Brain–Computer Interface for Fuzzy Position Control of a Robot Arm by Mentally Detected Magnitude and Sign of Positional Error

The paper addresses a novel approach to position control of a robot arm by utilizing three important brain signals, acquired with the help of an EEG interface. First, motor imagery signal is employed to activate the motion of a robotic link. Second, the error-related potential signal is acquired from the brain to stop the motion of the robotic link, when it crosses a predefined target position. Third, the approximate magnitude of the positional error is determined by steady-state visual evoked potential signal, acquired by noting the nearest flickering lamp that the robotic link has just crossed. The novelty of the present research is to decode the approximate magnitude of the positional error. Once the approximate magnitude and sign of the positional errors are obtained from the mental assessment of the experimental subject, the above two parameters are fed to a fuzzy position controller to generate necessary control commands to control the position of the end-effector of the robotic link around the predefined target position. Experiments undertaken confirm a low percentage of overshoot and small settling time of the proposed controller in comparison to those published in the current literature.