From raw sensor data to semantic trajectories

Abstract

GPS position are useful to analyse movements of mobile objects. Unfortunately, the outcome can be unsatisfactory due to imprecision and signal lost. Several sensors (generic or specific ones depending on the type of vehicle) are now included into mobile objects. This article describes a new generic model that enhances the semantic trajectory model and a process to extract semantic from GPS and other sensors. This process is based on the raw data from these sensors which help identify how, why and when mobile objects are moving, in order to add semantic information to the trajectories and then design new applications such as smart GPS, reporting systems or Remote Maintenance System. This process was successfully applied to the Indre et Loire fire department and its connected ambulances.