Towards Advanced Information Fusion for Driver Assistant Systems of Modern Vehicles

The ever increasing number of external and internal data available in modern cars can only become beneficial if the data are assessed correctly and presented according to the drivers' needs. This paper presents a well-designed method based on advanced information fusion that assembles and processes data centralized, which has been developed within the framework of the European funded research project I-WAY. A key task of this method is the reduction of the input data to a manageable amount that supports the unobtrusive delivery of useful and safety critical information to drivers. The solution proposed introduces the transformation of mainly cyclic arriving data from the time to the event domain. The main advantage is that the communication load is reduced to the minimum, as only relevant changes-events-must be processed in subsequent modules. The pre-selection of events thereby serves as data filter and data fusion, allowing for fast decision making. The use case at the end of the paper presents preliminary results on the application of this data fusion method gathered in a vehicle from video, radar, and weather information.