Highly automated vehicles require methods to represent unstructured environment features with high accuracy and low resource demands. In this thesis, a simplified map-based environment representation is applied to describe occupancy states and common object motion behavior. A concept to combine map-based representations with different contents and levels of detail is provided. The improvements to state-of-the-art approaches are illustrated by experiments with real sensor data and new evaluation techniques, which allow to measure map qualities.
«
Highly automated vehicles require methods to represent unstructured environment features with high accuracy and low resource demands. In this thesis, a simplified map-based environment representation is applied to describe occupancy states and common object motion behavior. A concept to combine map-based representations with different contents and levels of detail is provided. The improvements to state-of-the-art approaches are illustrated by experiments with real sensor data and new evaluation...
»