Road map assisted ground target tracking

Ground target surveillance is getting a more and more important task, for civil as for military applications. Ground targets may be bound to infrastructural constraints, like vehicles moving on roads. Therefore, incorporating road map information into the tracking process is a current topic of research. This paper presents a tracking algorithm, which considers road map information. The algorithm supports junctions as well as curves by applying a Gaussian sum algorithm which leads to a variable structure multiple model approach. The update is done using Kalman filtering. Beside the filtering, some aspects of the whole tracking framework will be presented and some simulation results are given.