The Control Method for the Robot Hand Based on the Fuzzy Theory

By moving muscles, the myogenic potential (Electromyogram: EMG) is observed on the surface of the living body. It is considered that the EMG is useful for controlling a robot hand. However, the EMG depends on physical conditions, the state of mind and so on. So, the original EMG will be not used for controlling the robot hand directly. In this study, it is considered that the EMG relating the motion of the human hand is analyzed by the fuzzy theory for making the robot hand performs the same motion as the human hand. EMG were measured under the following conditions. (1) opening the hand, (2) bending the thumb, (3) bending the middle finger, (4) bending the index finger, (5) closing the hand, (6) not move. Six production rules were made with fuzzificate data resulted from fourier transforming the EMG (30-band 1/3 octave analysis). Also the EMG measured by experimental motion of the human hand was transformed into the fuzzificate date. Rates of recognitions were calculated in comparison with the six production rules and the experimental data. And one production rule with highest rate of recognition was used for recognition of movement of the human hand in the computer. From the experimental results, about 90% of movement were recognized by the computer. The results were applied to control the robot hand.