The assignment of vehicles detected by distance sensors to lanes relative to the own vehicle is an important and necessary task for future driver assistance systems like Adaptive Cruise Control (ACC). The collective motion of objects driving in front of the vehicle allows a prediction of the vehicle's own driving course. The method uses not only data of the host vehicle to determine its own trajectory but as well data from a distance sensor supplying distances and angles of objects ahead of the vehicle to determine the trajectories of these objects. Algorithms were developed using an off-line simulation, which was fed with recorded data obtained from a real ACC vehicle. The results show a significant improvement in the quality of the predicted driving course compared to other methods solely based on data of the host vehicle. Particularly in situations of changing curvature, e.g. the beginning of a bend, the algorithm helps to improve the overall system performance of ACC.