BCMS

Buses often experience overcrowding and bunching along the route, leading to reduced seating chance, denied boarding, and long waiting and journey times. Currently no product exists for operators to control crowding along the route.

To address this, the project aims to commercialise Theoremus’s AI passenger crowding counting technology, which is currently a primitive and specialist solution, in use in GoogleMaps in Sofia. The project moves the image processing task from the central server to a computer on each bus, which removes the need of image data transmission, thereby reducing IT network requirements and increasing potential users. The crowding level of each bus (along with potential intervention measures to be suggested by UCL models) is shown on a map-based dashboard, and aids bus operators to effectively manage bus crowding along the route in real-time. The system will be tested in fleets of 20+ buses in Sofia and Bucharest in collaboration with the operators.