Stabilization and disturbance attenuation of uncertain discrete-time linear systems

Discrete-time systems in state space subject to uncertainties in the state and input matrices as well as to finite-energy additive disturbances are considered. No statistical assumptions nor matching conditions are required. Necessary and sufficient conditions for quadratic stability by linear state feedback and output-disturbance attenuation below a prescribed value are derived independently of the uncertainty structure. These conditions are then applied to the case of convex polyhedral uncertainly and the robust synthesis problem is cast into a mathematical programming problem. Additional constraints like robust pole assignment in subregions of the complex plane can easily be fit into the design procedure. Numerical examples are provided.