Path-following control for autonomous vehicles considering the uncertainties of yaw dynamics

With the rapid development of autonomous driving technology, achieving precise path-followingcontrol in uncertain environments has become a key challenge. To address the problem mentioned above, this paper proposes a path-following controller based on input-output feedback linearization and neural network-based approximation of uncertainties. It cancels out the nonlinear dynamics by designing a feed-forward control law, in which a neural network is employed to approximate the uncertainties in the vehicle yaw dynamics. High-fidelity co-simulations were carried out using CarSim and Simulink. The results show that after introducing neural network-based compensation, the steady-state error of path-following is reduced by 23%-30%, verifying the effectiveness of the proposed method.