Control of myoelectric prosthetic hand based on surface EMG

In this paper, a robot hand with two fingers, thumb and index finger, aiming at application to a myoelectric prosthetic hand was built to implement fundamental motions, such as grasping and holding, required in daily life. In order to control each finger of the developed robot hand independently, an identifier which recognizes the finger motions based on the surface EMG using the neural networks was constructed, and recognition rate of the finger motions was examined. Experiment for online finger operation of the built robot hand was carried out using the identifier, and fingers of the robot hand were able to be operated as intended.