Input-output feedback linearization control of uncertain systems using function approximation techniques

Abstract

In the input-output feedback linearization design, the time derivatives of the output function have to be taken successively for the presence of the control signal. Uncertainties in the system may appear in this process making the derivation infeasible. In this paper, the uncertainties are collected in each and every stage of the computation for the time derivatives. A transformation is then performed so that a simple cascade form can be obtained. In this representation, a lumped uncertainty will appear in the highest order sub-system and a function approximation based design can be employed to cover its effect. Simulation results for the input-output feedback linearization of an uncertain flexible-joint robot are given to verify the efficacy of the proposed method.