Section: New Results

Urban autonomous driving: dealing with intersections

Participants : Guillaume Tréhard, Evangeline Pollard, Fawzi Nashashibi.

The goal of this project, made in collaboration with Valeo is to develop a complete solution for autonomous driving on open roads. More specifically, IMARA's objectives are to provide the way to safely cross any kind of intersections for an autonomous vehicle in a urban context. Among the different relevant scenarios, we can notice:

• Intersection with different shapes: Roundabout, T junctions , X junctions;

• Intersection with different rules: With specific rules (traffic lights, main road...) or unspecified rules (“priority to the right”);

• Different traffic: Busy or empty intersections;

• Deal with abnormal situations: road works, policemen, firemen,...

Possible steps for this work can be listed as follows:

• Model the intersection: define relevant information, find a generic model for every intersection;

• Detect the intersection (shape, drivable area, traffic flows);

• Understand the priorities that rules it;

• Locate the car in the intersection by crossing it;

• Plan a path to get out of the intersection.