SAiBTraffics

1. Work Introduction and Presentation: SAiBtrafficS is a youth-led Cameroonian tech startup committed to solving eco-sustainability and mobility challenges through cutting-edge technologies. Established in 2024, Our mission is to build clean, smart, and inclusive technologies tailored to the realities of emerging economies like Cameroon and Africa. This project presents a Smart AI-based Traffic Light Control System that integrates artificial intelligence, hardware prototyping, and simulation. The system uses an OKDo Jetson Nano as the main controller, combined with cameras to capture real-time traffic data. A YOLOv5 model, optimized with the MindSpore framework, is used to detect and classify vehicles (cars, trucks, and emergency vehicles). Based on the detected vehicle density in each lane, the system dynamically adjusts traffic signal timing, allocating longer green lights to congested lanes and shorter ones to less busy lanes. An emergency override feature ensures that when an ambulance or fire truck is detected, the system automatically grants it a green signal for immediate passage. The system is validated using the CARLA simulation environment before implementation in a physical prototype. Additionally, it is powered by a solar energy module, allowing it to function independently of the electrical grid, making it ideal for regions with unstable power supply Problem to Be Solved Traffic congestion has become a persistent and serious challenge in many urban areas of the world, particularly in cities such as Yaoundé, Buea, and Douala. In some cases, even ambulances are forced to stop at red lights, wasting valuable time and delaying patient treatment. This occurs because most traffic lights operate using fixed timers that do not account for actual road conditions, leading to unnecessary delays, fuel waste, and increased pollution. 2.2 Stakeholder and Market Analysis 3 Innovating Factor • Traditional traffic light systems operate using fixed timing schedules that do not adapt to real-time traffic conditions. However, traffic flow is dynamic and varies with time, location, and environmental factors. As a result, fixed-time systems often lead to inefficiencies, such as granting green signals to empty roads while heavily congested roads experience unnecessary delays. • In many cities in Cameroon, traffic congestion is a serious issue. Drivers waste time at intersections, fuel consumption increases, and pollution worsens. Existing solutions such as road-embedded sensors are expensive and difficult to install. Traffic officers are overwhelmed in complex traffic situations. A more practical, affordable, and intelligent solution is needed one that can observe and react to traffic conditions in real time. Real-time adaptive signal control using AI-based vehicle density measurement, replacing fixed-timer logic with a responsive decision system • Automatic emergency vehicle detection and priority override using YOLOv5 classification, ensuring ambulances and fire trucks are never delayed at intersections • Integrated standalone solar power module enabling grid-independent operation a critical feature for regions with unstable electricity supply • On-device inference on the Jetson Nano using MindSpore Lite optimisation, ensuring low latency and privacy-preserving local processing without cloud dependency • Combined physical prototype and CARLA simulation validation methodology, enabling rigorous testing before real-world deployment

Built With

• python-package-index