Most capable robots combine a map-building method with local obstacle sensors. Indoor robots often use LiDAR or cameras with SLAM; outdoor mowers increasingly combine satellite positioning, RTK correction and vision.
Map, localise, plan
Navigation has three jobs: build or obtain a map, estimate the robot’s position on it, and plan a safe route. Good products constantly repeat all three rather than following a fixed line.
Indoor navigation
LiDAR gives consistent distance measurements in many lighting conditions. Cameras add object meaning and can recognise cables, shoes or pet waste, but raise different privacy and low-light questions.
Outdoor navigation
RTK corrects satellite-position errors using a local or network reference. Vision and LiDAR then help deal with trees, walls, animals and other obstacles that coordinates alone cannot explain.
Questions people ask
Does LiDAR mean a robot works in the dark?
Usually better than camera-only navigation, although obstacle-recognition features may still rely on visible or infrared imagery.
What is SLAM?
Simultaneous localisation and mapping: the robot builds a map while estimating its own changing position inside it.