Adaptive-neural-network-based robust lateral motion control for autonomous vehicle at driving limits

Parametric modeling uncertainties and unknown external disturbance are major concerns in the development of advanced lateral motion controller for autonomous vehicle at the limits of driving conditions. Considering that tyre operating at or close to its physical limits of friction exhibits highly nonlinear force response and that unknown external disturbance can be caused by changing driving conditions, this paper presents a novel lateral motion control method that can maintain the yaw stability of autonomous vehicle while minimizing lateral path tracking error at the limits of driving conditions The proposed control scheme consists of a robust steering controller and an adaptive neural network (ANN) approximator. First, based on reference path model, dynamics model and kinematics model of vehicle, the robust steering controller is developed via backstepping variable structure control (BVSC) to suppress lateral path tracking deviation, to withstand unknown external disturbance and guarantee the yaw stability of autonomous vehicle. Then, by combining adaptive control mechanism based on Lyapunov stability theory and radial basis function neural network (RBFNN), the ANN approximator is designed to estimate uncertainty of tyre cornering stiffness and reduce its adverse effects by learning to approximate arbitrary nonlinear functions, and it ensures the uniform ultimate boundedness of the closed-loop system. Both simulation and experiment results show that the proposed control strategy can robustly track the reference path and at the same time maintains the yaw stability of vehicle at or near the physical limits of tyre friction.