Key points are not available for this paper at this time.
Current research on advanced driver-assistance systems (ADASs) addresses the concept of highly automated driving to further increase traffic safety and comfort. In such systems, different maneuvers can automatically be executed that are still under the control of the driver. To achieve this aim, the task of assessing a traffic situation and automatically taking maneuvering decisions becomes significantly important. Thus, this paper presents a system that can perceive the vehicle's environment, assess the traffic situation, and give recommendations about lane-change maneuvers to the driver. In particular, the algorithmic background for this system is described, including image processing for lane and vehicle detection, unscented Kalman filtering for estimation and tracking, and an approach that is based on Bayesian networks for taking maneuver decisions under uncertainty. Furthermore, the results of a first prototypical implementation using the concept vehicle Carai are presented and discussed.
Schubert et al. (Mon,) studied this question.