Pedestrian Interaction Algorithm for Autonomous Vehicles Considering Crossing Intention and Driving Style

With the rise of automated driving, the interaction between autonomous vehicles and pedestrians is a difficult but vital problem to be tackled. A Pedestrian- Vehicle interaction model is proposed which takes pedestrian intention to cross the crossroad into account and can simulate different human driving style. Previous models mainly consider pedestrian detection results, meaning that the autonomous vehicle will take braking or turning actions to avoid collision as soon as it detects a pedestrian ahead, while human drivers tend to brake only when pedestrian shows intention to cross the crossroad. The proposed algorithm can improve the efficiency of autonomous vehicles passing intersection while taking safety into account, which is validated both in simulation and pedestrian-vehicle-in-the-loop experiments via virtual reality technology.