Localization of mobile robots by multiple landmark recognition

The paper proposes a method for localization of mobile robots by multiple landmark recognition. Landmarks are recognized in real time by Model-based matching method using Genetic Algorithms (GA). A mobile robot measures directions of landmarks while moving. Target landmarks are changed depending on the situation of the robot. The individuals of GA are allocated for recognizing multiple landmarks efficiently depending on the situation. The robot estimates its position by the directions of the two landmarks and its moving distance. We perform experiments using a mobile robot in a RoboCup soccer field. In the experiment, goalposts and corner posts are used as landmarks. Experimental results show the effectiveness of the method.