Similarities and differences between exosystem-model-based disturbance observer and model error compensator

This paper clarifies the theoretical similarities and differences between the exosystem-model-based disturbance observer and the model error compensator. In 1972, Johnson proposed the exosystem-model-based disturbance observer. This observer is constructed against the augmented system model, which consists of nominal plant and disturbance-generating models. Ohnishi et al. proposed to cancel out the effect of uncertainties with its estimated disturbance. Furthermore, in 2013, Okajima et al. proposed the model error compensator, which compensates for the error between the output of the nominal plant model and the measurement outputs of the actual plant. The theoretical differences among these control structures have yet to be fully understood. Therefore, this paper theoretically investigates the relation and shows the superiority of the exosystem-model-based disturbance observer to the model error compensator in its freedom to assign the behavior of the uncontrollable modes. This superiority enables the exosystem-model-based disturbance observer to suppress output fluctuations against uncertainties effectively. This effect is shown by a numerical example.