Wearable EEG-based human intention detection and its application in human care-robot systems

Abstract

This research developed a wearable EEG-based brain robot interface (WE-BRI) which detects the human subject's intention using steady-state visual evoked potential (SSVEP) and is applied in the food care tasks of robots. In detail, five types of visual stimuli with different frequencies are displayed on the different locations of a PC monitor. The human subject can select his/her intended food by focusing on a related specific visual stimulus for several seconds. A wearable EEG sensor then measures the subject's EEG and followed by on line frequency analysis. From the frequency analysis, the robot can detect the human subject's intention within limited time and can perform the task to move the related food to the subject. In this research, 5 health subjects are tested using this WE-BRI. It is found that in order to increase the success rate of intention detection, the subject should focus on a specific visual stimulus for more than 4 seconds. The highest success rate of the detection can reach to 92%. The developed WE-BRI is also expected to be applied in wider range of robotic human care tasks.