A kind of advanced brain fatigue early warning system, device and storage medium

Abstract

The present invention discloses an advanced brain fatigue early warning system, equipment and storage medium, comprising a data acquisition module, a preprocessing module, an adjustment capacity calculation module, a fatigue early warning module, a storage module and a communication module; the data acquisition module is used to acquire pupil diameter data of an object to be predicted; the preprocessing module is used to preprocess the pupil diameter data, remove invalid data, and perform linear interpolation; the adjustment capacity calculation module is used to calculate based on the preprocessed pupil diameter data, and correct it through the ambient light index to obtain a continuously updated brain fatigue adjustment capacity time series; the fatigue early warning module is used to establish an advanced prediction model according to the obtained fatigue adjustment capacity time series, and calculate the remaining time to brain fatigue by using the intersection of the prediction line obtained by the advanced prediction model and the brain fatigue alarm threshold line to start the alarm; the communication module uploads the brain fatigue adjustment capacity, alarm mode, etc. to a remote server for monitoring by the management party; and the storage module archives the intermediate data.

Description

Technical Field

[0001] The present invention relates to the field of neuroscience and signal processing technology, and in particular to an advanced brain fatigue early warning system, device and storage medium. Background Art

[0002] The fatigue of vehicle drivers, pilots, and traffic controllers is mainly manifested as brain fatigue. Brain fatigue is characterized by the masking effect of brain fatigue in the early stage and the sudden effect in the later stage. In layman's terms, although you don't look tired at the beginning of the task, brain fatigue has been accumulating, and you will be distracted and doze off unconsciously in the middle and late stages. The masking effect in the early stage and the suddenness in the middle and late stages bring about major hidden dangers of short warning time. The overt brain fatigue is the result of the combined action of two internal factors. One is the gradually accumulated physiological brain fatigue, which reflects the continuous consumption of cognitive resources, and the other is the regulatory system, which reflects the regulatory and compensatory effects of the higher centers. It can mask physiological fatigue in the early stage. Usually, after 30-120 minutes of the task, physiological fatigue can no longer be masked and will show brain fatigue. Conventional indicators are reflected through facial expressions, eyelid closure rate, EEG, etc., which belong to "post-fatigue measurement" and have serious safety hazards.

[0003] For example, the application number is 200920160060.4, which is a kind of oral fatigue warning device. Under normal circumstances, the teeth are clenched and no alarm is sounded. When tired or distracted, the alarm is released and the buzzer starts to alarm. This method requires the driver to bite the alarm at all times, which disturbs the driver. Moreover, when fatigue is found, the fatigue has reached a very dangerous level, and the reminder effect for safe driving needs to be improved. In the Chinese patent "Automobile Driver Fatigue State Prediction System Patent" with application number 201310229235.3, it uses the driver's eye closure time as a feature to start the alarm. It can be seen that the above methods all belong to "post-fatigue measurement". When fatigue is measured, brain fatigue has already manifested. At this time, fatigue has reached a very dangerous level, and there are serious safety hazards. The reminder effect for safe driving needs to be improved.

[0004] In summary, most of the current brain fatigue measurement methods reflect the external manifestations of brain fatigue from different aspects, reflecting the results of the regulation system on physiological brain fatigue (physiological fatigue is masked). Therefore, it can only be measured after physiological brain fatigue accumulates to a large enough level (the masking effect fails) and brain fatigue is manifested. The warning system based on this post-fatigue measurement method often has a warning lag and cannot provide intelligent warnings to drivers in advance. Summary of the invention

[0005] In response to the above-mentioned problems, the present invention aims to provide a brain fatigue early warning system to overcome the masking effect of brain fatigue. Based on the advanced prediction model, it can advance brain fatigue detection from the traditional "post-fatigue measurement" to "pre-fatigue measurement", and use the brain fatigue early warning system for intelligent prediction and intervention.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] An advanced brain fatigue warning system, characterized by comprising a data acquisition module, a preprocessing module, an adjustment capacity calculation module, a fatigue warning module, a communication module and a storage module;

[0008] A data acquisition module, used to acquire pupil diameter data of an object to be predicted;

[0009] A preprocessing module, used for preprocessing the pupil diameter data, removing invalid data, and performing linear interpolation to obtain preprocessed pupil diameter data;

[0010] An accommodation ability calculation module is used to calculate based on the preprocessed pupil diameter data and correct it through the ambient light index to obtain a continuously updated accommodation ability time series;

[0011] The fatigue warning module is used to perform mathematical modeling based on the obtained adjustment capacity time series, establish an advance prediction model, calculate the remaining time of brain fatigue and issue an alarm;

[0012] The communication module is used to upload the brain fatigue adjustment ability, alarm mode, etc. to the remote server for easy monitoring by the management party;

[0013] The storage module is used to archive intermediate data.

[0014] Furthermore, the fatigue warning module includes a brain fatigue occurrence time prediction submodule and an alarm severity setting submodule;

[0015] The brain fatigue occurrence time prediction submodule is used to establish an advance prediction model based on the obtained adjustment ability time series, and use the advance prediction model to obtain a prediction line that has not yet occurred, and then intersect the obtained prediction line with a predetermined threshold line to calculate the remaining time for brain fatigue to occur;

[0016] The alarm severity setting submodule is used to compare the remaining time for brain fatigue to occur with the threshold set by the user, trigger different alarm severities, and issue alarms according to different alarm severities.

[0017] Furthermore, the alarm severity is divided into several levels, and different levels of alarm severity correspond to different alarm sounds, voices and vibration intensities.

[0018] Furthermore, the method for constructing the lead prediction model includes:

[0019] Step 1: Obtain the pupil diameter data and ambient light data of the current person to be measured, preprocess the pupil diameter data, remove outliers, and perform linear interpolation on the vacancies after removing outliers to obtain the preprocessed pupil diameter data;

[0020] Step 2: Define an accommodation ability index based on the preprocessed pupil diameter data, and calculate the accommodation ability index to obtain the time series of the accommodation ability index;

[0021] Step 3: Calculate the ambient light index based on the ambient light data, and use an adaptive filter to remove ambient light interference to obtain the corrected time series of the accommodation ability index;

[0022] Step 4: Construct a brain fatigue accommodation ability prediction model based on the corrected time series of the accommodation ability index.

[0023] Furthermore, the specific steps of Step 2 are as follows:

[0024] Step 21: Denote the pupil diameter time series as x (1) (i), 1 ≤ i ≤ N, where N is the length of the time series, and take m points from it to form a vector u m (j):

[0025] u m (j) = [x (1) (j + k): 0 ≤ k ≤ m - 1];

[0026] Step 22: Define the distance between the vectors u m (i) and u m (j) as d:

[0027]

[0028] Step 23: Calculate the standard deviation of the time series x (1) (i) as r, and count the number of all groups of vectors u m (i) and u m (j) that satisfy d < r, and denote it as A m ;

[0029] Step 24: Calculate S (1) :

[0030]

[0031] Step 25: Average every two values of the pupil diameter time series x (1) (i) as new values to form a new time series x (2)(i), repeatedly execute steps 21 - 24 to obtain S (2) ;

[0032] Step 26: Repeatedly execute step 25 to successively obtain S (3) , S (4) , …, S (τ) , where τ is a number in the interval [10, 30];

[0033] Step 27: Define the calculation formula for the accommodation ability index as:

[0034]

[0035] Step 28: Calculate S using a sliding window, and the formula for setting the sliding window width is:

[0036]

[0037] where, F s is the sampling frequency of the pupil diameter;

[0038] Step 29: Obtain the time series S(n) of the accommodation ability index based on the calculated accommodation ability index.

[0039] Furthermore, the specific steps of step 3 are as follows:

[0040] Step 31: Collect the ambient light time series y(i), 1 ≤ i ≤ N through an environmental sensor, where N is the length of the time series, and take m points from it to form a vector u m (j):

[0041] u m (j) = [y(j + k): 0 ≤ k ≤ m - 1];

[0042] Step 32: Define the distance between vectors u m (i) and u m (j) as d:

[0043]

[0044] Step 33: Calculate the standard deviation of the time series y(i) as r, and count the number of all groups of vectors u m (i) and u m (j) that satisfy d < r, and denote it as A m ;

[0045] Step 34: Obtain T according to formula (1);

[0046] Step 35: Calculate T using a sliding window to obtain the ambient light index time series T1, T2, T3, ….

[0047] Step 36: Redefine the sliding window sequence of the adjustment capability index S(n) obtained in step 29 as d(n), n=1, 2, ...;

[0048] Step 37: Express d(n) as:

[0049] d(n)=S(n)+T1(n),n=1,2,…

[0050] Among them, T1(n) is the pupil accommodation index component caused by ambient light; S(n) is the real accommodation ability index component caused by the pupil;

[0051] e(n)=d(n)-z(n),n=1,2,…

[0052] Where, e(n) is the error between d(n) and the adaptive filter output z(n);

[0053] Step 38: The weight vector w of the adaptive filter can be obtained by iterative method:

[0054] w(n+1)=w(n)+μ[p-Rw(n)]

[0055] Here n represents the number of iterations, μ represents a positive constant, p is the cross-correlation vector of the data, and R is the correlation matrix of the data.

[0056] Step 39: Based on the solved e(n), obtain the time series of the corrected regulatory capability index.

[0057] Furthermore, the constructed prediction models include an advanced prediction model for predicting when the time of brain fatigue manifestation is greater than 5 minutes from the current time, and a near-threshold accurate prediction model for predicting when the time of brain fatigue manifestation is less than or equal to 5 minutes from the current time:

[0058] (1) The brain fatigue advance prediction model is used to predict the situation where the occurrence of brain fatigue manifestation is greater than 5 minutes from the current time. The specific steps are as follows:

[0059] Step 41: Define the loss function for the time series of the adjusted capability index:

[0060]

[0061] Step 42: For each input time point t n , the predicted estimate is:

[0062]

[0063] Step 43: Define the average loss function as:

[0064]

[0065] Step 44: Solve the parameters that minimize L(p,b) according to formula (7):

[0066]

[0067] and:

[0068]

[0069]

[0070] (2) The accurate prediction model of brain fatigue is used to predict the situation where the occurrence of brain fatigue manifestation is less than or equal to 5 minutes from the current time. The construction steps are as follows:

[0071] Step 45: During the real-time recording process, record for a period of time;

[0072] Step 46: According to the above steps, the time series of the moment-adjustment capability index (t1, S1), (t2, S2), ... (t N2 ,S N2 ), then an autoregressive model is established.

[0073] Step 47: The order of the autoregressive model can be selected as 50, the data window is 100s, and it is updated once every 1s, which is used to predict the waveform of 25s.

[0074] An electronic device, comprising:

[0075] at least one processor;

[0076] and a memory communicatively coupled to the at least one processor;

[0077] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps described in any one of claims 4 to 7.

[0078] A computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the steps according to any one of claims 4 to 7.

[0079] The beneficial effects of the present invention are:

[0080] First, the system proposed in the present invention can be directly installed on a transportation carrier or a workbench (such as a vehicle, airplane, ship, drone operating table, air traffic controller workbench, etc.), and the pupil diameter is collected by a non-contact, non-interference collection device, without the need for electrode contact or wear. In addition, the system also sets multiple levels of alarm intensity. According to the intensity set in advance by the user, an alarm can be issued with the corresponding alarm intensity when fatigue is detected.

[0081] Second, the present invention proposes an advanced prediction model to predict brain fatigue through the adjustment ability index of pupil diameter, directly measuring the size of cognitive adjustment ability, overcoming the limitations of traditional fatigue detection and prediction, and not being affected by masking effect and burst effect. Based on the prediction model, a brain fatigue early warning system is constructed. Through this early warning system, drivers of vehicles, airplanes, etc. or instrument operators can be warned by voice, vibration, etc., and corresponding warnings can be obtained before the personnel reach the dangerous brain fatigue threshold, thereby realizing pre-fatigue intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 This is the module architecture diagram of this system;

[0083] Figure 2 Flowchart of the system's alarming through the advance prediction model;

[0084] Figure 3 It is the test interface of the brain fatigue early warning system;

[0085] Figure 4 Divide the text display area into sub-areas. DETAILED DESCRIPTION

[0086] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0087] like Figure 1 As shown, the present invention proposes a brain fatigue early warning system, which includes a data acquisition module, a preprocessing module, an adjustment ability calculation module, a fatigue early warning module, a storage module and a communication module;

[0088] A data acquisition module is used to obtain pupil diameter data of the person to be predicted (vehicle driver, pilot, drone pilot, air traffic controller, etc.);

[0089] A preprocessing module, used for preprocessing the pupil diameter data, removing invalid data, and performing linear interpolation to obtain preprocessed pupil diameter data;

[0090] An accommodation capacity calculation module is used to calculate based on the preprocessed pupil diameter data and correct it through the ambient light index to obtain a continuously updated time series of brain fatigue accommodation capacity;

[0091] Fatigue warning module, including brain fatigue occurrence time prediction submodule and alarm intensity setting submodule;

[0092] The brain fatigue occurrence time prediction submodule is used to perform mathematical modeling based on the obtained adjustment ability time series, establish an advance prediction model, and use the advance prediction model to obtain the forecast complexity time series that has not yet occurred, and then intersect the obtained prediction line with the pre-determined threshold line to calculate the remaining time of the brain fatigue that will occur;

[0093] The alarm severity setting submodule is used to compare the remaining time for occurrence of brain fatigue with the threshold value set by the user (which can be obtained through experiments), trigger different alarm severity levels, and issue alarms according to different alarm severity levels;

[0094] The alarm severity that can be set in this system is alarm severity level 1, alarm severity level 2, alarm severity level 3, ...; different levels of alarm severity correspond to different alarm sounds, voices and vibration intensities. As a preferred embodiment, its alarm severity information is shown in Table 1.

[0095] Table 1 Examples of alarm severity classification

[0096]

[0097]

[0098] 2. Construction method of advanced prediction model

[0099] The advanced brain fatigue early warning system proposed by the present invention is implemented based on an advanced prediction model. Constructing the advanced prediction model includes the following steps:

[0100] Step 1: Obtain pupil diameter data of the driver to be monitored, and preprocess the obtained pupil diameter data: remove outliers, and perform linear interpolation on the vacancies after removing the outliers, so as to obtain preprocessed pupil diameter data;

[0101] Step 2: define the accommodation ability index, take a time window for the preprocessed pupil diameter data, obtain the corresponding pupil diameter time series data, calculate the pupil diameter time series data within the time window to obtain the time series of the accommodation ability index, and build a linear prediction model based on the time series; it specifically includes:

[0102] (1) Define the accommodation ability index based on the pupil diameter time series

[0103] First, denote the pupil diameter time series as x (1) (i), where 1 ≤ i ≤ N and N is the length of the time series. Select m points from it to form a vector u m (j):

[0104] u m (k) = [x (1) (j + k): 0 ≤ k ≤ m - 1]

[0105] where the j variable is the serial number of the points sequentially selected in the time series from front to back; the k variable is the offset with a range from 0 to m - 1; m is the dimension of the embedding vector u m (j), that is, the number of elements in the vector;

[0106] Second, define the distance between two vectors u m (i) and u m (j) as d,

[0107]

[0108] Third, calculate the standard deviation of the pupil diameter time series x (1) (i) as r, and count the number of all pairs of vectors u m (u m (i) and u m (j)) with a distance d < r (count 1 for a pair of vectors satisfying the above conditions), denoted as A m , and calculate S m based on A (1) :

[0109]

[0110] Third, calculate the average value of every 2 values of the pupil diameter time series x (1) (i), and use the obtained average value as the new value, thus forming a new pupil diameter time series x (2) (i). Repeat the above steps and obtain S (2) according to Formula 1, and so on, to obtain S (3) , S (4) , …, S (τ) where τ is a number between 10 and 30, and this value depends on the sampling rate of the pupil diameter. The higher the sampling rate, the larger the τ value. Preferably, when the sampling rate is 100 Hz, τ is set to 10.

[0111] Finally, define the accommodation ability index as:

[0112]

[0113] (2) Calculate the Adjustability Index through Sliding Window

[0114] The present invention uses a sliding window to calculate S, and the sliding window width is determined by the following formula:

[0115]

[0116] Among them, F s is the sampling frequency of pupil diameter;

[0117] The step length of the sliding window movement is between 0.1-10s. The shorter the step length, the more conducive it is to the subsequent prediction modeling (improving prediction accuracy), but the amount of calculation is also greater. Therefore, the step length of the present invention is set to 1s.

[0118] Through calculation, the time series S(n) of the adjustable capability index can be obtained.

[0119] (3) A linear model for predicting physiological brain fatigue regulation ability is constructed based on the continuous value of the regulation ability index to obtain fatigue-related parameters, which specifically include:

[0120] First, define the loss function:

[0121]

[0122] Secondly, for each input time point t n , the predicted estimate is:

[0123]

[0124] The loss function is used to calculate the difference between the predicted estimate and the true value:

[0125]

[0126] Again, the average loss function is defined as:

[0127]

[0128] And solve the parameters (p, b) that minimize L(p, b) using formula 7:

[0129]

[0130] get:

[0131]

[0132]

[0133] Among them, p * ,b * They are the slope and intercept of the linear model predicting the physiological system's brain fatigue regulation ability;

[0134] (4) The accurate prediction model of brain fatigue is used to predict the situation where the occurrence of brain fatigue manifestation is less than or equal to 5 minutes from the current time. Its construction steps are as follows: in the process of real-time recording, record a period of time; and according to the above steps, obtain the time series of the time-regulation ability index (t1, S1), (t2, S2),… (t N2 ,S N2 ), an autoregressive model is established. The order of the autoregressive model can be selected as 50, the data window is 100s, and it is updated every 1s to predict the waveform of 25s.

[0135] In summary, the linear models described in formula 5, formula 8 and formula 9 and the AR model described in (4) are the advance prediction models constructed by the present invention. The former is used for advance prediction, and the latter is used for more accurate prediction when approaching brain fatigue. The principle is to consider both the local periodicity and long-term attenuation trend of the adjustment ability curve. Physiological brain fatigue increases approximately linearly with the work task. At this time, the advance prediction model constructed by the present invention uses the negative slope and the starting value at the beginning of the task as linear prediction parameters.

[0136] Step 3: Predict brain fatigue based on the established advance prediction model and the preset threshold;

[0137] Based on the obtained linear prediction parameters, 3-5 (or more thresholds) can then be determined experimentally, for example, an early alarm threshold (20 minutes in advance), a mid-term alarm threshold (5 minutes), a late alarm threshold (1 minute) and a critical alarm threshold (20 seconds);

[0138] like Figure 2 As shown, the system issues an alarm based on the above method according to different final alarm thresholds.

[0139] Example

[0140] In order to verify the effectiveness of the system proposed in the present invention, its functions are verified.

[0141] 1. Test system description

[0142] In order to facilitate testing, the present invention designs a test interface for the brain fatigue early warning system. Figure 3 As shown. Figure 3 It can be seen that it is divided into chart area, text display area, button area, and text input area;

[0143] The area framed in red is the chart area, of which the main function of the left half is to draw a real-time pupil diameter image, that is, to obtain data from the eye tracker and draw an image of its relationship with time; the main function of the right half is to draw a physiological brain fatigue index image, that is, to analyze the pupil diameter data and then deduce its physiological brain fatigue index. The curve is drawn in red, and the remaining working time of the tested person is predicted. Two prediction methods are used, namely the linear model and the AR model. The first drawing will start 5 minutes after the start of the calculation, and the predicted images are drawn in blue and orange respectively.

[0144] The orange area is the button area, where the start button starts the eye tracker, and the eye tracker will receive the experimental data of the person being tested; the startall button starts drawing the pupil diameter curve and the MSE curve at the same time, which is equivalent to pressing the pupil diameter curve button and the MSE button at the same time; the pupil diameter curve button and the MSE button have the functions of starting drawing the pupil diameter curve and the MSE curve respectively; the calibration button starts the eye tracker calibration; the stop button is a stop button that can pause the program; the clear button mainly clears data, which is equivalent to resetting.

[0145] The green area is the text input area. Its main function is to save the data and images. You need to enter a file name.

[0146] The blue area is the text display area, which is divided into 5 areas, such as Figure 4 As shown, the value displayed in Area 1 is the number of pupil data, the value displayed in Area 2 is the size of the pupil, the value displayed in Area 3 is the size of the fatigue index, the value displayed in Area 4 is the point of the fatigue index, and Area 5 reports the remaining working time of the person being tested. The first prediction begins 5 minutes after the start of the calculation and is updated every 30 seconds.

[0147] 2. Test process

[0148] S1: Align the human eyeball with the eye tracker at a certain distance, and calibrate according to the eye tracker calibration software process to ensure that the experimental data collected is true and reliable;

[0149] S2: Use simulation driving software to perform simulated driving;

[0150] S3: Enter the test interface of the brain fatigue warning system, click the start button to start the eye tracker, then click the startall button or press the pupil diameter curve button and the MSE button in sequence. At this time, you can click the stop button to terminate the code according to user needs. After clicking the stop button, you can click the clear button to clear the data or save the data through the text input area according to needs;

[0151] S4: Preset the alarm threshold time, for example, 20 minutes. When the brain fatigue intelligent early warning system predicts that the time is less than 20 minutes, a voice reminder is given according to the reminder method of the corresponding alarm intensity, and at the same time, the change in the predicted fatigue time is observed to see whether it increases.

[0152] 3. Test results

[0153] After executing the above test process, it can be seen from the remaining working time area of ​​the test person in the test interface of the brain fatigue warning system that after the alarm intervention of this warning system, the predicted fatigue time of the test person increases and the curve also rises to a certain extent, which shows that the warning system of the present invention can perform intelligent intervention on the driver.

[0154] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An advanced brain fatigue early warning system, characterized in that: It includes a data acquisition module, a preprocessing module, an adjustment capacity calculation module, a fatigue warning module, a communication module and a storage module; A data acquisition module, used to acquire pupil diameter data of an object to be predicted; A preprocessing module, used for preprocessing the pupil diameter data, removing invalid data, and performing linear interpolation to obtain preprocessed pupil diameter data; An accommodation ability calculation module is used to calculate based on the preprocessed pupil diameter data and correct it through the ambient light index to obtain a continuously updated accommodation ability time series; The fatigue warning module is used to perform mathematical modeling based on the obtained adjustment capacity time series, establish an advance prediction model, calculate the remaining time of brain fatigue and issue an alarm; The communication module is used to upload the brain fatigue adjustment ability and alarm mode to the remote server for easy monitoring by the management party; The storage module is used to archive intermediate data.

2. The advanced brain fatigue early warning system according to claim 1, characterized in that: The fatigue warning module includes a brain fatigue occurrence time prediction submodule and an alarm intensity setting submodule; The brain fatigue occurrence time prediction submodule is used to establish an advance prediction model based on the obtained adjustment ability time series, and use the advance prediction model to obtain a prediction line that has not yet occurred, and then intersect the obtained prediction line with a predetermined threshold line to calculate the remaining time for the brain fatigue to occur; The alarm severity setting submodule is used to compare the remaining time for brain fatigue to occur with the threshold set by the user, trigger different alarm severities, and issue alarms according to different alarm severities.

3. The advanced brain fatigue early warning system according to claim 2, characterized in that: The alarm severity is divided into several levels, and different levels of alarm severity correspond to different alarm sounds, voices and vibration intensities.

4. The advanced brain fatigue early warning system according to claim 1, characterized in that: The method for constructing the advanced prediction model includes: Step 1: Obtain the pupil diameter data and ambient light data of the current person being measured, preprocess the pupil diameter data, remove outliers, and perform linear interpolation on the gaps after removing outliers to obtain preprocessed pupil diameter data; Step 2: Based on the preprocessed pupil diameter data, define the accommodation ability index, and calculate the accommodation ability index to obtain the time series of the accommodation ability index; Step 3: Calculate the ambient light index based on the ambient light data, use an adaptive filter to remove ambient light interference, and obtain a time series of the corrected adjustment capability index; Step 4: Construct a brain fatigue regulatory ability prediction model based on the time series of the corrected regulatory ability index.

5. The advanced brain fatigue early warning system according to claim 4, characterized in that: The specific steps of step 2 are: Step 21: Record the pupil diameter time series as ,in is the length of the time series, and takes Points form vectors : ； in, is the serial number of the points selected from the front to the back in the time series; is the offset; is the embedding vector The dimension of Step 22: Define the vector and The distance between : ； Step 23: Calculate the time series The standard deviation of , statistics satisfy All vectors of and and record it as ; Step 24: Calculation : (Formula 1) Step 25: Convert the pupil diameter time series to The average of every two values ​​is used as a new value to form a new time series , repeat steps 21-24 to obtain ; Step 26: Repeat step 25 to obtain ,in Numbers in an interval; Step 27: Define the calculation formula of the adjustment capability index as: (Formula 2); Step 28: Use sliding window calculation , and the setting formula of sliding window width is: (Formula 3); in, is the sampling frequency of pupil diameter; Step 29: Based on the calculated regulatory capability index, obtain the time series of the regulatory capability index .

6. The advanced brain fatigue early warning system according to claim 5, characterized in that: The specific steps of step 3 are: Step 31: Collecting Ambient Light Time Series Through Environmental Sensor ,in is the length of the time series, and takes Points form vectors : ； Step 32: Define the vector and The distance between : ； Step 33: Calculate the time series The standard deviation of , statistics satisfy All vectors of and and record it as ; Step 34: According to formula 1, we get ; Step 35: Use sliding window calculation , get the ambient light index time series ; Step 36: The adjustment capability index obtained in step 29 The sliding window sequence is defined as ; Step 37: It is expressed as: ； in, is the pupil accommodation index component caused by ambient light; The actual accommodation index component caused by the pupil; ； in, yes and the adaptive filter output The error of Step 38: The weight vector of the adaptive filter can be obtained by iteration : ； Here represents the number of iterations, represents a normal number, is the cross-correlation vector of the data, is the correlation matrix of the data; Step 39: Based on the solution , and obtain the time series of the corrected regulatory ability index.

7. The advanced brain fatigue early warning system according to claim 5, characterized in that: The constructed prediction models include an advanced prediction model for predicting when the time of brain fatigue manifestation is greater than 5 minutes from the current time, and a near-threshold accurate prediction model for predicting when the time of brain fatigue manifestation is less than or equal to 5 minutes from the current time: (1) The steps for constructing the brain fatigue advance prediction model are as follows: Step 41: Define the loss function for the time series of the adjusted capability index: ； Step 42: For each input time point , the predicted estimate is: ； Step 43: Define the average loss function as: ； step 44: Solve according to formula 7 Parameters to minimize: ； and: (Formula 8) (Formula 9); (2) The steps for constructing an accurate prediction model for brain fatigue are as follows: Step 45: During the real-time recording process, record for a period of time; Step 46: According to the above steps, the time series of the moment-adjustment capability index is obtained Then, an autoregressive model is established. Step 47: The order of the autoregressive model can be selected as 50, the data window is 100s, and it is updated once every 1s, which is used to predict the waveform of 25s.

8. An electronic device, characterized in that: include, at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps described in any one of claims 4 to 7.

9. A computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to make the computer execute the steps according to any one of claims 4-7.

Images