Entwicklung von Methoden zur fahrzeugübergreifenden Assoziation für die präventive Sicherheit (Diplomarbeit)
In cooperative perception systems, different vehicles share object data obtained by their local environment perception sensors, like radar or lidar, via wireless communication. To generate a complete and consistent representation of the vehicle's surrounding, objects, which are communicated by other vehicles are associated and fused with the locally perceived objects. However, if the vehicle self localizations of the communication partners are deficient, data association between communicated and local perception is a challenging task. The scope of this work is to use point matching algorithms to estimate the relative localization error between the communication partners by matching the communicated with the local object data in order to improve data association. Therefore, multiple point matching algorithms have been implemented and evaluated in a simulation environment and with real test vehicles.