ML-Based Traffic Detection and Autonomous Signal Operation

This research paper presents the development and real-time implementation of an AI-based traffic detection and autonomous signal operation system using the YOLOv5 object detection algorithm. The system aims to dynamically manage traffic flow based on vehicle density, thereby improving road efficiency and reducing congestion. Unlike prior approaches that relied on Raspberry Pi and external cameras, the proposed setup utilizes a standard laptop for processing live video input and an Arduino microcontroller to operate physical traffic signals via LEDs. The YOLOv5 model is employed for accurate and fast detection of vehicles such as cars, buses, trucks, and motorcycles using a webcam. Based on the detected vehicle count, green signal duration is adjusted dynamically in real-time. The Arduino board receives control commands (G, Y, R, and O) via serial communication to reflect the traffic signal states using LEDs. This low-cost, modular prototype offers a practical and scalable solution for smart traffic management in urban as well as rural environments. The paper also presents system architecture, hardware/software integration, and experimental results validating the effectiveness of the proposed solution