Research On Driverless Formula Racing Car Perception And Planning Based On Lidar

The rapid development of the automobile industry brings convenience to people’s lives,but also brings many traffic safety problems.Research on driverless technology is not only in line with the current development trend of the intelligent era,but also an important means to help solve road traffic safety problems.For unmanned vehicles,environment perception and path planning are two very important aspects,which play an indispensable role in ensuring the safety of the vehicle.The Autonomous Driving Formula Competition for college students organized by the Chinese Society of Automotive Engineering provides a broad and professional platform for university students to study autonomous driving technology.This article is based on the National Natural Science Foundation of China(51675257)and the Liaoning Provincial Key R&D Program Guidance Program(2017106020).Based on the robot operating system ROS platform,lidar sensor and integrated INS,this article aims at the environment perception and path planning problems in the calculation layer of the driverless equation car,Carried out the following research.(1)Research on obstacle detection methods.The original point cloud data collected by the lidar is used to retain the effective area of interest around the car(ROI)by setting an appropriate threshold for the three-dimensional coordinate axis;using the body-based grid algorithm to effectively reduce the number of point clouds;and using random sampling consistency(RANSAC).The algorithm filters out the point cloud data on the ground;the Euclidean distance clustering algorithm is used to cluster the target point cloud.The experimental results show that the algorithm can effectively obtain the central position coordinate information of the target cone barrel.(2)Research on localisation and mapping methods.In this paper,the vehicle body position coordinates are obtained by the combination of GPS and IMU positioning,and the laser radar is used to obtain the position information of the cone,and then the two data are merged together according to the principle of time synchronization.,Obtained the position information of the cone in the global coordinate system,and finally completed the construction of the track.(3)Research on the path planning method of local planning and global planning.The realization of this method is to obtain the racing position by combining inertial guidance,and then properly process the output data of the lidar,and obtain the position information of the cone barrel under the global coordinate system through coordinate conversion.In the eight-word surround and linear acceleration race,the combination of global planning and local planning is adopted to optimize the driving path through local planning on the basis of the car driving along the global path.In the high-speed track race,through the real-time perception of lidar,the optimal curve generation path planning algorithm is used to obtain the expected path.The experimental results show that different algorithms can effectively and reasonably plan the corresponding racing path for different tracks.The environment perception and path planning methods studied in the thesis have been applied to the driverless formula car,and their effectiveness and practicability have been verified in real car experiments.The car can safely and steadily leave the track according to the planned path to complete the race smoothly.