Real Time Traffic Congestion Management with AI & Arduino

Abstract

Traffic congestion is a serious issue in modern cities. It causes delays, higher fuel consumption, air pollution, and stress for commuters. The main problem is traditional traffic light systems. They are inefficient because they use fixed, preset timers that cannot adjust to changing traffic conditions. This results in unnecessary waits on empty roads while other lanes are overcrowded. This project proposes an intelligent and adaptive Real-Time Traffic Management System. It will combine AI (Artificial Intelligence) and micro-controllers like Arduino. The main goals are to design a real-time system that uses AI to adjust signal timings based on live traffic. This should reduce congestion and waiting time at intersections. It will also implement a low-cost, efficient signal control system using Arduino. The proposed solution combines AI and Arduino to manage a four-way traffic intersection. The system uses a camera to capture live video, which is processed by a Python script using a Computer Vision (CV) model like YOLO or Open-CV. This script detects and classifies vehicles in real-time, calculating traffic density in each lane. The density data informs a Decision-Making Algorithm, and the timing changes are sent via Serial Communication (USB/UART) to an Arduino Uno that controls the LED traffic lights. The Literature Survey supports this approach, citing studies that highlight the success of deep-learning models like Efficient-Det and YOLOv8 in real-time traffic monitoring. These models achieve high counting accuracy (around 96%) and effectively reduce congestion while prioritizing emergency vehicles. The Expected Outcomes are promising. The system aims to produce a functional prototype, optimize traffic flow through dynamic signal timing, and reduce congestion and waiting times. Additional benefits include better efficiency, faster vehicle flow, and decreased fuel use. It will also enhance safety by prioritizing emergency vehicles and lowering environmental impact through reduced idling and carbon emissions. This AI-based system will adapt in real time and manage congestion much more effectively than traditional fixed-time methods.