Guaranteed cost control of multimodeling systems

The guaranteed cost control problem for multiparameter singularly perturbed systems (MSPS) with uncertainties is investigated. The main contribution of this paper is that an /spl epsiv/-independent controller is newly derived by solving the reduced-order /spl epsiv/-independent slow and fast algebraic Riccati equations (AREs) which is extremely smaller than the dimension of full-order ARE. It is shown that if these AREs have positive definite stabilizing solution then the uncertain closed-loop system with the new /spl epsiv/-independent controller is quadratically stable and has the cost bound. As another important feature, our new results can be applied to the standard and the nonstandard MSPS because the nonsingularity assumption of the fast state matrices is not needed compared with the existing result.