Downloads & Free Reading Options - Results

An indirect adaptive control algorithm has been studied to control physical systems with parameter uncertainties. Although the particular algorithm investigated is applicable to a wide class of discrete time linear systems, parameter convergence and therefore global stability of the whole system is guaranteed only if the external excitation contains a sufficiently large number of frequency components. This thesis has also investigated the possibility of stopping the identification procedure when the parameter error becomes sufficiently small, so the controller automatically turns itself from adaptive to time invariant, while still guaranteeing global stability of the closed-loop system. In this way the adaptive controller might be activated only when its gains do not provide satisfactory performances. Computer programs of the indirect adaptive control have been written for simulation purposes using both recursive least squares and projection algorithms. (Author)