Towards Advanced Information Extraction in Cooperative Driving Support Systems

In this paper, a novel information extraction approach is presented as a part of a cooperative driving support system (I-WAY). The key features of the proposed solution are i.) a geo-reference events representation and merging with respect to geo-areas of interest and ii.) the efficient management of low-confident information through geo-referenced information fusion and the extraction of advanced high-confident information exploiting events interrelationships. The events interrelationships are represented as a Bayesian network constructed upon expert knowledge and statistical data. The new information produced is forwarded to the driver and to the vehicle’s telematic platform for communication to other vehicles.