Integration of Ontological Scene Representation and Logic-Based Reasoning for Context-Aware Driver Assistance Systems

Abstract

Co-operative driver assistance systems share information about their surrounding with each other, thus enhancing their knowledge and their performance. For successful information exchange and interpretation, a common domain understanding is needed. This paper first presents an ontology-based context-model for driving scene description, including next to spatio-temporal components also additional context information like traffic signs, state of the driver and the own-vehicle. For traffic rules, we integrate the ontological scene description with a logic programming environment, to enable complex and powerful reasoning on the given information. The proposed ontology is discussed with respect to a set of validation criteria. For integration with logic programming a prototypical development of an overtaking assistant is shown to demonstrate the feasibility of the approach.