Abstract: This paper focuses on the trajectory tracking problem of unmanned vehicles. On the basis of selecting model predictive control theory as the theoretical basis of this study, firstly set the front wheel angle of the vehicle as the corresponding control quantity, and set the lateral position and yaw angle of the vehicle as the state quantity to establish the three-degree-of-freedom dynamic model of the vehicle. Then, combined with the magic tire formula, in order to ensure the accuracy and effectiveness of the model, the existing model is linearized and discretized to complete the preliminary conception of the prediction model. On this basis, the relaxation factor and soft constraint are added to complete the establishment of the incremental linear time-varying model trajectory. After the model optimization is completed, it is verified through the Simulink/CarSim joint simulation platform. By setting different simulation conditions, simulated the complex working conditions encountered by the driverless vehicle as much as possible. After several simulation experiments, the trajectory tracking controller designed in this paper is under good road conditions. The maximum safe longitudinal driving speed is 68 km/h, which provides a data basis for subsequent in-depth research on driverless vehicles.

Key words: Driverless vehicles; Model predictive control; Trajectory tracking optimization; Simulink/ CarSim co-simulation

摘要： 文章重点研究无人驾驶车辆的轨迹跟踪问题，选取模型预测控制理论作为此次研究的 理论基础。首先设定汽车的前轮转角作为相应的控制量，将汽车的横向位置和横摆角设置为 状态量，建立起汽车的三自由度动力学模型，然后结合魔术轮胎公式，为保证模型的精度有 效性，对现有模型进行线性化和离散化处理，完成预测模型的初步构想，在此基础上增加松 弛因子和软约束，从而完成增量线性时变模型轨迹的建立。模型优化完成后，通过 Simulink/ CarSim 联合仿真平台进行验证，设定不同仿真工况，尽可能模拟无人驾驶车辆遇到的复杂工 况，经多次仿真实验，文章设计的轨迹跟踪控制器在良好路面条件下，最高安全纵向行驶速 度为 68 km/h，为后续无人驾驶深入研究提供数据基础。

关键词: 无人驾驶车辆；模型预测控制；轨迹跟踪优化；Simulink/CarSim 联合仿真