EEG Analysis Of The Operator In The Human Machine Interaction System

Among the research processes of the human-machine interaction system(HMIS),a lot of researcher mainly do the research of machine systems,and neglect the role of the machine operators or human.In fact,a proficient operator can operate HMIS bitterly.A training process can improve the operating quality or enhance the operation level.So,it is important that characteristics of the operator in the HMIS.Because the electroencephalogram(EEG)is safe,convenient,noninvasive and cheap,this paper use EEG to reveals the characteristics of human in a HMIS.In view of the shortage of existing methods,this paper proposes the research method that includes the idea in which human are the center of the HMIS.In the research of HMIS,the video and force feedbacks and the time delay are influence factor of the study.This paper by the help of video feedback,the changes of operator's EEGs are researched during the period from a unskillful or novice to a proficient,A HMIS working platform was build that includes a virtual inverted pendulum and a joystick controller.The joystick controls the virtual pendulum with speed commands.The virtual inverted pendulum model is solved with the fourth-order Runge-Kutta algorithm,the EEG signals and pendulum data are recorded simultaneously.On the platform,an EEG data acquisition or signal sampling system was designed and built up.Subsequently,novice and proficient,who have low and high operating level respectively,were asked to operate the HMIS and their EEG data were recorded and studied,The independent component analysis(ICA)and ADJUST algorithms were used to remove artifacts to get high SNR EEG signals.Wavelet analysis algorithm is adopted to extract ??????? data respectively,then power spectrum,coherence analysis and sample entropy is utilized to analyze human cognitive behaviors in different proficient levels.Experimental results indicate the pendulum bar change in a smaller range when the operator is skilled while a relatively large range with unskilled operators.Also,compared with novice operators,the EEG signals from proficient operators have smaller sample entropy.Experiment results show that there is great difference of power spectrum in different channels between novice and proficient operators.Compared to novice operators,power spectrum change remarkably in frontal(FP)region and occipital(O)region.By wavelet coherence analysis,the difference of coherence is the most remarkable with ? rhythm.The proficient operators possess higher ability to adapt to human-machine interaction system,using their knowledge and experience to adjust to the control state accurately.This work has important significance to build novel control methods for robotic systems and rehabilitation systems.