Adaptive Path Tracking Controller for Intelligent Driving Vehicles for Large Curvature Paths

In this article, we use MPC algorithm to design an adaptive path tracking controller based on the vehicle coordinate system, which is effectively applicable to path tracking scenarios with different vehicle speeds and large path curvatures. To reduce the lateral position error and heading angle error, a fitting function learned through a large number of simulations is used to adaptively adjust the prediction horizon parameter and a compensation strategy of steering angle increment designed based on fuzzy control algorithm is used to reduce the influence of model mismatch and low modeling accuracy on the path tracking control effect, then the front wheel steering angle is calculated and output to the vehicle model for path tracking. In this article, multi-scenario simulations are conducted in Simulink and CarSim environments to verify the performance of the proposed controller. The result shows that the adaptive path tracking controller proposed in this article achieves a more satisfactory path tracking control effect than that of fixed-parameter MPC controller. In the simulations using proposed controller, the maximum lateral position error does not exceed 2 cm, the average lateral position deviation does not exceed 1 cm, the average heading angle error does not exceed 0.06 rad, and the error state amount keeps in a reasonable range, which ensures safe and stable tracking under scenarios with different vehicle speeds and path curvatures.