Event-based NMPC for networked control systems over UDP-like communication channels

Networked controlled systems have recently received attention from the industry since they allow for flexibility and cost reduction. However, due to the fact that communication media can be subject to random delays, packet dropouts, jitters and other uncertainties, destabilization of the closed loop system can occur. Model predictive control has demonstrated to be a valid solution to cope with these issues. On the other hand, it typically relies on TCP-like (or connection oriented) protocols, i.e. either the received or the lost information is acknowledged. In this work, we propose an event-based model predictive control algorithm for nonlinear continuous time systems subject to state and input constraints which is based on UDP-like communication. We show that without the use of any acknowledgment or error message we can derive a compensation algorithm, which used in combination with the controller, under mild conditions, guarantees closed loop stability. The solution is applied to a continuous stirred tank reactor where an exothermic irreversible reaction takes place. The simulations show the effectiveness of the presented algorithm.