[Back]

T. Thalmann, A. Abdalla:
"Assessing the Influence of Preprocessing Methods on Raw GPS-Data for Automated Change Point Detection";
Talk: International Conference on Geographic Information Science (AGILE 2014), Castellón; 06-03-2014 - 06-06-2014; in: "Connecting a Digital Europe Through Location and Place", J. Huerta, S. Schade, C. Granell (ed.); Springer, (2014), ISBN: 978-90-816960-4-3; 123 - 139.

The automated recognition of transportmodes from GPS data is a problem that has received a lot of attention from academia and industry. There is a comprehensive body of literature discussing algorithms and methods to find the right segments using mainly velocity-, acceleration- and accuracy-values. Less work is dedicated to the derivation of those variables. The goal of this chapter is to identify the most efficient way to preprocess GPS trajectory data for automated change-point (i.e., the points indicating a change in transportation mode) detection. Therefore the influence of different kernel based smoothing methods as well as an alternative velocity derivation method on the overall segmentation process is analyzed and assessed.

The automated recognition of transportmodes from GPS data is a problem that has received a lot of attention from academia and industry. There is a comprehensive body of literature discussing algorithms and methods to find the right segments using mainly velocity-, acceleration- and accuracy-values. Less work is dedicated to the derivation of those variables. The goal of this chapter is to identify the most efficient way to preprocess GPS trajectory data for automated change-point (i.e., the points indicating a change in transportation mode) detection. Therefore the influence of different kernel based smoothing methods as well as an alternative velocity derivation method on the overall segmentation process is analyzed and assessed.

GPS, Change Point Detection, Mode of Transport