A quasi-infinite horizon nonlinear model predictive control scheme with guaranteed stability (Q1298283)

A computationally attractive quasi-infinite horizon nonlinear model predictive control (MPC) scheme is introduced. In the proposed MPC, as in conventional MPC schemes, the closed-loop control is calculated by solving the optimization problem for each time step on-line. The setup differs from the standard MPC. A terminal state penalty term is added to the finite horizon objective functional and an additional terminal inequality constraint has also to be satisfied. The special linear state feedback loop is used to determine these quantities off-line. As a result, the proposed nonlinear MPC has a quasi-infinite prediction horizon, but the input profile which is determined on-line is of finite nature. The proposed MPC scheme guarantees asymptotic stability and is computationally attractive as shown by the given numerical examples.