Controle por modelo de referência de sistemas incertos baseado no algoritmo

Abstract

In this dissertation, a higher order sliding mode controller is proposed for uncertain and monovariable systems with relative degree one. The solution is based on the super-twisting algorithm and on the model reference adaptive control paradigm, considering only output-feedback. The control algorithm is implemented using variable gains to deal with a wide class of uncertainties/disturbances. First order approximation filters are used to enable a solution that do not depend on the availability of the system states for the control implementation. The proposed strategy guarantees exact tracking to a reference model with global convergence propoerties in finite time. Furthermore, this approach based on the model reference adaptive control provide a natural structure for the application of nominal values to uncertain parameters of the plant, providing a control effort reduction. Simulation results exemplify the effectivity of the proposed control strategy. In addition, a comparison to the recently proposed in the literature output-feedback variable gains super-twisting algorithm solution is made. In the latter, the control is obtained using the normal form of the system and estimating the norm of unmeasured states.