Stabilization of Neuro-Control Structure using Lyapunov Functional Based Approach

Abstract

In this paper, a neuro-control structure is proposed for the speed control of a nonlinear motor drive system. The neural network is trained to learn the inverse dynamics of the considered system from observation of the input-output data. After achieving the training process, a direct adaptive control approach with a model following controller is performed. The Lyapunov BackPropagation algorithm (LBP) is developed and used to adjust online the neural network so that the neural model output follows the desired one and maintains the stability of the neurocontrol scheme for a large variation of the motor speed drive system. The obtained simulation results verify the effectiveness of the developed Lyapunov BackPropagation algorithm to have a fast error convergence and highlight its performance to guarantee the stability of the designed approach