Research On Intelligent Vehicle Autonomous Navigation Path Planning And Tracking Control Algorithm

In recent years,my country’s car ownership has grown rapidly.While automobiles bring convenience to human beings,they also cause many problems such as traffic accidents.As an effective way to solve these problems,intelligent vehicles are gradually being sought after by people.As the key technology of autonomous driving,path planning and tracking control have become a hot spot in today’s research.Based on the National Natural Science Foundation of China project "Research on Deep Level Perception and Understanding Methods of Intelligent Vehicles in Complex Dynamic Environments"(Project Number:U1864204),the research on the path planning and tracking control methods of intelligent vehicles is carried out.The specific content of the paper is as follows:(1)This paper studies the method of global path planning for intelligent vehicles,based on the Floyd algorithm to plan the global shortest path through the necessary points.The principle of inertial integrated navigation and its importance to autonomous driving are introduced.According to the key nodes of the semantic topology map,RTK＿SINS is used to collect the coordinates of each road section.The coordinates are transformed into the local coordinate system through Gaussian transformation,coordinate transposition,and coordinate translation to complete the establishment of a global high-precision map.Establish a weighted directed graph and plan the shortest path through Floyd algorithm.Finally,a smooth and continuous global path is obtained by fitting the cubic quasi-uniform B-spline method.(2)The method of local path planning for is studied.A hybrid algorithm of artificial potential field and A* is proposed.On the basis of improving the repulsive function,the original path constraint,dangerous driving boundary constraint and vehicle kinematics constraint are introduced to complete the optimization of the artificial potential field method.Aiming at the local optimal problem where vehicles,obstacles,and target points co-exist on the same line,the A* algorithm is first used to plan the intermediate target points with sudden heading angle changes,and then the artificial potential field method is used to complete the planning according to the target points.Simulation experiments show that the hybrid algorithm successfully plans a continuous smooth,collision-free local path in the face of static obstacles,and no local optimum occurs.(3)The method of intelligent vehicle path tracking control is studied,and a multi-constrained LVT MPC path tracking controller is designed.Through reasonable assumptions and simplifications,a vehicle dynamics model is established,and then linearized and discretized to obtain a linear time-varying model.Using the front wheel angle as the control quantity,the LVT MPC path tracking controller is designed.And constrain the center of mass slip angle,lateral acceleration,and wheel slip angle.Finally,a simulation test of low,medium and high vehicle speeds and low,medium and high adhesion coefficient roads is carried out.The results show that the vehicle can follow the target path smoothly and accurately,and the steering changes smoothly,which verifies the effectiveness,adaptability and robustness of the controller in different driving conditions.(4)Complete vehicle data verification and analysis of local path planning and tracking control algorithm.The design scheme of the intelligent vehicle experiment platform and the steering actuator are introduced.Based on the experimental platform,the experimental data is collected,and the algorithm is verified and analyzed.The results show that the hybrid algorithm can successfully plan a smooth and collision-free local path when facing static obstacles without causing the vehicle to fall into the local optimum.The LVT MPC controller can control the vehicle to accurately drive along the target path,and the steering changes smoothly,which meets the requirements of path tracking for control accuracy and smooth driving.