An efficient approach for EMG-based robot control

Electromyography (EMG) is a physiological signal generated during muscle contraction. It implicates, to certain extent, the human intention for movement. Being intuitive and natural, it has been applied for the control of the rehabilitation device, human-assisting robot, and others. However, due to the time-varying, highly nonlinear characteristics inherent in the EMG signal, its analysis and processing poses severe challenges. In this paper, we propose a simple and effective approach for an upper arm EMG-based robot control system, in which the EMG signals measured from the human upper arm are used to govern the robot manipulator. To achieve real-time motion governing, we adopt a very simple method, which selects certain critical values for motion classification, to determine the mapping between the upper arm EMG signals and corresponding robot movements. Both the empirical method and fuzzy system are used to select the critical values. And, experiments have been performed to validate the proposed approach, in which the fuzziness present in the individuals has been investigated and tackled.