Abstract
This paper proposes a robust design of the discrete time model reference adaptive control system in the presence of modeling errors and shows the results of the feasibility test applied to position control of a DC-motor. In this scheme, the dominant parameters of the discrete time plant-model expressed in terms of δ-operator are estimated by an adaptive law with dead zone and then the current estimates are used to determine the controller parameters by calculation. In the synthesis of the control input, the output error is fed back through a fixed compensator to reduce the tracking error due to the parameter adjustment-mismatch. Stability of the adaptive system is analytically proved to be guaranteed by setting the width of the dead zone of adaptive law appropriately. The feasibility test has verified that the proposed scheme can realize satisfactory control performance in spite of the presence of modeling errors.
