Robust explicit MPC based on approximate multi-parametric convex programming

Many robust Model Predictive Control (MPC) schemes require the on-line solution of a convex program, which can be computationally demanding. For deterministic MPC schemes, multi-parametric programming was successfully applied to move most computations off-line. In this paper we adopt a general approximate multi-parametric algorithm recently suggested for convex problems and propose to apply it to a classical robust MIPC scheme. This approach enables one to implement a robust MPC controller in real time for systems with polytopic uncertainty, ensuring robust constraint satisfaction and robust convergence to a given bounded set.