Self-localization is critical in the mobile hu-manoid soccer robot systems. Robot can use the robot vision system to position itself, to recognize objects, to measure distance, and to find the coordinate of the objects. In this paper, we propose a visual self-localization algorithm for humanoid soccer robot. The proposed self-localization me-thod has the advantages of reducing complexity and higher localization ratio. Furthermore, the performance of the self-localization method can be easily improved by searching time. The experimental results indicate that the self-localization method can increase the sample accurate rate at the localization
Proceedings of the 2008 CACS International Automatic Control Conferenc, 5p.