Event-based sampling strategies in networked control systems

A postulate of resource-aware design of modern control systems drives an intensified research towards more extensive usage of event-based models. The objective of most research works on event-triggered control is positioned in the context of networked sensor and control systems due to perspective of better bandwidth utilization and reducing computational complexity. The aim of this paper is to review event-based sampling strategies in terms of analytical evaluation of the mean traffic rate produced to the communication channel. The focus is put on the analysis of threshold-based sampling criteria including send-on-delta reporting strategy (Lebesgue sampling), and its extensions (send-on-area, send-on-energy). Next, the Lyapunov sampling as a scheme that addresses control system stability is surveyed. The overview of research studies shows that in many application scenaria the event-based sampling schemes can provide around one order of magnitude of traffic reduction compared to conventional periodic sampling while still keeping similar system performance.