selftriggered

Self-triggered, in the context of control theory, refers to a sampling strategy where the controller determines the next update time autonomously based on the current state and a plant model. It is part of the broader family of event-driven approaches and is often contrasted with traditional time-triggered (periodic) control and with purely reactive event-triggered schemes that rely on external conditions to trigger updates. The aim is to reduce unnecessary sampling and communication while maintaining desired stability and performance.

The typical self-triggered design uses a model of the plant and a performance or stability criterion, such as a Lyapunov function, to predict how the system will evolve after an update. From the present state, the controller computes the latest time at which the next update must occur to preserve stability, bound the tracking error, or limit disturbance amplification. The computed time becomes the next scheduling point, and updates occur only at these self-determined instants. In practice, the resulting inter-sampling time is a state-dependent quantity that may be conservative or adaptive to changing conditions.

Applications of self-triggered control appear in networked control systems, embedded and cyber-physical systems, and multi-agent coordination, where communication and computation resources are constrained. It is closely related to event-triggered control, with differences in how triggering instants are generated and enforced. Extensions address disturbances, model uncertainties, delays, and packet losses to preserve guarantees under more realistic operating conditions.