2A2-F04 A clustering method to estimate friend or foe for Robocup Middle Size League(RoboCup and Robot Contest)
直也 平尾, Hiroyuki MIYAMOTO · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2011
RoboCup is an international project to promote AI,robotics,and related field. "Hibikino-Musashi" is a team participates in Middle Size League of Robocup. "Hibikino-Musashi" robots have a camera as sensor and they communicate mutually. They estimate position of themself and obstacles from omni-directional camera images while playing succer and understand friend position from communication. But they don't identify obstacles are friend or foe. If positions of friend and obstacles are accurate data, identification is very easy but positions of obstacles are inaccurate data because omni-directional camera has attribute about distance error. In this paper approach that clustering method of position data of obstacles to identify obstacles are friend or foe discount distance error and create world model of soccer field where friend and foe are to communicate data of position of itself and obstacles.