Localization is the problem of estimating a robot's pose relative to a map of its environment.
Localization is the problem of estimating a robot's pose relative to a map of its environment. There are a number of types of robot localization. The easiest is position tracking, where the initial pose of the robot is known and the localization system must correct for small errors generated by the robot's odometry. A more difficult problem is the global localization problem, where the robot is not told its initial position and must determine it. Localization is a sub-problem of the more complex field of Simultaneous Localization and Mapping (SLAM), where the localization system must build the map and locate the robot in the map at the same time.
One of the key sub-problems is the kidnapped robot problem. After the robot has localized, the robot is transported to a different location without being informed of the transition. This problem tests the localization system's ability to recover from complete failure.