Improved optimisation approach to the robust H2/H∞ control problem for linear systems (original) (raw)

A strategy for robust H 2 /H 1 state-feedback control synthesis, with regional pole placement, applied to continuous-or discrete-time linear time-invariant uncertain systems is presented. It is based on a multiobjective optimisation over the space of the controller parameters. In the case of systems with polytope-bounded uncertainty, the H 2 and H 1 norms, calculated in all polytope vertices and in possible 'worst case' interior points are taken as the optimisation objectives. An a posteriori exact cost norm computation based on a branch-and-bound algorithm is applied for closed-loop performance assessment. The proposed strategy is applied to continuousand discrete-time examples, including the design of a decentralised controller, and the results are compared with LMI-based formulations. E.N. Gonçalves is with the