Situation analysis for driver assistance systems at urban intersections

Intersections are one major accident location type. This paper presents a situation analysis that combines a driver intent prediction and probabilistic reachable sets to determine the criticality of a scene. The intent prediction algorithm is presented and its evaluation in a study in real world driving. It is explained how the reachable sets are calculated and the collision probability is defined. These approaches use detailed digital maps and a graph based environmental model to perceive these complex situations.