Research On Mental Fatigue Detection Method Based On Flexible Wearable Two Channel EEG Acquisition Equipment

Mental fatigue will cause inattention and cognitive response speed to reduce,cause serious harm to social production and living.The pilot and the driver’s mental fatigue easily lead to major accidents,is considered to be one of the main causes of catastrophic accidents.It is of great significance to explore a feasible fatigue detection method for avoiding human error in various work and eliminating safety hidden trouble in fine operation.Among the numerous fatigue detection methods,EEG has become a hot research direction because it can directly reflect the state of brain activity and its irreplaceable advantages in the field of wearable fatigue detection.Existing EEG acquisition equipment is mainly aimed at scientific research and medical purposes,and has problems such as complex operation and poor wearing experience,so it cannot be applied to fatigue detection in actual scenes.The design of a few wearable EEG acquisition devices mostly relies on the flexible EEG cap to achieve its wearable performance,which still has shortcomings such as many wires and the need to carry a signal acquisition box.Due to the application requirements of EEG in fatigue detection,we hope that the EEG acquisition device can collect EEG signals from the hairless areas of the human body as much as possible,with smaller volume and fewer leads.Therefore,a flexible wearable dual-channel wireless EEG acquisition scheme was proposed in this paper.This scheme abandoned the traditional design of wire and acquisition box,and achieved a wearable EEG acquisition device with convenient wearing,less limited use scenarios and high comfort while ensuring the accuracy and accuracy of acquisition.There are some inherent problems in using dual-channel EEG acquisition equipment to collect human frontal EEG,such as fewer channel leads and introducing a large number of electroocular artifact,which brings some difficulties to preprocessing frontal EEG signals.On forehead brain electrical characteristics,this paper puts forward two kinds of artifact removal based on complete empirical mode decomposition method,respectively is the use of constant false alarm rate judge the position and remove eye electric CEEMDAN-CFAR algorithm and using complexity as artifact after ICA decomposition by interpolating CEEMDAN-ICA algorithm,two kinds of algorithm is checked by practice,all can better remove the eye electricity in the brain electrical signal artifact component and low frequency noise,at the same time,to a great extent,has kept the details of the original signal,proved to be of two kinds of effective low channel EEG preprocessing method.Many studies have proved that fatigue dichotomy based on EEG can achieve high accuracy,which verifies the feasibility of fatigue detection method based on EEG.However,algorithms with better classification performance often rely on multi-channel EEG data or a large number of calculations.Since this study was carried out on a dual-channel wireless EEG acquisition device with a small number of channels,it is more necessary to combine the advantages of features to obtain a higher classification accuracy.In previous studies,researchers often choose a certain kind of feature extraction algorithm to explore the neural mechanism of fatigue,which makes the author of this paper unable to obtain the dominant features applicable to the forehead according to previous research data.Therefore,in this paper,a variety of algorithms including power spectrum,frequency domain features,nonlinear dynamics,wavelet entropy and so on were used to extract the fatigue EEG features,and the recursive feature elimination method based on SVM model was used to obtain the dominant features.In order to verify the reliability of the dominant features,two fatigue models,emergency night doctor and fatigue driving,were selected in this paper for comparison,and six relatively stable dominant features were finally found.The selected advantage features may be applied in the future rapid fatigue detection based on frontal EEG.However this study take advantage of classification of fatigue characteristics,is the relative fatigue state change of binary classification based on machine learning,how to get the competitive advantages for the fatigue degree of quantitative is still a difficult problem,especially in a lot of research on brain electrical characteristics changing with the fatigue of the universality of the law has not been unified conclusion.In order to solve this problem,it may be necessary to further explore the changes of EEG characteristics to reveal the neural mechanism of fatigue.On the basis of searching for the dominant characteristics,this paper further explores the change rule of the dominant characteristics,hoping to provide some help for the later exploration of the correlation between EEG characteristics and fatigue.