Path Tracking of Unmanned Vehicles Based on the Time-Varying Local Model
BAI Guo-xing,ZHOU Lei,MENG Yu,LIU Li,GU Qing,WANG Guo-dong
DOI: https://doi.org/10.13374/j.issn2095-9389.2022.03.18.003
2023-01-01
Abstract:The development of unmanned vehicles has been extremely rapid in recent years. Unmanned vehicles require path tracking control. Based on mature mathematical modeling methods for unmanned vehicles, path tracking control research using model-based control methods, such as feedback linearization control, optimal control, and model predictive control, is very common. Currently, two types of model-based control methods are commonly used in the path tracking control of unmanned vehicles: based on global and local models. The path tracking control based on the global model has a coupling relationship between the longitudinal speed of the unmanned vehicle and the lateral displacement error and longitudinal displacement error in the global coordinate system. Furthermore, this coupling relationship varies with the heading angle, making the controller unable to take the longitudinal speed as a control input to improve the accuracy of path tracking control. Path tracking controllers based on local models usually use errors as reference models, making the controller less accurate when the curvature of the reference path greatly varies. To address the above issue, an unmanned vehicle path tracking control method based on a time-varying local model is proposed considering the principle of rolling optimization of nonlinear model predictive control. Specifically, a time-varying local coordinate system is first established based on the time-varying pose of the vehicle. Then, a reference path in front of the vehicle is entered into this local coordinate system. The model-based iterative prediction is completed in this local coordinate system, and finally, the control is achieved using the optimization solution. The proposed control method is verified by co-simulation using MATLAB and CarSim. The simulation conditions include low-speed and high-adhesion road conditions, low-speed and low-adhesion road conditions, and high-speed low-adhesion road conditions. The simulation results show that the path tracking controller based on the time-varying local model outperforms the path tracking controller based on the global model, the path tracking controller based on the local model, and the Stanley path tracking controller. The maximum absolute value of the displacement error of the proposed controller does not exceed 0.3342 m under all simulation conditions, and the maximum absolute value of the heading error does not exceed 0.0913 rad. Moreover, the proposed controller can still complete the path tracking in situations where other controllers fail, such as high-speed and low-adhesion road conditions.