Abstract
Authors have already given state feedback decoupling algorithms which are based on Hessenberg forms. But it is often impossible to place poles which correspond to the intersection subspace of controllability subspaces. For controllable linear systems, by adding additional dynamics, it is possible to place all poles while preserving decoupling. Several papers have already given design methods for such a dynamic compensator. But these are not suitable for computer implementation. This paper gives a new design method for a dynamic compensator which stabilizes the decoupled system. Since this method places only unstable poles, the order of the compensator is lower. This method is based on the QR method and the standard controllability form. Accordingly, it is more efficient and accurate than the other methods.
