CN114224359B - Electroencephalogram fatigue index judging method for high-speed rail dispatcher - Google Patents

Electroencephalogram fatigue index judging method for high-speed rail dispatcher Download PDF

Info

Publication number
CN114224359B
CN114224359B CN202111556302.3A CN202111556302A CN114224359B CN 114224359 B CN114224359 B CN 114224359B CN 202111556302 A CN202111556302 A CN 202111556302A CN 114224359 B CN114224359 B CN 114224359B
Authority
CN
China
Prior art keywords
electroencephalogram
dispatcher
fatigue
index
wave
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202111556302.3A
Other languages
Chinese (zh)
Other versions
CN114224359A (en
Inventor
张光远
王亚伟
王灿
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Jiaotong University
Original Assignee
Southwest Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest Jiaotong University filed Critical Southwest Jiaotong University
Priority to CN202111556302.3A priority Critical patent/CN114224359B/en
Publication of CN114224359A publication Critical patent/CN114224359A/en
Application granted granted Critical
Publication of CN114224359B publication Critical patent/CN114224359B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A61B5/372—Analysis of electroencephalograms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A61B5/372—Analysis of electroencephalograms
    • A61B5/374—Detecting the frequency distribution of signals, e.g. detecting delta, theta, alpha, beta or gamma waves
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device

Abstract

The invention discloses an electroencephalogram fatigue index judging method for a high-speed railway dispatcher, which comprises the following steps of: s1, selecting an experimenter to perform a high-speed railway driving scheduling experiment; s2, constructing an electroencephalogram processing data set; s3, constructing an overall dispatcher brain electrical fatigue model; s4, judging the brain electrical fatigue indexes of the individual schedulers. The invention combines the operation and physiological characteristics of the high-speed railway dispatcher, and obtains the brain electric fatigue index capable of reflecting the reliability of the overall dispatcher and the individual dispatcher, thereby providing a more scientific method and theoretical basis for the fatigue judgment of the high-speed railway dispatcher and ensuring the safety and the reliability of railway transportation.

Description

Electroencephalogram fatigue index judging method for high-speed rail dispatcher
Technical Field
The invention relates to the field of high-speed railway dispatching, in particular to an electroencephalogram fatigue index judging method for a high-speed railway dispatcher.
Background
The high-speed railway dispatcher is used as a central of the dispatching command system, and can effectively avoid decision errors and driving safety accidents caused by fatigue by measuring the fatigue degree in the working process, thereby having important significance on the safety reliability of the whole railway transportation system. In the current fatigue measurement method, electroencephalogram measurement is considered as the most accurate way of measuring fatigue. Researchers monitor the extent of fatigue by brain waves (EEG) and can judge the mental status of a person (delta wave, 1-3.5Hz, theta wave, 3.5-7.5Hz, alpha wave, 7.5-12.5Hz, beta wave, 12.5-30 Hz) from four typical brain waves.
Currently, the related research on the brain electrical fatigue index in the field of high-speed rail schedulers is less common. In other fields, the main discussion of researching fatigue indexes by utilizing brain electricity includes the relation between a power spectrum mean value and fatigue degree, the relation between fatigue characteristics and brain electricity signals, the effectiveness of brain electricity equations in fatigue and the like. The existing research modes mainly aim at a certain aspect of the electroencephalogram index, and the difference between the population and the individual is ignored, so that the reliability of fatigue indexes obtained by most electroencephalogram researches is not strong.
Since the reliability of the electroencephalogram fatigue index changes when the fatigue of the population and the individual is reflected, there is a very good literature to study the electroencephalogram index affecting the fatigue of the population and the individual of the high-speed railway dispatcher respectively. Therefore, it is highly desirable to provide a research method for the brain electrical fatigue index of the high-speed railway dispatcher, find the reliability brain electrical index which can reflect the fatigue of the overall and individual high-speed railway dispatcher, and provide a more scientific method and theoretical basis for the fatigue judgment of the high-speed railway dispatcher.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides the method for judging the brain electrical fatigue index of the high-speed railway dispatcher, combines the operation and physiological characteristics of the high-speed railway dispatcher, and obtains the brain electrical fatigue index capable of reflecting the reliability of the overall dispatcher and the individual dispatcher, thereby providing a more scientific method and theoretical basis for judging the fatigue of the high-speed railway dispatcher and ensuring the safety and the reliability of railway transportation.
The aim of the invention is realized by the following technical scheme: an electroencephalogram fatigue index judging method for a high-speed railway dispatcher comprises the following steps:
s1, selecting an experimenter to perform a high-speed railway driving scheduling experiment;
s2, constructing an electroencephalogram processing data set;
s3, constructing an overall dispatcher brain electrical fatigue model;
s4, judging the brain electrical fatigue indexes of the individual schedulers.
In the step S1, 32 national iron group dispatchers with ages of 28-38 are selected to train staff as experimental staff, and the selected staff meets the following requirements: the sleeping bag has the advantages of no psychological or mental diseases, good sleeping quality, no color weakness or color blindness, and proficiency in mastering the dispatching operation skills of the high-speed railway; the functional beverage of caffeine or alcohol cannot be drunk in the evening before the experiment;
preferably, since the working time of the high-speed rail dispatcher is generally up to 12 hours, the daytime tiring period 11 of the high-speed rail dispatcher is selected in step S1: 00-16:00 an experiment was performed.
In the step S1, the process of performing the high-speed rail formation scheduling experiment includes:
s101, 24 hours before the start of an experiment, enabling 32 tested schedulers to be in an easy fatigue state, and explaining the operation of an experiment flow, an experiment task and an experiment platform to the tested schedulers;
s102, 15 minutes before the start of the experiment, enabling a tested dispatcher to firstly perform a simulation experiment, and familiarizing with the actual working environment and specific working tasks of the dispatcher;
s103, after the experiment is started, working events randomly occur in the system every 15 minutes, a tested dispatcher executes simulated dispatching operation tasks including conversation operation, operation and recording operation according to railway technical management rules, and monitoring operation is executed in the rest time; recording the original brain electrical data of the tested dispatcher at the moment by using a 64-channel Neuroscan brain electrical instrument, and simultaneously recording KSS (K-S) scale values, wherein each tested is 20 times in total;
s104, after the experiment is finished, removing impurities and noises from the data, removing power frequency electricity and myoelectricity interference, preprocessing the data, and laying a foundation for the subsequent construction of an electroencephalogram data set.
Meanwhile, the work events which occur randomly are faced to in addition to the monitoring task in the experimental process of the tested dispatcher, so that the brain electrical load of the tested dispatcher when the tested dispatcher monitors and the brain electrical load of the tested dispatcher when the tested dispatcher faces to the random events are calculated by using a DORATASK method based on the work task. The brain electrical load is used as a part of a fusion output end and as brain electrical reaction time in brain electrical fatigue index analysis of a subsequent individual dispatcher when an overall dispatcher brain electrical fatigue model based on a neural network is constructed:
the DORATASK method mainly converts load pressure values of a dispatcher into working time to measure, and calculates an electroencephalogram load value T during monitoring of the high-speed railway dispatcher through the following steps w And brain electrical load value T in the face of random events p ：
Figure BDA0003419164610000021
Wherein T is 1 Monitoring time, T 2 Talk time, T 3 -operation recording time, T 4 -brain recovery time, delta T-recording time interval taken as 15 minutes; the mental recovery time is set to 2-4 seconds according to the characteristics of the high-speed rail dispatcher, and the tested dispatcher is recorded every 15 minutesBrain electrical load during monitoring;
Figure BDA0003419164610000022
wherein T is p1 Reaction time, T p2 Wave time, T p3 The recovery time is taken as 200ms and the Δt-recording time interval is taken as 15 minutes; the stimulation time of one electroencephalogram experiment is 900ms, the sampling rate is 1000Hz, and the electroencephalogram load of a tested dispatcher facing a random event is recorded every 15 minutes.
Preferably, in step S2, considering that the slow wave of the brain electrical signal gradually increases and the fast wave gradually decreases when the brain transitions from the normal state to the fatigue state, wherein the slow wave includes a delta wave, the theta wave includes an alpha wave and a beta wave, and the delta wave does not appear in a deep sleep state; only alpha, beta and theta waves are considered; alpha wave, beta wave, theta wave peaks and amplitudes under different electrodes are extracted through an electroencephalograph, and power P of three waveforms is extracted through Fourier transformation α 、P β 、P θ On the basis, the variation coefficients of the three waveforms are extracted and calculated by using a variation coefficient model to form an electroencephalogram index variation coefficient data set, and the obtained electroencephalogram power is used to form an electroencephalogram index equation data set.
Preferably, said step S2 comprises the following sub-steps:
s201, collecting 20 brain wave data of each tested dispatcher to obtain 20 groups of brain wave data, wherein each group of brain wave data comprises alpha wave, beta wave and theta wave data; removing impurities and noise, removing interference of power frequency electricity and myoelectricity, and adding an original electroencephalogram index data set;
s202, carrying out Fourier transform on the denoised brain wave data to obtain the power P of three waveforms of the brain wave in each acquired data α ，P β ，P θ The fourier transform is:
Figure BDA0003419164610000031
x(n)＝x a (nT)| t＝nT (T＝1/f s ,f s ≥2f c ) (4)
Figure BDA0003419164610000032
the sample points collected in the formula are N, wherein N is the total number of the sample points collected in each experiment, N is 20T according to the experiment requirement s =nt denotes the duration of the finite length signal, x a (t) is the finite length signal after filtering, x a (T) duration of T s The highest frequency is f c ；
S203, calculating the occurrence rate of three waveforms in each acquisition process:
Figure BDA0003419164610000033
s204, calculating the mean square error of the occurrence rate of the three waveforms in each acquisition process:
Figure BDA0003419164610000034
s205, calculating variation coefficients of three waveforms in each acquisition process to form an electroencephalogram index variation coefficient data set:
Figure BDA0003419164610000035
in order to establish the relation between the original electroencephalogram data and the variation coefficients and fatigue degrees of the three waveforms, providing a curve change rate index; by describing the curve of the brain wave index change, including the brain wave curve of theta, alpha and beta waves, and the occurrence rate curve and the variance curve obtained by Fourier transform based on a variance coefficient model, and assuming that two points of P, Q are any two points on three curves, the unit daily curve change rate is expressed as:
Figure BDA0003419164610000041
the relationship between the degree of fatigue and fatigue is represented by a mapping:
F(R)＝{A|A:R→[0,1]} (10)
wherein the domain R is defined as fatigue, and the fuzzy subset of the fatigue degree is A, mu A Membership function of A, wherein
Figure BDA0003419164610000046
x→μ A ；
The fatigue degree is represented by a curve change rate index, and the curve change rate index at this time refers to a unit daily change rate index, and the relationship between the fatigue degree of the dispatcher and the unit daily curve change rate is characterized by:
Figure BDA0003419164610000042
for experimental sample X on any curve describing brain wave index change i ＝p*(x 1 ,x 2 ,...,x 20 ；y i ) The method comprises the steps of carrying out a first treatment on the surface of the Wherein y is i The amplitude of the waveform in the ith sampling is represented, p is the number of people of the tested dispatcher, and the number is taken as 32; when taking the interval [ a, b ]]At the time, for any point coordinate P (epsilon, delta), epsilon>0，δ>The rate of change of 0 is expressed as
Figure BDA0003419164610000043
The change curve of the occurrence rate of the alpha wave and the theta wave is in an ascending trend, and the change curve is increased along with the deepening of the fatigue degree; the change curve of the occurrence rate of the beta wave shows a decreasing trend, and the change curve decreases with the deepening of the fatigue degree; the state of the dispatcher is continuously changed under the vigilance and relaxation under the influence of working fatigue and decision-making behaviors in the fatigue resistance stage; meanwhile, the alpha CV, the beta CV and the theta CV are greatly changed along with the increase of the fatigue degree, which shows that the fatigue degree is obviously influenced by the variation coefficients of three waveforms and corresponds to the change curves of the occurrence rate of three typical waves;
in addition, for the data obtained after each acquisition and denoising, the brain wave power P of the three waveforms obtained by fourier transform in step S202 α ，P β ，P θ And forming an electroencephalogram index equation data set by using the ratio and summation ratio of the three powers: θ/β, θ/(α+β), (θ+α)/β, (α+θ)/(α+β).
Figure BDA0003419164610000044
Preferably, the step S3 includes:
taking ten electroencephalogram indexes of alpha wave, beta 0CV, beta 1CV, theta CV and theta/beta, theta/(alpha+beta), (theta+alpha)/beta, (alpha+theta)/(alpha+beta) in an original electroencephalogram index data set and a coefficient of variation data set of the electroencephalogram indexes as input ends, and integrating subjective values, namely KSS values; electroencephalogram load value T during monitoring in experimental process w And brain electrical load value T in the face of random events p As an output end, a two-layer neural network model containing hidden layers is established, and the number of neurons of the hidden layers is 10;
the transfer function of the hidden layer of the neural network is selected as a sigmod function, which can be expressed as an f function
Figure BDA0003419164610000045
The transmission function of the output layer is selected as a purelin function, and the training function is selected as a tranlm function;
for a given m training samples { (x (1), y (1)), (x (2), y (2)), (x (m), y (m)) }, d (i) is the desired output of the corresponding input x (i), given an error function of:
Figure BDA0003419164610000051
let W be ij B is the weight of the connection between the jth neuron and the ith neuron of the hidden layer i For the bias of the ith neuron, net i Is the input to the ith neuron. The BP algorithm calculates weights and offsets according to the following formula:
h(i)＝f(net i ) (15)
Figure BDA0003419164610000052
let W be T As a weight matrix, a i For inputting characteristic parameters, the output result can be expressed as
Figure BDA0003419164610000053
The judgment idea for the reliability of the brain fatigue index of the overall dispatcher is as follows: dividing the average value of the tested dispatcher data into a training set, a verification set and a test set, and obtaining the best BP neural network model through training a model and adjustment; and judging ten electroencephalogram indexes by using the model, taking an R value representing the correlation degree between the electroencephalogram indexes and fatigue and an MSE mean square error representing the difference between a sample predicted value and an actual value as criteria, and if the R value of the index is closer to 1 and the MSE is smaller, considering that the reliability of the index is higher when the electroencephalogram fatigue of the overall dispatcher is judged.
Preferably, the step S4 includes:
s401, randomly selecting electroencephalogram indexes and electroencephalogram reaction time in the experimental process of n high-speed rail scheduling experimenters as statistics (X) 1 ,Y 1 ),(X 2 ,Y 2 ),...,(X n ,Y n ) The method comprises the steps of carrying out a first treatment on the surface of the Assuming that independent variables and dependent variables have linear correlation, setting X and Y as samples extracted from two different populations X and Y, wherein X is from ten electroencephalogram indexes in building a population electroencephalogram fatigue model, and Y is from a calculated electroencephalogram load value T of a tested dispatcher in the face of a random event p . With an observation value of x 1 ,x 2 ,...,x n And y 1 ,y 2 ,...,y n Pairing them to form (x 1 ,y 1 ),(x 2 ,y 2 ),...,(x n ,y n ) Checking whether the electroencephalogram index is related to the electroencephalogram reaction time or not;
let us assume H 0 : x and Y are uncorrelated.
Figure BDA0003419164610000054
H 1 : x and Y are related.
S402, calculating test statistics, spearman grade correlation coefficient r s The method is used for measuring the important index of the correlation degree of two samples, and the calculation formula is as follows:
Figure BDA0003419164610000055
wherein r is s The value range of (C) is [ -1,1]When |r s The closer to 1, the higher the correlation between samples is; conversely, when |r s The closer to 0, the lower the correlation between samples is;
s403, making a decision, when
Figure BDA0003419164610000056
When rejecting H 0 Conversely, H cannot be rejected 0 。
Therein, wherein
Figure BDA0003419164610000061
Is a critical value, related to the number of observations of the sample n, the alternative hypothesis and the given level of significance, when n takes 8 ∈>
Figure BDA0003419164610000062
The judgment idea for the reliability of the brain electrical fatigue index of the individual dispatcher is as follows: calculating the correlation coefficient between the electroencephalogram index and the electroencephalogram reaction time of each tested dispatcher by a Spearman rank test method, and transversely comparing the correlation coefficient of each type of electroencephalogram index to obtain the correlation coefficientThe index with a larger value can be considered to have higher reliability when judging the brain electrical fatigue of the individual dispatcher; if the correlation coefficient between a certain electroencephalogram index and the electroencephalogram reaction time is larger than
Figure BDA0003419164610000063
Then hypothesis H may be rejected 0 Accept H 1 The electroencephalogram index is considered to have correlation with the electroencephalogram reaction time.
The beneficial effects of the invention are as follows: according to the invention, the brain electrical fatigue index research experiment is carried out by combining the operation and physiological characteristics of a high-speed rail dispatcher, so as to construct a brain electrical index research total system; and under the condition of a variation rule of brain wave indexes based on a variation coefficient model, respectively carrying out brain wave fatigue index analysis on an overall dispatcher and an individual dispatcher by using a neural network model and a spline rank test model to obtain brain wave fatigue indexes capable of reflecting the reliability of the overall dispatcher and the individual dispatcher, and obtaining the brain wave fatigue indexes capable of reflecting the reliability of the overall dispatcher and the individual dispatcher, thereby providing a more scientific method and theoretical basis for fatigue judgment of the high-speed railway dispatcher and ensuring the safety and the reliability of railway transportation.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a flow chart of an experimental procedure of the present invention;
FIG. 3 is a flow chart for constructing an electroencephalogram processing dataset;
fig. 4 is a graph showing waveforms (left), occurrence (middle) and coefficient of variation (right) of three brain waves;
FIG. 5 is a graph showing the relationship between the brain electrical equation and fatigue for moving averages of brain regions;
FIG. 6 is a schematic diagram of an overall dispatcher fatigue model;
fig. 7 is a schematic diagram of an individual dispatcher electroencephalogram fatigue index determination.
Detailed Description
The technical solution of the present invention will be described in further detail with reference to the accompanying drawings, but the scope of the present invention is not limited to the following description.
As shown in fig. 1, the method for determining the brain fatigue index of the high-speed railway dispatcher comprises the following steps:
s1, selecting an experimenter to perform a high-speed railway driving scheduling experiment;
s2, constructing an electroencephalogram processing data set;
s3, constructing an overall dispatcher brain electrical fatigue model;
s4, judging the brain electrical fatigue indexes of the individual schedulers.
As shown in fig. 2, in the step S1, 32 iron group dispatchers of ages 28-38 are selected to train the staff. The selected test has no any psychological or mental diseases, good sleep quality, no color weakness or color blindness and is skilled in the operation skill of the high-speed railway dispatching. The functional beverage of caffeine or alcohol cannot be drunk in the evening before the experiment, and an informed experiment book is signed.
The working time of the high-speed rail dispatcher is generally as long as 12 hours, so the experiment selects a period 11 of easy fatigue of the high-speed rail dispatcher in the daytime: 00-16:00 an experiment was performed.
24 hours before the experiment starts, 32 tested schedulers are in an easily fatigued state, and the operation of an experiment flow, an experiment task and an experiment platform is explained to the tested.
15 minutes before the experiment starts, the tested dispatcher firstly carries out a simulation experiment and is familiar with the actual working environment and specific working tasks of the dispatcher.
After the experiment is started, working events randomly occur in the system every 15 minutes, a tested dispatcher executes simulated dispatching operation tasks including conversation operation, operation and recording operation according to railway technical management rules, and monitoring operation is executed in the rest time; recording the original brain electrical data of the tested dispatcher at the moment by using a 64-channel Neuroscan brain electrical instrument, and simultaneously recording KSS (K-S) scale values, wherein each tested is 20 times in total;
after the experiment is finished, the data are subjected to impurity removal and noise removal, power frequency electricity and myoelectricity interference are removed, data preprocessing is performed, and a foundation is laid for the subsequent construction of an electroencephalogram data set.
Meanwhile, the work events which occur randomly are faced to in addition to the monitoring task in the experimental process of the tested dispatcher, so that the brain electrical load of the tested dispatcher when the tested dispatcher monitors and the brain electrical load of the tested dispatcher when the tested dispatcher faces to the random events are calculated by using a DORATASK method based on the work task. The brain electrical load is used as a part of a fusion output end and as brain electrical reaction time in brain electrical fatigue index analysis of a subsequent individual dispatcher when an overall dispatcher brain electrical fatigue model based on a neural network is constructed:
the DORATASK method mainly converts load pressure values of a dispatcher into working time to measure, and calculates an electroencephalogram load value T during monitoring of the high-speed railway dispatcher through the following steps w And brain electrical load value T in the face of random events p ：
Figure BDA0003419164610000071
Wherein T is 1 Monitoring time, T 2 Talk time, T 3 -operation recording time, T 4 -brain recovery time, delta T-recording time interval taken as 15 minutes; the mental recovery time is set to 2-4 seconds according to the working characteristics of the high-speed rail dispatcher, and the brain electrical load when the dispatcher to be tested monitors is recorded every 15 minutes;
Figure BDA0003419164610000072
wherein T is p1 Reaction time, T p2 Wave time, T p3 The recovery time is taken as 200ms and the Δt-recording time interval is taken as 15 minutes; taking the stimulation time of one electroencephalogram experiment as 900ms, the sampling rate as 1000Hz, and recording the electroencephalogram load of a tested dispatcher facing a random event every 15 minutes;
in step S2, slow waves (delta waves, theta waves) of the brain electrical signal are gradually increased and fast waves (alpha waves, beta waves) are gradually decreased in consideration of the transition of the brain from the normal state to the fatigue state. In general, the delta wave occurs in a state of deep sleep, and thus is not discussed herein.
The alpha wave, the beta wave and the wave crest and the amplitude of the theta wave under different electrodes are extracted through an electroencephalograph, the variation coefficients of the three waveforms are extracted and calculated by utilizing Fourier transform based on a variation coefficient model to form an electroencephalograph index variation coefficient data set, and the specific process is shown in figure 3:
said step S2 comprises the sub-steps of:
s201, collecting 20 brain wave data of each tested dispatcher to obtain 20 groups of brain wave data, wherein each group of brain wave data comprises alpha wave, beta wave and theta wave data; removing impurities and noise, removing interference of power frequency electricity and myoelectricity, and adding an original electroencephalogram index data set;
s202, carrying out Fourier transform (FFT) on the denoised brain wave data to obtain the power P of three waveforms of the brain wave in each acquired data α ，P β ，P θ The fourier transform is:
Figure BDA0003419164610000081
x(n)＝x a (nT)| t＝nT (T＝1/f s ,f s ≥2f c ) (4)
Figure BDA0003419164610000082
the sample points collected in the formula are N, wherein N is the total number of the sample points collected in each experiment, N is 20T according to the experiment requirement s =nt denotes the duration of the finite length signal, x a (t) is the finite length signal after filtering, x a (T) duration of T s The highest frequency is f c ；
S203, calculating the occurrence rate of three waveforms in each acquisition process:
Figure BDA0003419164610000083
s204, calculating the mean square error of the occurrence rate of the three waveforms in each acquisition process:
Figure BDA0003419164610000084
s205, calculating variation coefficients of three waveforms in each acquisition process to form an electroencephalogram index variation coefficient data set:
Figure BDA0003419164610000085
the variation coefficient waveform is shown in figure 4;
in order to establish the relation between the original electroencephalogram data and the variation coefficients and fatigue degrees of the three waveforms, providing a curve change rate index; by describing the curve of the brain wave index change, including the brain wave curves of theta, alpha and beta waves, and the occurrence rate curve and the variance curve obtained by Fourier transform based on the variance coefficient model, and assuming that two points of P, Q are any two points on three curves (curves in fig. 4), the unit daily curve change rate is expressed as:
Figure BDA0003419164610000091
the relationship between the degree of fatigue and fatigue is represented by a mapping:
F(R)＝{A|A:R→[0,1]} (10)
wherein the domain R is defined as fatigue, and the fuzzy subset of the fatigue degree is A, mu A Membership function of A, wherein
Figure BDA0003419164610000095
，x→μ A ；
The fatigue degree is represented by a curve change rate index, and the curve change rate index at this time refers to a unit daily change rate index, and the relationship between the fatigue degree of the dispatcher and the unit daily curve change rate is characterized by:
Figure BDA0003419164610000092
for experimental sample X on any curve describing brain wave index change i ＝p*(x 1 ,x 2 ,...,x 20 ；y i ) The method comprises the steps of carrying out a first treatment on the surface of the Wherein y is i The amplitude of the waveform in the ith sampling is represented, p is the number of people of the tested dispatcher, and the number is taken as 32; when taking the interval [ a, b ]]At the time, for any point coordinate P (epsilon, delta), epsilon>0，δ>The rate of change of 0 is expressed as
Figure BDA0003419164610000093
As can be seen from fig. 4, the change curve of the occurrence rate of the α wave and the θ wave is in an ascending trend, which increases with the increase of the fatigue degree; the change curve of the occurrence rate of the beta wave shows a decreasing trend, and the change curve decreases with the deepening of the fatigue degree; the state of the dispatcher is continuously changed under the vigilance and relaxation under the influence of working fatigue and decision-making behaviors in the fatigue resistance stage; meanwhile, the alpha CV, the beta CV and the theta CV are greatly changed along with the increase of the fatigue degree, which shows that the fatigue degree is obviously influenced by the variation coefficients of three waveforms and corresponds to the change curves of the occurrence rate of three typical waves;
in addition, for the data obtained after each acquisition and denoising, the brain wave power P of the three waveforms obtained by fourier transform in step S202 α ，P β ，P θ And forming an electroencephalogram index equation data set by using the ratio and summation ratio of the three powers: θ/β, θ/(α+β), (θ+α)/β, (α+θ)/(α+β).
Figure BDA0003419164610000094
The relation between the brain electrical equation of the brain partition moving average value and fatigue is shown in fig. 5, and the equation fatigue index is shown to be severely changed along with the increase of the brain fatigue degree according to the formula (9), so that the equation fatigue index has obvious change trend. It can be seen from the calculation of the formula (12) that the changes of θ/β and (α+θ)/(α+β) are small, the changes of θ/(α+β) are not large, and the (θ+α)/β shows significant activity differences, which indicates that the index may be superior to other electroencephalogram equation indexes in reflecting the fatigue degree.
The step S3 includes: ten electroencephalogram indexes such as an original electroencephalogram index data set (alpha wave, beta wave and theta wave), an electroencephalogram index variation coefficient data set (alpha CV, beta CV and theta CV), an electroencephalogram index equation data set (theta/beta, theta/(alpha+beta), theta+alpha)/beta, (alpha+theta)/(alpha+beta)) and the like are taken as input ends, fused subjective values (KSS values, dispatcher workload DORATASK values and dispatcher electroencephalogram load DORATASK values) are taken as output ends, and a two-layer neural network model containing hidden layers is established, wherein the number of neurons of the hidden layers is 10 as shown in figure 6.
The transfer function of the hidden layer of the BP neural network is selected as a sigmod function, and can be expressed as an f function
Figure BDA0003419164610000101
The transfer function of the output layer is selected as purelin function, and the training function is selected as tranlm function.
For a given m training samples { (x (1), y (1)), (x (2), y (2)), (x (m), y (m)) }, d (i) is the desired output of the corresponding input x (i), given an error function of:
Figure BDA0003419164610000102
let W be ij B is the weight of the connection between the jth neuron and the ith neuron of the hidden layer i For the bias of the ith neuron, net i Is the input to the ith neuron. The BP algorithm calculates weights and offsets according to the following formula:
h(i)＝f(net i ) (15)
Figure BDA0003419164610000103
let W be T As a weight matrix, a i To input specialThe sign parameter, the output result can be expressed as
Figure BDA0003419164610000104
The judgment idea for the reliability of the brain fatigue index of the overall dispatcher is as follows: dividing the average value of the tested dispatcher data into a training set, a verification set and a test set, and obtaining the best BP neural network model through training a model and adjustment; and judging ten electroencephalogram indexes by using the model, taking an R value representing the correlation degree between the electroencephalogram indexes and fatigue and an MSE mean square error representing the difference between a sample predicted value and an actual value as criteria, and if the R value of the index is closer to 1 and the MSE is smaller, considering that the reliability of the index is higher when the electroencephalogram fatigue of the overall dispatcher is judged.
As shown in fig. 7, the step S4 includes:
randomly selecting electroencephalogram indexes and electroencephalogram reaction time in the experimental process of n high-speed rail dispatching experimental personnel as statistics (X) 1 ,Y 1 ),(X 2 ,Y 2 ),...,(X n ,Y n ). Assuming that there is a linear correlation between the independent and dependent variables, let X, Y be samples taken from two different populations X, Y, with observations of X 1 ,x 2 ,...,x n And y 1 ,y 2 ,...,y n . Pairing them to form (x 1 ,y 1 ),(x 2 ,y 2 ),...,(x n ,y n ) And (5) checking whether the electroencephalogram index is related to the electroencephalogram reaction time.
S401 proposes hypothesis H 0 : x and Y are uncorrelated.
Figure BDA0003419164610000105
H 1 : x and Y are related.
S402, calculating test statistics, spearman rank correlation coefficient r s Is an important index for measuring the correlation degree of two samples. The calculation formula is as follows:
Figure BDA0003419164610000111
note that: r is (r) s The value range of (C) is [ -1,1]When |r s The closer to 1, the higher the correlation between samples is; conversely, when |r s The closer to 0, the lower the correlation between samples.
S403 makes a decision when r s ≥r s α When rejecting H 0 Conversely, H cannot be rejected 0 。
The judgment idea for the reliability of the brain electrical fatigue index of the individual dispatcher is as follows: calculating a correlation coefficient between each tested dispatcher electroencephalogram index and electroencephalogram reaction time by a Spearman rank test method, and transversely comparing the correlation coefficient of each class of electroencephalogram indexes, wherein the index with the larger correlation coefficient value can be considered to have higher reliability when judging the electroencephalogram fatigue of an individual dispatcher; if the correlation coefficient between a certain electroencephalogram index and the electroencephalogram reaction time is larger than
Figure BDA0003419164610000112
Then hypothesis H may be rejected 0 Accept H 1 The electroencephalogram index is considered to have correlation with the electroencephalogram reaction time.
Therein, wherein
Figure BDA0003419164610000113
Is a critical value, related to the number of sample observations n, the choice hypothesis and the given significance level, according to the check threshold table of Spearman rank correlation coefficient, when n is 8 #>
Figure BDA0003419164610000114
In the embodiment of the application, in the overall dispatcher electroencephalogram fatigue model construction, ten electroencephalogram indexes such as an original electroencephalogram index data set (alpha wave, beta wave and theta wave), an electroencephalogram index variation coefficient data set (alpha CV, beta CV and theta CV), an electroencephalogram index equation data set (theta/beta, theta/(alpha+beta), theta+alpha)/beta (alpha+theta)/(alpha+beta) are used as input ends of a neural network, namely, the electroencephalogram fatigue index of the overall dispatcher is studied.
After the model is built, the overall average of 32 schedulers is taken, 70% of data is used as a training set, 15% of data is used as a verification set, and 15% of data is used as a test set. And (5) continuously training and adjusting, and evaluating a final model by using the thought of the BP neural network. The previous ten fatigue indexes are judged and analyzed through a final model, and the (theta+alpha)/beta, beta CV and theta CV are better as indexes when the overall dispatcher brain electrical fatigue is evaluated.
For individuals, the Spearman rank test is used for carrying out hypothesis test on ten electroencephalogram indexes and electroencephalogram loads affecting the electroencephalogram fatigue, and whether the ten indexes affect the fatigue of individual schedulers or not and the influence degree is large or not is checked, and the data used at this time are the average value of each scheduler. Through inspection analysis, beta CV and theta CV are found to be still significant in individual fatigue detection.
In the embodiment of the present application, in order to determine that the selected index in the training set has significance in reflecting the fatigue degree, a single-factor analysis of variance is performed on each index, and the analysis result of each index is shown in table 1.
TABLE 1 training set index single factor analysis of variance table
Figure BDA0003419164610000115
Figure BDA0003419164610000121
In general, if the p value is less than 0.01, the significance level assumption may be considered satisfied. As can be seen from Table 1, the p-values of the selected indices are less than 0.01, so that ten indices, namely, the original electroencephalogram data set index and the electroencephalogram processing data set index, can be used as the judging index of the fatigue degree of the dispatcher.
Brain electrical fatigue index analysis of overall high-speed rail dispatcher
The average value of the data of each measurement of 32 tested persons is calculated, 70% of the data is used as a training set, 15% of the data is used as a verification set, and 15% of the data is used as a test set. The training times were set to 10000 times, and the learning rate was set to 0.01. Data analysis for each waveform index was obtained as shown in table 2.
TABLE 2 analysis of fatigue determination index for brain electrical data set neural network
Figure BDA0003419164610000122
The R value represents the degree of correlation between the electroencephalogram index and the fatigue, and the closer the R value is to 1, the greater the degree of correlation is; MSE is mean square error, represents the difference between the predicted value and the actual value of the sample, and continuously reduces the loss function by continuously changing all parameters in the neural network, so that a neural network model with higher accuracy is trained.
As can be seen from table 2, the following R values and mean square deviations can reach significance level, and from analysis of the electroencephalogram equation, the (θ+α)/β equation is most active, the R value can reach 0.94, and the best performance is β CV and θcv, the R value can reach 0.96, and the error peak tends to be 0, which indicates that the three indexes are better than other indexes in the fatigue determination process.
Brain electrical fatigue index analysis of individual high-speed rail dispatcher
The Spearman rank test method can calculate the correlation coefficient between the electroencephalogram index and the electroencephalogram reaction time of each scheduling experimenter, randomly select 8 tested persons from 32 tested schedulers to test, and the result is shown in table 3.
TABLE 3 electroencephalogram measurement index System and electroencephalogram reaction time correlation coefficient Table
Figure BDA0003419164610000123
Figure BDA0003419164610000131
Note that: when the confidence is 0.05, the correlation is significant; when the confidence is 0.01, the correlation is significant.
By calculating the correlation coefficient, it can be found that the difference between the subjects leads to a change in brain wave activity. However, only at least one index of the electroencephalogram variation coefficients among the original electroencephalogram indexes, the electroencephalogram variation coefficients and the electroencephalogram equation shows significance to reflect fatigue, and the significance of beta CV and theta CV is highest. The significance of the (theta+alpha)/beta equation is only-0.43, which indicates that the fatigue verification index of the overall average value of the dispatcher is not necessarily suitable for the individual fatigue verification index. In the above, it was explained that βcv and θcv are most reliable in individual fatigue determination.
While the foregoing description illustrates and describes a preferred embodiment of the present invention, it is to be understood that the invention is not limited to the form disclosed herein, but is not to be construed as limited to other embodiments, but is capable of use in various other combinations, modifications and environments and is capable of changes or modifications within the spirit of the invention described herein, either as a result of the foregoing teachings or as a result of the knowledge or skill of the relevant art. And that modifications and variations which do not depart from the spirit and scope of the invention are intended to be within the scope of the appended claims.

Claims (4)

1. A brain electrical fatigue index judging method for a high-speed rail dispatcher is characterized in that: the method comprises the following steps:
s1, selecting a tested dispatcher to perform a high-speed railway driving dispatching experiment;
in the step S1, the process of performing the high-speed railway driving dispatching experiment includes:
s101, 24 hours before the start of an experiment, enabling 32 tested schedulers to be in an easily fatigued state, and explaining the operation of an experiment flow, an experiment task and an experiment platform to the tested schedulers;
s102, 15 minutes before the start of the experiment, enabling a tested dispatcher to firstly perform a simulation experiment, and familiarizing with the actual working environment and specific working tasks of the dispatcher;
s103, after the experiment is started, working events randomly occur in the system every 15 minutes, a tested dispatcher executes simulated dispatching operation tasks including conversation operation, operation and recording operation according to railway technical management rules, and monitoring operation is executed in the rest time; recording the original brain electrical data of the tested dispatcher at the moment by using a 64-channel Neuroscan brain electrical instrument, and simultaneously recording KSS (K-S) scale values, wherein each tested dispatcher is 20 times in total;
s104, after the experiment is finished, removing impurities and noises from the data, removing power frequency electricity and myoelectricity interference, preprocessing the data, and laying a foundation for the subsequent construction of an electroencephalogram data set;
meanwhile, the work event which occurs randomly is faced in addition to the completion of the monitoring task in the experimental process of the tested dispatcher, so that the DORATASK method based on the work task is used for calculating the brain electrical load of the tested dispatcher when monitoring and the brain electrical load when facing the random event; the calculated brain electrical load is used as a part of a fusion output end and is used as brain electrical reaction time in brain electrical fatigue index analysis of a subsequent individual dispatcher when an overall dispatcher brain electrical fatigue model based on a neural network is constructed:
the DORATASK method mainly converts the load pressure value of the dispatcher into working time to measure, and calculates the brain electrical load value T when the tested dispatcher monitors the load value T through the following steps w And brain electrical load value T in the face of random events p ：
Figure FDA0004248440740000011
Wherein T is 1 Monitoring time, T 2 Talk time, T 3 -operation recording time, T 4 -brain recovery time, delta T-recording time interval taken as 15 minutes; the mental recovery time is set to 2-4 seconds according to the working characteristics of the tested dispatcher, and the brain electrical load when the tested dispatcher monitors is recorded every 15 minutes;
Figure FDA0004248440740000012
wherein T is p1 Reaction time, T p2 Wave time, T p3 The recovery time is taken as 200ms and the Δt-recording time interval is taken as 15 minutes; taking the stimulation time of one electroencephalogram experiment as 900ms, the sampling rate as 1000Hz, and recording the electroencephalogram load of a tested dispatcher facing a random event every 15 minutes;
s2, constructing an electroencephalogram processing data set;
the alpha wave, the beta wave, the wave peak and the amplitude of the theta wave under different electrodes are extracted by an electroencephalograph to obtain an original electroencephalograph index data set, and the power P of three waveforms is extracted by Fourier transformation α 、P β 、P θ Extracting and calculating the variation coefficients of the three waveforms by using a variation coefficient model on the basis to form an electroencephalogram index variation coefficient data set, and forming an electroencephalogram index equation data set by using the obtained electroencephalogram power;
s3, constructing an overall dispatcher brain electrical fatigue model;
the step S3 includes:
taking ten electroencephalogram indexes of alpha wave, beta 0CV, beta 1CV, theta CV and theta/beta, theta/(alpha+beta), (theta+alpha)/beta, (alpha+theta)/(alpha+beta) in an original electroencephalogram index data set and a coefficient of variation data set of the electroencephalogram indexes as input ends, and integrating subjective values, namely KSS values; electroencephalogram load value T during monitoring in experimental process w And brain electrical load value T in the face of random events p As an output end, a two-layer neural network model containing hidden layers is established, and the number of neurons of the hidden layers is 10;
the transfer function of the hidden layer of the neural network is selected as a sigmod function, which can be expressed as an f function
Figure FDA0004248440740000021
The transmission function of the output layer is selected as a purelin function, and the training function is selected as a tranlm function;
for a given m training samples { (x (1), y (1)), (x (2), y (2)), (x (m), y (m)) }, d (i) is the desired output of the corresponding input x (i), given an error function of:
Figure FDA0004248440740000022
let W be ij B is the weight of the connection between the jth neuron and the ith neuron of the hidden layer i For the bias of the ith neuron, net i An input for the ith neuron; the BP algorithm calculates weights and offsets according to the following formula:
h(i)＝f(net i )
Figure FDA0004248440740000023
let W be T As a weight matrix, a i For inputting characteristic parameters, the output result can be expressed as
y＝f(W e T (f(W t T a i )+b t )+b e )
The judgment idea for the reliability of the brain fatigue index of the overall dispatcher is as follows: dividing the average value of the tested dispatcher data into a training set, a verification set and a test set, and obtaining the best BP neural network model through training a model and adjustment; judging ten electroencephalogram indexes by using the model, taking an R value representing the correlation degree between the electroencephalogram indexes and fatigue and an MSE mean square error representing the difference between a sample predicted value and an actual value as criteria, and if the R value of the index is closer to 1 and the MSE is smaller, considering that the reliability of the index is higher when judging the electroencephalogram fatigue of a general dispatcher;
s4, judging brain electrical fatigue indexes of individual schedulers;
the step S4 includes:
s401, randomly selecting electroencephalogram indexes and electroencephalogram reaction time in the experimental process of n high-speed rail dispatching testees as statistics (X) 1 ,Y 1 ),(X 2 ,Y 2 ),...,(X n ,Y n ) Assuming that there is a linear correlation between the independent variable and the dependent variable; let x, y beSamples extracted from two different populations X and Y, wherein X is from ten electroencephalogram indexes in building a population electroencephalogram fatigue model, and Y is from a calculated electroencephalogram load value T of a tested dispatcher in the face of a random event p The method comprises the steps of carrying out a first treatment on the surface of the With an observation value of x 1 ,x 2 ,...,x n And y 1 ,y 2 ,...,y n Pairing them to form (x 1 ,y 1 ),(x 2 ,y 2 ),...,(x n ,y n ) Checking whether the electroencephalogram index is related to the electroencephalogram reaction time or not;
let us assume H 0 : x and Y are uncorrelated;
Figure FDA0004248440740000031
H 1 : x and Y are related;
s402, calculating test statistics, spearman grade correlation coefficient r s The method is used for measuring the important index of the correlation degree of two samples, and the calculation formula is as follows:
Figure FDA0004248440740000032
wherein r is s The value range of (C) is [ -1,1]When |r s The closer to 1, the higher the correlation between samples is; conversely, when |r s The closer to 0, the lower the correlation between samples is;
s403, making a decision, when
Figure FDA0004248440740000033
When rejecting H 0 Conversely, H cannot be rejected 0 ；
Therein, wherein
Figure FDA0004248440740000034
Is a critical value, related to the number of observations of the sample n, the alternative hypothesis and the given level of significance, when n takes 8 ∈>
Figure FDA0004248440740000035
The judgment idea for the reliability of the brain electrical fatigue index of the individual dispatcher is as follows: calculating a correlation coefficient between each tested dispatcher electroencephalogram index and electroencephalogram reaction time by a Spearman rank test method, and transversely comparing the correlation coefficient of each class of electroencephalogram indexes, wherein the index with the larger correlation coefficient value can be considered to have higher reliability when judging the electroencephalogram fatigue of an individual dispatcher; if the correlation coefficient between a certain electroencephalogram index and the electroencephalogram reaction time is larger than
Figure FDA0004248440740000036
Then reject hypothesis H 0 Accept H 1 The electroencephalogram index is considered to have correlation with the electroencephalogram reaction time.
2. The method for determining the brain fatigue index for the high-speed rail dispatcher according to claim 1, wherein the method comprises the following steps: in the step S1, 32 national iron group schedulers with ages of 28-38 are selected to train staff as tested schedulers, and the selected persons meet the following requirements: the sleeping bag has the advantages of no psychological or mental diseases, good sleeping quality, no color weakness or color blindness, and proficiency in mastering the dispatching operation skills of the high-speed railway; the functional beverage of caffeine or alcohol cannot be drunk in the evening before the experiment; and, in the step S1, a period 11 in which the tested dispatcher is fatigued in the daytime is selected: 00-16:00 an experiment was performed.
3. The method for determining the brain fatigue index for the high-speed rail dispatcher according to claim 1, wherein the method comprises the following steps: in the step S2, considering that when the brain transits from the normal state to the fatigue state, the slow wave of the brain signal gradually increases, the fast wave gradually decreases, wherein the slow wave comprises a delta wave, the theta wave comprises an alpha wave and a beta wave, and the delta wave only appears in a deep sleep state; only alpha, beta and theta waves are considered; alpha wave, beta wave, theta wave peaks and amplitudes under different electrodes are extracted through an electroencephalograph, and power P of three waveforms is extracted through Fourier transformation α 、P β 、P θ On the basis, the variation coefficients of the three waveforms are extracted and calculated by using a variation coefficient model to form an electroencephalogram index variation coefficient data set, and the obtained electroencephalogram power is used to form an electroencephalogram index equation data set.
4. The method for determining the brain fatigue index for the high-speed rail dispatcher according to claim 3, wherein the method comprises the following steps: said step S2 comprises the sub-steps of:
s201, collecting 20 brain wave data of each tested dispatcher to obtain 20 groups of brain wave data, wherein each group of brain wave data comprises alpha wave, beta wave and theta wave data; removing impurities and noise, removing interference of power frequency electricity and myoelectricity, and adding an original electroencephalogram index data set;
s202, carrying out Fourier transform on the denoised brain wave data to obtain the power P of three waveforms of the brain wave in each acquired data α ，P β ，P θ The fourier transform is:
Figure FDA0004248440740000041
x(n)＝x a (nT)| t＝nT (T＝1/f s ,f s ≥2f c ) (4)
Figure FDA0004248440740000042
the sample points collected in the formula are N, wherein N is the total number of the sample points collected in each experiment, N is 20T according to the experiment requirement s =nt denotes the duration of the finite length signal, x a (t) is the finite length signal after filtering, x a (T) duration of T s The highest frequency is f c ；
S203, calculating the occurrence rate of three waveforms in each acquisition process:
Figure FDA0004248440740000043
s204, calculating the mean square error of the occurrence rate of the three waveforms in each acquisition process:
Figure FDA0004248440740000044
s205, calculating variation coefficients of three waveforms in each acquisition process to form an electroencephalogram index variation coefficient data set:
Figure FDA0004248440740000045
in order to establish the relation between the original electroencephalogram data and the variation coefficients and fatigue degrees of the three waveforms, providing a curve change rate index; by describing the curve of the brain wave index change, including the brain wave curve of theta, alpha and beta waves, and the occurrence rate curve and the variance curve obtained by Fourier transform based on a variance coefficient model, and assuming that two points of P, Q are any two points on three curves, the unit daily curve change rate is expressed as:
Figure FDA0004248440740000046
the relationship between the degree of fatigue and fatigue is represented by a mapping:
F(R)＝{A|A:R→[0,1]} (10)
wherein the domain R is defined as fatigue, and the fuzzy subset of the fatigue degree is A, mu A Membership function of A, wherein
Figure FDA0004248440740000054
x→μ A ；
The fatigue degree is represented by a curve change rate index, and the curve change rate index at this time refers to a unit daily change rate index, and the relationship between the fatigue degree of the dispatcher and the unit daily curve change rate is characterized by:
Figure FDA0004248440740000051
for experimental sample X on any curve describing brain wave index change i ＝p*(x 1 ,x 2 ,...,x 20 ；y i ) The method comprises the steps of carrying out a first treatment on the surface of the Wherein y is i The amplitude of the waveform in the ith sampling is represented, p is the number of people of the tested dispatcher, and the number is taken as 32; when taking the interval [ a, b ]]At the time, for any point coordinate P (epsilon, delta), epsilon>0，δ>The rate of change of 0 is expressed as
Figure FDA0004248440740000052
The change curve of the occurrence rate of the alpha wave and the theta wave is in an ascending trend, and the change curve is increased along with the deepening of the fatigue degree; the change curve of the occurrence rate of the beta wave shows a decreasing trend, and the change curve decreases with the deepening of the fatigue degree; the state of the dispatcher is continuously changed under the vigilance and relaxation under the influence of working fatigue and decision-making behaviors in the fatigue resistance stage; meanwhile, the alpha CV, the beta CV and the theta CV are greatly changed along with the increase of the fatigue degree, which shows that the fatigue degree is obviously influenced by the variation coefficients of three waveforms and corresponds to the change curves of the occurrence rate of three typical waves;
in addition, for the data obtained after each acquisition and denoising, the brain wave power P of the three waveforms obtained by fourier transform in step S202 α ，P β ，P θ And forming an electroencephalogram index equation data set by using the ratio and summation ratio of the three powers: θ/β, θ/(α+β), (θ+α)/β, (α+θ)/(α+β).
Figure FDA0004248440740000053
CN202111556302.3A 2021-12-18 2021-12-18 Electroencephalogram fatigue index judging method for high-speed rail dispatcher Active CN114224359B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111556302.3A CN114224359B (en) 2021-12-18 2021-12-18 Electroencephalogram fatigue index judging method for high-speed rail dispatcher

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111556302.3A CN114224359B (en) 2021-12-18 2021-12-18 Electroencephalogram fatigue index judging method for high-speed rail dispatcher

Publications (2)

Publication Number Publication Date
CN114224359A CN114224359A (en) 2022-03-25
CN114224359B true CN114224359B (en) 2023-06-30

Family

ID=80758526

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111556302.3A Active CN114224359B (en) 2021-12-18 2021-12-18 Electroencephalogram fatigue index judging method for high-speed rail dispatcher

Country Status (1)

Country Link
CN (1) CN114224359B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115408891B (en) * 2022-11-02 2023-02-07 深圳联友科技有限公司 Modeling method for user experience percentile measurement containing multiple electroencephalogram cognitive indexes

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106691474A (en) * 2016-11-25 2017-05-24 中原电子技术研究所（中国电子科技集团公司第二十七研究所） Brain electrical signal and physiological signal fused fatigue detection system
CN107280663A (en) * 2017-07-07 2017-10-24 南京邮电大学 A kind of method of the tired brain electrical feature research based on different experiments difficulty

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
NZ630770A (en) * 2013-10-09 2016-03-31 Resmed Sensor Technologies Ltd Fatigue monitoring and management system
CN103989485A (en) * 2014-05-07 2014-08-20 朱晓斐 Human body fatigue evaluation method based on brain waves
CN107095671A (en) * 2017-05-27 2017-08-29 西南交通大学 Driving fatigue recognition methods and system
CN107095670A (en) * 2017-05-27 2017-08-29 西南交通大学 Time of driver's reaction Forecasting Methodology
US20200151474A1 (en) * 2017-07-31 2020-05-14 Alcohol Countermeasure Systems (International) Inc. Non-intrusive assessment of fatigue in drivers using eye tracking
CN108272463A (en) * 2017-08-07 2018-07-13 上海交通大学 A kind of flyer's fatigue detection method based on EEG signals
CN109770924B (en) * 2019-01-24 2020-06-19 五邑大学 Fatigue classification method for building brain function network and related vector machine based on generalized consistency

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106691474A (en) * 2016-11-25 2017-05-24 中原电子技术研究所（中国电子科技集团公司第二十七研究所） Brain electrical signal and physiological signal fused fatigue detection system
CN107280663A (en) * 2017-07-07 2017-10-24 南京邮电大学 A kind of method of the tired brain electrical feature research based on different experiments difficulty

Also Published As

Publication number Publication date
CN114224359A (en) 2022-03-25

Similar Documents

Publication Publication Date Title
Hu et al. Do entrepreneurial alertness and self-efficacy predict Chinese sports major students' entrepreneurial intention?
CN105877766B (en) A kind of state of mind detection system and method based on the fusion of more physiological signals
JP4590555B2 (en) Sensitive state discrimination method and apparatus
JP3310498B2 (en) Biological information analyzer and biological information analysis method
CN109646022B (en) Child attention assessment system and method thereof
CN106504475B (en) Method for detecting fatigue driving based on EEG signals
CN110604565A (en) Brain health screening method based on portable EEG equipment
CN114224359B (en) Electroencephalogram fatigue index judging method for high-speed rail dispatcher
CN111150410B (en) Psychological pressure evaluation method based on fusion of electrocardiosignals and electromyographic signals
Talebinejad et al. Fatigue estimation using a novel multi-fractal detrended fluctuation analysis-based approach
CN113499035B (en) Pain identification system based on confidence interval fusion threshold criterion
CN109443766A (en) A kind of heavy-duty vehicle gearbox gear Safety Analysis Method
CN114847958A (en) Stress and fatigue monitoring method and system based on electrocardiosignals
CN112057087A (en) Method and device for evaluating autonomic nerve function of high-risk schizophrenic population
CN113143208A (en) Pain sensitivity assessment system and method based on multi-dimensional measurement
Bernabei et al. Automatic detection of tic activity in the Tourette Syndrome
Cipresso et al. Continuous measurement of stress levels in naturalistic settings using heart rate variability: An experience-sampling study driving a machine learning approach
CN110680285A (en) Anesthesia degree monitoring device based on neural network
CN108324272A (en) A kind of device for checking and examining anxiety degree based on LZC complexities
Kang et al. Driver emotional and perceptual evaluation over various highway horizontal curves
CN113729708B (en) Lie judgment method based on eye movement technology
CN111696011B (en) System and method for monitoring, regulating and controlling student autonomous learning
CN115153549A (en) BP neural network-based man-machine interaction interface cognitive load prediction method
Yulmetyev et al. Manifestation of chaos in real complex systems: Case of parkinson’s disease
CN106344008A (en) Method and system for detecting waking state in sleep state analysis

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant