Prediction of stopping distance for autonomous emergency braking using stereo camera pedestrian detection

Abstract

Autonomous vehicle (AV) attracts a lot of research interest because of their ability to reduce road fatalities and save people's lives. Pedestrian detection and collision avoidance are the crucial parts of AVs. The existing pedestrian detection systems don't guarantee better accuracy under complex scenarios, poor light conditions, and low complexity overhead. Moreover, these systems rely on monocular camera object detection. Hence, this paper aims to propose a stereo-vision based pedestrian detection and collision avoidance system for AVs. It uses two cameras fixed at a specific distance apart to scan the environment. Once a pedestrian is detected, the system calculates the distance. The Automatic Emergency Braking System (AEBS) controller algorithm will activate the AEB if the estimated distance is less than 3.3 m, considered safe. MATLAB is used for the implementation, and the experimental results reveal that the proposed method is promising in terms of prediction accuracy and minimizing fatalities.