2A2-H04 Study on Automatic Braking Control for Collision Avoidance Based on Pedestrian Motion Prediction(Car Robotics & ITS)

Abstract

This thesis focuses on collision avoidance automatic braking control system based on pedestrian motion prediction in order to enhance the effectiveness of automatic emergency braking. In the situation that a pedestrian suddenly change his moving direction, current active safety systems cannot avoid collisions in the critical situations due to physical limits in braking capability. Therefore, in order to develop high-performance advanced driver assistance systems, it is important to predict pedestrian motion. This research focuses on the situation that there is the parked vehicle in front of pedestrian who will avoid the parked vehicle and change his moving direction to the center of road. In order to predict pedestrian motion correctly, pedestrian motion model is formulated based on experimental result by using LAIDAR. In addition, an automatic braking control method is constructed based on pedestrian motion prediction.