Exploring a New Model for Mobile Positioning Based on CDR Data of The Cellular Networks

Abstract

The emerging technologies related to mobile data especially CDR data has great potential for mobility and transportation applications. However, it presents some challenges due to its spatio-temporal characteristics and sparseness. Therefore, in this article, we introduced a new model to refine the positioning accuracy of mobile devices using only CDR data and coverage areas locations. The adopted method has three steps: first, we discovered which model of movement (Move, Stay) is associated with the coverage areas where the mobile device was connected using a Kalman filter. Then, simultaneously we estimated the location or the position of the device. Finally, we applied map-matching to bring the positioning to the right road segment. The results are very encouraging; nevertheless, there is some enhancement that can be done at the level of movement models and map matching. For example by introducing more sophisticated movement model based on data-driven modeling and a map matching that uses the movement model type detected by matching "Stay" location to buildings and "Move" model to roads.