Research On Local Path Planning Of Driverless Vehicle

With the increase of the number of cars,traffic congestion occurs from time to time and traffic safety accidents occur frequently.The safety situation of the road is increasingly severe.The application of driverless vehicle provides a new idea for solving these problems.Path planning is the core technology of driverless vehicle and this paper studies the local path planning of driverless vehicle.Aiming at the local path planning in static scene,an improved DRRT algorithm is proposed to generate the path that conforms to the vehicle kinematics model.Aiming at the local path planning in dynamic scene,a multi constraint artificial potential field is constructed to improve the disadvantages of target unreachable and falling into local optimum.The specific research contents are as follows.The theory and method of local path planning for driverless vehicle.The influencing factors of local path planning are analyzed in detail,including external environmental factors such as road,traffic sign,pedestrian and the factor of their own vehicle.The global coordinate system and natural coordinate system used in the path planning system are introduced.The kinematic model of the vehicle is described in detail and the formula reasoning is made,which provides a theoretical basis for the path planning to meet the curvature requirements of the vehicle.Local path planning of driverless vehicle based on DRRT algorithm in static scene.The path planning principle of basic RRT algorithm is described and the shortcomings of the algorithm are the random tree growth has no direction and the planned path does not conform to the kinematic model of driverless vehicle.Aiming at these shortcomings,an improved path planning algorithm DRRT is proposed.The sampling method of random nodes is improved and the selection condition of adjacent nodes is redefined.The redundant path is trimmed and the 4-order Bézier curve is used to smooth the broken line path.The static simulation environment is established in MATLAB platform,and the proposed DRRT algorithm is verified by simulation.Through the comparative experiments with basic RRT algorithm,target biased RRT algorithm and RRT~* algorithm,the effectiveness of the proposed DRRT algorithm is proved,and the final path can ensure the safety of driving.Local path planning of driverless vehicle based on improved artificial potential field method in dynamic scene.The path planning principle of the basic artificial potential field method is described in detail,and the main shortcomings of the algorithm are that the target is not reachable and it is easy to fall into local optimum.Aiming at the shortcomings of the algorithm and the suitable dynamic environment,the function of obstacle artificial potential field is improved and path curvature constraint and dynamic obstacle constraint are added.A multi constrained artificial potential field is constructed for local path planning.The dynamic simulation environment is established in MATLAB platform,and the improvement measures are simulated to verify that the planned path meets the driving requirements of vehicles and can avoid obstacles smoothly.Simulation verification of path tracking based on MATLAB and PreScan.The static and dynamic scenes are built in PreScan and the validity of the proposed algorithm is verified by checking the accuracy of path tracking.The experiments show that the actual trajectory of the driverless vehicle is basically consistent with the target path.The error of path tracking is within 0.3 m and the change of lateral acceleration is within 0.4 g,which proves that the algorithm has good applicability for local path planning of driverless vehicle.