An Adaptive Tracking Controller For A Mobile Robot Using Neural Networks

Abstract

In this paper a design procedure for an adaptive tracking controller for a mobile robot subject to kinematic constraints is presented. The dynamics of the mobile robot is assumed to be completely unknown, and is on-line identified using neural network based estimators. Both the form of the controller and the adaptation laws of neural network weights are derived from a Lyapunov analysis of stability. Under certain conditions, the tracking stability of the closed loop system, and the convergence of the neural network weight updating process are guaranteed. No preliminary learning stage of neural network weights is required. Computer simulations conducted in the case of a mobile robot with two independently actuated wheels are included to demonstrate the performances of this neural network controller.