Research On Local Obstacle Avoidance Path Planning And Path Tracking Control For Autonomous Vehicle Based On MPC

Obstacles pose a great threat to the safety of autonomous vehicles during the driving process.It is necessary to replan the reference path to ensure that the reference path can avoid obstacles and control the vehicle to strictly track the reference path so as to avoid accidents.This thesis takes autonomous vehicles in the obstacle scenes as the research object.Local obstacle avoidance path planning and path tracking control were researched based on MPC theory to ensure the driving safety and handling stability of autonomous vehicles,and the main research contents are as follows.The nonlinear model predictive control algorithm was adopted to design the local obstacle avoidance path planner for the problem of local obstacle avoidance path planning in scenes with obstacles.The vehicle point quality model was adopted to build the predictive model of path planner.To solve the limitations of the traditional obstacle avoidance function in excessive obstacle avoidance,the objective function including the new obstacle avoidance function was designed,and the discrete points of the local planning path were obtained by solving the nonlinear optimization problem with constraints.The fifth-order polynomial was adopted to fit the discrete points of the local path,and the polynomial coefficients were import the controller,and the local obstacle avoidance path planner was designed.A two-degree-of-freedom nonlinear dynamics model of the vehicle was found,and the path tracking controller predictive model was discreted and approximately linearized.According to the requirements of the controller,an objective function with constraints was designed,and transformed the function into a quadratic programming problem to solve the control parameters and the path tracking controller was designed.The simulation experiments of double-lane conditions under different velocities was carried out.The results show that the vehicle tracking deviation is large and the stability of vehicle is reduced at high velocity.In view of the above problems,the influence of horizon parameters on the controller was further researched.Through theoretical analysis combined with simulation experiments,the influence of horizon parameters on controller was analyzed,and a dual-horizon-parameters evaluation index considering tracking accuracy and driving stability was proposed.The optimal dual-horizon-parameters at different velocities were obtained according to the evaluation index.Then horizon parameters update strategy was proposed.The prediction horizon and control horizon were updated by real-time detection of vehicle velocity,and the adaptive prediction horizon and control horizon path tracking controller was designed.The simulation experiment of the controller under different velocities verifies that the adaptive horizon parameters path tracking controller has better tracking accuracy and driving stability.The local obstacle avoidance path planner adopting the new obstacle avoidance function and the adaptive horizon parameters path tracking controller were integrated,and a joint simulation platform integrating the planner and controller was built.The simulation experiments under multiple obstacles condition were carried out.The simulation result shows that the integrated structure of planner and controller can replan safe paths and achieve precise path tracking control in the scene with multiple obstacles and realize safe driving in obstacle scenes.