Strategy Of Inteligent Vehicle Path Following And Obstacle Avoidance Based On GPS And Lidar

With the increase in car ownership,traffic accidents and traffic congestion are becoming more and more serious,and unmanned vehicle technology has become an important method to improve traffic safety.Navigation positioning is an important part of driverless technology.The autonomous navigation of unmanned vehicles can be divided into two parts: global navigation and local navigation.Global navigation uses high-precision GPS to collect path information and construct an electronic map to plan the global path.Local navigation uses Lidar,camera,millimeter wave radar and other sensory sensors to acquire and integrate the surrounding environment information,and uses the local path planning algorithm to search for the travelable area to achieve dynamic path adjustment and realize obstacle avoidance and other functions.In this paper,a path following and obstacle avoidance strategy for unmanned vehicles based on GPS and Lidar is proposed.Firstly,the overall system design framework of the unmanned vehicle is determined,and the relevant hardware equipment is selected according to the design requirements.The software and hardware modules for the navigation and obstacle avoidance of the unmanned vehicle are built for the requirements of this paper,and the hardware equipment is tested.At the software level,functions such as data communication,algorithm implementation,and human-machine information are completed.Secondly,the unmanned vehicle navigation vector map is established and optimized.The differential GPS system extracts the effective information to filter the useless data,and uses the differential GPS to establish the lightweight vector map of the unmanned vehicle.Traffic lights,crosswalks,speed limit signs and other elements are integrated into the vector map.The map is optimized and complemented by the piecewise cubic Hermite polynomial interpolation method to ensure the integrity and high precision of the map.Thirdly,the unmanned vehicle global path planning and path following scheme are determined,the roads are classified and numbered,and the way tag and waypoint arrays are established as the links and decision behaviors of each road,and then a global path planning strategy is proposed.The second-order Bezier curve is used to curve the intersection to ensure the accuracy of path tracking.Based on this,an unmanned vehicle path tracking strategy based on vehicle lateral error and angle error fitting is proposed.The real-time and accuracy of the tracking strategy is verified by experiments.Finally,the dynamic path adjustment and obstacle avoidance scheme of unmanned vehicles is determined.An obstacle avoidance strategy based on the combination of differential GPS global path and Lidar is proposed.In the unstructured road,the vector field histogram algorithm is used to avoid barrier.The differential GPS obstacle avoidance strategy is adopted in the structured road.The lanes are switched by calculating the heading angle of adjacent lanes.The feasibility and reliability of the above strategies are verified by real vehicle test.In this paper,the innovations are as follows.Firstly,a vector map based unmanned vehicle path tracking strategy is proposed.Compared with the traditional path tracking strategy,this method removes the large-scale 3D point cloud data and introduces lightweight.Vector maps add road information to the map in the form of labels,reducing storage space and enhancing the ability of unmanned vehicles to understand road information.Secondly,an obstacle avoidance strategy based on differential GPS and Lidar is proposed.Compared with the traditional obstacle avoidance strategy,this method can quickly select the obstacle avoidance strategy for different road conditions.Lightweight storage makes its decision making more rapid,meets the driving requirements of the road,and further improves the probability of obstacle avoidance to ensure driving safety.