Estimação robusta em controle adaptativo.

Abstract

Estimation algorithms used in adaptive control systems should be able to detect and track plant parameter variations. They should also be able to deal with unmodelled dynamics and measurement noise, yielding good estimates even in the presence of periods of low excitation. This work aims at the aspects of robust estimation in adptive control. Here, modifications are made in the incremental least squares algorithm in order to increase its parameter tracking ability and decrease the effects of the disturbances. Two adaptive controllers are presented that use the resulting estimation algorithm. The first one uses the conventional pole placement technique. The second one uses a frequency domain approach for controller redesign. Simulation examples are presented illustrating the performance of the new algorithm and its potential use in adaptive control applications.