MODEL-BASED EVENT-TRIGGERED PREDICTIVE CONTROL FOR NETWORKED SYSTEMS WITH DATA DROPOUT

This paper proposes a novel structure of networked control systems (NCSs) with communication logic, which incorporates model-based networked control systems (MB-NCSs), predictive control, and an event-triggered communication scheme into a unified framework to consider the bandwidth reduction of the network communications. Within this framework, first an event-triggered communication scheme at the sensor side is introduced to determine whether or not the sensor measurement should be transmitted to the controller through the imperfect forward paths. Second, at the controller, a model of the plant is used to predict future state behavior of the plant, and based on the predicted state between successful transmission instants, a novel predictive event-triggering scheme is proposed to compress the size of the packetized control signals transmitted from the controller side to the actuator side through feedback paths. Finally, a unified model of NCSs is established. Based on this model, a codesign condition of the two event-triggering schemes and state feedback controller gain in terms of linear matrix inequality will be derived, using a common quadratic Lyapunov approach. The advantage of the model presented in this paper in comparison to the existing event-based models is that it can compensate for the effect of data-dropout on the closed-loop system actively while requiring less bandwidth without sacrificing stability of the closed-loop system, which will be demonstrated by a practical example at the end of this article.