-1
(local)
Region-2
(local)
Global scenario of vehicular network
Figure 5. Case study of proposed model

2653
(a) (b)
(c) (d)
(e)
Figure 6. Evaluation of graphical outcome of proposed model: (a) consistency in relaying clearance signal,
(b) duration for route selection by controller, (c) duration of relaying optimized signal, (d) number of relayed
clearances, and (e) number of cleared vehicles
Table 1. Comparison with existing literature
Approaches Methodology Network type Limitation
Ali et al. [16] Conventional routing Centralized Limited coverage
Rai et al. [17] Signaling-based Centralized Small network
Chowdhury et al. [18] Vehicle prioritization Peer-to-peer Small area coverage
[19]-[22], [26], [30],
[31]
Machine learning, Markov
modelling, deep learning
Centralized Cannot respond emergency event
[23]-[25], [29] Analytical, data driven Centralized Cannot meet dynamic demands
Ma and Zou [27] Particle swarm optimization Centralized Extensive computational demand
Ma et al. [28] Game theory Centralized Cannot model complex logic
[32], [33] SDN Centralized Cannot cater up dynamic demands
Proposed Analytical Distributed, decentralized,
large network
Demands more calibration towards
optimization

 ISSN: 2252-8938
2654
5. CONCLUSION
The proposed study has presented a novel SDN-based framework which can acquire contextual
information both from local and global environment in order to undertake a decision of service relaying. The
work has been carried out considering VANET deployment to support functionally working in large ITS
environment. The summary of contribution and novelty associated with the proposed study are as follows:
i) a faster and contextual decision-making support system has been presented which can meet the routing
demands in complex traffic environment, ii) the proposed scheme introduces a collaborative method which
entitles the SDN controller system to access both local and global information in order to undertake decision
making for clearing the awaiting vehicles in observation point, iii) different from conventional traffic
signalling process, the proposed scheme offers a novel feature of relaying a discrete set of decision for all the
vehicles uniquely such that it can address the dynamic and stochastic properties of the vehicular movement in
ITS, and iv) the proposed scheme introduces a complete decentralized strategy towards traffic management
with lesser dependencies on extensive number of traffic data to undertake decision. The future work of the
proposed system will be continued in the direction of further optimizing the traffic management module by
involving various challenging environment. The idea of next-level of work will also be to investigate the
impact of various network-related artifacts on the performance attributes. This agenda of work will be
towards a better calibrated traffic management model in urban transportation system.
REFERENCES
[1] E. Bowlin, M. S. Khan, B. Bajracharya, B. Appasani, and N. Bizon, “Challenges and solutions for vehicular ad-hoc networks
based on lightweight blockchains,” Vehicles, vol. 5, no. 3, pp. 994–1012, 2023, doi: 10.3390/vehicles5030054.
[2] A. G. Ismaeel et al., “Enhancing traffic intelligence in smart cities using sustainable deep radial function,” Sustainability, vol. 15,
no. 19, p. 14441, 2023, doi: 10.3390/su151914441.
[3] M. A. Rahman, A. A. Hossain, B. Debnath, Z. M. Zefat, M. S. Morshed, and Z. H. Adnan, “Intelligent vehicle scheduling and
routing for a chain of retail stores: a case study of Dhaka, Bangladesh,” Logistics, vol. 5, no. 3, 2021, doi:
10.3390/logistics5030063.
[4] C. E. Cortés and B. Stefoni, “Trajectory simulation of emergency vehicles and interactions with surrounding traffic,” Journal of
Advanced Transportation, vol. 2023, pp. 1–23, 2023, doi: 10.1155/2023/5995950.
[5] C. Ren et al., “Boundary guidance strategy and method for urban traffic congestion region management in internet of vehicles
environment,” Journal of Advanced Transportation, vol. 2023, pp. 1–18, 2023, doi: 10.1155/2023/4866326.
[6] K. Ahmad, H. Khujamatov, A. Lazarev, N. Usmanova, M. Alduailij, and M. Alduailij, “Internet of things-aided intelligent
transport systems in smart cities: challenges, opportunities, and future,” Wireless Communications and Mobile Computing,
vol. 2023, pp. 1–28, 2023, doi: 10.1155/2023/7989079.
[7] N. E. Mohamed and I. I. Radwan, “Traffic light control design approaches: a systematic literature review,” International Journal
of Electrical and Computer Engineering, vol. 12, no. 5, pp. 5355–5363, 2022, doi: 10.11591/ijece.v12i5.pp5355-5363.
[8] J. C. Marasigan, G. P. Mayuga, and E. Magsino, “Development and testing of braking and acceleration features for vehicle
advanced driver assistance system,” International Journal of Electrical and Computer Engineering, vol. 12, no. 2, pp. 2047–2057,
2022, doi: 10.11591/ijece.v12i2.pp2047-2057.
[9] A. A. Musa, S. I. Malami, F. Alanazi, W. Ounaies, M. Alshammari, and S. I. Haruna, “Sustainable traffic management for smart
cities using internet-of-things-oriented intelligent transportation systems (ITS): challenges and recommendations,” Sustainability,
vol. 15, no. 13, p. 9859, 2023, doi: 10.3390/su15139859.
[10] E. I. A. Kareem and H. K. Hoomod, “Integrated tripartite modules for intelligent traffic light system,” International Journal of
Electrical and Computer Engineering, vol. 12, no. 3, pp. 2971–2985, 2022, doi: 10.11591/ijece.v12i3.pp2971-2985.
[11] D. Oladimeji, K. Gupta, N. A. Kose, K. Gundogan, L. Ge, and F. Liang, “Smart transportation: an overview of technologies and
applications,” Sensors, vol. 23, no. 8, Apr. 2023, doi: 10.3390/s23083880.
[12] I. Wahid et al., “Vehicular ad hoc networks routing strategies for intelligent transportation system,” Electronics, vol. 11, no. 15,
2022, doi: 10.3390/electronics11152298.
[13] I. Laña, J. J. S. -Medina, E. I. Vlahogianni, and J. D. Ser, “From data to actions in intelligent transportation systems: a
prescription of functional requirements for model actionability,” Sensors, vol. 21, no. 4, pp. 1–34, Feb. 2021, doi:
10.3390/s21041121.
[14] L. Nkenyereye, L. Nkenyereye, S. M. R. Islam, Y. H. Choi, M. Bilal, and J. W. Jang, “Software-defined network-based vehicular
networks: a position paper on their modeling and implementation,” Sensors, vol. 19, no. 17, Aug. 2019, doi: 10.3390/s19173788.
[15] A. M. Kirthima and S. K. Pushpa, “Evolving technological advancements and its effectivity study in intelligent traffic system,”
Lecture Notes in Networks and Systems, vol. 232 LNNS. Springer International Publishing, pp. 447–457, 2021, doi: 10.1007/978-
3-030-90318-3_37.
[16] A. M. Ali, M. A. Ngadi, I. I. A. Barazanchi, and P. S. JosephNg, “Intelligent traffic model for unmanned ground vehicles based
on DSDV-AODV protocol,” Sensors, vol. 23, no. 14, Jul. 2023, doi: 10.3390/s23146426.
[17] S. C. Rai, S. P. Nayak, B. Acharya, V. C. Gerogiannis, A. Kanavos, and T. Panagiotakopoulos, “ITSS: an intelligent traffic
signaling system based on an IoT infrastructure,” Electronics, vol. 12, no. 5, 2023, doi: 10.3390/electronics12051177.
[18] A. Chowdhury, S. Kaisar, M. E. Khoda, R. Naha, M. A. Khoshkholghi, and M. Aiash, “IoT-based emergency vehicle services in
intelligent transportation system,” Sensors, vol. 23, no. 11, Jun. 2023, doi: 10.3390/s23115324.
[19] K. Nazar et al., “Towards intelligent zone-based content pre-caching approach in VANET for congestion control,” Sensors,
vol. 22, no. 23, Nov. 2022, doi: 10.3390/s22239157.
[20] A. A. Ahmed, S. J. Malebary, W. Ali, and O. M. Barukab, “Smart traffic shaping based on distributed reinforcement learning for
multimedia streaming over 5G-VANET communication technology,” Mathematics, vol. 11, no. 3, 2023, doi:
10.3390/math11030700.
[21] J. Guo, L. Cheng, and S. Wang, “CoTV: cooperative control for traffic light signals and connected autonomous vehicles using
deep reinforcement learning,” IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 10, pp. 10501–10512, 2023,

2655
doi: 10.1109/TITS.2023.3276416.
[22] H. Beenish, T. Javid, M. Fahad, A. A. Siddiqui, G. Ahmed, and H. J. Syed, “A novel markov model-based traffic density
estimation technique for intelligent transportation system,” Sensors, vol. 23, no. 2, Jan. 2023, doi: 10.3390/s23020768.
[23] S. Kannadhasan, M. Shanmuganantham, and R. Nagarajan, “System model of VANET using optimization-based efficient routing
algorithm,” IOP Conference Series: Materials Science and Engineering, vol. 1119, no. 1, 2021, doi: 10.1088/1757-
899x/1119/1/012021.
[24] M. A. Makridis, J. Schaniel, and A. Kouvelas, “Rule-based on-off traffic control strategy for CAVs on motorway networks:
assessing cooperation level and driving homogeneity,” IEEE Access, vol. 11, pp. 35111–35121, 2023, doi:
10.1109/ACCESS.2023.3265470.
[25] P. Zibner, P. H. L. Rettore, B. P. Santos, J. F. Loevenich, and R. R. F. Lopes, “DataFITS: a heterogeneous data fusion framework
for traffic and incident prediction,” IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 10, pp. 11466–11478,
2023, doi: 10.1109/TITS.2023.3281752.
[26] Y. Gao, C. Zhou, J. Rong, Y. Wang, and S. Liu, “Short-term traffic speed forecasting using a deep learning method based on
multitemporal traffic flow volume,” IEEE Access, vol. 10, pp. 82384–82395, 2022, doi: 10.1109/ACCESS.2022.3195353.
[27] Q. Ma and Z. Zou, “Traffic state evaluation using traffic noise,” IEEE Access, vol. 8, pp. 120627–120646, 2020, doi:
10.1109/ACCESS.2020.3006332.
[28] Q. Ma, S. Zhang, Y. Qiao, and M. Feng, “Traffic guidance self-organization method for neighbor weaving segments based on
self-organized critical state,” IEEE Access, vol. 8, pp. 171784–171795, 2020, doi: 10.1109/ACCESS.2020.3025006.
[29] C. B. Rafter, B. Anvari, S. Box, and T. Cherrett, “Augmenting traffic signal control systems for urban road networks with
connected vehicles,” IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 4, pp. 1728–1740, 2020, doi:
10.1109/TITS.2020.2971540.
[30] S. S. Sepasgozar and S. Pierre, “Fed-NTP: a federated learning algorithm for network traffic prediction in VANET,” IEEE Access,
vol. 10, pp. 119607–119616, 2022, doi: 10.1109/ACCESS.2022.3221970.
[31] R. Huang, W. Guan, G. Zhai, J. He, and X. Chu, “Deep graph reinforcement learning based intelligent traffic routing control for
software-defined wireless sensor networks,” Applied Sciences, vol. 12, no. 4, 2022, doi: 10.3390/app12041951.
[32] B. Kurungadan and A. Abdrabou, “Using software-defined networking for data traffic control in smart cities with wifi coverage ,”
Symmetry, vol. 14, no. 10, 2022, doi: 10.3390/sym14102053.
[33] F. Masood, W. U. Khan, S. U. Jan, and J. Ahmad, “AI-enabled traffic control prioritization in software-defined iot networks for
smart agriculture,” Sensors, vol. 23, no. 19, Oct. 2023, doi: 10.3390/s23198218.
BIOGRAPHIES OF AUTHORS
Arjunan Mari Kirthima received B.E. degree in Computer Science and
Engineering from M. K. University, and the M.E. in Computer Networking Engineering from
Anna University. Currently, she is an assistant professor at the Department of Computer
science and Engineering, BMSIT and also research scholar in VTU research center. Her
research interests include VANET, computer networks, machine learning, and artificial
intelligence. She can be contacted at email: krithi.a@bmsit.in.
Dr. Pushpa Sothenahalli Krishnaraju received B.E. degree in Computer
Science and Engineering from SCJCIT, and the M.Tech. degree in Computer Science and
Engineering from Bangalore University. She holds a Ph.D. in CSE from Vinayaka Missions
University. Currently, she is a professor at the Department of Information Science and
Engineering, BMSIT. She comes with 19 years of teaching experience in Academia. She has
published around 10 international journals, and 20 international conference papers. Her
research interests include wireless sensor networks, big data, data analytics, and artificial
intelligence. She can be contacted at email: pushpask@bmsit.in.

Software defined network-based controller system in intelligent transportation system

More Related Content

Similar to Software defined network-based controller system in intelligent transportation system (20)

More from IAESIJAI (20)

Recently uploaded (20)

Software defined network-based controller system in intelligent transportation system