A control method of power-assisted robot for upper limb considering intention-based motion by using sEMG signal

As power-assisted robots such as exoskele-ton robot have been widely used in eclectic applications, the robot becomes more interactive than industrial robots. More specifically, the power-assisted robot for rehabilitation requires to enhance power with respect to the intended motion. To do that, the power-assisted robot should recognize which part of interaction is based on human-intention. In this paper, a new classifier, which consists of force information measured by F/T sensor on the robot and sEMG signals from muscle activation, is proposed to extract human-intention under interaction including external force. The proposed classifier can be applied to estimate the external force level generated due to the interaction. Based on the proposed classifier, a simple control method to enhance power to assist the intention-based motion is developed to validate the proposed approach. For the simplicity and clarity of the approach, 1DOF testbed robot is used to demonstrate the proposed approach.