SLAM: How Robots Map a Room They Have Never Seen
SLAM builds a map and works out the robot’s place in it at the same time, using what the sensors see as it moves.
The robot repeats this loop many times a second. Tap a step to pause on it.
To know where it is, a robot needs a map. To make a map, it needs to know where it is. SLAM solves both together.
As the robot moves it matches new sensor readings with the map so far, adds new walls and features, and corrects its position when it recognises a place it has seen before.
A vacuum or warehouse robot can enter a new building and start working, with no one drawing a map first.
It would rely on counting wheel turns, which drift, and the map and position would slowly fall apart.
Empty corridors, moving people and mirrors confuse SLAM, and large maps need a lot of computing.
What does SLAM do at the same time?