Robust constrained predictive control using comparison model (Q705445)

A robust model predictive control (MPC) method for constrained linear systems with disturbances and model uncertainties is proposed. In addition to the nominal reference model, a ``comparison'' model for the worst-case analysis (which is based on a robust control Lyapunov function for unconstrained system) is incorporated into the MPC scheme. The ``comparison'' model enables the authors to transform the robust MPC problem to a nominal one without uncertain terms, and to obtain the missing terminal constraint through the use of linear inequalities. Based on the terminal constraint, a condition for initial states (under which the ultimate boundedness of the closed loop is guaranteed without violating state and control constraints) is derived. Since the terminal condition is given by linear constraints, the original min-max optimization problem is reduced to the standard quadratic programming problem.