Trajectory Tracking Of Autonomous Vehicle Based On Model Predictive Contour Control

Autonomous driving technology can effectively alleviate traffic congestion and traffic safety problems,and provide a strong impetus for industrial upgrading.Unmanned vehicles have become one of the hot research technologies.Model predictive control has the ability to predict information and handle multiple constraints,and can control target vehicles in multiple scenarios and multiple targets.Therefore,it is of great significance to study model predictive control in the field of autonomous driving.Firstly,the road model and vehicle model used in model predictive control are constructed.Based on Frenet’s formula,a method to obtain the correspondence between centerline and road boundary is used to parameterize lane lines.Combining the monorail twodegree-of-freedom dynamics model and the kinematics model,a hybrid model is adopted,which can give consideration to the high speed and low speed accuracy of the model.Secondly,a path following control method for autonomous vehicle based on model predictive contour control is proposed.The method combines path planning and path tracking together and realizes path tracking by inputting lane line and vehicle status into a single-layer controller.The vehicle stability envelope and tire friction circle constraint were combined to improve the lateral and longitudinal stability of tracking respectively.The nonlinear objective function is transformed into a quadratic programming problem by using the linear time-varying approximation method,which improves the efficiency and stability of the solution.Then,a new method based on lane boundary constraint is proposed,and lane boundary constraint is added into the model predictive contour control to realize the obstacle avoidance function of the single-layer controller.The improved grid method was used to establish an environmental model to identify the obstacle area,and the feasible area of the vehicle was identified by solving the feasible area of the convex approximate obstacle avoidance,and a new lane boundary was obtained to realize the obstacle avoidance function.Finally,the proposed controller is verified by the Car Sim /Simulink co-simulation platform.The simulation results show that the proposed controller has good autonomous planning ability,can plan the optimal path with the shortest time,and can better realize the obstacle avoidance function to maintain the stability and safety of the vehicle.