Robust predictive control with optimal load tracking for critical applications. Final report

This report derives a multi-input multi-output (MIMO) version of a two-degree-of-freedom receding-horizon control law based on mixed H2/H∞ minimization. First, the integrand in the frequency domain representation of the MIMO performance criterion is decomposed into disturbance and reference spectra. Then the controller is derived which minimizes the peak of the disturbance and reference spectra. Then the controller is derived which minimizes the peak of the disturbance spectrum and the integral of the reference spectrum on the unit circle. The resulting two-degree-of-freedom MIMO control strategy, referred to as the minimax predictive multivariable control (MPC), is shown to have worst-case-disturbance-rejection and robust-stability properties superior to those of purely H2-optimal controllers, such as Generalized Predictive Control (GPC), for identical horizons. An attractive feature of the receding horizon structure of MPC is that it can, in ways similar to GPC, directly incorporate input constraints and pre-programmed reference inputs, which are nontrivial tasks in the standard H∞ design