teleoperation
Teleoperation is controlling a robot from a distance, with a human doing the deciding in real time while the robot does the doing somewhere else. The person watches through the robot's cameras and sensors, then sends commands — pushing a joystick, moving their own hand, or pressing buttons — and the robot copies those commands far away. It is the opposite end of the scale from autonomy: here the robot supplies the body and the reach, but a human still supplies the brain.
The reason to do this is usually that the place is too dangerous, too far, too small, or too delicate for a person to be there in the flesh. A bomb-disposal robot lets an operator handle explosives from behind cover; an underwater robot lets a pilot on a ship inspect a pipeline a kilometer down; a surgeon can guide tiny instruments through incisions far smaller than a hand. The human's judgment goes where the human's body safely cannot.
Teleoperation, full autonomy, and shared control form a spectrum rather than rivals. Pure teleoperation puts the human fully in charge; full autonomy takes the human out of the loop; and shared control sits between, letting the robot handle steady, fast, or repetitive parts while the human supplies high-level intent. A delay in the signal — even a fraction of a second — makes pure teleoperation hard, which is one big reason robots working far away, like on other planets, must lean on autonomy instead.
After the Fukushima accident, engineers sent teleoperated robots into reactor buildings too radioactive for people, steering them through rubble from a safe control room.
The robot goes where a human cannot, but a human is still the one deciding each move.
Teleoperation is sometimes called telerobotics, especially when the robot also does a little of its own automatic stabilizing between the human's commands.