Y. Bai

Pages: 3-14

Abstract
Aiming at the problems of poor prediction effect and long time in obstacle prediction method in unmanned vehicle driving, an accurate obstacle prediction method in unmanned vehicle driving is proposed. Collect obstacle data with the help of lidar in unmanned vehicle driving obstacle avoidance system; The vehicle coordinate system and lidar coordinate system are established respectively, the obstacle data is transformed into the vehicle coordinate system, and the data noise is removed by Kalman filter; By setting the clustering region to cluster the obstacle data, the obstacle points are fitted into a straight line to extract the obstacle features; Match the obstacles with the maximum similarity, determine the Kalman filter, determine the motion state of the obstacles, and complete the obstacle prediction in the operation of unmanned vehicle driving. The experimental results show that the prediction effect of the proposed method is good.
Keywords: unmanned vehicle driving; obstacle prediction; lidar; kalman filter; multi-feature fusion