Hamilton–Jacobi inequality robust neural network control of a mobile wheeled robot

Abstract

The work presented here presents a new approach to determine a robust neural network follow-up motion control of a mobile wheeled robot. The solution is a result of a Hamilton–Jacobi inequality, enabling synthesis of control of a non-linear object in terms of input to output stability. By applying Lyapunov’s theory of stability, it was demonstrated that all signals are limited, while the determined control provides a relatively high accuracy of actuated motion. The weights of the neural network were updated in real time and online. The produced simulation solutions confirmed the efficiency of the approach contemplated here.