Gain scheduled controllers with guaranteed linearization properties

Gain scheduling based on plant linearizations about equilibria implicitly imposes a linearization requirement on the controller. This paper presents an approach for gain scheduling that yields controllers that automatically possess the appropriate linearization property and that can also assume a linear parameter-varying (LPV) form. The approach is a modification of the so-called 𝒟 method described in the literature wherein the need for time-derivatives of certain controller inputs is avoided. A gain scheduled controller is designed for the ball and beam experiment to illustrate the technique