particle-filter-based localization algorithm [28] was applied,
which was provided by the robot software. The localization
algorithm minimizes errors by fusing data from wheel encoders
and the data from an LRF, along with a preloaded map of the
room. Before the experiment, a map of the room with wall
positions was created, and this blue print was saved within the
robot. The accuracy of the position was within 5 cm with both
robots, which was estimated by repeated measurements. The
corrected position data of the robot were recorded in a log file.