A neuro-sliding control approach for a class of nonlinear systems

Abstract

This paper proposes a learning method for the compensation of uncertainties, for a class of nonlinear systems. A sliding model control strategy is used for the robust control design after a prior stable learning phase. Gaussian networks are used to identify the uncertainties during this learning phase. Learning and control bounds are guaranteed by properly constructing the training structure. The proposed technique has been validated using a hardware example case of an electromechanical system. Experiments have shown that the inclusion of the proposed learning technique in the robust control design results in improved system performance.