Event-Triggered Control
Event-triggered control updates the control action only when a state-dependent condition signals that an update is needed, saving computation and communication.
Beyond periodic sampling
Traditional digital control samples and updates on a fixed clock, whether or not anything has changed. Event-triggered control instead monitors the system continuously and recomputes the control only when a triggering condition fires, typically when the error between the current state and the state at the last update grows too large. During quiet periods no updates occur, conserving processor cycles and network bandwidth.
The triggering rule
A common rule triggers an update when the norm of the measurement error exceeds a fraction of the norm of the current state. This threshold is derived from a Lyapunov analysis so that, between events, the Lyapunov function still decreases enough to guarantee stability. The rule guarantees performance close to continuous feedback while updating far less often.
Avoiding Zeno behavior
A key requirement is a positive minimum time between events; without it the trigger could fire infinitely fast, which no hardware can serve. For many linear designs a strictly positive minimum inter-event time can be proven, ruling out this Zeno behavior. The related self-triggered scheme goes further, computing the next update time in advance so continuous monitoring is unnecessary.
- Update only when a state-dependent condition fires
- Threshold derived to preserve Lyapunov decrease
- Reduces computation and network load versus periodic control
- Must guarantee a positive minimum inter-event time
Event-triggered control is especially valuable in networked and distributed systems where communication is the scarce resource, and in embedded systems where processor time is limited. It trades a small performance margin for large savings in resource use.
For a design-stage distributed control network with many nodes sharing a link, event triggering would cut message traffic while maintaining stability, evaluated in simulation. It is a leading approach to resource-aware control.