The Subway Navigator application was developed for a Human Computer Interaction class. It is meant to demonstrate the use of audio recognition as a method of stop detection in public transport vehicles, especially subways, as traditional navigation solutions do not work underground.
The application uses audio fingerprinting for stop detection. By constantly monitoring the filtered audio signal specific sounds such as the beeping sound the doors make when closing are detected. This is used by a probabilistic model to infer the current motion state of the train. The necessary sound databases are implemented for the Glasgow subway, certain ScotRail services and the London underground.