Authors : M. Sedraoui and S. Gherbi
Abstract: This study presents a contribution to the study of the robustness of RST controller obtained via the law of standard GPC (Generalized-Predictive-Control). The GPC is a very recent technique of control system, it is the subject of many research during these last years and allready applied successfully in industry. It is based on a predictive model CARIMA (Controlled Auto Regressive Integrated Moving Average), his role is to predict the future behavior of the system in an extended time horizon, thus it is based on a minimization of a quadratic criterion (performance criterion) to obtain a control law minimizing the error between the output of the system and the reference input. A considerable advantage of the law of GPC is that we can transform it in a RST polynomial form, which allow us to have the possibility of checking the robustness of the controller obtained via this law on the frequential domain. There are several methods of robust control which guarantee the robustness of stability and performances, among which: LQG: Linear-quadratic-Gaussian, the H2 method, the H method, these methods give very effective results but in a limited frequential domain. In the following, we will stain to show that method RST-GPC can be as robust and powerful as that introduced by the robust control methods (in particular the method H in the example which will follow).
M. Sedraoui and S. Gherbi , 2008. The Robustness of the RST Controller Obtained VIA the Law of Generalized Predictive Control . International Journal of Soft Computing, 3: 78-83.