Car navigation systems are nowadays widely used in the everyday traffic. Their technology and visualization for better navigations are developed rapidly e.g. high accuracy GPS receivers, visualization technique in three dimensions from SiemensVDO or Volkswagen and Google. The navigation systems should also be made more easy for machines to notice, understand and record some essential traffic situations in everyday use like 'experienced' drivers do in their routine driving: recognizing these conflict situations and reacting almost unconsciously. The semantic informations extracted from traffic scenes are proposed to enhance the capacity of on-the-shelf navigation systems and made more suitable for autonomous vehicles. The additional classifications of essential traffic informations and estimation of the driving visibility are the basis for making decision to navigation tasks. Furthermore, the semantic structure can be also supporting to the behaviour-based networks and traffic scene interpretations. The aim of this approach is the information abstraction from traffic scenes and evaluation of 'driving experiences' working correctly together with informations from the navigation systems and how to organize them in database. Consequently, an autonomous vehicle should be shown or warned for danger sites or essential traffic events just in time, so that the self-planning of routes and traveling time with performance and safety is possible.
Semantic road maps for autonomous vehicles
Semantische Straßenkarten für autonom entscheidende Fahrzeuge
2007
7 Seiten, 6 Bilder, 13 Quellen
Aufsatz (Konferenz)
Englisch