A feedback loop is a system that adjusts its behavior based on measurements of its output. The output feeds back into the input. If the output is too low, the system increases effort. If it is too high, it decreases effort. The loop keeps the system near the target.

Feedback loops are everywhere in robotics. A motor controller measures position and adjusts voltage. A drone measures tilt and adjusts rotor speeds. A robot arm measures force and adjusts grip. Without feedback, the system runs open-loop and drifts with changing conditions.

Parts of a feedback loop

• Sensor: measures the output.
• Comparator: compares measurement to setpoint.
• Controller: computes the correction.
• Actuator: applies the correction.
• Plant: the system being controlled.

Negative feedback stabilizes. Positive feedback amplifies. Most control systems use negative feedback because the goal is to reduce error, not increase it. The gain of the loop determines how aggressively it corrects. High gain corrects quickly but can oscillate. Low gain is stable but sluggish. Tuning the loop is a core skill in robotics. It is also a useful way to think about systems in general. Any process that measures its results and adjusts is running a feedback loop. The concept is simple. Getting it right is not.