Glossary term
Feedback
A control mechanism in which information about a system's output is used to adjust its input or internal behaviour.
Definition
conceptFeedback is the use of information about a system's output — or its deviation from a desired value — to adjust the system's input or control action, with the goal of reducing error and maintaining a desired behaviour.
Feedback is the architectural principle that makes automatic control possible. By continuously measuring the output and comparing it to the reference (setpoint), a feedback controller can correct for disturbances, compensate for model uncertainty, and drive the system toward the desired operating condition without requiring a perfectly accurate model of the plant. Feedback is distinct from feedforward, which anticipates disturbances using a model rather than reacting to measured error.
Feedback is the principle by which a system uses information about its own output to modify its behaviour. It is the foundational idea of automatic control: instead of prescribing every action in advance, a feedback system continuously observes what is happening and adjusts accordingly. This makes control robust — capable of responding to disturbances, uncertainties, and model errors that would defeat any purely pre-programmed strategy.
The closed-loop structure
A feedback control system has four essential elements. The plant is the process or physical system whose behaviour is to be controlled — a motor, a chemical reactor, an aircraft, a heating system. The sensor measures the output of the plant and converts it into a signal that can be processed. The controller computes a control action based on the difference between the measured output and the desired reference value. The actuator converts the controller’s command into a physical input to the plant.
The error signal e(t) is the difference between the reference r(t) and the measured output y(t):
The controller processes this error and produces a control signal u(t) that drives the plant. Because the output is subtracted from the reference — so that the loop acts to reduce the error — this architecture is called negative feedback. Negative feedback stabilises systems and reduces sensitivity to plant variations and external disturbances. It is the dominant paradigm in engineering control.
Negative versus positive feedback
While negative feedback stabilises and corrects, positive feedback amplifies deviations from equilibrium. Positive feedback is undesirable in most control applications — it is the mechanism behind oscillations and runaway processes. It appears intentionally in bistable circuits, oscillator designs, and certain biological signalling pathways where hysteresis or regenerative amplification is the desired behaviour.
Feedback across engineering domains
The concept of feedback extends across every engineering discipline. A thermostat, a cruise control system, a fly-by-wire flight control system, a voltage regulator, and a blood glucose regulation mechanism all rely on feedback to compare actual behaviour with desired behaviour and make corrective adjustments. The mathematical tools used to analyse and design feedback systems — transfer functions, stability margins, sensitivity functions, and controller design methods such as PID — are treated in dedicated entries.