Research On Trajectory Tracking Of Unmanned Vehicle Based On Model Predictive Control

With the rapid development of computer information processing,artificial intelligence,big data,sensors,and so on.Intelligence and unmanned technology are developing rapidly in various fields,and the research of driverless cars has also become a hot topic.Trajectory tracking control,as the last technology of unmanned vehicles,is also a key step.It is the main evaluation standard and manifestation of the performance of unmanned vehicles.Therefore,studying the trajectory tracking control and finding a method that can track the reference trajectory quickly and stably has important practical significance for the realization of unmanned driving.The article studies the application of Model Predictive Control(MPC)method in the trajectory tracking of unmanned vehicles.The main contents of this paper are as follows:(1)From a geometric point of view,study the law of vehicle motion,analyze the geometric relationship of the controlled vehicle in space,such as position,speed,and heading angle,establish a kinematic model of the vehicle,and clarify the kinematic geometric constraints that the vehicle should meet during driving.As the basis of dynamic analysis and the guarantee of trajectory tracking control.(2)By analyzing the forces of the vehicle in the three directions of longitudinal,transverse and yaw,a six degree of freedom dynamic model of the vehicle is established.A tire model with lateral and longitudinal joint tire force is used to obtain more accurate tire force,and then a more accurate dynamic model is obtained.(3)In order to prevent extreme conditions(such as sideslip)of the vehicle from occurring,the constraints of control quantity,control increment,tire slip angle and vehicle center of mass slip angle are added to the constraints of the predictive model at the same time.Make full use of the advantages of the MPC control method to deal with constraints,and construct an MPC controller by analyzing and designing a reasonable cost function.(4)The Car Sim/Simulink/Matlab co-simulation platform was built.Since only the simulation of the vehicle in Simulink could not explain the complex structure of the vehicle,Car Sim was introduced.Car Sim has a detailed description of the overall structure of the vehicle,and the motion characteristics of the almost real vehicle.In the Simulink/Car Sim co-simulation platform built,verify the performance of each controller in tracking control.Through the Car Sim/Simulink/Matlab co-simulation platform built,on the basis of the above work,the simulation verification and comparison of the improved MPC controller in a variety of situations under the limit of double-line shifting conditions have been completed.The good performance of the trajectory tracking of the controller is verified,it can adapt to the changes of vehicle speed and environment,and has good robustness and environmental adaptability.