Abstract. Propagation models are very important for the development and deployment of wireless communication networks. They are able to predict the path loss for different propagation conditions, but cannot include all propagation phenomena in detail. This fact leads to variations between predicted and measured field strengths. These variations can be reduced by calibrating some parameters of the propagation models with the help of exact measurement data. However, two problems occur when applying measurement data. On the one hand, the maps used for the prediction have only a limited resolution. On the other hand, the GPS data are erroneous due to the limited GPS accuracy and due to sampling errors. These errors can lead to variations up to 200 m between the measured positions and the possible positions on the road network. Therefore, a map-matching algorithm has to be applied which projects the wrong GPS positions automatically onto the street vectors used for the predictions. Thus, a good basis of data for calibration can be created.
Analysis of the impact of map-matching on the accuracy of propagation models