Uma proposta de controlador-preditor baseado em observadores aumentados para sistemas com atraso na entrada

Abstract

In this work, a control structure based on augmented observer with a scheme predictor is proposed for systems with input delay. An analysis of the characteristics and performance of the proposal is carried out through a new methodology based on filtered perturbations, moreover comparisons are made with other techniques in the literature. This strategy allows to mitigate completely perturbations of known dynamics at steady state from the output of the system and is also capable of dealing with some types of modeling error, either in the process or in the disturbance dynamics. For this, an augmented state observer is used that includes a filter dynamics and a predictor that uses observed states to obtain predictions in time and therefore compensates the feedback branch delay. Tuning is done from obtaining the state feedback gain and the observer gain, which can be obtained using a synthesis procedure based on LMIs defined in the work. The proposed tuning is done using a closed-loop system approximation through to consider diferential delay equantions and then, using Lyapunov-Krasovski functions, it’s obtained a convex optimization problem that reduces the H∞ norm of relation of the perturbation to output, therefore it obtains the tuning and guarantees the stability of the closed-loop system. The proposed control structure is checked against presented techniques and, even though it is simpler and more versatile, has superior performance.