A novel robust adaptive neuro-sliding mode steering controller for autonomous ground vehicles /

Abstract

This paper intends to harness recent advances in intelligent commands to develop a novel steering control strategy for autonomous ground vehicles (AGVs) maneuvering problems. In this research, the vehicle is embodied by a single track-model (ST) taking into consideration its lateral dynamics. Besides, it is subject to unknown disturbances and uncertainties, associated with the tire/road adhesion coefficient, which are upper bounded by a positive limit. An on-line Radial Basis-function Neural-Networks (RBNN) adaptive mechanism is designed to estimate this upper limit in the sense of Lyapunov. On this basis, a novel robust adaptive neuro-sliding mode steering angle controller (AN-SMC) is introduced. Whilst the effectiveness of the controller is impacted by its parameters, a new neural-network optimization algorithm (NNA) is employed to provide optimal values for these settings. The proposed approach is evaluated through a single lane-change-maneuver planned out by a new methodology. The steering controller revealed a powerful path-tracking via the comparisons carried out in support of the developed strategy.