Reduced-order solutions to H/sub /spl infin// and LPV control problems involving partial-state feedback

We solve a parameter-dependent control problem for a linear parameter-varying (LPV) plant in which both noisy outputs and/or exact measurements of some plant states are fed back. Known bounds on the parameters' rates of variation are used to reduce conservatism. We give necessary and sufficient conditions, expressed as linear matrix inequalities (LMIs), for solvability by full- and reduced-order controllers. If the desired controller order equals the plant order minus the number of exactly measured states, then the intractable rank condition currently required for reduced-order synthesis can be avoided; moreover, this controller order is the largest needed for a solution. The solutions of the solvability LMIs are used to derive explicit controller formulae.