Networked event-triggered control of spatially distributed processes using a dual-mode communication strategy

This work focuses on model-based networked control of spatially-distributed processes modeled by uncertain nonlinear parabolic PDEs with multiple spatially-distributed actuator/sensor configurations that communicate over a resource-constrained network. Based on an approximate finite-dimensional model of the infinite-dimensional system, a set of stabilizing distributed model-based controllers are initially designed. Communication between the control configurations is realized through an event-triggered strategy which suspends and restores the sensors' transmission over the network when certain stability thresholds are met or violated. The communication strategy is dual-mode in that it switches between two complementary modes of transmission: a decentralized mode in which local communication triggers control the broadcast events of a given configuration, and a centralized mode in which the broadcast events of all configurations are coordinated by a high-level supervisor in response to a global communication trigger. It is shown that the decentralized mode reduces the network load during the transient phase, whereas the centralized mode ensures ultimate boundedness of the closed-loop trajectories under non-vanishing perturbations. The results are illustrated using a simulation example.