Implementation of a Smart IoT-Based Trotro Passenger Pickup System in Ghana

As the number of people living in cities increases, there is a need to improve the transportation system. In Ghana, the public transport systems in the urban areas are faced with longer waiting times, overcrowding at bus stops, and a lack of a communication system between bus stops and the drivers of the vehicles. Recent developments in the Internet of Things (IoT), Artificial Intelligence (AI), location management microservices, computer vision, cloud computing, and embedded systems offer novel opportunities to overcome these challenges within the public transportation system. In this research, a low-cost passenger counting and tracking system is developed and integrated into a Trotro public transport system in Ghana. The system is designed using a Raspberry Pi with a camera module attached to it and a portable power bank. The system uses a computer vision model (YOLOv11) trained to count passengers at a bus stop by recognizing human heads and bodies in their entirety. This dual detection method gives the system robustness against overcrowding. When only heads can be seen, it counts by the number of heads. When full bodies can be seen, it counts by the number of bodies. When both can be seen, the model ensures that the head and body bounding boxes are paired one-to-one to prevent overcounting. After counting, the data is sent to a web server where it is stored and processed for use in a mobile application. The system has an in-built Bluetooth-enabled speaker that announces the driver’s trip information to passengers at the bus stop. The research demonstrates how technologies can improve the transportation system in Ghana.