Non-Iterative Distributed Model Predictive Control with Event-Triggered Communication

This paper examines event-triggered communication of state information in linear time-invariant distributed control systems subject to constraints implemented over communication networks using non-iterative distributed model predictive control (DMPC). The approach utilizes a combination of an event-trigger, a neighbor model and a predictor in every subsystem. The predictor and neighbor models are updated with new measurements in an event-triggered fashion meaning that the subsystems only communicate when it is required thus leading to a reduced communication effort. Recursive feasibility of the control law and convergence of the closed-loop system are ensured using a robust controller design. A simulation example illustrates that the proposed method can reduce the network utilization significantly.