Research On Trajectory Tracking Control For Intelligent Vehicles

Safety is one of the most important issues in the development process of vehicle industry. In the loop of driver-vehicle-environment, due to distraction, fatigue, drunk-driving and lack of experience, driver was considered as the most uncertain part, but also the main cause of frequent traffic accidents. As intelligent vehicles partially and totally excluded the human factor, the traffic accidents can be greatly avoided by developing them.The main research of intelligent vehicles can be divided into three parts:(1) Environment perception, aimed to detect traffic signs, obstacles and other vehicles on road. An environmental sensing system composed of radars, cameras, GPS, etc. are usually used to achieve this goal. It provides a dynamic map of the near environment of the autonomous vehicle;(2) Trajectory planning and decision, which forms a cluster of safe and feasible trajectories, and then determine an optimal one(reference trajectory) in the available space.(3) Vehicle control system, aimed to control vehicles to travel along the reference trajectory(trajectory tracking) using actuators like brakes or powertrain and steering wheel. Vehicle control includes path following and speed following. In this paper, it focuses on the trajectory tracking control for intelligent vehicles.Firstly, the optimal preview theory was used to deal with the trajectory tracking control for intelligent vehicles. Under the assumption that the longitudinal speed is a constant, the longitudinal and lateral vehicle dynamic were decoupled and the trajectory tracking problem was decomposed in the speed tracking problem and path following. A feedforward-feedback control strategy has been proposed for lateral path following. The effectiveness of the controller has been analyzed in the simulation environment. Because of unable to consider the character of the actuator and tire, the optimal preview controller can’t assure the stability at the handling limits.In view of that, a model-based predictive control(MPC) scheme is investigated. A three degree of freedom vehicle model has been use to reflect the coupled longitudinal and lateral vehicle dynamic characteristic. A MPC-based trajectory tracking controller was proposed with the steering angle and the longitudinal acceleration as the control input. As the longitudinal speed was mainly determined by the longitudinal vehicle dynamic, the longitudinal acceleration can be computed by the longitudinal controller based on the optimal preview theory,thus at every predictive horizon, the longitudinal acceleration was considered as a known variable. In this case, the MPC problem was simplified by converted a Multiple-Input Multiple-Output(MIMO) system into a Single-Input Multiple-Output(SIMO) system.Further, In order to reduce the computational burden, the NMPC is converted to a linear time varying(LTV) MPC based on successive online linearization of the nonlinear system model. By considering the constraints of the tire slip angle, the stability of the vehicle was improved.Finally, in order to verify the effectiveness of the proposed methods, the simulations on different conditions have been conducted. Meanwhile, the optimal preview controller was tested on an experimental vehicle.