ISSN 2347 - 3983

Oguz Aydıner et al., International Journal of Emerging

VolumeTrends in Engineering
9. No.11, NovemberResearch,
2021 9(11), November 2021, 1387 – 1394
International Journal of Emerging Trends in Engineering Research
Available Online at http://www.warse.org/IJETER/static/pdf/file/ijeter049112021.pdf
https://doi.org/10.30534/ijeter/2021/049112021

An IoT Based Intelligent Traffic Management System and Its

Implementation on Cupcarbon for Smart Cities
Oguz Aydıner1, Adnan Kavak 1,2
1
Yeditepe Univ., Electrical and Electronics Eng. Dept., 34000 İstanbul, Turkey,
[email protected]
2
Kocaeli University, Computer Eng. Dept, 41380 İzmit, Turkey, [email protected]

Received Date : October 03, 2021 Accepted Date : October 25, 2021 Published Date : November 07, 2021

 The congestion in traffic is essentially because of high

ABSTRACT demand on traffic, which occurs in different times of the day
with different contributing factors. Most of the congestions
This paper proposes an internet of things (IoT) based are caused by bottlenecks on the traffic [3]. They are mostly
intelligent traffic management system that can aid the reason why traffic even gets congested because the
problematic traffic situations in smart cities by classifying following reasons of congestions are secondary to bottlenecks.
congestions via sensory data, and then controlling traffic The poor signal timing or even worse, non-signaling causes
lights and creating alternate routes for incoming vehicles to an interesting phenomenon called “Phantom Traffic Jam”
the congested zones in order to relieve or avoid congestions which occurs when the slightest tipping point of the traffic is
completely. The proposed intelligent traffic management reached when the road is heavily dense with vehicles, but still
system consists of different subsystems such as Test flowing nonetheless [4]. The traffic incidents play a similar
Operation, Supervisory, Traffic Light, and Pathfinder role to the work zones as they too create bottlenecks and
subsystems. The system is represented by flowcharts with restrict the “airflow” of the roads so the traffic gets congested
their explanations and its operation with some defined very quickly. Work zones usually creates bottlenecks on the
scenarios is validated with the CupCarbon simulation roads since they need to close down that area of maintenance
environment. to the traffic and the five-lane road goes down to three, there
is indeed going to be bottlenecks. Bad weather is also a cause
Key words : Cupcarbon, IOT, Smart City, Traffic for traffic congestion simply because it increases the demand
Management, 5G for use of transportation via traffic.

1. INTRODUCTION Todays’ mega-cities which are prime contenders to become

the future’s smart cities are bound to have above mentioned
The traffic congestion is one of the major problems in today’s traffic problems. Description of a framework for utilizing the
mega-cities. There are of course outliers to the congestion various sources efficiently in the context of traffic
problems but some cities have it worse than others. As an management is addressed in [5]. One countermeasure of the
example, Istanbul has the second most congested and traffic congestion problems of such cities is the design and
sudden-stopping traffic in the world [1]. Istanbul’s traffic installation of intelligent traffic systems that can be brought
congestions are very often, emissions are high, incidents are by the advancement of technology such as IOT networks
regular and road rages are a part of daily life. To address the enabled by fifth generation (5G) mobile networks [6]. In [7],
traffic problem in the big and highly populated cities such as Intelligent traffic control system to solve crossroad problems
Istanbul, heavy infrastructure investments can be made which and parking space problems using wireless sensor network
will cost a lot of taxpayer money, such as bridges and (WSN) to collect road data and available parking space in
additional roads. However, these costly investments do not smart cities are given.
only cost a large amount of taxpayer money, they also come
with additional costs such as taking up free space, In this paper, as a solution to the congestion problem in smart
environmental damage and prevent further infrastructure cities, we propose a method for intelligent traffic management
projects. Therefore, installing intelligent systems to manage based on IOT sensor data. Our solution addresses a possible
traffic that can adjust themselves based on the vehicle density solution to traffic congestion problems. We perform
are beneficial in terms of the best use of resources [2]. simulations of the proposed system via CupCarbon IOT
simulator.

1387
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

2. INTELLIGENT TRAFFIC MANAGEMENT

SYSTEM BASED ON IOT NETWORK The Supervisory Subsystem’s (SS) flowchart is given in
Figure 2. Its purpose is to tie the whole Traffic Management
In this paper, we propose a traffic management system to System together with its subsystems. It starts off with sending
operate in smart cities based on IOT sensor data. The a “situation report” signal and receive feedback.
proposed system consists of four subsystems namely Test Classification of road congestion levels within the IoT
Operation, Supervisory, Traffic Light, and Pathfinder Network area is firstly done by the Supervisory System to
subsystems. Below, we explain operation of each of these determine the situation of the traffic and decide whether to
subsystems. contact a relevant subsystem or end operation. If it decides to
activate a subsystem, it then needs to decide which one to
wake up. So, there has to be a classification about the road
congestion levels.

Figure 1: Test Operation Subsystem of the Proposed Traffic

Management System

Figure 1 shows the “Test Operation Subsytem” of the

proposed IOT Traffic Management System, which is good to
have for a realistic system’s operations. It could save a lot of
the hassle and time-waste of trying to detect the erroneous Figure 2: Supervisory Subsystem of the Proposed Traffic
nodes in a system where there is not a safety measure built in. Management System
As we can see, it firstly orders the transmitter to send a test
signal to all of the IOT sensor nodes to check which are Level_0 is “no congested traffic” detected case, which means
responsive and which are not. The nonresponsive nodes are that the system detected no congestion as in the detected
marked with their location data for the maintenance crew. object count and their detection timeframes are as expected on
The responsive nodes pass the first test and continue to the the area where IoT network is deployed. So, it ends its
second step for testing their functionality, as in checking if operation and starts its timer for the duration of 15 minutes.
they are doing what they are supposed to do. In the last step Because it’s dealing with traffic and traffic never stops so the
they are tested for their most essential function; congestion system shouldn’t stop either. That is why after setting its
sensing. The test system sends a signal for their sensing and timer, it waits. Level_1 is “heavy traffic” detected case, which
reporting capabilities and either confirms they are indeed means that the vehicle count and their detection timeframes
operational or decides that the data they are reporting is are between minimum and maximum parameters and the
erroneous and marks them for maintenance. If the test is traffic is congested. So, the Supervisory Subsystem activates
passed by flying colors and there are neither errors in the the Traffic Light Subsystem (TLS) on the road that are
operation of the nodes nor non-responsive nodes, the system determined to be Level_1 congestion. The TLS operates as
ends operation and notifies the tester that the system is shown in Figure 3. Level_2 is “deadlock traffic” detected
healthy. case, which means that the traffic is extremely congested. In

1388
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

this situation, the IoT Network’s vehicle count is very high the current congestion. So, the system halts its operation,
and the more deciding factor is, their detection timeframes are contacts Pathfinder (PATH) subystem and ends all
very high as well, meaning they are barely moving or operations.
stationary. So, if this happens to be the case, SS proceeds as
shown in Figure 2 and activates PATH system to map Figure 4 shows the flowchart of the Pathfinder (PATH)
alternative routes to incoming vehicles in order to decrease Subsystem in detail. Similar to other subsystems, PATH too
the congestion on the road that is already reached Level_2 sends a signal to all nodes of the IoT network to receive
congestion. After some iterations when the congestion level is information about the current status of the road congestions
decreased, the operating subsystems notify the Supervisory from the sensors. Then the system classifies the received
System right before ending its operation, and the system starts feedback from the sensors and sets a priority of Level_0 over
its timer for when it is over. Level_1 road when mapping alternative routes. This is done
in order to both optimize the routes of incoming vehicles to
their destinations and divert the traffic to more unused routes
to help decrease the congestion of the deadlock road. Meaning
a Level_0 road can potentially host more vehicles than a
Level_1 road because it’s further away from becoming
another deadlock. After the priority phase, the system reaches
out to the driver’s GPS to request location. After receiving, it
then notifies the driver about the upcoming congestion and
requests a manual input regarding their destination.

Figure 3: Traffic Light Subsystem of the Proposed Traffic

Management System

The flowchart of the Traffic Light Subsystem (TLS) is shown

in Figure 3. It first sends a signal like in Test Operation
Subsystem to all nodes to receive a feedback about the current
congestion levels of the region wherein the sensors are placed.
When the feedback is received, the system locates the area of
congestion, and shortly after that, it locates the traffic lights
that effect those particular roads. Then the system sends a
signal to increase the duration of green light by 15 seconds,
which is a placeholder value and would change according to Figure 4: Pathfinder (PATH) Subsystem of the Proposed Traffic
the circumstances. After increasing the green light duration, Management System
the system waits and observes the congestion levels by
sending a signal again to the IoT network nodes at the end of When the destination is set, the system maps an alternative
the wait duration which is another placeholder value as 5 route for the driver. Then constantly begins to check the
minutes. If the congestion level is dropped to Level_0 from alternative route’s congestion levels to see if there is any
Level_1, the system sends a confirmation signal to the SS and change in the congestion and if there is no change it then
ends it operation. But if the congestion level is still the same, checks vehicle location to see if it arrived its destination. If the
as in Level_1, the system goes back a few steps in the system senses a change of congestion in the alternative route,
operation and increases the green light duration for another it then asks if this change is an unwanted change as in an
15 seconds. However, if the congestion level is raised to increase in the congestion levels. If that is not the case and the
Level_2 even if the system has increased the green light change is for example, Level_1 to Level_0, then it does
duration, then it means the TLS will not be enough to resolve nothing since this case benefits everyone and ends its

1389
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

operation after notifying the Supervisory Subsystem.  Depending on the congestion levels, The system
However, if the change is in an unwanted direction, as in an communicates with the related subsystems to resolve it.
increase in the congestion levels, then PATH goes back to the  The two deciding factors of the congestion levels are
beginning of the operation to map a new route for the driver. vehicle count and detection timeframes. We chose to explain
The operation is only resolved when the congested road’s them without giving them numerical values as it would not be
levels have been reduced as there is no other means to deal accurate and would differ heavily according to the IoT
with it better than the PATH system so the cycle repeats until Network’s placement.
when the levels are dropped.
3.1. IOT Node Communication in Intelligent Traffic
3. CUPCARBON SIMULATIONS AND RESULTS Management System

Proposed traffic management system is simulated using For the demonstration of the operation of our proposed
CupCarbon IOT Simulator [8],[9]. It is a communications system, we have chosen the location as “Bostancı Bridge” in
simulation tool to visualize the cases created, specifically Istanbul, because the congestion problems are quite common
designed for IoT Networks and Smart Cities with built-in in that region. It gets congested every rush hour and even on
OpenStreetMap. It is very helpful to visually explain the the weekends regularly. It was also quite handy for drawing
workings of communications networks such as sensors and routes as it is right above the D100 highway in Istanbul so that
decision support system of the project. The program has two we can have more real estate for the sensor placements. The
different environments for simulation. One is for events first simulation is designed to better understand how the
which include mobility like traffic and natural events. Second communication between the IoT Network nodes occur and
environment is for discrete events of the IoT Networks, which how do the systems send a check signal to nodes for them to
also encompasses the first environments.The platform is still check the traffic congestion levels and report back their
being developed and the users have limited uses for their gatherings. In Figure 5, the system is in its initialized state.
simulation scenarios due to the complex structure of the IoT The IoT Network is made up of 100 nodes which are scattered
Networks on the OpenStreetMap frame. around the area of selection at random. The system is in its
inactive state so we are not seeing any communication arrows
We have used several objects in CupCarbon simulations to popping out of them. The transmitter node is at the
accurately describe traffic congestions and aid the bottom-right of the network if you look closely, which will
understanding of the system which is represented with send a signal to the system and receive a feedback throughout
flowcharts. Sensors are modules which has the purpose of the simulation.
detecting objects or changes within it’s sensing radius.
Sensors can be actually linked together by just placing them
close enough and assigning the correct scripts with their roles
as in receiver/sensor/transmitter to make them communicate.
Base station’s function is very similar to what they do in real
life, which is transmitting and receiving signals and being a
bridge between node-system communications. Base station
locations need to be in the most optimal places to better
communicate with the network. Mobile objects are to
visualize moving vehicles / drones on the OpenStreetMap
within the assigned routes. The mobile object also has a
sensing radius similar to the sensor nodes on the simulation.
This is because the object can be scripted to interact with the Figure 5. A Screenshot of CupCarbon Simulation at Initial
nodes it comes across to change its route. The basic principles Condition for Node Communication in Intelligent Traffic
of the CupCarbon simulations of proposed traffic Management. Total of 100 nodes are assumed to be deployed
management systems are as follows; randomly around Bostancı Bridge in Istanbul
 The IOT Network refers to a network of sensors scattered
around a designated area which communicate between We can see from Figure 6 that the signal has been sent and it’s
themselves and the base stations for the management. being transmitted from one node to another within their
 The system never stops and when it comes to the end of the radius in a wave-like formation. The closest nodes to the
operation flow, it starts a timer for when it ends, the operation transmitter are not showing any arrow signs because they are
repeats. already checked, given up their information and marked. We
 The system sends signal to all of the nodes of IoT Network can see that transmitted signal is in the middle of the IoT
and checks the congestion. network and progressing towards the outer edges. The nodes
 The system classifies the traffic congestion into three levels are rapidly checking and confirming each other’s signals and
as no traffic, heavy traffic and deadlock. relaying them to the unchecked nodes so on and so forth. Each
node gets checked multiple times to see if it’s marked and if

1390
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

they are, they are left alone so that the signal wave can So as can be seen in Figure 8, the feedback message
continue towards the outer edge. The nodes get marked once transmission has begun from the outermost responsible node.
they have relayed their information to their corresponding They are individual nodes following a straight line and not all
neighbor node. When transmitted signal reaches the very of them at once, because that would be inefficient to have all
edge nodes, they will signal the responsible node between nodes communicate back while we can have them report to
them to relay a feedback signal to the transmitter. You can see their responsible nearest node and still have the same result
the receiver script in Figure 7, which runs at every node in the but faster. We can see that the feedback signal is moments
simulation except the transmitter node which is located down away from reaching the transmitter and the signaling
below in Figure 6. In the node script, the process is easy to operation is about to be finalized. You can see the transmitter
follow as nodes very quickly ask their neighboring nodes script down below. As can be seen in Figure 9, basically the
when they receive the signal if they are marked. If they are not transmission occurs at the point where the “send” command
marked it means they are not checked, meaning the signaling is and then it waits until receiving the feedback from the
should progress towards them, so they get the information nodes and checks if it matches its type, which it will, then
from their neighboring marked nodes and with adding their marks it and ends the process.
own, relay it to them, then the receiving node marks the
neighboring node to keep them out of the checking loop and
allowing the signal to progress like a wave towards the edges
of the network.

Figure 8. A Screenshot of CupCarbon Simulation Showing

Feedback Message Tranmission Towards Transmitter Source Node
for Node Communication in Intelligent Traffic Management System.

Figure 6. A CupCarbon Simulation Screenshot After a While When

Node Communication in Intelligent Traffic Management is Started.

Figure 9. The CupCarbon Script Running at the Individual

Transmitter Nodes for Node Communication in Intelligent Traffic
Management System

3.2. Route Changing with Location Signaling in

Intelligent Traffic Management System

This simulation is combined with various features to better

visualize the PATH subystem in our proposed intelligent
traffic management system. The operation will start with our
mobile object interacting (representing a vehicle) with
assigned roadside IOT sensors to receive “change route”
Figure 7. The CupCarbon Script Running at the Individual Receiver commands and reaching its final destination. Upon reaching,
Nodes for IOT Node Communication in Intelligent Traffic it will actively signal its location to the IoT network
Management System established on the map.

1391
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

In Figure 10, we can see that the simulation is in its initial

phase, the mobile object is not moving and the system is not
active but we can still see the pre-drawn routes for the mobile
object to follow. It really is not possible in a CupCarbon
simulation to literally draw alternative routes during the
simulation progress and have that kind of processing
capability but it’s still quite alright to simulate a system like Figure 12. The CupCarbon Simulation Script of the Sensor1 for
PATH as shown in the figure to give a vague sense of Route Following and Changing with Location Signaling
operation and actually explain the system in flowcharts.
In Figure 13, we can see that our mobile object (vehicle) has
Figure 11 shows that the simulation has begun, mobile object reached the second sensor for route changing and its script is
has quickly followed it’s assigned marker route and interacted almost identical to the first one, only difference being is the
with the first route changing sensor, which signals it to follow sent signal as in the name of the new route, which is m3 in
the node m2 as in the other route in its radius. So, mobile this case, the name of the third marker route. The one minor
changes its route and proceeds to follow m2. In Figure 12, you difference is the tag name changes from A to B, which the
can see the script of the first sensor node. It simply reads a tag sensor is checking, that is simply to avoid confusion as you
and sends a route change signal as m2 which is the name of will see in the script of the mobile object. The first sensor
the second route. demands A tag, when receives, commences to send signal
change route, the second sensor demands B tag, so on and so
forth.

Figure 10. Initial Screenshot on CupCarbon for the Route Following Figure 13. Third Screenshot on CupCarbon for the Route Following
and Changing with Location Signaling Simulation and Changing with Location Signaling Simulation

Figure 14 shows that the mobile object has reached its last
stop regarding the route changer sensors, which is the third
sensor. It will then send a signal to change its route from m3
to m4 which is the fourth and final route before arriving its
destination.

Figure 11. Second Screenshot on CupCarbon for the Route

Following and Changing with Location Signaling Simulation

Figure 14. Fourth Screenshot on CupCarbon for the Route

Following and Changing with Location Signaling Simulation

1392
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

Figure 15 shows that the mobile has reached its final

destination but it’s not the end of the simulation. On the
contrary, another interesting operation starts to take place,
which is the location signaling from the end point to the IoT
network for the Supervisory System to get out of the loop.
When the mobile object has reached to the region shown by
the fire sign at the end point, the sensors there are linked
together and scripted to sense this particular occurrence and
mark each other just like in the Node Communications
simulation prior to this, only difference is that the signals are
relayed by the base stations so they visually seem more
orderly.

Figure 15. Fifth Screenshot of the Route Following and Changing

with Location Signaling Simulation
Figure 20. The Script of the Mobile Object (Vehicle) for the Route
Following and Changing with Location Signaling Simulation

4. CONCLUSION
In this paper, we have proposed a system for intelligent traffic
management to overcome traffic congestion problems in
smart cities. The system is composed of four subsystems: Test
Operation Subsystem, Supervisory Subsystem, Traffic Light
Subsystem, and Pathfinder (PATH) Subsystem subsystems.
Figure 16. The Script of the Sinks of the Route Following and Algorithms of each of these subsystems have been given and
Changing with Location Signaling Simulation simulation of this system with an emphasis on Route
Following and Changing with Location Signaling scenario
has been made on CupCarbon platform which is an IOT
In Figure 16, the script of the sink which is a base station is network simulation tool.
given. They are linked to many sensors and control the flow of
signals along the network to properly use the data when the The systems that are designed here are raw and core concepts
events occur such as the marked vehicle reaching its and there is vast amount room for improvement with
destination point. possibilities for potential development in the future smart
cities. For example, the IoT Network’s sensors can be
In Figure 17, the script of the main player of the simulation upgraded and reinforced with image processing which can
which is the mobile object is given. As explained priorly at the enable the system to easily recognize the problems that are
sensor scripts, you can see that the tag changes depending on occurring and swiftly work on solutions. Or rather, the system
the sensor. Which puts the mobile’s operation in another “if” can be combined with artificial intelligence which can be
condition, allowing it to easily proceed to change its route trained to manage the traffic flow entirely. These are all
depending on the sensor demanding the signal and incoming futuristic possibilities that have enormous potential but they
command. can indeed be realized and they would solve a vast spectrum of
problems.

1393
Oguz Aydıner et al., International Journal of Emerging Trends in Engineering Research, 9(11), November 2021, 1387 – 1394

REFERENCES
1. https://www.tomtom.com/en_gb/traffic-index/ranking/?po
pulation=MEGA,LARGE,SMALL [Accessed 10 2021].
2. H. Terzioglu, A. A. M. Harb and A. Durdu, Design and
Implementation of Real-Time System with Fuzzy Logic
Method For Junction Traffic Signalization Control,
European Journal of Science and Technology, no. Special
Issue, pp. 490-497, 2019.
3. A. Koźlak and D. Wach, Causes of Traffic Congestion in
Urban Areas. Case of Poland, SHS Web of Conferences,
vol. 57, 2018.
4. K. Goldmann and G. Sieg, Economic Implications of
Phantom Traffic Jams: Evidence From Traffic
Experiments, Transportation Letter: The International
Journal of Transportation Research, vol. 12, no. 6, pp.
386-290, 2020.
5. A. Allström, J. Barceló, J. Ekström, E. Grumert, D.
Gundlegård and C. Rydergren, Traffic Management for
Smart Cities, Designing, Developing, and Facilitating
Smart Cities: Urban Design to IoT Solutions, pp. 211-240,
2017.
6. L. Chettri and R. Bera, UL-DMRS Based NB-IoT Uplink
System and its Performance Analysis Toward 5G
Machine Type Communications, International Journal of
Emerging Trends in Engineering Research, vol. 9, no.1, pp.
6-14, January 2021.
7. A. Hilmani, A. Maizate and L. Hassouni, Automated
Real-Time Intelligent Traffic Control System for Smart
Cities Using Wireless Sensor Networks, Wireless
Communications and Mobile Computing, vol. 2020, pp. 1-28,
September 2020.
8. CupCarbon IoT Simulator, http://cupcarbon.com.
9. M. Akkurt and K. Küçük, Simulation of Smart City
Applications Based on IoT Technologies with
CupCarbon, In Proc. 3rd International Conference on
Computer Science and Engineering, 2018, pp. 179-184.
10.

1394