Adaptive neural network tracking control of an omnidirectional mobile robot

Abstract

The trajectory tracking control problem of an omnidirectional mobile robot is studied in this article. The omnidirectional mobile robot is adsorbed by suction cups for aircraft skin inspection, which is a typical nonholonomic system. In the control design part, a novel adaptive neural network control scheme is presented in the presence of uncertainty and external disturbance. The adaptive neural network has the characteristics of weights online updating. The neural network is applied to estimate uncertainty online to obtain the desired tracking performance. The weights online updating algorithm contains a correction term, which is an improved algorithm to ensure robustness. On the basis of Lyapunov theory, the closed-loop system can converge to an arbitrarily small domain containing origin. This illustrates that the closed-loop system is globally asymptotically bounded stable. Excellent control performance can be obtained by selecting design parameters reasonably. A simulation example of tracking an eight-shape trajectory is given to verify the effectiveness of the proposed control scheme.