ROBOT WITH AUTOMATIC SELECTION OF TASK-SPECIFIC REPRESENTATION FOR IMITATION LEARNING

摘要

PROBLEM TO BE SOLVED: To provide a method for controlling a robot excellent in imitation learning. SOLUTION: The method includes steps of: observing a movement of an entity in the environment of a robot; recording the observed movement using a sensorial data stream and representing the recorded movement in a different task space representation; and selecting a part of the task space representation for the imitation learning and reproducing the movement to be imitated. The part of the task space representation includes a step of including one task space representation or including a series of different task space representations. COPYRIGHT: (C)2010,JPO&INPIT