Modern mobile devices with embedded sensors and GPS receivers allow us to collect all sorts of data, that we can use to gain information about users. The goal of this thesis is to develop a mobile app service for detecting user transportation mode. Our set of selected transportation modes consists of city bus, tram, and train. We used accelerometer, gyroscope, and GPS receiver for data collection. With the use of collected data, geographical railway data and different classification methods we developed a prediction model. Random forest performed best with 97 % accuracy.
|
