Improved Information Processing for Cooperative Vehicle Safety Applications

Abstract

Current Vehicle-to-Vehicle and Vehicle-to-Roadside communication based on IEEE 802.11 cannot guarantee a reliable dissemination of situational information which is required for Situation-Aware Driver Assistance Systems. By using a local Knowledge Base which manages the situational information, the drawbacks of the absence of congruous situational information for the hazard detection can be mitigated. Our work identifies four requirements which have to be fulfilled by this Knowledge Base. The objective of this paper is to show the potentials and benefits of using Bayesian Networks for the processing of uncertain and missing information. Bayesian Networks provide an optimal solution for the management of situational information in local Knowledge Bases and thus we suggest to use Bayesian Networks for the early and reliable hazard detection in Situation-Aware Driver Assistance Systems.