Model-based low-order control design for mechanical servo-systems

Detailed models of mechanical servo-systems can be used for low-order control design if we invoke closed-loop reduction in an iterative process of model reduction - optimal control synthesis - controller reduction. The twin feedback configuration is the key to formulate closed-loop reduction for both the mechanical model and the controller. Closed-loop balanced reduction is a straightforward extension of balanced reduction that can be applied in each step of the reduction process. Using a dedicated MATLAB toolbox, the change in performance after reduction can be monitored and by means of graphical manipulation of extra frequency weights the reduction process can be controlled. In many examples the final controller was close to the optimal full-order controller. The proposed approach is a step towards optimal fixed-order control of high-order systems.