Research And System Development Of Path Planning Algorithm For Autonomous Vehicles Based On High-definition Map

With the continuous development and progress of Internet technology,the development of autonomous vehicles is also changing with each passing day.In the field of autonomous driving,path planning plays a vital role.In most researches and applications,the path planning problems involved are based on grid map or road-level navigation map.These two maps have poor semantics and cannot support the complex driving rules of autonomous vehicles.High-definition maps can well solve the above-mentioned problems.Compared with other map formats,high-definition map has the characteristics of rich semantics and accurate road descriptions,which can better meet the path planning needs of autonomous vehicles.Therefore,the research on path planning based on high-definition map is very meaningful.This paper proposes a path planning algorithm for autonomous driving vehicles based on high-definition maps,and constructs a path planning system for autonomous driving.The main work carried out is as follows:1.Preparation of high-definition mapsA relatively low-cost high-definition map production technology is proposed.The lidar and IMU sensors on the experimental vehicle are used,and the lidar SLAM algorithm is selected to generate a dense lidar point cloud map through SLAM technology.Label,customize high-definition map standard semantic information such as roads,lanes,and lane lines,and generate high-definition map in standard Open DRIVE format.2.Improvement of path planning algorithm and construction of path planning system based on high-definition mapConstruct the path planning system,analyze the static files of the local high-definition map,design the lane connection method in the high-definition map according to the characteristics of the Open DRIVE high-definition map,generate a structured lane-level road network in the system,and generate it according to the high-definition map system modeling.Aiming at the lane-level global path planning of the structed lane-level road network,an optimized A~* algorithm is proposed,which is specially adapted to the high-definition map.Improve the heuristic function in the A~* algorithm,integrate map characteristics(such as driving cost,lane direction,etc.)into the heuristic function of the algorithm,and optimize the node expansion method in the original A~* algorithm,and support such as Path planning for special scenarios such as lane changes and U-turns.Compared with the original A~* algorithm,the efficiency is greatly improved,and the obtained path is guaranteed to be the optimal solution.2.Simulation and real vehicle test of path planning systemSimulation and real vehicle testing of the path planning system.A trajectory planning algorithm based on high-definition maps is implemented to adapt to the path planning system.Deploy the path planning system and trajectory planning module on the Carla automatic driving simulation platform for joint simulation.Finally,the path planning system is tested on the actual vehicle in the experimental field.Simulation and real-vehicle test results show that the real-time performance of the path planning system is good,and the generated path can be better applied in the automatic driving system.Observation of experimental data shows that the path planning system is highly feasible,and the path planning algorithm based on high-definition map performs better than traditional path planning algorithms on autonomous vehicles.