A path planning algorithm based on fusing lane and obstacle map

This paper proposes a path planning algorithm for autonomous driving in urban environments. The processing of video and Velodyne pointcloud provides information about the positions of lane markers and obstacles in the local map, which are then converted to a lane costmap and obstacle costmap. The referenced GIS follow line is used for generating a series of offset curves, and the best follow line is selected according to a combination of lane, obstacle and background cost. Additional handling of planning path and maximum speed is provided. Our planning algorithm can handle various road types such as U-turn, intersections, and different driving behaviors including passing over or following front vehicles, etc. The proposed navigation framework is implemented on an autonomous vehicle, which exhibits good performance on Future Challenge 2013, Changshu, China.