Distributed data fusion: application to confidence management in vehicular networks

This paper presents a method for the management of information dissemination in a vehicular network (VANET). Due to the particularities of the application (ad hoc network, dynamical nodes, broadcast messages), an algorithm has been developed to fuse and combine data in distributed systems. Matching spatial information is made easier by the use of a numerical map as support of a database. A model of confidence management based on the belief function framework is then described considering spatial dispersion of data sources, delays due to the multi-hop transmission and dependency between sources. Preliminary results are presented based on simulated messages referenced on a real map data.