Abstract
Self-localization is one of the most important subjects for mobile robots. The robot position is a primary information for collision avoidance, path planning, decision making for RoboCup soccer robots. The environment of RoboCup robots changes dynamically as other robots are always moving, and the accuracy of self-localization have been affected by lighting condition, wheel slipping, etc, so that single robot position estimation has limitation. In the proposed method, the ball is set as the landmark, and by sharing the landmark positions of other robots, the robot estimates its position. The proposed algorithm is introduced into robots and evaluated in the robot field.
