Abstract:The off-line robust model predictive control (MPC) offline calculates a state feedback law sequence with corresponding domains of attraction. On-line, the real-time control law is chosen as the off-line feedback law corresponding to the domain of attraction in which the current state lies. or as a linear interpolation of two offline feedback laws corresponding to two domains of attraction between which the current state lies. An existing, off-line robust MPC achieved quadratically stable off-line feedback laws. In this paper, however, the parameter-dependent Lyapunov function (PDLF) is introduced which obtains poly-quadratically stable offline feedback laws to improve the feasibility and optimality. The improved controller preserves the original controller's stability, and continuity of the control law with respect to the system state.

Improving Off-Line Synthesis Approach Of Robust Model Predictive Control