Output Constraint Handling in Analytical MPC Algorithms Based on Hammerstein Models with Presumed Trajectory of Future Control Changes
Novel numerically efficient, analytical MPC algorithms based on Hammerstein (nonlinear) models are proposed in the paper. They use the idea consisting in the assumption that the shape of a trajectory of future changes of a control signal is presumed in advance. In the proposed algorithms it is relatively easy to take output constraints into consideration on the entire prediction horizon. Thus, the constraint handling is very efficient because the control action can be appropriately modified many sampling instants before the potential constraint violation can occur.