1

Revolutionizing Urban Transportation: IoT-Driven

Traffic Management and Smart Parking for
Sustainable Smart Cities
Your Name, Member, IEEE,

Abstract—Urban transport and car parking monitoring in II. R ELATED W ORKS

Indian metropolitan cities face significant challenges, such as
low speeds, long wait times, and prolonged vehicle idling. As The proposed system in [1] integrates the YOLO model
the number of Smart Cities increases, so does the demand for
innovations in urban transport. The existing traffic management with a traffic congestion analysis module to optimize traffic
system struggles with heavy congestion due to the rising number flow during peak hours. Using a dataset from an Indian driving
of vehicles and complex traffic conditions, as it is static in nature. dataset and Google images, the YOLO-based CNN detects and
This article aims to propose a dynamic traffic management identifies vehicles, employing YOLO Darknet architecture
system that includes technologies such as the Internet of Things with OpenCV in Python. Achieving precision rates of
(IoT), object detection algorithms, and mathematical approaches
to make data-driven decisions to adjust traffic signals accord- 76.70%-89.24% (YOLO v3) and 83.08%-92.65% (YOLO
ingly, which improves traffic flow, enhances air quality, and also v4), the model analyzes traffic congestion every second and
reduces carbon emissions, contributing to a more sustainable performs well in adverse weather and also predicts future
urban environment. Parking is another noteworthy issue that traffic patterns for better peak hour management.
contributes to urban congestion. A combination of GPS and The paper [2] suggests a two-tier edge computing architecture
parking sensors, along with an application that shows available
parking spaces in real time, will cut wait times, decrease the for traffic monitoring and leverages computer vision for
likelihood of wasting space, and reduce the time taken to find a speed and congestion identification. It uses edge nodes
spot, enhancing urban mobility. (cloudlets) and centralised traffic management centres
(TMCs) algorithms. While the edge algorithms are optimised
Index Terms—Intelligent Transportation System (ITS) Internet
of Things (IoT) Object detection Real-time Traffic Control for limited resources, the TMC handles more complex
Sustainable Urban Mobility. tasks. This adaptive method improves performance during
bad weather by reducing data transfer to the TMC in
response to network conditions. This hybrid approach, which
leverages computer vision, outperforms traditional techniques,
I. I NTRODUCTION according to testing results. This is a significant advancement
in IoT-driven traffic control.

I NDIA is starting to see significant problems with traffic

congestion; several of its cities are among the most clogged
worldwide. A research by the Boston Consulting Group es-
The proposed system in the research [3], updates traffic
light timers based on traffic density, which is measured
using infrared sensors. These sensors, controlled by an
timates that by 2030, the cost of India’s traffic congestion ESP8266-based Node MCU, send data to a central cloud
problems could reach 37 billion dollars. This means that the system. RFID tags enable automatic toll billing and alert
problems with parking management and urban mobility need nearby traffic police of speed violations. The implementation
creative, quick fixes. The ineffectiveness of current parking includes a layout of a four-way junction (East, West, North,
management systems and the absence of networking across South) with ultrasonic sensors in all directions, connected via
traffic intersections are the main causes of the technology an IP network.
gap. This gap can be filled by integrating solutions with IoT In [4], The RFID reader scans the RFID tag on the ambulance
technology, which can make decisions in real-time for non- and updates the upcoming traffic light to green, displaying a
linear traffic conditions. Our suggested solution consists of an message for vehicles ahead to provide a ”Green Corridor”
inventive traffic switching algorithm that enables computers by shifting lanes. Each ambulance is equipped with an
to make judgements instantly and provides real-time traffic RFID tag, and traffic signals have RFID readers that store
control based on data acquired. Poor parking space manage- the tag information and send it to the server via cloud
ment is another factor contributing to today’s traffic, as it computing. The implementation involves an LCD connected
causes clogged roadways and more vehicles. With the goal of to an Arduino Mega. When an ambulance passes the RFID
reducing the average amount of time spent looking for parking reader, a message on the LCD changes from ”WELCOME
and easing traffic congestion, an interactive smart parking TO NH-24” to ”AMBULANCE COMING - GO TO THE
application and a scalable, Internet of Things-based smart LEFT LANE,” and the red LED turns green.
parking infrastructure deliver complete real-time information
on parking spot availability and precise positions.
 2

The paper [5] develops a predictive risk analysis model parking slots.It introduces a new deep learning architecture,
combining deep learning and mathematical modeling. It VPS-Net, specifically designed for vacant parking slot
gathers traffic camera feeds, GPS data, and historical patterns detection. The architecture combines a multi-object detection
to design a network for complex interactions. The model network and a classification network to improve accuracy
assesses risks by tracking five objects (car, motorcycle, cycle, and robustness.The proposed method achieves accurate
rickshaw, human), their positions, and velocities using LSTM, localization of parking slots by using geometric information
YOLO, and SORT algorithms. The risk prediction process and spatial constraints.
is divided into path prediction and risk analysis modules, The paper[11] presents a detailed case study of a smart parking
assigning dynamic risk values to object pairs. This approach system implemented in the city of Melbourne, Australia. The
demonstrates effective predictive risk assessment in chaotic system utilizes a combination of technologies, including IoT
Indian traffic scenarios using deep learning and rule-based sensors, cloud computing, and mobile applications, to address
modeling. the challenges of parking in urban areas.It provides a valuable
The paper[7] proposes a reservation-based parking system case study of a smart parking system in a large urban city,
that aims to address the challenges of parking in urban areas. demonstrating its practical applications and benefits. The
The system uses mobile and web technologies to provide paper highlights the environmental benefits of smart parking,
real-time information about parking availability, facilitate including reduced emissions and fuel consumption.this paper
reservations, and guide drivers to available parking spaces. offers a valuable perspective on the practical implementation
The paper introduces a novel reservation-based parking of smart parking systems and their potential to improve urban
system, allowing drivers to reserve parking spaces in advance, mobility and sustainability.
reducing the time spent searching for available parking.The
proposed system aims to improve parking efficiency by III. P ROPOSED W ORK : S MART T RAFFIC M ANAGEMENT
reducing the time drivers spend searching for parking and A. Vehicle Detection Module
minimizing traffic congestion.The system leverages mobile
Real-time traffic footage is recorded by CCTV cameras
and web technologies to provide a user-friendly interface and
positioned at traffic signals and is subsequently retained for
enable real-time information access.
further processing. The number of vehicles on the road is
A novel approach to address the challenges of parking
detected and counted using the YOLO v8 object detection
in urban areas is presented in [8] . The proposed system
technique.This approach uses a single neural network to eval-
leverages the capabilities of cloud computing and Raspberry
uate the entire image, splits it into areas, and then forecasts
Pi devices to provide real-time information on parking
probabilities and bounding boxes for each region. These an-
availability and facilitate efficient parking management.The
ticipated probabilities are used to weight the bounding boxes.
system effectively combines the Internet of Things (IoT) and
Using the graphical image annotation application LabelIMG,
cloud computing to create a scalable and efficient solution.The
photos that were scraped from Google were manually labelled
system utilizes sensors and cloud-based analytics to provide
to create the dataset for the model. The DNN module of
real-time updates on parking occupancy, enabling drivers to
OpenCV was used to load the pre-trained YOLO model
find available parking spaces quickly.By leveraging cloud-
weights. Yolov3.cfg, the configuration file, was utilised in
based computation, the system can handle large volumes of
accordance with the model’s specifications. Using pre-trained
data and scale as needed to accommodate growing urban
weights, the model recognises many vehicle types, including
areas.The use of Raspberry Pi devices as edge computing
Car, Bike, Bus/Truck, and Motorbike. The process of detection
nodes helps to reduce hardware costs and improve system
involves feeding a picture into the model, which uses blob
efficiency.
format to process the image. As shown in Fig. 1, the model
The authors of paper[9] identify the key components of
produces bounding boxes and class probabilities. A threshold
smart parking systems, such as sensors, communication
was established for the minimal level of confidence needed
networks, and data management systems.The paper explores
for a successful detection. Bounding boxes are drawn on the
the different technologies used in smart parking, including
images using OpenCV depending on the coordinates and labels
IoT, cloud computing, and machine learning. It also explores
that are recognised. Following the results’ saving to an output
the different technologies used in smart parking, including
picture file, a count of every vehicle class that was identified
IoT, cloud computing, and machine learning.It provides
is given back.
a comprehensive overview of the field of smart parking
systems. The authors discuss the different types of smart
parking systems, their components, and the technologies B. Traffic Simulation Module
used to implement them. They also explore the challenges Pygame was used to create a traffic simulation from scratch,
and opportunities associated with the development and as seen in Fig. 2, which mimics actual traffic circumstances.
deployment of smart parking systems. An illustration of a 4-way intersection with four traffic signals
A novel method for detecting vacant parking slots in images is shown by the simulation. A fixed timer, set to 25 seconds,
captured from a surround-view camera system in proposed indicates how long it will take for the light to change from
in [10]. The proposed method, named VPS-Net, combines green to yellow, yellow to red, or red to green on each signal.
the advantages of a multi-object detection network with a Every signal also shows the quantity of cars that have passed
classification network to accurately locate and classify vacant through the intersection next to it. Cars, bikes, buses, lorries,
 3

traffic signal system to dynamically adjust by optimizing the

assigned green time for different approaches.
1) Initial Setup and Green Passage Time (GPT)
calculation: At a four-way intersection, let the four
roads be Up, Down, Right, and Left. Initially, snapshots
of all four roads are taken, and the road with the highest
density is set as the priority and chosen for the green
light first. Assuming the ”Up” road has the highest
density, it is given the green light. As the green light
phase for the current priority road (“up”) approaches its
last five seconds, new snapshots of the other three roads
(down, left, and right) are taken. The traffic density on
these roads is re-evaluated using the GPT formula (see
Equation 1).
Fig. 1. YOLO-Based Vehicle Detection and Classification with Bound-
ing Boxes

Fig. 2. Traffic Simulation Using Pygame

and rickshaws are among the many vehicles that approach the
crossing from all angles. Certain cars in the rightmost lane are
set to turn and cross the intersection to increase realism.When
a vehicle is generated, it is randomly determined whether it
turns or stays straight. A timer included into the simulation
displays the amount of time that has passed since it began. By
using this configuration, the traffic system may be seen and
contrasted with the current static system.
Simulated features include:
1. Fixed Timer Traffic Lights: These traffic lights alternate
between green and yellow, red and green, and both on a set Fig. 3. Dynamic Signal Switching
25-second cycle.
2. Vehicle Diversity: To replicate real-world traffic, the simu- 2) Heirarchical Traffic Lane Prioritization : Traffic sig-
lation includes a variety of vehicle types, such as automobiles, nal adjusts dynamically in a hierarchical manner. Firstly,
bikes, buses, trucks, and rickshaws. 3. Turning Vehicles: To the roads are assessed based on their Green Passage
enhance the realism of the simulation, vehicles in the rightmost Time (GPT), which reflects current traffic conditions.
lane have the opportunity to turn at the intersection. The choice Then, other factors, such as waiting time and recently
to turn is made at random. 4. Vehicle Count Display: Next to selected roads, are considered to maintain fairness and
each traffic signal, the number of vehicles that have crossed responsiveness. Once a road is selected for the green
the intersection is shown. 5. Elapsed Time Display: The total light, it is added to a queue called ”Recent Choices.” The
amount of time since the simulation started is displayed by a road with the highest GPT is chosen for the next green
timer. light phase, provided its count in the Recent Choices
queue is not more than two, ensuring an equitable
distribution of green light time across all directions.
C. Dynamic Signal Switching To address potential inefficiencies, a ”waiting queue”
Dynamic Signal Switching Module consists of several com- is integrated, which tracks each road’s waiting time
ponents as in Fig. 3 - Green Passage Time Calculation, since its last green light phase. A threshold, such as
Hierarchical Traffic Lane Prioritization, Dynamic Queue Man- 90 seconds, is set; roads exceeding this threshold are
agement Unit, and Real-time Density Update Unit. These added to the waiting queue. When selecting the next road
elements work together as shown in Algorithm 1 to make the For a green light, the algorithm first checks the waiting
 4

queue. If it is not empty, the road at the front of the Queues) according to the current situation. A road’s po-
queue—indicating the longest waiting time—receives sition is reset, and the queue is updated when it receives
the green light. If the waiting queue is empty, the algo- a green signal, ensuring equitable traffic distribution.
rithm proceeds with the standard procedure to calculate 4) Real-time Density Update Unit: The vehicle count
the GPT based on real-time traffic conditions and other detection module takes a snapshot of each road’s traffic
relevant factors, ensuring a balanced and efficient traffic every five seconds and processes it to determine vehicle
flow at the intersection. The GPT calculates the weighted density in each lane. The vehicle density is then updated
total time for all vehicle types by multiplying the count in real-time.
of each vehicle type with its average time, summing 5) Signal Transition Unit: The next selected road first
these products, and scaling the result by 0.25, which is receives an orange light phase of 5-10 seconds. This
a factor calculated by assuming the no. of lanes as 4. phase serves two purposes:
X  • It allows vehicles to prepare to move, overcoming
GPT = 0.25 × vCountvType × avgTimevType (1) the initial inertia.
vType
• It provides buffer time for vehicles on the current

Algorithm 1 Dynamic Signal Switching Algorithm green light road to clear the intersection.
Input: Density Up, Density Down, Density Right, Den- After the orange light phase, the selected road is given a
sity Left, Threshold Waiting Time green light for the calculated GPT. This process repeats
Output: Green Light Sequence in a continuous loop, ensuring dynamic signal switching.
1: Densities ← [ ]
2: Recent Choices ← [ ] D. Emergency Vehicle (EV) Priority Module
3: Waiting Queue ← [ ]
Emergency vehicle (EV) response time is critical in saving
4: Priority Road ← max(Densities)
lives and mitigating damage during emergencies. Current
5: Green Light(Priority Road)
traffic signals often operate on fixed cycles or adaptive systems
6: while True do
that fail to effectively prioritize emergency vehicles (EVs).
7: if Green Light(Priority Road) ≤ 5 then
This can result in delayed green signals when EVs approach
8: Updated Density ← Capture Densities({Down, Left,
intersections, causing critical hold-ups. On the other hand,
Right})
activating green lights too early or for too long leads to
9: for road in {Down, Left, Right} do
unnecessary delays for non-emergency vehicles, contributing
10: GPT[road] ←Cal GPT(Updated Density[road])
to congestion and inefficiency. The EV Priority Module is
11: end for
aimed at integrating EV detection and dynamic traffic signal
12: end if
adjustment to improve emergency response times, reduce
13: Recent Choices.append(Priority Road)
congestion, and optimize overall traffic flow by the integration
14: if Waiting Queue is not empty then
of RFID and GPS.
15: Next Road ← Waiting Queue.pop(0)
16: else 1) RFID-Based Detection: Emergency vehicles are
17: Next Road ← max(GPT) equipped with RFID tags that can be detected by
18: while Recent Choices.count(Next Road) > 2 do readers installed near intersections.
19: Remove Next Road from GPT 2) GPS Data Retrieval: The system gathers real-time loca-
20: Next Road ← max(GPT) tion data of the approaching EV.
21: end while 3) Arrival Time Calculation: Based on the current speed
22: end if and position of the EV, the system calculates the time
23: for road in {Down, Left, Right} do the vehicle will take to reach the intersection.
24: if Waiting Time > Threshold Waiting Time then 4) Traffic Signal Adjustment: Traffic signals are dynami-
25: Waiting Queue.append(road) cally controlled to prioritise the approaching emergency
26: end if vehicle by providing it with a green light at the right
27: end for moment.
28: Step 4: Transition to Green Light RFID readers are to be installed at key intersections. When
29: Orange Light(Next Road) {Transition to orange for 5- an EV approaches an intersection, the vehicle’s RFID tag is
10 seconds} detected by the RFID reader. The detected vehicle’s ID and
30: Green Light(Next Road, GPT[Next Road]) timestamp are sent to a central server by the RFID reader.
31: Priority Road ← Next Road This data transmission initiates the next steps of the EV
32: end while prioritization process. Once an RFID reader detects an EV,
the real-time GPS data of the EV is requested by the central
server. Using the retrieved GPS data, the expected arrival time
3) Dynamic Queue Management Unit: After reevaluat-
at the intersection is calculated by the system with the formula
ing the waiting time, number of vehicles, and priority
for each road, the Dynamic Queue Management Unit Distance
updates the traffic queues (Recent Choices and Waiting Arrival time =
Speed
 5

The calculated arrival time is used by the system to adjust Data Transmission and Processing:
the traffic signal at the relevant intersection. The traffic signal Central Server: The Arduino interfaces with the application
is set to turn green for the EV just in time for its arrival, to transmit distance measurements and LED status via serial
minimizing wait time and ensuring a smooth passage. The communication or wireless modules. The central server
system switches back to normal signal operation once the EV collects and processes this data in real-time from all parking
has safely passed the intersection. spots, updating a cloud-hosted database to maintain current
parking availability, ensuring scalability and reliability.
IV. P ROPOSED W ORK : S MART PARKING
To address the inefficiencies in parking management and Mobile Application Development
provide a seamless user experience, we propose an IoT-based 1) Login and Registration:
smart parking system integrated with a mobile application. a) Includes secure login and registration functionality,
The Algorithm 2 details the step-by-step hardware based allowing users to create and manage their accounts.
distance measurement and LED control. This system combines This ensures that personal data and preferences are
hardware, software, and network components to offer real-time stored securely and accessed only by authorized
monitoring and data-driven insights into parking availability. users . Home page as depicted in Fig 5 shows a
Hardware Components: logged in user page
• Ultrasonic Sensors (HC-SR04): These sensors are in-
stalled in each parking spot to detect the presence or
absence of a vehicle. The HC-SR04 sensor works by
emitting ultrasonic waves and measuring the time it takes
for the waves to bounce back after hitting an object,
allowing for accurate distance measurement.
• Arduino Microcontroller: The Arduino microcontroller
acts as the central processing unit for the sensors. It
receives signals from the HC-SR04 sensors, processes the
data to determine the occupancy status of each parking
spot, and controls the corresponding LED indicators.
• LED Indicators: Each parking spot is equipped with an
LED indicator. The LED turns on or off based on the
occupancy status, providing immediate visual feedback
to drivers—green for vacant and red for occupied.
• Communication Module (Wi-Fi/Bluetooth): The Ar-
duino is connected to a communication module, such as
the ESP8266 (for Wi-Fi) or HC-05 (for Bluetooth). This
module enables the Arduino to transmit the processed
data to a central server.
Fig. 5. Park Ease Home Page

Fig. 4. Smart Parking Proposed Work Fig. 6. Live Map Showing Available Parking Spaces
 6

2) User Dashboard: 6) Notifications and Alerts:

a) Add and Manage Vehicles: Users can register and a) Availability Updates: Users receive notifications
manage a list of their cars and bikes, linking them about parking spot availability and reservation con-
to their profiles for easy tracking and management firmations.
as depicted in Fig 5. b) Expiration Reminders: Alerts for expiring
reservations or upcoming payments.
b) View Previous Parking Slots: Users can access a
history of their previous parking slots, including 7) Navigation Assistance:
details such as location and time of use as shown a) Directions: Provides turn-by-turn navigation to the
in Fig 7. reserved parking spot, ensuring users can easily
find their way as depicted in Fig 6.
Algorithm 2 Distance Measurement and LED Control
Input: TRIG pin, ECHO pin, LED pin
Output: LED state
1: Initialize Pins:
2: SetPin(TRIG, OUTPUT)
3: SetPin(ECHO, INPUT)
4: SetPin(LED, OUTPUT)
5: Trigger Sensor:
6: SendPulse(TRIG, 10 microseconds)
7: Measure Distance:
8: Duration ← MeasureTime(ECHO)
9: Distance ← CalculateDistance(Duration)
10: Control LED:
11: if measured distance ≤ 70 cm then
12: Turn on LED
13: LogEvent(LED turned on due to proximity)
14: else
15: Turn off LED
16: LogEvent(LED turned off due to distance)
17: end if
Fig. 7. Real-Time Parking Information and Reservation Page

3) Real-Time Parking Information:

a) Spot Availability: Provides a real-time map
showing the status of parking spots—available,
occupied, or reserved—helping users find parking
quickly.

4) Transaction History:
a) Track Transactions: Users can view a history of
their parking transactions, including payments,
reservations, and any associated fees.

5) Reservation and Payment:

a) Reserve Spots: Allows users to book parking spots
in advance and make payments directly through the
app.
b) Payment Processing: Supports various payment
methods, including credit/debit cards and digital Fig. 8. Park Ease Payment Page
wallets, for a seamless transaction experience.
Fig 8 shows a sample payment page for the
reservation of a parking slot
V. E XPERIMENTS AND R ESULTS
Smart Traffic :
In this experiment using Pygame, three signal control strate-
gies were compared: Static Signal Control, Dynamic Signal
 7

Control (without queue consideration), and Dynamic Signal

Switching with Queue Consideration. The results demonstrate
that the Dynamic Signal Switching with Queue Consideration
consistently outperforms the other two methods, as seen in Fig.
4. On average, this algorithm clears 20-30% more vehicles
per cycle compared to the Static Control, which lacks the
flexibility to adapt to changing traffic conditions. The Static
Control approach is the least efficient, while the Dynamic
Signal Control (No Queue) sits in between, failing to fully
capitalize on traffic management by ignoring queue size. The
graph Fig.5. shows the Total Waiting Time, where the dynamic
queue-based algorithm maintains waiting times comparable to
the static and non-queue dynamic approaches. Although the
algorithm clears significantly more vehicles, it manages traffic
efficiently without substantially increasing the total waiting Fig. 9. Comparison of Vehicle Clearance Efficiency Across Signal
Control Strategies
time. There are occasional spikes, but overall, the method per-
forms effectively, balancing the flow of vehicles and minimiz-
ing congestion. Table 1 compares fuel consumption (in liters)
between a static algorithm and a dynamic signal switching
algorithm over 12 simulation cycles. Average fuel consump-
tion data for different classes of vehicles in Indian cities is
sourced from reference[2]. The percentage improvement of the
dynamic algorithm over the static algorithm is approximately
35%.This shows that Dynamic Signal Switching with Queue
Consideration is clearly the most effective approach for traffic
management.
Smart Parking :
TABLE I
F UEL C ONSUMPTION C OMPARISON FOR D IFFERENT A LGORITHMS

Cycle Static Algo (ltrs) Proposed Algo (ltrs)

1 20.348581 13.226578 Fig. 10. Comparison of Waiting Time Across Different Signal Control
2 29.387478 19.101861 Strategies
3 30.564661 19.867029
4 17.004261 11.052769
5 34.453041 22.394477
6 26.520631 17.375211
3) Real-time Feedback:
7 14.676345 9.539624
8 15.318711 9.957162 The system responded accurately to changes in the
9 19.001000 12.208976
10 19.174005 12.762924 obstacle’s position. Adjusting the distance of the
11 14.450680 9.855845 simulated car in the Wokwi simulator led to the
12 15.012696 9.281160 expected changes in the LED status, demonstrating
the real-time capability of the system.

1) Distance Measurement Accuracy:

The HC-SR04 sensor accurately measured the dis-

tance to the obstacle (simulated as a car). The
measured distance was consistent with the object’s
position in the simulation.

2) LED Functionality:

Obstacle ¡ 30 cm: The LED turned on, providing

a visual indication that the object was within the
detection range.
Obstacle ¿= 30 cm: The LED turned off, confirm-
ing that the object was outside the detection range. Fig. 11. Fig: LED Turned ON when Obstacle ¡ 30 cm
 8

VI. C ONCLUSION [11] Khan, Aftab Kulkarni, Parag Shergold, Ian Jones, Michael Dogramadzi,
Marko Carnelli, Pietro Sooriyabandara, Mahesh. (2020). Reducing Park-
ing Space Search Time and Environmental Impacts: A Technology Driven
The smart traffic and parking system in this project aim to Smart Parking Case Study. IEEE Technology and Society Magazine. 39.
make everyday urban mobility smoother and more efficient. 62 - 75. 10.1109/MTS.2020.3012329.
Using YOLO v8 for precise vehicle detection and a dynamic [12] L. F. P. de Oliveira, L. T. Manera and P. D. G. D. Luz, ”Development
of a Smart Traffic Light Control System With Real-Time Monitoring,” in
signal switching algorithm, the traffic system is able to clear IEEE Internet of Things Journal, vol. 8, no. 5, pp. 3384-3393
20-30% more vehicles per cycle compared to static signals, [13] S. Deshumkh, M. Gayakwad, N. S. More, R. Jadhav, K. Kadam and
while also managing waiting times and improving the overall H. Magar, ”Smart Traffic Management System Using RFID System,”
2024 MIT Art, Design and Technology School of Computing Interna-
flow. At the same time, the smart parking solution, which uses tional Conference (MITADTSoCiCon), Pune, India, 2024, pp. 1-4, doi:
IoT tools like ultrasonic sensors and Arduino microcontrollers, 10.1109/MITADTSoCiCon60330.2024.10575670.
helps drivers easily find and reserve parking spots in real-time [14] A. Hazarika, N. Choudhury, M. M. Nasralla, S. B. A. Khattak and
I. U. Rehman, ”Edge ML Technique for Smart Traffic Management in
through a user-friendly mobile app. Together, these systems Intelligent Transportation Systems,” in IEEE Access, vol. 12, pp. 25443-
may help in reducing congestion and make both traffic and 25458, 2024, doi: 10.1109/ACCESS.2024.3365930.
parking more efficient. Looking ahead, linking communication [15] Nijim, M., Kanumuri, V., Alaqqad, W., Albataineh, H. (2023). Advanced
Traffic Management System for Smart Cities. In: Daimi, K., Al Sadoon,
between traffic junctions could improve coordination across A. Proceedings of the 2023 International Conference on Advances in
intersections and may lead to even smoother traffic flow. Giv- Computing Research (ACR’23). ACR 2023. Lecture Notes in Networks
ing priority to emergency vehicles through intelligent signal and Systems, vol 700. Springer, Cham.
[16] Zhang, L.; Huang, J.; Li, X.; Xiong, L. Vision-based Parking-slot
preemption can help them pass through traffic faster during Detection: A DCNN-based Approach and A Large-scale Benchmark
critical situations. Incorporating vehicle-to-infrastructure com- Dataset. IEEE Trans. Image Process. 2018, 27, 5350–5364
munication will further boost the system’s efficiency, making [17] Loeffler, A.; Ronczka, J.; Fechner, T. Parking lot measurement with
24 GHz short range automotive radar. In Proceedings of the 16th
urban transport smarter and more responsive. Bringing vehicle- International Radar Symposium (IRS), Dresden, Germany, 24–26 June
to-infrastructure communication into the system will make it 2015; pp. 137–142.
work even better, helping urban transport become smarter and [18] Lee, S.; Hyeon, D.; Park, G.; Baek, I.-J.; Kim, S.-W.; Seo, S.-W.
Directional-DBSCAN: Parking-slot detection using a clustering method
more responsive to real-time traffic conditions. in around-view monitoring system. In Proceedings of the 2016 IEEE
Intelligent Vehicle Symposium (IV), Gotenburg, Sweden, 19–22 June
2016; pp. 349–354.
[19] Loeffler, A.; Ronczka, J.; Fechner, T. Parking lot measurement with
R EFERENCES 24 GHz short range automotive radar. In Proceedings of the 16th
International Radar Symposium (IRS), Dresden, Germany, 24–26 June
[1] R.B.Jha, A.Rai and R.Kala, ”Predictive Risk Analysis using Deep Learn- 2015; pp. 137–142.
ing in Indian Traffic,” 2021 IEEE International Intelligent Transportation [20] Zhang, L.; Li, X.; Huang, J.; Shen, Y.;Wang, D. Vision-Based Parking-
Systems Conference (ITSC), Indianapolis, IN, USA, 2021, pp. 258-263, Slot Detection: A Benchmark and a Learning-Based Approach Symmetry
10.1109/ITSC48978.2021.9564406. 2018, 10, 64.
[2] G. Liu et al., ”Smart Traffic Monitoring System Using Computer Vi- [21] ”Evaluation of Idling Fuel Consumption of Vehicles Across Different
sion and Edge Computing,” in IEEE Transactions on Intelligent Trans- Cities”, Kumar, PSharma, Niraj, Chalumuri, Ravi Sekhar, July 2015,
portation Systems, vol. 23, no. 8, pp. 12027-12038, Aug. 2022, doi: Conference: Recent Advznces in Treffic Engg,(RATE 2015)At: SVNIT ,
10.1109/TITS.2021.3109481. Surat, Gujrat(india).
[3] M.Amer and R.Maher, ”PV/IoT based Smart Traffic Management Sys-
tem,” 2022 International Conference on Electrical, Computer, Communi-
cations and Mechatronics Engineering (ICECCME), Maldives, Maldives,
2022, pp. 1-6, 10.1109/ICECCME55909.2022.9988110.
[4] B.N.Rao et al., ”Smart Traffic Management System using IoT,” 2022
IEEE International Symposium on Smart Electronic Systems (iSES),
Warangal, India, 2022, pp. 627-630, 10.1109/iSES54909.2022.00138.
[5] V.Bali, S.Mathur, V.Sharma and D.Gaur, ”Smart Traffic Manage-
ment System using IoT Enabled Technology,” 2020 2nd Interna-
tional Conference on Advances in Computing, Communication Control
and Networking (ICACCCN), Greater Noida, India, 2020, pp. 565-
568,10.1109/ICACCCN51052.2020.9362753
[6] A. R, J. U, Y. H, A. A and M. J, ”Design and Development of a Real-
Time Traffic Congestion Analysis System,” 2022 International Conference
on Computer Science and Software Engineering (CSASE), Duhok, Iraq,
2022, pp. 230-235,10.1109/CSASE51777.2022.9759708.
[7] El-Seoud, Samir El-Sofany, Hosam Taj-Eddin, Islam. (2016). Towards
the Development of Smart Parking System using Mobile and Web
Technologies. 10.1109/IMCTL.2016.7753762.
[8] A. Gupta, P. Rastogi and S. Jain, ”Smart Parking System using Cloud
based Computation and Raspberry Pi,” 2018 2nd International Conference
on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC)I-
SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC), 2018
2nd International Conference on, Palladam, India, 2018, pp. 94-99
[9] K. Hassoune, W. Dachry, F. Moutaouakkil and H. Medromi, ”Smart
parking systems: A survey,” 2016 11th International Conference on
Intelligent Systems: Theories and Applications (SITA), Mohammedia,
Morocco, 2016, pp. 1-6
[10] Li, W., Cao, L., Yan, L., Li, C., Feng, X. and Zhao, P., 2020. Vacant
parking slot detection in the around view image based on deep learning.
Sensors, 20(7), p.2138.