Research On Local Path Planning And Path Tracking Control Of Intelligent Vehicle

With the vigorous development of automobile manufacturing industry,car ownership has increased dramatically,resulting in increasingly serious road traffic safety problems.Intelligent Transportation System(ITS)is the development direction of future transportation.which is conducive to solving traffic safety problems.As an important part of ITS,intelligent vehicles contribute to the development of intelligent transportation,and its realization depends on many key technologies.Local path planning and path tracking control are the key research contents.In this paper,the DWA local path planning algorithm based on improved RRT and the path tracking controller based on state extended MPC and corner compensation are designed to solve the problem that the vehicle is difficult to follow the global path in the obstacle avoidance scene and the real-time and tracking accuracy are difficult to guarantee in the process of path tracking.Aiming at the local path planning problem,the RRT global path planning algorithm is designed.On this basis,the sampling method of RRT is improved by setting the middle sampling points of the path,and the global planning path is smoothed and optimized by cubic non-uniform B-spline curve.The 3-DOF kinematics model of vehicle is established,and the local path planning algorithm based on DWA is designed.The global path key points planned by the improved RRT are extracted as the sub-goal points of the DWA algorithm,and the DWA local path planning algorithm based on the improved RRT is designed through the trajectory evaluation function of the improved DWA algorithm.The effectiveness of the designed local path planning algorithm is verified in four different obstacle scenarios.For the problem of path tracking control,the tire dynamics model using magic formula is established and the tire characteristics are analyzed.On this basis,the vehicle yaw dynamics model is established and linearized according to the state trajectory method.The model state variables are extended and converted,and the state extended MPC controller is designed.The front wheel angle is solved by designing the objective function and constraint conditions to track the target path.The vehicle-path tracking error model is established,and the angle compensation fuzzy controller is designed to correct the front wheel angle according to the vehicle lateral and heading deviation.Verify the real-time performance and tracking accuracy of the path tracking controller under three different speeds.The MATLAB/CarSim co-simulation platform is built to verify the effectiveness of the planning and tracking method in single and multiple obstacle avoidance scenarios.The simulation results show that in the four obstacle scenarios of path planning,compared with the DWA algorithm based on A*,the planning path distance and planning time of DWA local path planning algorithm based on improved RRT are reduced.Among them,the planning path distance is reduced by 7.04%and 5.1%in complex obstacle scenarios and indoor space scenarios,and the planning time is reduced by 16.6%and 13.3%,respectively.Compared with the traditional MPC controller,the path tracking controller based on state-extended MPC and angle compensation has better tracking performance under the standard double lane change condition of path tracking,and the maximum lateral and heading deviation of path tracking are reduced by 23%and 17%respectively.The average control increment solving time of state extended MPC controller is reduced by more than 14%.In the single and multiple obstacle avoidance simulation experiments,the designed local path planning algorithm has smooth and smooth obstacle avoidance path.The path tracking controller maintains that the maximum lateral deviation of the vehicle is less than 0.5m and the maximum lateral acceleration is about 0.03g.