Event Based Controller
Event Based Controller
In time-triggered control the sensing, control, and actuation are driven by a clock.
In event-triggered control the input is held constant, not periodically, but while
performance is satisfactory.
• The controller or the control loop is connected to the system via sensors . The sensors
continuously monitors the output of the system.
• If the output signal is well under the limit then the control loop will remain disconnected or we
can say the controller will not act.
• If there is any error sensed by the the sensor then it sends a signal to the controller which then
get connected and start acting as feedback loop to resolve the issues.
• So here the sensor’s act is time triggered but the controller act is event triggered.
Time triggered Event triggered
AD
EVENT PID
AD LOGIC
Input
signal
Using the error e(t ) = x (τi ) − x (t ) we can rewrite the closed loop dynamics as: time
x˙ (t ) = Ax (t ) + BKx (τi )
= Ax (t )+BKx (t ) − BKx (t ) + BKx (τi )
= (A + BK )x (t ) + BKe(t )
Since A + BK is a stable matrix there exists a Lyapunov function V whose derivative along the closed
loop dynamics satisfies:
V˙ ≤ −a ǁxǁ2 + b ǁxǁǁeǁ (1)
Assume now that we can enforce the inequality:
ǁeǁ ≤ σ ǁxǁ (2)
for some σ satisfying −a + bσ < −a’ with a’ > 0. It would then follow from (1) and (2):
V˙ ≤ −aǁxǁ2 + bǁxǁ ǁeǁ ≤ −aǁxǁ2 + bσǁxǁ2 ≤ −a’ǁxǁ2
This suggests that ǁeǁ ≤ σǁx ǁ could be used to decide when to execute the control task. By
executing the control task at t = τ we obtain e(t ) = x (τ )-x (t ) = 0 which implies (2).