How a robot 'sees' the world
A robot has no built-in understanding of the room it's in. Everything it knows comes from sensors converting the physical world into numbers it can process. A camera turns light into pixel values. A microphone turns air pressure changes into a waveform. An ultrasonic sensor turns the time it takes sound to bounce back into a distance. None of these are 'seeing' in the human sense, they're just streams of numbers that software has to interpret.
This is why the sensor a robot uses shapes what it's capable of. A robot with only a bump sensor can tell it hit something, but not what it hit or how far away it was. A robot with a camera can recognize a face, but might completely miss a pane of glass right in front of it. Choosing sensors is really choosing which parts of reality your robot is allowed to know about.