抄録
The paper proposes a method for acquisition of knowledge by interaction between human and robot for constructing intelligent environment. At first, a robot acts according to instruction from human, and acquires knowledge of environment for action from the instruction. If the instruction is too defectiveness for the robot to act, the robot requires more detailed instructions from human. Then the robot stores the acquired knowledge into the RF-ID tags attached to the objects in the environment. As the robot receives various instructions, various properties of the objects are gradually accumulated into the RF-ID tags. Finally, the robot will be able to act without detailed instruction from human, since the robot can obtain the knowledge for the action from the RF-ID tags on the spot where the robot should act.
