“…A widely used technique is RBF neural networks (RBF NNs) which approximates a nonlinear function up to some error tolerance. Amongst the works that use RBF NNs are, in 1999, Ge Besides, in many real cases, the system states are not available for measurement so, if the system can be expressed in normal form [15] and in presence of dynamics uncertainties, high-gain observers [4], [6], [11], [14], are the most suitable option to estimate the unavailable states of the plant. The disturbance rejection property of high-gain observers makes the adaptive controller more robust against the errors induced by compensation of the system uncertainties.…”