This paper presents a self-localization system using multiple RFID reader antennas and High-Frequency RFID-tag textile floor for an indoor autonomous mobile robot. Conventional self-localization systems often use vision sensors and/or laser range finders and an environment model. It is difficult to estimate the exact global location if the environment has number of places that have similar shape boundaries or small number of landmarks to localize. It tends to take a long time to recover the self-localization estimation if it goes wrong at once. Vision sensors work hard in dark lighting condition. Laser range finder often fails to detect distance to a transparent wall. In addition, the self-localization becomes unstable if obstacles occlude landmarks that are important to estimate position of the robot. Door opening and closing condition affects the selflocalization performance. Self-localization system based on reading RFID-tags on floor is robust against lighting condition, obstacles, furniture and doors conditions in the environment. Even if the arrangement of the obstacles or furniture in the environment is changed, it is not necessary to update the map for the self-localization. It can localize itself immediately and is free from well-known kidnapped robot problem because the RFID-tags give global position information. Conventional self-localization systems based on reading RFID-tags on floor often use only one RFID reader antenna and have difficulty of orientation estimation. We have developed a self-localization system using multiple RFID reader antennas and High-Frequency RFID-tag textile floor for an indoor autonomous mobile robot. Experimental results show the validity of the proposed methods.

Omni-vision system using an omni-mirror is popular to acquire environment information around an autonomous mobile robot. In RoboCup soccer middle size robot league in particular, self-localization methods based on white line extraction on the soccer field are popular. We have studied a self-localization method based on image features, for example, SIFT and SURF, so far. Comparative studies with a conventional self-localization method based on white line extraction are conducted. Compared to the self-localization method based on white line extraction, the method based on image feature can be applied to a general environment with a compact database.

No abstract

No abstract

No abstract