抄録
It is known in psychology that the human visual space and kinesthetic space do not completely coincide. Human subject's response undershoots the hand reaching point in depth under the visual control of reaching movement without vision of his limb. In this paper, six neural network models of sensory integration of binocular view space and kinesthetic space using hand position are presented to find how the joint angles of the right arm and eye movements are integrated into the human space perception. Visual reaching experiments revealed that one of the presented models showed a characteristics which was similar to those of the human subjects. The structure of the model is equivalent to the structure of human sensory integration. In this sensory integration, the signal space for the perception of the positional coincidence is that of the arm joint angles. The signals of eye movements are transformed into the equivalent signals of the arm joint angles which are similar to the signals for the arm control. It suggests that such a structure is organized for the needs of the arm control in this type of sensory motion control.
