Smart Urban Traffic Management Using IoT, YOLO-Based Vehicle Detection, and Real-Time User Guidance

One of Egypt's severe challenges is traffic congestion, which usually happens when the number of vehicles exceeds the capacity, leading to an increase in the idling of these vehicles on the road, causing longer trip times, high fuel consumption, and air pollution in the crowded areas. To overcome this issue, we created an IoT-based traffic management system. The system consists of an ESP32-Wrover module combined with OV5640 cameras, LED strips, an AI model, and an application. After the selection of our solution, we started to construct it. A maquette of the road network was created using a wooden sheet; the wooden sheet was then covered by the road network design as a sticker. We connected the camera and LED to the ESP in order to construct our hardware base. The AI model was created, including the CV2 (OpenCV) library for object vision and the YOLO8 model for car count and better image processing. The application was created, and then the OpenStreetMap API was added to it in order to enable navigation. Our app user interface is well organized and user-friendly. The system detects traffic and guides users via the application and LED strips. In case of incidents, users can report to authorities using our app. The cost of our prototype is low and affordable (only 1115 L.E.), which is a powerful point of our project. After testing, we found that our project has achieved the design requirements, as its accuracy reached 93.75%, its time response is 0.84 sec, and it decreases the trip time by 20 min. Fuel consumption is also reduced from 0.3 liter to 0.1 liter. To sum up, our smart traffic management system meets the objectives of the challenge, ensuring safety and minimizing trip time, fuel consumption, and air pollution with the highest possible accuracy.