エアホッケーロボットにおける対戦者のデータに基づく攻撃動作決定

抄録

In this paper, we propose an algorithm for an air-hockey robot to decide its optimal motion based on data of human responses. The trajectories of the puck in a lot of games by humans are recorded, and they are approximated by Gaussian Mixture Models. The model is used to predict human motions, and it is integrated into a 2-step look ahead algorithm. The proposed method is verified by numerical simulation.