Robust control of SISO systems subject to hard input constraints, Report No. 2349

The design of robust controllers that guarantee prescribed hard bounds on the control signal when facing model uncertainty is considered. Our approach to avoid a conservative design is to describe the external signals quite accuratly by means of “admissible sets”, hard bounded in amplitude and rate. The core problem to solve is the calculation of the maximum output amplitude of a set of systems. This problem will be solved for a certain parameterisation of the set in a computationally attractive way. Controller design can then be formulated by employing a double Youla parameterisation of uncertain plant and stabilising controllers. We demonstrate, how the employed model uncertainty can be identified by means of standard identification techniques. Moreover, we present a combination with an MPC scheme, that allows us to obtain faster time responses and invites the possibility to encounter for time domain and frequency domain performance specifications.