You hop into the shower. Eee, that’s cold! Your hand flips it up to hot. A few seconds (or minutes, depending on your geographical location and the power of your water heater) later and now it’s omg hot hot hot. Somewhere in there you manage to flip it back to a tolerable level and finish your shower.
You are actually acting as a sophisticated controller/feedback system. Your skin is serving as a temperature sensor and feeding back information to your brain that tells you if it’s too hot or too cold. Your hand is serving as the motor that uses signals from your brain to tell you which direction to flip the nozzle in and by how much. Control systems are everywhere. They control everything from the total travel and position of a complicated robot arm in an automotive factory to the flap position of an aileron on an aircraft you fly.
The purpose of a control system is to alter the dynamic behavior of a system or reduce the effect of disturbances. Your car’s anti-lock brakes, cruise control, and emissions system all use some form of a controller. In a position control experiment the effect is obvious. The controller can specify the speed of the motor or positioning device based on the feedback it gets from the system. Some controllers are simple, or proportional controllers. Others are PD controllers: proportional derivative. Or the more complicated PID: proportional integral derivative.
Negating oscillatory behavior of mechanical devices is one common use of a PID controller and might take you from something that looks like this:
To something that once you’ve implemented the feedback control looks like something more like this:
The overshoot is when your hand moves too far on the nozzle and in an attempt to make the water warmer you make it too warm as well as trying to not have to spend too long oscillating the nozzle to find the ideal spot. So think of that next time you’re in the shower and the next time you slam your foot on the brakes or watch a commercial jetliner adjust its wing or tail flaps. The human brain as the first sophisticated control algorithm!

Woo! Bring on the Bode plots!
PID has it’s place and is very good as a control system. However the maths can be hard to het your head around and the system will need careful setting up. Yes it’s very good but has it’s place.
Using PI or P control will often give enougth control where some overshoot is allowed, not critical or filtered by the process under control.
When designing a control system the designer must consider whith method is best.
We refer to man in the loop as “Meat Servos”…
I like the view of the human controller.
A human also has excellent feed-forward control systems.
For example, when closing a door, you measure the friction of the door and deduce the amount of force it requires to close it without slamming it too hard all in a fraction of a second.