This work presents the IRKA method for reduction or approximation of control systems according to the ${\cal H}_2$ norm. The basic definitions of control system theory are presented first, followed by the basic idea of system projection methods. Next, the connection between projection and interpolation of transfer function is proven and presented. Then the ${\cal H}_2$ norm is presented. Several results associated with the ${\cal H}_2$ norm, including Meier-Luenber conditions for ${\cal H}_2$ optimal approximation of the system are presented. These conditions are a link between the approximation of systems and the interpolation of transfer functions. Finally the IRKA method and the proof of convergence in the case of stable SSS systems are presented. In the end, several examples of using the IRKA method in practice are shown.
|