Research On Self-tracking Of Driverless Vehicle Based On Lidar

Urban congestion and traffic accidents have become the major problems that need to be solved urgently.With the rapid development of sensor and computer technology,driverless vehicles are considered one of the solutions to the problem.Self-tracking function which mainly relies on GPS technology to locate and navigate vehicles guarantees safe driving of driverless vehicles.However,GPS signals are easily disturbed by tall and large objects during transmission.It is difficult to maintain the differential positioning status with high precision.In order to solve the problem of vehicle tracking control when GPS positioning is inaccurate,this paper proposes a self-tracking method based on Lidar.The idea of this paper is to identify the boundary of the structured road by using the point cloud collected by Lidar,and then to control the steering system and driving system of the driverless vehicle to keep it at a certain distance from the boundary and track ahead.The main research contents are as follows:(1)Research on the performance parameters of the Lidar(HDL-64E)and determine the effective data range for road boundary recognition by analyzing the distance resolution of the laser point cloud in the radial direction and the circumferential direction.Then analyze the structure of the UDP packets and deduce the calculation formula for computing the 3D coordinates of laser point by using measured distance and angle data.(2)Propose a method of road boundary recognition based on image processing.Firstly,the original data of lidar is processed by filtering the isolated noise points and array interpolation.Multi-feature filtering is used to filter out road points and preserve road boundary points.Then convert 3D point cloud to 2D image by discarding altitude information,and use pixel stretching and closing operation to solve the problem of boundary pixel discontinuity.Then line by line scanning algorithm is used to determine the specific location of the boundary,and finally using the inherent properties of the road to exclude interference factors.(3)The tracking control of the vehicle is researched on two aspects of the upper computer decision and the lower computer control.The upper computer generates the target path by offsetting the right boundary to the vehicle position,and calculates the target front wheel angle and speed of the vehicle from the relative position relationship between the target path and the vehicle.The lower computer adopts an improved PID algorithm to control the motor of the electric power steering.The braking control adopts a hierarchical fuzzy strategy and a detailed throttle and brake switching rules are formulated.(4)For the four typical scenes that often appear in the driverless vehicle race,carried out the road boundary recognition test,and a small curvature road for 500 meters was selected to carry out self-tracking experiments based on lidar.Through the analysis of the results verify the feasibility of the method in this paper.