Autonomous mobile soccer robot localization using particle filter through Omni-vision

摘要

"Where am I?" is the primary question, which is the representation of localization, that the mobile soccer robot must be answered. Dead reckoning is the most popular method used in a wheeled mobile robot. However, the increasing position error is the main topic of the dead reckoning method. Furthermore, in this paper, the localization of a mobile soccer robot using a particle filter through Omnivision is proposed. The sensor model and motion model of particle filter are also discussed, where the sensor model is obtained from segmentation and feature extraction of the soccer field landmark. The experimental result showed that the proposed method estimated the robot position accurately with 15% error.