Neural Petri Control: an application on a mobile robot

Abstract

In the present work, an innovative nonlinear controller of nonholonomic mechanical systems, characterized by a dynamic not well known model a priori, using a new neural model obtained by the combination of a Petri net with a neural network, is proposed. The performances of the control algorithm are evaluated for tasks of tracking of time trajectories. The study of the stability of the total system to closed loop is based on the Lyapunov theory. Simulation experiments, made taking into consideration a nonholonomic mobile robot, to two wheels, allowed to verify the theoretical results.