Feedback control theory: The idea
In plain words
The idea explained without jargon
Feedback control keeps something on target by measuring it and correcting it. A thermostat measures the room, compares the reading with the setting and switches the heating to shrink the difference. That difference is the error, and the controller turns the error into a correction. The hard part is not reacting but reacting well: a correction that is too strong or too late can make the system swing wider and wider, so stability matters, and Maxwell derived the conditions for it in 1868 [1]. Minorsky's study of 1922 recommended mixing proportional, derivative and integral action [1]. Black showed in 1927 that negative feedback can reduce distortion at the cost of reducing overall gain [1].
as of 2026-11-12
Key equation
The equation that sums it up, and what each symbol means
- Equation
- e(t) = r(t) - y(t); u(t) = Kp * e(t) + Ki * (integral of e(t) dt) + Kd * (rate of change of e(t))
- Symbols
- {"r(t)":"the wanted value, such as the set temperature","y(t)":"the measured value, such as the room temperature","e(t)":"the error: wanted minus measured","u(t)":"the correction sent to the actuator, such as the heater power","Kp":"how strongly the controller reacts to the error now (proportional action)","Ki":"how strongly it reacts to the error built up over time (integral action)","Kd":"how strongly it reacts to how fast the error is changing (derivative action)"}
- Note
- This is the three-term (PID) controller, whose mix of proportional, derivative and integral action Minorsky recommended in 1922 [1]. Feedback theory as a whole has no single formula; its core result is a test for stability, which Maxwell first derived for governors [1].
as of 2026-11-12
An analogy
A picture from everyday life that makes the idea easier to hold
Steering a car down a lane is feedback control: your eyes measure how far you are from the line, and your hands correct. Steer too hard and the car weaves, too softly and it drifts. The picture stops being true because a driver also looks ahead and plans, whereas a basic controller only reacts to the error it measures now and its recent history, and delay in the loop can make it overshoot.
as of 2026-11-12
Sources
- A History of Automatic Control C. C. Bissell (chapter in a Springer handbook, ISBN 978-3-540-78830-0; Lund University course copy)