Mobility Episode Discovery in the Mobile Networks Based on Enhanced Switching Kalman Filter

The usage of mobile phones has become an important activity in our lives. The passive mobile positioning of mobiles provides large-scale data about human mobility. Hence, in this paper, we are presenting a technique based on continuous time switching Kalman filter to efficiently detect stop and move episodes. The technique has practical and theoretical advantages as the model is more closely related to measure human mobility characteristics and less sensitive to variations in radio network operations. The technique was tested on real radio network data and the results indicated significant improvement with respect to the model performance and the literature.