open-loop vs closed-loop control

These are the two basic ways a machine can carry out a command. Open-loop control means commanding blindly: you send out an instruction and simply trust it works, never checking the result. Closed-loop control means measuring what actually happened and feeding that measurement back in to adjust the command until reality matches the goal. The vivid contrast is making toast two ways — setting a timer and walking off (open loop: it might come out pale or burnt, you'll find out later) versus standing there watching the bread and pulling it the instant it turns golden (closed loop).

Open loop is simpler, cheaper, and faster because it needs no sensors and never second-guesses itself — a microwave running for ninety seconds, a sprinkler on a clock, a stepper motor told to take exactly two hundred steps. It works beautifully as long as the world behaves exactly as assumed. But it has no idea when it is wrong: if the food was frozen, the soil already wet, or the motor slipped, an open-loop system sails on cheerfully producing a bad result, because the 'loop' from output back to input is left open and broken.

Closed loop costs more — it needs sensors, comparison, and careful tuning — but it earns that cost by staying accurate in a messy, changing world. The word loop refers to the circle of information: command goes out, the sensor reads the result, that reading comes back and reshapes the next command, closing the circle. Almost every robot that has to be precise or safe runs closed loop; open loop survives where the task is cheap, predictable, and forgiving.