A portable artificial robotic hand controlled by EMG signal using ANN classifier

This paper aims at building a portable robotic hand for physically disabled people to perform basic hand movements. Surface Electromyography(EMG) signal is collected from muscles of human forearm to extract the subject´s intentions of action, where six kinds of gestures are selected for discussion. An Artificial Neural Network(ANN) is trained and utilized to distinguish the desired movement according to the features picked up from the myoelectric signal. A simple robotic hand with seven degrees of freedom has been built and hardware circuits including signal acquisition, power management, and microprocessor are designed with no wire connecting to computer, making it compact and convenient to use. At last, experiments have been conducted to verify the validity of the whole system. The results show an efficient and relatively accurate recognition performance of this work.