Informationsfusion für die kooperative Umfeldwahrnehmung vernetzter Fahrzeuge

Cooperative vehicle decisions require a common, consistent perception. For this purpose, fusion methods for communicated self- and environmental information are developed and tested. The registration in common coordinates is carried out after localisation by merged GPS and dead reckoning data. Decentralized fusion combines recursive multi-object tracking and grid-based mapping. Thus negative sensor evidence can be described and inconsistencies can be checked for plausibility.