An improved off-line dynamic output feedback robust MPC design for ellipsoidal estimation error sets

For the system with both polytopic uncertainty and bounded disturbance, an improved off-line approach to dynamic output feedback robust model predictive control for an ellipsoidal estimation error set is considered. To reduce the on-line computational burden, a look-up table is constructed off-line for on-line searching the real-time controller parameters. In the off-line optimization stage, for the pre-specified estimated states and the fixed ellipsoidal estimation error set, a sequence of controller parameters and the corresponding regions of attraction are calculated and stored in a look-up table. Compared with the previous results, the additional estimated state constraint for future time in the off-line optimization is removed from the existing algorithms, which decreases the conservatism of the off-line optimization problem. In the on-line searching stage, the real-time controller parameters are searched in the look-up table according to the region of attraction with the closest containment of the real-time estimated sate. The searched controller parameters can guarantee the convergence of the real-time augmented state. A numerical example is given to illustrate the effectiveness of the approach.