In event-driven control approaches, control updates are triggered by event conditions that are often characterized by different types of boundaries defined in the state-space domain. In this paper we define an event-driven control approach where boundaries are manifolds characterized as invariant sets. With such a boundary, a control update will only be activated when the system trajectory intersects the boundary. And the system trajectory must intersect again the boundary to activate the next control update. For linear systems, we derive an scaling property for these boundaries that permit to regulate the accuracy of the control without altering the timing offered by the scaled boundary.

Peer Reviewed