Physical AI Lab // Episode 1.2 // Quiz
Control: PID and control rate
Getting the motors to catch up with the target, steadily
01
A robot vacuum follows a pre-programmed route and reads no sensors. On a rug its wheels slip; it drifts off the route and carries on anyway. What kind of case is this?
Closed loop, just with P set too low
Closed loop, just with the control rate too low
Open loop: it doesn't measure, so it can't correct
Open loop, and adding I would bring it back on track
02
The inverted pole in the page no longer wobbles, but it always settles just short of the target. Which term should you adjust first?
Raise P further so it pushes harder
Add a little I to make up the gap
Raise D so it brakes even harder
Lower the control rate to calm the response
03
The pole keeps shooting past the target and being pulled back, oscillating without settling. What is the most likely situation and fix?
The target angle is wrong; set a new target
D is too high; lower D so it moves more freely
The control rate is too high; turn it down
P is too high; add a little D to damp it
04
A well-tuned PID is balancing the inverted pole. After you turn the control rate right down, it sways and falls. Why?
Corrections are so far apart the error grows
At a low rate, the P, I and D terms reset to zero
A low rate saves power, so the motor gets weaker too
PID can only be computed at one thousand hertz
05
You want to cite “about 90% of industrial controllers are PID-type” in a report. Which wording is most appropriate?
One study estimates about 90% are PID-type
Official vendor statistics show 90% of controllers are PID
PID can solve 90% of industrial control problems
With 90% on PID, it suffices for complex systems too
Check