Robust model predictive control for nonlinear systems based on Volterra series
Master of Science
In this thesis we develop a Nonlinear Robust Model Predictive Control (NRMPC) algorithm for nonlinear plants modeled by second order Volterra series. Robust stability is achieved through the addition of cost function constraints that prevent the sequence of the optimal controller costs from increasing for the true plant. Model uncertainty is parameterized by ellipsoid bounds on the plant parameters. The same approach is used to reject constant output disturbances. If some certain assumptions concerning the disturbance and plant parameters are satisfied, the plant is guaranteed to reach an offset free steady state.