Intelligent Vehicle Trajectory Planning And Tracking Control Based On Model Predictive Control

Intelligent vehicles are machines that can work automatically or semi-automatically in complex environment.As an indispensable part of an intelligent transportation system,they can greatly improve driving safety,greatly improve the efficiency of highway traffic,and reduce resource consumption.Trajectory planning and tracking control are important parts to ensure that the intelligent vehicle can run normally.This thesis mainly studies the trajectory planning and trajectory tracking of the intelligent vehicle.Based on the model predictive control algorithm,the controller is designed by combining the primal dual neural network optimization and artificial potential field method to realize the trajectory planning and tracking control of the intelligent vehicle.The existing problems such as multi-state variable constraints and poor real-time performance are solved.The main research contents of this thesis are divided into the following:1.Aiming at the kinematics model of the vehicle,a predictive controller is designed based on the model predictive control algorithm to achieve tracking of the desired trajectory.The external disturbance is considered in the vehicle kinematics model,and a controller combining feedforward and feedback controller is adopted.In order to improve the realtime robustness,a linear tracking error model is adopted,which keeps the high-order terms while linearizing.The objective function is determined,and the control increment constraint and relaxation factor are set in the objective function.Then,the objective function and constraints are iterated to the constrained quadratic programming problems,and the primal dual neural network is used to solve the rolling optimization.Finally,the stability of the proposed controller is analyzed by Lyapunov theory,and the effectiveness of the proposed controller for reference trajectory tracking is verified by simulation analysis.2.Aiming at the dynamics model of the vehicle,the trajectory tracking problem of model predictive control is studied.First of all,the nonlinear dynamic model of vehicle is established,and the tire model is used to analyze the performance of vehicle tire characteristics,fit the lateral and longitudinal forces of the vehicle tire,and simplify the vehicle dynamic model under the theory of small angle hypothesis.According to the state space equations of the vehicle dynamics model,the linear time-varying model and the prediction equation are further derived.In order to determine the objective function,in addition to considering the constraints such as the control amount and control increment,dynamic constraints such as the center-of-mass lateral angle and road surface adhesion coefficient are added to ensure the safety and stability of driving.Then the optimization problem is transformed into a quadratic programming problem,and the primal dual neural network is used to solve the quadratic programming problem to improve the calculation rate.Finally,simulation analysis.3.Based on the vehicle kinematics model,the controller is designed by combining the artificial potential field method and the model predictive control to realize intelligent vehicle trajectory planning and tracking control.Firstly,the driving environment is modeled by artificial potential field method,which mainly includes environmental factors such as roads,environmental vehicles,and target points.According to the interaction between them and intelligent vehicle,the corresponding potential field functions are established respectively.Then,combined with the model prediction control algorithm,the environmental potential field is added to the objective function.In addition to the constraints on the input control quantity and the input control increment of the intelligent vehicle,the safety zone constraints updated with time are also set,and the optimal trajectory selection is realized through the optimization algorithm in the model predictive control.Finally,the proposed method is verified and analyzed by simulation.