MPC-based Path Tracking And Local Path Planning Controller Design

With the substantial increase in the number of cars,consumers have also put forward higher requirements for car products,promoting digital transformation and upgrading of cars,making autonomous driving a research hotspot today.In addition,traffic accidents have been increasing in recent years,and autonomous driving technology and assisted driving technology can reduce human-caused traffic accidents.Path tracking and path planning,as key links in autonomous driving technology,have also attracted more and more attention.First,the vehicle kinematics model and dynamics model are established.And when the tire slip angle is small and the slip rate is small,the formulas of t ire longitudinal force and lateral force are simplified,and the model is further simplified based on the assumption of small angle to obtain a simplified vehicle dynamics model.Then take the vehicle dynamics model as an example,the state equation is continuous,discretized,and then iteratively recursed to obtain a new state equation.The state equation can predict the state of the vehicle in the predicted time domain by controlling the increment of the variable.Then determine the objective function of the model predictive control,and then transform the objective function into the standard form of the quadratic programming problem and solve it.Then,the kinematics model controller and the dynamics model controller are established respectively,and the builtup controller model and the Carsim vehicle model are used for joint simulation under double-line shifting conditions.The results show that the tracking performance of the controller is good when the vehicle speed is low.However,when the vehicle sp eed is high,the path tracking controller established by the kinematics model at the position of the road with greater curvature will have a larger deviation from the reference path.On the whole,the tracking stability of the dynamic model path tracking c ontroller is better.Then use the established vehicle dynamics model to design the path tracker.For the moving obstacle,determine all the position points of the obstacle in the predicted time domain through its current position and movement speed,and th en add the vehicle position calculated by the controller to calculate the overall distance between the vehicle and the obstacle in the predicted time domain.The overall distance is used to evaluate the likelihood of a collision between the vehicle and an obstacle,and this is added to the objective function of the model predictive control.Then a co-simulation model is established,which uses the path following controller based on the dynamic model in Chapter 3.Then the simulations are carried out under t he double-line shifting working condition and the conic working condition respectively,and the algorithm can achieve obstacle avoidance and path tracking well.And by changing the parameters of the quadratic curve to generate different reference curves,i t is verified that the algorithm can achieve obstacle avoidance and path tracking well under different reference curves.Finally,the objective function of the integrated controller is designed.The relationship between the possibility of collision between the vehicle and the obstacle and the control quantity is first analyzed,and then the obstacle avoidance function is established,so that the overall objective function can be transformed into a quadratic programming In the standard form,the quadratic pr ogramming solution function is used to optimize the solution.Then a co-simulation model of the integrated controller is established,and the algorithm is verified by simulation under the double-line shifting working condition and the conic working conditi on.However,when the parameters change under the conic working condition,the integrated controller will appear that the minimum distance between the vehicle and the center point of the obstacle is very small and the vehicle ordinate and the ordinate of t he reference path have a large offset,and the weight parameters need to be adjusted.optimize.Then the Simulink optimization toolbox is used to optimize the weight parameters.After optimization,the performance of the integrated controller is obviously better.Finally,it is verified that the integrated controller can also achieve obstacle avoidance and path tracking well when the vehicle meets the obstacle laterally.