基于人-车-路闭环系统的路径跟踪模型预测控制

Model Predictive Control of Path Tracking Based on Human-Vehicle-Road Closed-Loop System

为提高人-车-路闭环系统的路径跟踪能力以及操纵稳定性能,利用Matlab/Simulink软件,搭建了3自由度汽车动力学模型以及基于车辆侧向加速度反馈修正的预瞄驾驶员模型,并与Carsim模型进行对比验证,检验了上述模型的正确性;根据模型预测控制理论,搭建了基于主动转向的路径跟踪模型预测控制器,采用双移线路径作为仿真工况,分别在有/无模型预测控制的情形下进行了仿真。结果表明:该模型预测控制器可将路径跟踪的误差均值控制在0.1m以内,将瞬态误差极值降低50%左右,并消除了路径跟踪后期的震荡现象;此外,可保持汽车侧向加速度、横摆角速度值等操纵稳定性指标在合理范围之内。研究结果可为智能汽车的人-机共驾研究以及无人驾驶方向的进一步研究提供参考。

In order to improve the path tracking ability and handling stability of Human-Vehicle-Road Closed-Loop System,a three-degree-of-freedom vehicle dynamics model and a preview driver model based on vehicle lateral acceleration feedback correction are built by using MTALAB/SIMULINK software. The correctness of the above model is verified by comparing with CARSIM model. According to the model predictive control theory,a model-based vehicle dynamics model is built. The path tracking model predictive controller of active steering is simulated with or without model predictive control using double-shift path as the simulation condition. The results show that the model predictive controller can control the mean error of path tracking within 0. 1 m,reduce the extreme value of transient error by about 50%,and eliminate the oscillation phenomenon in the later stage of path tracking. In addition,it can keep the vehicle lateral acceleration and yaw angular velocity within a reasonable range. The research results can provide reference for the study of human-vehicle co-driving of intelligent vehicles and the further study of the direction of unmanned driving.