Neural Network Based Sliding Mode Lateral Control For Autonomous Vehicle

Nowadays, autonomous driving represents a major challenge for automobile manufacturers in order to reach the latest levels of autonomy. Any autonomous vehicle development project focuses on three fundamental phases; environmental perception, trajectory planning and path pursuit which including control and command as an integral part. This paper presents a modified Sliding Mode Controller based on the Radial Basic Function Neural Networks (SMC_RBNN) able to control the lateral dynamics of the vehicle. For a sinusoidal reference path, the proposed control strategy, SMC_RBNN, showed better results than those obtained with a conventional Sliding Mode Controller (SMC), in terms of lateral tracking error.