Robust Tracking by Reduced-order Disturbance Observer: Linear Case

The disturbance observer (DOB) is one of the widely-used control methods for robust tracking under plant uncertainties and external disturbances. Typically it is implemented by two (so-called) Q-filters as well as an inverse model for the nominal plant. Observing that these two Q-filters have identical dynamics, we propose a reduced-order implementation of the conventional DOB configuration. Moreover, we analyze this newly proposed configuration and claim that the robust stability condition, which has been found for the conventional configuration, still holds for the reduced-order case, and the robust transient performance can also be maintained as before with a saturation function introduced in the feedback loop.