Method for predicting and judging psychological fatigue of dispatcher based on physiology and response time indexes

Abstract

The invention relates to a dispatcher psychological fatigue prediction and discrimination method based on physiology and reaction time indexes. Firstly, psychological fatigue state indexes are constructed; by designing a test program, a psychological fatigue state test webpage system is constructed; collecting test data of the indexes during objective reaction and performing data processing; calculating influence probabilities of subjective and objective influence factors in different psychological fatigue states, and constructing a reaction time prediction model; according to the psychological fatigue state index threshold, constructing a courier psychological fatigue state discrimination model; and judging the psychological fatigue state of the dispatcher according to three states of no fatigue, moderate fatigue and severe fatigue in combination with test data and a prediction result. According to the method, quantitative evaluation of the psychological fatigue state of the courier is realized; the system is simple and easy to deploy; the problems of large size, high cost, complicated operation and the like of traditional equipment are solved; the autonomous guarantee requirement of high-reliability operation of civil aviation is met; the method has a wide application prospect in the aspect of aviation operation safety management monitoring.

Description

Technical Field

[0001] The present invention relates to a method for predicting and identifying the mental fatigue state of civil aviation dispatchers, and particularly to a method for predicting and discriminating the mental fatigue of dispatchers based on physiological and reaction time indicators, belonging to the technical field of human factors engineering and intelligent safety guarantee in aviation operation. Background Art

[0002] As one of the core positions in civil aviation operation control, the responsibilities of dispatchers include monitoring flight dynamics, assessing operation risks, coordinating operation resources, etc. During the flight operation, due to the influence of complex weather, airspace restrictions, emergencies, etc., dispatchers need to quickly complete flight plan adjustment and safety judgment within a short time, highly relying on their sustained attention and reaction speed.

[0003] The work characteristics such as long-term high-intensity duty and day-night shift rotation make dispatchers prone to mental fatigue, manifested as decreased vigilance, prolonged reaction time, and increased operation errors. In severe cases, it may pose a hidden danger to aviation operation safety.

[0004] Existing research mostly focuses on the fatigue formation mechanism and the change trend of single physiological parameters, lacking systematic task design and quantitative standards. The testing tools mostly rely on expensive and bulky laboratory equipment, making it difficult to achieve on-site application in the post. Existing methods are mostly static identification of fatigue state, lacking the ability to predict the fatigue change trend, difficult to give early warning and intervention in time, limiting their application value in the actual operation guarantee scenario. At the same time, the fatigue state classification method is highly subjective, lacking accurate discrimination based on multiple indicators and influencing factor data. Summary of the Invention

[0005] The present invention aims to provide a method for predicting and discriminating the mental fatigue of dispatchers based on physiological and reaction time indicators, which can achieve dynamic prediction and hierarchical discrimination of fatigue state suitable for on-site application in the post. This method integrates electroencephalogram signals, eye movement parameters and behavioral response data to construct a structured mental fatigue discrimination system.

[0006] First, determine the mental fatigue state index constructed from physiological indicators and objective reaction time indicators; screen out eight discriminant indicators sensitive to fatigue, including electroencephalogram, eye movement, and various reaction time parameters; construct a mental fatigue test task system to collect electroencephalogram, eye movement, and individual influencing factor data and automatically record reaction time information; complete data preprocessing through steps such as filtering and abnormal elimination; construct a Bayesian network model, infer the reaction time growth rate based on personal influencing factors, and recursively generate hourly prediction values starting from the first test data; use Gaussian kernel density estimation, sliding window, and clustering methods to determine the discriminant thresholds of each indicator between different fatigue states; combine the artificially set thresholds and the clustering results of the Gaussian mixture model to achieve the discrimination of non-fatigue, moderate fatigue, and severe fatigue of the mental fatigue state based on the multi-weight fusion method.

[0007] The technical solution adopted by the present invention is: A method for predicting and discriminating dispatchers' mental fatigue based on physiological and reaction time indicators includes the following steps: Step 1: Determine the mental fatigue state index constructed from physiological indicators and objective reaction time indicators; the physiological indicators are the power spectral density of the α wave, the power spectral density of the θ wave, the fixation speed, and the blink frequency; the objective reaction time indicators are the psychomotor vigilance reaction time, the sustained attention reaction time, the fine attention reaction time, and the visual fatigue reaction time.

[0008] Step 2: Determine that the dispatcher's mental fatigue test consists of three parts: physiological indicator test, objective reaction time indicator test, and subjective and objective influencing factor test of mental fatigue. By designing a test program, construct a mental fatigue state test web page system, and the mental fatigue state test web page system connects Java and the background database through PHP.

[0009] Step 3: Collect the test data of the objective reaction time indicators through the mental fatigue test web page system, and collect the physiological indicators through an electroencephalogram device and an eye movement device respectively, and then perform data processing on the collected test data.

[0010] Step 4: Based on the processed test data, construct a dispatcher Bayesian network, calculate the influence probabilities of subjective and objective influencing factors on the sustained attention reaction time, the fine attention reaction time, and the visual fatigue reaction time under different mental fatigue states, and then construct prediction models for the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time.

[0011] Step 5: Based on the prediction models of the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time, calculate the corresponding reaction time prediction values for each hour after the first test through the reaction time data of each item, which are the prediction results.

[0012] Step 6: Determine the thresholds of each mental fatigue state indicator by analyzing the test data of each item.

[0013] Step 7. Based on the threshold of the psychological fatigue status index, a model for distinguishing the psychological fatigue status of the dispatcher is constructed. Based on the model for distinguishing the psychological fatigue status of the dispatcher, combined with the test data and the prediction results, the psychological fatigue status of the dispatcher is distinguished according to the three states of no fatigue, moderate fatigue and severe fatigue.

[0014] The design principle of the present invention is to build a psychological fatigue prediction and discrimination model with reaction time as the core based on the job characteristics of civil aviation dispatchers' high-intensity mental labor. Dispatchers need to continuously receive and process a large amount of dynamic information during their work, such as flight information, ground-air communications and flight operation data. The reception, discrimination, judgment and feedback of information are highly dependent on the individual's attention maintenance and rapid reaction ability. Reaction time, that is, the time from external stimulation to response behavior, is a key indicator to measure the coordination of the neuromuscular system and work agility, reflecting the changes in their psychological fatigue state.

[0015] Combined with the task requirements of the dispatcher, the present invention constructs a reaction time prediction system from three aspects: sustained attention, subtle discrimination ability and visual fatigue. Among them, sustained attention affects the efficiency of monitoring external information, subtle discrimination ability is related to the judgment of complex or small changes, and visual fatigue limits the accuracy and response speed under long-term high-load conditions. The system collects multimodal physiological indicators such as EEG and eye movements, extracts the discrimination threshold of each indicator based on a data-driven method, and divides the sample space in combination with a clustering algorithm to achieve the discrimination of the current state of mental fatigue. Finally, a prediction model based on the growth trend of reaction time is constructed, and a psychological fatigue state discrimination model is constructed in combination with physiological and objective reaction time indicators, corresponding to the two major functional modules of psychological fatigue development prediction and state discrimination.

[0016] The effects of the present invention and the beneficial effects produced are: 1. Numerical identification and trend prediction: The present invention realizes the quantitative evaluation of the mental fatigue state, predicts the development of the dispatcher's mental fatigue state based on the change trend during reaction time, and completes the identification of the mental fatigue state in combination with physiological data, assisting in job risk identification and response strategy formulation.

[0017] 2. The system is simple and easy to deploy: The test platform has a compact structure and a convenient collection process. It can be deployed and applied in the actual working environment, overcoming the problems of large size, high cost, and cumbersome operation of traditional equipment, and avoiding the defects of questionnaire evaluation that are easily affected by subjective bias.

[0018] 3. Strong technological autonomy: Break the reliance on foreign key technologies, build a full-process controllable psychological fatigue testing and prediction method system, and meet the independent guarantee needs of high-reliability operation of civil aviation.

[0019] 4. Wide range of applications: It can be extended to the psychological fatigue assessment tasks of dispatchers, air traffic controllers, safety inspectors, etc., and has broad application prospects in aviation operation safety management and human factor monitoring. Description of the Drawings

[0020] Figure 1 It is a flow chart of the dispatcher's psychological fatigue prediction and discrimination method based on physiological and reaction time indexes in the embodiment of the present invention; Figure 2 It is an example diagram of the continuous attention reaction time data test interface in the embodiment of the present invention; Figure 3 It is an example diagram of the fine attention reaction time data test interface in the embodiment of the present invention; Figure 4 It is an example diagram of the visual fatigue reaction time data test interface in the embodiment of the present invention; Figure 5 It is a flow chart of the psychological fatigue influencing factor data test program in the embodiment of the present invention; Figure 6 It is a flow chart of the program for constructing the prediction models of continuous attention, fine attention, and visual fatigue reaction time in the embodiment of the present invention; Figure 7 It is a flow chart of the program for determining the threshold of the psychological fatigue state index in the embodiment of the present invention; Figure 8 It is a flow chart of the psychological fatigue state discrimination program in the embodiment of the present invention. Detailed Embodiment

[0021] The present invention will be further described below in conjunction with the drawings and embodiments.

[0022] As Figure 1 shown, a dispatcher's psychological fatigue prediction and discrimination method based on physiological and reaction time indexes includes the following steps: Step 1. Determine the psychological fatigue state indexes constructed by physiological indexes and objective reaction time indexes; the physiological indexes are alpha wave power spectral density, theta wave power spectral density, fixation speed, and blink frequency; the objective reaction time indexes are psychomotor vigilance reaction time, continuous attention reaction time, fine attention reaction time, and visual fatigue reaction time.

[0023] The initial selection test indicators cover electrocardiogram signal indicators, electrodermal signal indicators, electromyogram signal indicators, electroencephalogram signal indicators, eye movement indicators, psychomotor vigilance test indicators, sustained attention test indicators, fine attention test indicators, and visual fatigue test indicators. The electroencephalogram signal indicators include delta wave power spectral density, theta wave power spectral density, alpha wave power spectral density, beta wave power spectral density, and gamma wave power spectral density. The eye movement indicators include fixation speed, blink frequency, fixation duration, fixation count, total fixation time, saccade amplitude, saccade speed, blink duration, and percentage of closed-eye time.

[0024] After actual measurement and further screening among the above initial selection indicators based on the sensitivity, specificity, and repeatability of each psychological fatigue indicator of the tested dispatchers, eight indicators including physiological indicators and objective reaction time indicators are finally determined as the psychological fatigue state indicators. The physiological indicators are alpha wave power spectral density, theta wave power spectral density, fixation speed, and blink frequency. The objective reaction time indicators are psychomotor vigilance reaction time, sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time.

[0025] Step 2: Determine that the psychological fatigue test for dispatchers consists of three parts: physiological indicator test, objective reaction time indicator test, and subjective and objective influencing factors test of psychological fatigue. By designing a test program, a psychological fatigue state test web system is constructed. The psychological fatigue state test web system connects Java and the background database through PHP.

[0026] Use the psychological fatigue test web system to test the psychomotor vigilance, sustained attention, fine attention, and visual fatigue reaction time of dispatchers; the psychological fatigue test web system automatically records the start time, average reaction time, number of errors, and data on psychological fatigue influencing factors for each test, and generates a time stamp. The reaction time is measured in milliseconds.

[0027] When testing dispatchers, obtain electroencephalogram signal data by wearing an electroencephalogram device and obtain eye movement data by installing an eye tracker; obtain data on psychological fatigue influencing factors and objective reaction time data through the psychological fatigue test web system and the test program.

[0028] According to the psychological fatigue test web system, as Figures 2 to 5 shown, the objective reaction time indicator test is divided into four task tests: psychomotor vigilance, sustained attention, fine attention, and visual fatigue. The test procedures for the four task tests all perform the following operations: A1. Before the test starts, a prompt for the judgment and clicking method of the test and a start test button are presented in the center of the screen.

[0029] A2. After the tested dispatcher clicks the start test button, several groups of test questions are presented in the center of the screen, and the tested dispatcher makes judgments and clicks the button according to the prompt before the test starts.

[0030] A3. After the tested dispatcher makes a judgment on each set of test questions presented on the screen and clicks the button for operation, the psychological fatigue test web system automatically jumps to the next set of test questions until several sets of test questions have been completed.

[0031] A4. After the psychological fatigue test web system has completed all the presented test questions, the test data is automatically saved to the background database.

[0032] A5. Based on the judgments made by the tested dispatcher on the test questions and the button click operations, the number of correct and incorrect answers for each set of test questions is counted.

[0033] A6. Calculate the reaction time R for each completed set of test questions k : (1); Where k is the objective reaction time index, is the time stamp when the test question is displayed, is the time stamp of the button operation.

[0034] A7. Store the test data of the tested dispatcher each time. After the test is completed, calculate the average reaction time A k : (2); Where R k is the reaction time, and G k is the total number of operations for the k-th type of objective reaction time index test.

[0035] Take the average reaction time of psychomotor vigilance as the reaction time data of psychomotor vigilance, take the average reaction time of sustained attention as the reaction time data of sustained attention, take the average reaction time of fine attention as the reaction time data of fine attention, and take the average reaction time of visual fatigue as the reaction time data of visual fatigue.

[0036] As Figure 2 shown, for the test of reaction time data of sustained attention: Use the average reaction time judged by the orientation of the aircraft nose as the test data.

[0037] As Figure 3 shown, for the test of reaction time data of fine attention: Use the average reaction time judged by the flight number matrix matching as the test data.

[0038] As Figure 4 shown, for the test of reaction time data of visual fatigue: Use the average reaction time selected by route tracking as the test data.

[0039] As Figure 5 shown, for the test of data on influencing factors of psychological fatigue: Collect data on various influencing factors of psychological fatigue as the test data.

[0040] Step 3: Collect the test data of the objective reaction time index through the mental fatigue test web system, and collect the physiological indexes through an electroencephalogram device (frontal patch electroencephalogram signal recorder) and an eye movement device (eye movement signal tracking recorder) respectively, and then process the collected test data of each item.

[0041] After completing the collection of the test data, filter and denoise the obtained electroencephalogram signal data, extract the power spectral density of the α wave and the power spectral density of the θ wave through fast Fourier transform and remove the outliers through the quartile method, calculate the mean value of the power spectral density of the α wave and the θ wave per hour respectively, extract the fixation speed and the number of blinks from the obtained eye movement data and remove the outliers through the quartile method, calculate the mean value of the fixation speed per hour and the mean value of the blink frequency respectively, remove the outliers from the collected reaction time data through the quartile method, and directly perform time alignment processing on the test data after processing each index according to the recorded timestamp.

[0042] Step 4: Based on the processed test data of each item, construct a dispatcher Bayesian network, calculate the influence probabilities of the subjective and objective influencing factors on the sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time under different mental fatigue states respectively, and then construct a prediction model for the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time.

[0043] As Figure 6 shown, the steps to construct a prediction model for the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time are as follows: 1. Calculate the reaction time growth rate based on the test data, and use piecewise linear fitting to determine the growth rate stage and the corresponding growth rate under different mental fatigue states.

[0044] The growth rate stage and the corresponding growth rate of each index are shown in Table 1: Table 1 Growth rate stage and corresponding growth rate of each index

[0045] 2. Construct a flight dispatcher Bayesian network, and analyze and calculate the influence probabilities of the subjective and objective influencing factor data of mental fatigue on sustained attention, fine attention, and visual fatigue.

[0046] 3. After inputting the test data, infer the growth rate stage of the corresponding sustained attention reaction time index, fine attention reaction time index, and visual fatigue reaction time index per hour based on the influence probability of the subjective and objective influencing factors of mental fatigue on sustained attention, fine attention, and visual fatigue. Finally, starting from the first test data of the sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time, recursively generate the predicted reaction time values for each subsequent hour in combination with the growth rate: (3); Wherein, is the predicted reaction time value of the k-th objective reaction time index after the t-th hour, is the first test reaction time data of the k-th objective reaction time index, n is the number of tests to be predicted, is the growth rate stage inferred from the i-th test of the k-th objective reaction time index, is the end time of the i-th test, is the start time of the i-th test.

[0047] Step 5. According to the prediction models of the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time, calculate the predicted reaction time values corresponding to each subsequent hour through the reaction time data of each first test, which are the prediction results.

[0048] Step 6. Determine the thresholds of each mental fatigue state index by analyzing each test data.

[0049] As Figure 7 shown, the mental fatigue state index threshold determination process performs the following operations: According to each test data, determine the thresholds of each mental fatigue state index respectively. The thresholds of the EEG indexes of α-wave power and θ-wave power are determined by extracting the distribution valley values through Gaussian kernel density estimation; the eye movement indexes of fixation speed and blink frequency are analyzed by a sliding window to extract the extreme points of the coefficient of variation as the reference boundaries; the reaction time indexes respectively perform K-means clustering and density-based spatial (DBSCN) clustering to divide the agile, slightly sluggish, and severely sluggish response groups, and calculate the mean value of the clustering boundary as the threshold; obtain the discrimination thresholds of the mental fatigue state discrimination indexes in the three states of no fatigue, moderate fatigue, and severe fatigue, and calculate the ratio of the between-class variance to the within-class variance of each mental fatigue state discrimination index to obtain the clustering boundary discrimination degree. The discrimination thresholds of each mental fatigue state discrimination index in the three states of no fatigue, moderate fatigue, and severe fatigue are shown in Table 2: Table 2 Discrimination Thresholds

[0050] Use the discrimination thresholds of the various psychological fatigue state discrimination indicators given in Table 2 above in the three states of no fatigue, moderate fatigue, and severe fatigue as a reference in the construction process of the discrimination model.

[0051] Step 7: Construct a dispatcher's psychological fatigue state discrimination model based on the thresholds of the psychological fatigue state indicators. Based on the dispatcher's psychological fatigue state discrimination model, combine the test data and the prediction results, and discriminate the dispatcher's psychological fatigue state according to the three states of no fatigue, moderate fatigue, and severe fatigue.

[0052] As Figure 8 shown, the psychological fatigue state discrimination program process performs the following operations: According to the discrimination thresholds of the various psychological fatigue state discrimination indicators in the three states of no fatigue, moderate fatigue, and severe fatigue, first use the Gaussian mixture model to cluster the test data and perform a preliminary division to obtain a pseudo-classification result. Based on the pseudo-classification result, by calculating the ratio of the between-class variance to the within-class variance of each psychological fatigue state indicator under the pseudo-classification result, obtain the rule path discrimination degree of each psychological fatigue state indicator. Normalize the rule path discrimination degree of each psychological fatigue state indicator by standard deviation, and set the fusion weight according to the clustering boundary discrimination degree and the rule path discrimination degree of each psychological fatigue state indicator as the threshold weight, that is, construct a dispatcher's psychological fatigue state discrimination model. The dispatcher's psychological fatigue state discrimination model outputs the three states of no fatigue, moderate fatigue, and severe fatigue respectively.

[0053] Taking the test data of a certain dispatcher under test after 7 hours of work as an example, calculate the test data after acquisition and processing through formulas (1) and (2). According to the psychological fatigue state discrimination model, obtain the fatigue state discrimination result. The test data and discrimination results of each index of the dispatcher under test are shown in Table 3: Table 3 Test Data and Discrimination Results

[0054] According to formula (3), obtain the prediction results of the sustained attention, fine attention, and visual fatigue reaction times of the dispatcher under test after 2 hours. Further obtain the prediction discrimination results according to the prediction data, as shown in Table 4: Table 4 Prediction Results and Prediction Discrimination Results

[0055] According to the test data shown in Table 2, use formulas (1), (2), and (3) to predict the sustained attention, fine attention, and visual fatigue reaction times of the dispatcher under test after 2 hours of work respectively, and complete the fatigue state discrimination of this test of the dispatcher under test through the psychological fatigue state discrimination model.

[0056] As shown in Table 2, after the test data collection of various index data of the tested dispatcher, the obtained electroencephalogram (EEG) signal data is filtered and denoised. The power spectral density of the alpha wave and the power spectral density of the theta wave are extracted by fast Fourier transform, and outliers are removed by the quartile method. The mean values of the power spectral density of the alpha wave and the theta wave per hour are calculated respectively. The mean value of the power spectral density of the theta wave is -0.47, and the mean value of the power spectral density of the alpha wave is -1.58. The fixation speed and the number of blinks are extracted from the obtained eye movement data, and outliers are removed by the quartile method. The mean value of the fixation speed and the mean value of the blink frequency per hour are calculated respectively, and the fixation speed is 1.96 pixels / ms, and the blink frequency is 29 times / minute. According to the psychological fatigue test web system, outliers are removed from the collected reaction time data by the quartile method. The number of errors in sustained attention is 5, the number of errors in fine attention is 2, and the number of errors in visual fatigue is 1. After removing outliers from the collected reaction time data, according to Formula (1) and Formula (2), the reaction time of sustained attention is 0.59 s, the reaction time of fine attention is 5.36 s, and the reaction time of visual fatigue is 0.29 s. Based on the prediction models of sustained attention, fine attention, and visual fatigue reaction time, according to Formula (3), it is predicted that after 2 hours, the reaction time of sustained attention will reach 0.62 s, the reaction time of fine attention will reach 5.49 s, and the reaction time of visual fatigue will reach 0.38 s. The processed test data of each index is directly aligned in time according to the recorded timestamp. According to the psychological fatigue state discrimination model, this group of data conforms to the characteristic performance of severe psychological fatigue. Therefore, the test data of this tested dispatcher is judged to be in a severe fatigue state.

[0057] Based on the multi-modal indicators such as EEG, eye movement, physiology, and reaction time, through the discrimination of various characteristic items and the model prediction results, it can be concluded that the tested dispatcher has shown a relatively obvious severe psychological fatigue state.

[0058] According to the "Supervision Manual on Mental Fatigue Management Practices" (Document No. 9966) issued by the International Civil Aviation Organization, personnel in a psychological fatigue state above moderate level shall not perform safety-related duties.

Claims

1. A method for predicting and discriminating dispatchers' mental fatigue based on physiological and reaction time indicators, characterized in that, The method includes the following steps: Step 1: Determine the mental fatigue state index constructed from physiological indexes and objective reaction time indexes; the physiological indexes are alpha wave power spectral density, theta wave power spectral density, fixation velocity, and blink frequency; the objective reaction time indexes are psychomotor vigilance reaction time, sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time; Step 2: Determine that the dispatcher mental fatigue test consists of three parts: physiological index test, objective reaction time index test, and subjective and objective influencing factors test of mental fatigue. By designing a test program, construct a mental fatigue state test web system, and the mental fatigue state test web system connects Java and the background database through PHP; Step 3: Collect the test data of the objective reaction time index through the mental fatigue test web system, and collect the physiological indexes through an electroencephalogram device and an eye movement device respectively, and then perform data processing on the collected test data; Step 4: Based on the processed test data, construct a dispatcher Bayesian network, calculate the influence probabilities of subjective and objective influencing factors on the sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time under different mental fatigue states respectively, and then construct prediction models for the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time; Step 5: Based on the prediction models of the dispatcher's sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time, calculate the corresponding reaction time prediction values for each hour after the first test through the reaction time data of the first test, which is the prediction result; Step 6: Determine the thresholds of each mental fatigue state index by analyzing the test data; Step 7: Based on the thresholds of the mental fatigue state indexes, construct a dispatcher mental fatigue state discrimination model. According to the dispatcher mental fatigue state discrimination model, combined with the test data and the prediction result, discriminate the dispatcher's mental fatigue state into three states: no fatigue, moderate fatigue, and severe fatigue.

2. The dispatcher psychological fatigue prediction and discrimination method based on physiological and reaction time indexes according to claim 1, characterized in that When testing the dispatcher, obtain the electroencephalogram signal data by wearing an electroencephalogram device, and obtain the eye movement data by installing an eye tracker; obtain the mental fatigue influencing factor data and objective reaction time data through the mental fatigue test web system and the test program; After completing the collection of test data, filter and denoise the obtained electroencephalogram signal data, extract the alpha wave power spectral density and theta wave power spectral density and remove the outliers, calculate the mean values of the alpha wave and theta wave power spectral density for each hour respectively, extract the fixation velocity and blink count from the obtained eye movement data, and remove the outliers, calculate the mean value of the fixation velocity and the mean value of the blink frequency for each hour respectively, remove the outliers from the collected reaction time data, and directly perform time alignment processing on the test data after processing each index according to the recorded timestamp.

3. The dispatcher psychological fatigue prediction and discrimination method based on physiological and reaction time indexes according to claim 2, characterized in that, Test the psychomotor vigilance, sustained attention, fine attention, and visual fatigue reaction time of the dispatcher through the mental fatigue test web system; the mental fatigue test web system automatically records the start time, average reaction time, number of errors of each test, and the mental fatigue influencing factor data, and generates a timestamp.

4. The method for predicting and discriminating the psychological fatigue of dispatchers based on physiological and reaction time indexes according to claim 3, characterized in that, According to the mental fatigue test web system, the objective reaction time index test is divided into four task tests: psychomotor vigilance, sustained attention, fine attention, and visual fatigue. The test procedures for the four task tests all perform the following operations: A1. Before the test starts, the test judgment and click method prompts and the start test button are presented in the center of the screen; A2. After the tested dispatcher clicks the start test button, several groups of test questions are presented in the center of the screen, and the tested dispatcher makes judgments and clicks the button operations according to the prompts before the test starts; A3. After the tested dispatcher makes a judgment and clicks the button operation for each group of test questions presented on the screen, the mental fatigue test web system automatically jumps to the next group of test questions until several groups of test questions are completed; A4. After the mental fatigue test web system completes all the presented test questions, the test data is automatically saved to the background database; A5. According to the judgments and click button operations made by the tested dispatcher on the test questions, the number of correct and incorrect answers for each group of test questions is counted; A6. Calculate the reaction time for completing each set of test questions R k : （1）； wherein, k is the objective reaction time index, is the timestamp when the test question is displayed, is the timestamp of the key operation; A7. Store the test data of each tested dispatcher, and calculate the average reaction time after the test is completed. A k : （2）； In the formula, R k is the reaction time, G k is the k total number of operations for testing the 5. The dispatcher psychological fatigue prediction and discrimination method based on physiological and reaction time indexes according to claim 4, characterized in that The steps to construct the prediction models for the sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time of dispatchers are as follows: Calculate the reaction time growth rate based on the test data, and use piecewise linear fitting to determine the growth rate stage and the corresponding growth rate under different mental fatigue states. Construct a Bayesian network for flight dispatchers, analyze and calculate the influence probabilities of the subjective and objective influencing factors of mental fatigue on sustained attention, fine attention, and visual fatigue. After inputting the test data, infer the growth rate stage of the corresponding sustained attention reaction time index, fine attention reaction time index, and visual fatigue reaction time index per hour through the influence probabilities of the subjective and objective influencing factors of mental fatigue on sustained attention, fine attention, and visual fatigue. Finally, starting from the first test data of the sustained attention reaction time, fine attention reaction time, and visual fatigue reaction time, generate the predicted reaction time values for each subsequent hour by combining the growth rate recurrence: （3）； Wherein, is the predicted reaction time value of the k th objective reaction time index after the t th hour, is the reaction time data of the first test of the k th objective reaction time index, n is the number of tests to be predicted, is the growth rate stage inferred from the k th objective reaction time index at the th test, i is the end time of the th test, and is the start time of the th test.

Images