The majority of localization algorithms start at a known position and add internal movement data and
external environment data to this position each cycle. If the robot isreplaced or the sensor data quality is too
low, these algorithms are usually not able to recover to a useful position estimation Members of these so-called
local approaches are the linear least squares estimator and the Kalman filter. Robots equipped with global
localization algorithms like Markov localization and particle filter are able to localize themselves even under
global uncertainty. This Article focuses on local and global localization, static environments andpassive
approaches. Active approaches have to be discussed along with the decision making. To be able to cope with
dynamic environments, map building is necessary. Both topics are not within the scope of this work.