A new adaptive second-order non-singular terminal sliding mode lateral control combined with neural networks for autonomous vehicle

Abstract

This paper presents a novel adaptive second-order non-singular terminal sliding mode control combined with a radial basic neural network structure and a triangular neural observer to model and control an autonomous vehicle. Firstly, the dynamics of the vehicle are presented. Secondly, the control strategy is designed, more precisely unmodelled dynamics of the system are estimated by using artificial neural networks to model them and inject them into the control law taking into account the system stability by using the Lyapunov function. Thirdly, as it is not always possible to access all the state variables representing the system, an auxiliary dynamic system, called an observer combined with the neural networks is used, to estimate the state of some dynamic variables of the vehicle which are basically not measurable. Lastly, the efficiency and superiority of the proposed method are proved by simulation results performed using MATLAB.