The design and development of a real-time computer program for controlling the anthropomorphic six degrees of freedom robotic hand developed in the Robotics Laboratory at San Diego State Univesity is presented. The program allows the hand to be controlled either manually or via EMG signals received from electrodes attached to the user's forearm. The robot hand's joints can be operated independently or ordered to perform synergetic grasp motions. In particular, this research continues the previous work conducted at the lab of using feature extraction and classification of EMG signals in real-time to determine the user's hand position and command the robot hand to achieve the same motion. The grasp modes studied include: cylindrical, spherical, point, and lateral hand positions, and a first order rational approximation function is used to achieve the synergetic motions of the robot hand. The results demonstrate that feature extraction and classification of EMG signals serve as an adequate method for controlling the motions of a robot hand and that such control systems may be utilized in areas such as prostheses.
