抄録
In this paper, we propose a method for constructing a teleoperating system of a humanoid robot that combines the controllability and autonomy.We use a motion capture system to operate the humanoid robot. The intension of the operator is estimated using Hidden Markov Model to generate the movement of the humanoid robot.We make the six types of Hidden Markov Model about reaching behaviors, and set the target position of the hand each behavior.First, the humanoid robot generates the reaching behavior based on operator's behavior. Second, it moves automatically reaching toward the target position based on the result of estimation. We did some experiments about that using humanoid HRP-4.
