International Journal on Science and Technology (IJSAT)

E-ISSN: 2229-7677 ● Website: [Link] ● Email: editor@[Link]

Smart Traffic Management System

Abhishek N M1, Nikhil2, Sunil B P3
1,2
Student, Department of CSE, Government Polytechnic Mirle College, Karnataka, India
3
Guide and HOD of Department of computer science and engineering collage of government
polytechnic, Mirle

Abstract
Traffic congestion is a growing concern in urban areas, leading to increased travel time, fuel consumption,
and environmental pollution. Traditional traffic light systems operate on fixed timers, which often fail to
adapt to real-time traffic conditions, causing unnecessary delays. This project aims to develop a smart
traffic management system that dynamically adjusts traffic signals based on vehicle density. The system
utilizes a USB webcam to capture live road images, which are processed using YOLOv5 object detection
to count the number of vehicles on each road. An Arduino UNO controls the traffic lights based on the
detected vehicle count, ensuring efficient traffic flow. IR proximity sensors are integrated as a secondary
detection mechanism to improve accuracy, especially in low-light conditions. The system also includes an
OLED display to show real-time traffic signal status and countdown timers for improved user awareness.
By implementing this intelligent traffic control mechanism, the system reduces waiting time, prevents
unnecessary stoppages, and improves overall road efficiency. The use of machine learning for vehicle
detection and a microcontroller-based traffic light system makes it a cost-effective and scalable solution
for modern cities. This project contributes to the advancement of smart city infrastructure, promoting
better traffic management and reducing congestion-related issues.

Introduction
Traffic congestion is one of the biggest challenges in urban transportation, leading to delays, increased
fuel consumption, and environmental pollution. Conventional traffic light systems operate on predefined
timers, which do not consider real-time traffic conditions, often causing inefficient traffic flow. To address
this issue, smart traffic management systems are being developed to dynamically adjust traffic signals
based on real-time vehicle detection.
This project introduces a smart traffic light system that uses a USB webcam, YOLOv5 object detection,
and an Arduino UNO to control traffic signals efficiently. The USB webcam captures live traffic images,
which are processed using YOLOv5, a deep learning-based vehicle detection model, to count the number
of vehicles on each road. The Arduino UNO then adjusts the traffic light durations accordingly to optimize
traffic movement. Additionally, IR proximity sensors serve as backup detection units to enhance accuracy,
especially in low-visibility conditions. An OLED display is used to show real-time traffic signal status and
countdown timers, improving driver awareness.
The system is designed to reduce waiting times, improve traffic flow, and minimize congestion, making it
an efficient solution for modern cities. By integrating computer vision, microcontroller-based automation,
and real-time traffic analysis, this project contributes to the development of intelligent transportation
systems that can be easily implemented in urban environments.

IJSAT25012979 Volume 16, Issue 1, January-March 2025 1

 International Journal on Science and Technology (IJSAT)
E-ISSN: 2229-7677 ● Website: [Link] ● Email: editor@[Link]

Literature Review
1. Smart Traffic Management System
Authors: Bhuvan S T, Manjunath H.R, Abhiman H.R, Ranjan Kumar, Sachin G Rao
Summary: This paper highlights IoT integration in traffic management to alleviate congestion. It proposes
a hybrid model combining centralized and decentralized traffic control strategies, utilizing AI-based
predictive techniques for improved efficiency.
2. Vehicle Detection Algorithms (YOLO and Others)
Authors: Md. Jahin Alam et al.
Summary: The paper explores modern vehicle detection using machine learning models such as YOLO.
The YOLOv5 Nano version provides real-time, efficient object detection, making it ideal for embedded
systems like ESP32-CAM.
[Link] Traffic Control Using AI
Authors: Paul Shruti Kanaila et al.
Summary: This research discusses video-based AI traffic analysis, improving congestion management
through real-time vehicle categorization and signal optimization.

Methodology
The smart traffic management system dynamically adjusts traffic signals based on real-time vehicle
detection. The working process includes the following steps:
Image Capture & Vehicle Detection
• A USB webcam captures live traffic images.
• The images are processed using YOLOv5, which detects and counts vehicles.
Traffic Signal Control
• The vehicle count data is sent to the Arduino UNO, which adjusts signal durations.
• The road with higher traffic density gets a longer green light.
Backup Detection using IR Sensors
• IR sensors detect vehicles as a backup in case of low visibility.
Real-Time Display Updates
• An OLED display shows traffic signal status and countdown timers.
Continuous Monitoring & Safety
• The system continuously updates signals based on traffic conditions.
• Proper wiring and power management ensure stable and safe operation.
This approach improves traffic flow, reduces waiting times, and enhances urban mobility efficiently.

Existing System
Conventional traffic management systems operate on fixed-time signal cycles, which do not adjust based
on actual traffic conditions. These systems follow a predefined schedule, often causing unnecessary
delays and increased congestion. Some traffic control methods rely on Inductive Loop Sensors embedded
in roads to detect vehicles, but these solutions are expensive to install and maintain. In certain areas,
manual traffic control by police officers is still used, but this approach is inefficient and subject to human
error. Overall, the current system lacks real-time adaptability, AI-driven decision-making, and
efficient traffic flow management, making it ineffective in handling unpredictable traffic variations.

IJSAT25012979 Volume 16, Issue 1, January-March 2025 2

 International Journal on Science and Technology (IJSAT)
E-ISSN: 2229-7677 ● Website: [Link] ● Email: editor@[Link]

Fig: 1 Existing System

Disadvantages of the Existing System

• Predefined Signal Cycles: Traffic lights operate on fixed timing, causing inefficiencies when some
lanes experience heavy congestion while others remain empty.
• Limited Vehicle Sensing: IR sensors only detect vehicles near the stop line, making them unsuitable
for monitoring traffic flow over long distances.
• Lack of AI-Based Processing: The system does not utilize real-time image recognition, preventing it
from distinguishing vehicle types or accurately assessing congestion.
• Poor Scalability: Expanding the system to accommodate new roads or modifying traffic rules requires
manual adjustments, making it less adaptable.
• No Intelligent Prioritization: While emergency vehicle detection can be integrated, the system lacks
AI-driven recognition to automatically prioritize essential vehicles in real time.

Proposed System
The proposed AI-Based Traffic Light Control System utilizes image-based vehicle detection to optimize
traffic signal timing dynamically. A laptop camera captures live traffic footage, and an AI model
processes the images to detect and count vehicles in real time. Based on the vehicle density, an Arduino
UNO adjusts the green light duration for each road while maintaining synchronization with the red light
on the opposite [Link] system alternates signal control between the two roads, ensuring fair traffic
distribution. A countdown timer displays the remaining signal time, providing clear information to
drivers. By continuously monitoring traffic conditions, this system enhances road efficiency, reduces
congestion, and improves overall traffic management.

Advantages
• Real-Time Traffic Optimization: The system uses a laptop camera with AI-based vehicle
detection to dynamically adjust signal timing based on traffic density, reducing unnecessary wait
times.
• Efficient Traffic Flow: By prioritizing the road with higher vehicle density, the system ensures
smooth and balanced traffic movement.
• Accurate Vehicle Detection: Image-based processing provides precise vehicle counting, improving
traffic signal efficiency.

IJSAT25012979 Volume 16, Issue 1, January-March 2025 3

 International Journal on Science and Technology (IJSAT)
E-ISSN: 2229-7677 ● Website: [Link] ● Email: editor@[Link]

• Energy Efficient & Cost-Effective: Unlike fixed-timing systems, dynamic signal control optimizes
energy consumption and reduces idle time at signals.
• Scalability & Flexibility: The system can be expanded to multiple intersections and adapted to
different traffic patterns without major reconfiguration.
• Improved Road Safety: AI-driven monitoring helps reduce congestion and potential accidents
caused by inefficient signal timing.

FLOW CHART
The AI-Based Traffic Light Control System begins by capturing road images using a laptop camera,
which are processed by an AI model to detect and count vehicles. The count is sent to Arduino UNO,
which checks if the vehicle number exceeds a set threshold. If traffic is heavy, the green signal duration
is extended; otherwise, the default timing is applied. Finally, the Arduino updates the traffic lights,
ensuring real-time, dynamic, and efficient traffic management.

Fig: 2 Flow chart

Use Case Diagram

The UML use case diagram represents the AI-Based Traffic Light Control System, where a laptop
camera captures real-time traffic images, and an AI model processes and detects vehicle count. The data
is sent to Arduino UNO, which analyzes traffic density to determine the appropriate signal timing. Finally,
the Arduino dynamically adjusts the traffic lights, displaying the countdown timer to optimize traffic
flow efficiently.

IJSAT25012979 Volume 16, Issue 1, January-March 2025 4

 International Journal on Science and Technology (IJSAT)
E-ISSN: 2229-7677 ● Website: [Link] ● Email: editor@[Link]

Fig:3 Use Case Diagram

Conclusion
The AI-Based Traffic Light Control System successfully enhances traffic management by utilizing real-
time image processing and AI-based vehicle detection. By dynamically adjusting signal timings based on
traffic density, the system reduces congestion, minimizes wait times, and improves road efficiency. The
integration of Arduino UNO and a laptop camera ensures a cost-effective and scalable solution adaptable
to various intersections. This project demonstrates the potential of AI-driven automation in smart city
infrastructure, leading to safer, smarter, and more efficient traffic control systems.

References
1. Asha, C.S.; Narasimha Dhan, A.V. Vehicle Counting for Traffic Management System using YOLO
and Correction filter. In Proceedings of the 2018 IEEE International Conference on Electronics,
Computing and Communication Technologies (CONECCT), Bangalore, India, 16–17 March 2018; p.
18149219.
2. Song.; Liang, H.; Li, H.; Dai, Z.; Yun, X. Vision-based vehicle detection and counting system using
deep learning in highway scenes. Eur. Transp. Res. Rev. 2019, 11, 51. [CrossRef]
3. A. Kumar, R. Singh, and P. Sharma, "Smart Traffic Management System Using IoT," CiteSeerX, 2023.
[Online].
4. I. A. Ashqar, A. M. Ghaith, and M. A. Sharadqah, "Design and Implementation of a Smart Traffic Light
Management System Controlled Wirelessly by Arduino," ResearchGate, May 2020. [Online].
5. A. Gupta and S. R. Joshi, "IoT-Based Smart Traffic Management System," International Journal of
Emerging Technologies and Innovative Research (JETIR), vol. 6, no. 6, pp. 62–65, Jun. 2019.
[Online].

IJSAT25012979 Volume 16, Issue 1, January-March 2025 5