world frame vs body frame
These are the two most important coordinate frames in robotics, and the whole trick is that one stays still while the other moves. The world frame (also called the global or fixed frame) is bolted to the environment — a fixed corner of the room or a spot on the map — and it never budges, no matter what the robot does. The body frame (also called the local or moving frame) is glued to the robot itself and travels everywhere the robot goes, turning when the robot turns. They are simply two different vantage points for describing the same scene.
The difference shows up the moment a robot gives a command. 'Drive to the door' is a world-frame thought — the door has a fixed spot on the map. But 'turn left and go forward two metres' is a body-frame thought — 'left' and 'forward' mean whatever the robot is currently facing, so the very same command sends it somewhere different depending on which way it started. It is exactly the gap between 'meet me at the north entrance' (world) and 'it's on your right' (body): one is anchored to the map, the other to you.
Good robots fluently translate between the two, dozens of times a second. A camera on the robot naturally sees the world in the body frame (the cup is up-and-to-the-left of the lens), but the planner usually thinks in the world frame (the cup sits at a fixed map location). Converting a measurement from one frame to the other is one of the most common operations in all of robotics, and forgetting which frame you are in is one of the most common bugs — a robot that confuses 'left of me' with 'west on the map' will wander off in a confidently wrong direction.
A drone hovering in gusty wind reports its ground velocity as (0, 0, 0) in the world frame — it is holding its spot over the ground — yet its airspeed is far from zero: the body frame feels wind rushing past, and the motors work hard to cancel each gust so the drone stays put.
Zero ground velocity in the world, but a strong airflow felt in the body: the same drone, two viewpoints.
Many systems use several body frames at once — one for the base, one for the arm's wrist, one for the camera — all moving, all kept track of relative to the single fixed world frame.