A systematic and complete method to design robust predictive controlle
rs for unconstrained linear systems is proposed. The synthesis procedu
re is based on rigorous theoretical foundations, without resorting to
approximations or ad hoc design guidelines, yet it remains a viable to
ol for practical application. A significant feature is that the robust
predictive controller retains the servo performance of a nominal pred
ictive controller designed using conventional methods. In addition, th
e robust predictive controller can be designed to guarantee perfect st
eady-state rejection of asymptotically constant disturbances. The robu
st design method is developed for systems affected by unmodeled dynami
cs and is based on solving a discrete-time model-matching problem. It
is shown that the robust controller can be classified legitimately as
a predictive controller because it minimizes the same performance func
tional as the nominal predictive controller. An illustrative design ex
ample is given.