Research On Automatic Parking Path Planning And Tracking Control Algorithm Based On Random Sampling And Optimization

With the development of the social economy and the automobile industry,vehicle ownership in China increased year by year,traffic congestion and parking difficulties and other problems have become increasingly prominent.Automatic parking system can reduce the operational burden of the driver in the parking process,improve parking security,therefore,automatic parking system has been widespread concern.In order to solve the current existing automatic parking system’s strict requirements on the starting position and the discontinuous curvature of the reference trajectory,which makes it difficult to track,the parking trajectory planning algorithm and the trajectory tracking control algorithm are designed.Firstly,the relevant parameters under low-speed parking conditions were analyzed to establish a vehicle kinematics model as a reference model for parking trajectory planning algorithm,and compares the model with the output of the Carsim dynamics model at different vehicle speeds to verify the accuracy of kinematics model.After that,the common parking scenes were simplified to establish a parking environment model,and a collision constraint model based on the distance between convex sets and a vehicle collision detection method based on a rapid repulsion-straddle experiment were established.Aiming at the daily parallel parking and vertical parking scenarios,a hierarchical planning algorithm based on the improved RRT* algorithm(Rapidly-exploring Random Tree Star)of Reeds-Shepp curve combined with nonlinear optimization is proposed.First,improve the RRT*RS algorithm from the two goals of improving the efficiency of path planning and reducing the cost of parking,and then for the shortcomings of the Reeds-Shepp curve that meets the minimum turning radius but not the continuous curvature of the path,a nonlinear optimization problem based on convex set obstacle constraints is constructed.For the constrained nonlinear optimization problem,the nonlinear optimization problem is solved by taking the RRT*RS planning result as the initial solution to realize the hierarchical planning with continuous curvature of the parking trajectory.And for a variety of parking scenarios,the proposed planning algorithm is simulated for trajectory planning.The simulation results prove that the proposed hierarchical planning algorithm based on RRT*RS algorithm and nonlinear optimization can be used in different starting positions and parking types.In order to realize the tracking control of the parking trajectory,the longitudinal tracking controller and the lateral tracking controller were researched and analyzed respectively,and the longitudinal speed controller based on PID algorithm and the lateral tracking controller based on pure pursuit and kinematics MPC were established.The tracking effects of the two types of lateral control are compared,and the results show that the MPC control algorithm based on the kinematic model has better path adaptability and stability under low-speed and large-curvature driving conditions.Finally,in order to verify the effect of the parking trajectory planning algorithm and tracking controller proposed in this thesis,a parking simulation scene is built based on the Carsim and Matlab/Simulink co-simulation platform,and the parking planning and control algorithm proposed in this thesis was carried out.Co-simulation verification results show that the parking trajectory planned by the planning method in this thesis can ensure that the vehicle realizes the continuous rotation of the front wheel angle during parking,and the vehicle can still maintain a stable output of the control variable at the reversing point.The controller can accurately follow the parking trajectory with high robustness.