Abstract :
What is a decisive difference between a mechanical hand and our human hand? The robot hand can not be a versatile tool, even if it is made to resemble a human hand exactly in shape and mechanism. Even at the present age of robotics, it is too clumsy to perform a variety of ordinary tasks that a human encounters in his or her everyday life. In this talk, I claim that the clumsiness manifests the lack of knowledge of everyday physics and as well the lack of scheme for designing an artificial CNS (Central Nervous System) for the robot so as to cope with its sophisticated interplays with environments of everyday tasks. Here, the term "everyday physics" is used as a scientific domain related to accountability of dexterous accomplishment of ordinary tasks through manipulating things, with or without sensing and with or without consciousness as seen in ordinary human life. Here, design of an artificial CNS should be a domain of science and engineering that should be called "human robotics". It should be a core of robotics that attempts to unveil secrets of human intelligent behaviors from perception to action and vice versa. Actually, a half century ago N. A. Bernstein was enthusiastic in perceiving the mysteries of dexterity of human movements with redundant Degrees-of-Freedom regardless of illposedness of inverse kinematics.
