Detection and Tracking Algorithm of Front Vehicle Based on Laser Radar 2015-01-0307

Nowadays active collision avoidance has become a major focus of research, and a variety of detection and tracking methods of obstacles in front of host vehicle have been applied to it. In this paper, laser radars are chosen as sensors to obtain relevant information, after which an algorithm used to detect and track vehicles in front is provided. The algorithm determines radar's ROI (Region of Interest), then uses a laser radar to scan the 2D space so as to obtain the information of the position and the distance of the targets which could be determined as obstacles. The information obtained will be filtered and then be transformed into cartesian coordinates, after that the coordinate point will be clustered so that the profile of the targets can be determined. A threshold will be set to judge whether the targets are obstacles or not. Last Kalman filter will be used for target tracking. To verify the presented algorithm, related experiments have been designed and carried out. The results of the experiments show that the algorithm can identify and track vehicles critically and has comparatively high reliability.