Robust model predictive control under limited capacity communication constraints

This paper deals with robust model predictive control (MPC) of an input and state constrained linear system, when the system and controller are separated by a digital communication channel of limited capacity. By extending some existing results within the robust MPC literature, the MPC scheme proposed in this paper employs time-variant "safety margins" within the finite horizon optimization problem, in order to ensure "robustness" against the uncertainty introduced by the communication channel. Stability and feasibility of the proposed method are established. The theoretical results are illustrated with a simple example, and some conclusive remarks are made, regarding the computational complexity, and future extensions of the proposed MPC scheme.