Optimization of control transmissions by event-driven model prediction

In this paper we study the optimization of control transmissions in event-driven control systems, with respect to the control performance and number of communication events. The control signal transmissions are decided based on optimization in a model predictive framework, called Model Predictive Triggering (MPT). MPT is developed because the typical send-on-delta approach does not take into account control performance. Several transmission cost functions criteria are compared. We show and compare the performance of the different alternatives using simulation, including a conventional model predictive control approach. The MPT method is shown to be the effective considering the transmission count and control performance, with tuning parameters to select a suitable trade-off between communication and control.