Abstract
In this article, the case is considered where a regulated output of an uncertain system with unknown disturbance input is constrained to lie between two specified time-varying bounds. It is supposed a controller has already been designed using any desired technique. This controller may not satisfy expected output constraints. The objective is to design saturations on the control signal so that the closed-loop is altered only when those specifications are not met. The theory is presented in the linear case with linear dependence on an unknown disturbance. Application to an uncertain launcher linear model with unknown wind disturbance is presented.