Intuitive robot navigation using wireless EMG and acceleration sensors on human arm

In this work, we propose a new method of robot navigation control through electromyogram (EMG) and acceleration sensors attached to human arms. This method allows the user to control the robot remotely by his/her intuitive motion like car steering. EMG signal processing decides whether to control the robot or not. Then, the robot takes a motion out of 4 possible motions (Forward, Backward, Left turn, Right turn) that is inferred from acceleration sensor. The accuracy of the motion discrimination has the success rate of 99% without any time delay. The whole system was implemented and we verified its utility by humanoid robot demonstration.