Event-Driven Model Predictive Controller for State Constrained Systems: An Input Signal Reconstruction Method

A novel event-driven model predictive control (MPC) technique is constructed for perturbed nonlinear system including both input and state constraints. The primary merit of the developed method is the event-driven input signal is constructed based on sparse control samples, and only the selected samples, rather than a continuous control signal, need to be transmitted through the network. Given such a framework, a tightened constraint is designed to satisfy the robust requirement, and an event-driven scheme is designed to lessen the computational as well as communication load of the considered MPC system. Then, theoretical requirements for guaranteeing the MPC feasibility and system convergence are figured out. Finally, proposed input signal reconstruction based MPC method is tested via simulation experiments as well as comparative study.