AI BASED SMART TRAFFIC MANAGEMENT SYSTEM

Abstract

Traffic congestion has become a major issue in urban areas due to the fixed-time traffic signal system that cannot adopt to real-time traffic conditions. The proposed project, AI Powered Adaptive Traffic Management System, aims to solve this problem by using Artificial Intelligence and Computer Vision to dynamically control traffic signals. The main problem with traditional systems is long waiting times, fuel consumption, air pollution, and delays for emergency services. The proposed project, AI Based Smart Traffic Management System, aims to address these problems by implementing an intelligent system that can monitor and control traffic signals dynamically based on live traffic conditions. The core problem statement of this project is the lack of adaptability in current traffic control system. This project rectifies the problem by using computer vision and machine learning traffic in real-time. The system uses YOLOv8 object detection algorithm with OpenCV to detect and count different types vehicles such as cars, buses , trucks, and motorcycles from live CCTV camera feeds. After counting, the traffic density is classified into Low, Medium, and High categories. Based on this classification, an algorithm calculates the optimal green signal time for each lane, instead of using a fixed timer. The further improve efficiency, a separate CNN model is integrated to detect emergency vehicles like ambulances and fire trucks. When an emergency vehicle is detected, the system automatically provides priority by giving a green signal to that lane immediately. KEYWORDS:AI Traffic Management, YOLOv8, Computer Vision, Real-time Traffic Analysis, Deep Learning, Vehicle Detection, Smart City.