Recently, studies of power assistive device controlled by noninvasive Brain-Machine Interface (BMI) have been developed. However, The discrimination accuracy for the control of power assist device by Electroencephalograph(EEG) signals is still lower than electromyograph(EMG) signals. One reason is that the control system can’t discriminate the differences of different mental states exactly. This study is aiming to analyze the EEG signals related to tense in operating state when wearing an exoskeleton robot. Here, We try to design an experiment to investigate EEG variations when subjects feel the different intensities of tense state.
