The BMW SURF Project: A Contribution to the Research on Cognitive Vehicles

Within the last years, research in the field of cognitive vehicles has evolved to an important activity especially for car manufactures in the premium segment. Motivated by the robustness and flexibility of inter-human communication, highly integrated contextual awareness is being seen as a key component to provide an intuitive human-machine interface to the wide range of functions the driver is being confronted with in a modern driver's working place. In this work, we propose a highly flexible software architecture to develop and evaluate various services in the context of cognitive vehicle behaviour. The system has been implemented in an experimental car that is equipped with a wide range of sensors to acquire information from three sources: the car, the environment and the driver. Special efforts have been made to evaluate the performance and the potential of different driver monitoring approaches for the use in adaptive automotive applications, in both advanced driver assistance systems (ADAS) and in-vehicle information systems (IVIS). To evaluate the performance of our approach in detail, we have chosen an adaptive lane departure warning system. Different strategies for driver intention analysis are compared with regard to several criteria in a small evaluation study. The results clearly shows the importance and the potential of a complex contextual analysis for the emergence of future intelligent vehicle applications.