CN111803065A - Dangerous traffic scene identification method and system based on electroencephalogram data - Google Patents

Abstract

The invention discloses a dangerous traffic scene identification method and system based on electroencephalogram data, wherein the dangerous traffic scene identification method based on the electroencephalogram data comprises the following steps: acquiring a stimulus object and an experimental variable based on a traffic scene; completing driving test of a driver through a stimulating object and an experimental variable, and acquiring EEG data of the driver containing an electroencephalogram signal within a preset time; preprocessing the acquired electroencephalogram signals, and removing signal noise in the electroencephalogram signals; extracting characteristic indexes from the preprocessed electroencephalogram signals, and analyzing by adopting a mathematical statistics method to obtain characteristic indexes with significance; calculating the difference between the predicted value and the true value of the power spectral density with the significant characteristic indexes; and identifying the danger degree of the traffic scene according to the difference value. The method can scientifically measure the dangerous degree of the traffic scene, accurately evaluate the mental load of the driver, is beneficial to controlling the mental load of the driver in a reasonable interval, and effectively reduces the driving risk.

Description

Dangerous traffic scene identification method and system based on electroencephalogram data

Technical Field

The invention relates to the field of traffic driving safety, in particular to a dangerous traffic scene identification method and system based on electroencephalogram data.

Background

With the rapid development of social economy, the automobile reserves in China are rapidly improved, the road construction is continuously promoted, meanwhile, the traffic scenes are increasingly rich, the road traffic participants are continuously increased, the road traffic environment is increasingly complex, and the traffic safety accident occurrence amount is increased year by year. The danger degree of the traffic scene can directly influence the physiological and psychological states of a driver, and when the driver is in a dangerous traffic scene, the mental load borne by the driver is large, so that the phenomena of slow decision making and wrong operation of the driver are easily caused, and traffic accidents are easily caused. In the face of dangerous road traffic scenes, the method scientifically measures the degree of danger of the traffic scenes, accurately evaluates the mental load of the driver, adopts reasonable measures to control the mental load of the driver in a reasonable interval, reduces the driving risk and becomes an important problem.

The research of traffic scene danger degree can not only effectual promotion human driver's driving safety, reduces the incidence of road traffic accident, but also can provide theoretical basis for the research of intelligent driving and vehicle safety technique.

For an automatic driving vehicle, the functions of intelligent driving are finally judged, the performance of the vehicle is achieved, whether the vehicle can be driven in a daily traffic environment or not is judged, the technical assessment of the vehicle must be completed, and an important link in examination and check is the test of a traffic scene. Because a negative correlation exists between the algorithm performance of the unmanned vehicle and the traffic scene risk degree, the vehicle environment cognition and the algorithm understanding performance are poor due to the test scene with high difficulty, and the vehicle performance cannot be accurately evaluated, so the scene difficulty directly influences the test result of the vehicle, and the research of the unmanned vehicle is very important for the evaluation of the intelligent vehicle. However, the existing research on the identification of the risk level of the traffic scene is relatively lacked, and the commonly used research method lacks certain objectivity and quantitative data support.

The conventional method for researching the traffic scene danger degree mainly selects factors related to the traffic scene difficulty, uses the factors as evaluation indexes, and qualitatively and quantitatively describes the factors so as to realize the evaluation of the traffic scene danger degree. The selected index can be classified into a static index and a dynamic index according to its characteristics. The static indexes mainly comprise subjective opinions of related professionals, factors which have large influence on the road traffic environment and road traffic accident data. The dynamic index mainly uses a factor reflecting the state of motion of the vehicle or the driver as an evaluation index. The static index or dynamic index method is mainly started from a factor end, namely, environmental factors influencing the dangerous degree of a traffic scene are used as evaluation indexes. The determination of the degree of danger of the traffic environment should be based on the awareness and behavior of the driver and be dynamically variable. The existing research methods fail to define and describe the traffic scene danger degree quantitatively and objectively from the view point of drivers.

Disclosure of Invention

In order to solve the technical problems, the invention aims to provide a dangerous traffic scene identification method and system based on electroencephalogram data.

According to one aspect of the invention, a dangerous traffic scene identification method based on electroencephalogram data is provided, which comprises the following steps:

acquiring a stimulus object and an experimental variable based on a traffic scene;

completing driving test of a driver through a stimulating object and an experimental variable, and acquiring EEG data of the driver containing an electroencephalogram signal within a preset time;

preprocessing the acquired electroencephalogram signals, including removing signal noise in the electroencephalogram signals;

extracting characteristic indexes from the preprocessed electroencephalogram signals, and analyzing by adopting a mathematical statistics method to obtain characteristic indexes with significance;

calculating the difference between the predicted value and the true value of the power spectral density with the significant characteristic indexes;

and identifying the danger degree of the traffic scene according to the difference value.

Furthermore, preprocessing is carried out on the acquired electroencephalogram signals, and signal noise in the electroencephalogram signals is removed, wherein the signal noise comprises notch and band-pass filtering denoising, ICA (independent component analysis) method-based electro-oculogram denoising, experimental variable stimulation segment data extraction and baseline correction.

Further, extracting characteristic indexes of the preprocessed electroencephalogram signals comprises the following steps: the power spectral density of several rhythmic waves is extracted from the EEG data as a characteristic index.

Further, a characteristic index with significance is obtained by analyzing by a mathematical statistics method, and the method comprises the following steps:

extracting alpha waves, beta waves, theta waves and waves of the whole brain to obtain 4 characteristic indexes;

dividing a brain channel into four brain areas according to frontal lobe, parietal lobe, occipital lobe and temporal lobe, and extracting power spectral densities of brain waves, theta waves, alpha waves and beta waves from the four brain areas respectively to obtain 16 characteristic indexes;

analyzing the electroencephalogram signal of a driver in the driving process by adopting a power spectrum estimation method,

wherein the AR power spectrum estimation model is as follows:

the order p in the AR power spectrum estimation model is estimated by using a cost function, -jtw is a complex exponential signal, and the coefficient c of the AR power spectrum estimation model_piAnd σ²Solving by using a Burg algorithm to minimize the sum of powers of the prediction errors of the items before and after the AR model;

checking and analyzing the P (w) difference of the collected characteristic indexes by adopting a normality check;

carrying out one-way anova on the characteristic indexes which obey the requirements of normal distribution and homogeneity of variance, and carrying out nonparametric anova on the characteristic indexes which do not obey the normal distribution or do not meet the requirements of homogeneity of variance;

and determining the characteristic index with the P (w) smaller than the preset value as the characteristic index with significance.

Further, the method for calculating the power spectral density predicted value with the significant characteristic index comprises the following steps:

dividing the preset time into n time windows, wherein the EEG signal data with the power spectral density of the significant characteristic indexes corresponding to the time windows are as follows:

x⁽⁰⁾＝(x⁽⁰⁾(1),x⁽⁰⁾(2),…,x⁽⁰⁾(n)), where n is the number of electroencephalogram signal data.

The primary accumulation generation sequence of the original data is as follows:

x⁽¹⁾＝(x⁽¹⁾(1),x⁽¹⁾(2),…,x⁽¹⁾(n))

wherein:

based on x⁽¹⁾Establishing a first-order gray differential equation of a GM (1,1) model:

and calculating the parameters a and u to be identified of the equation by using a least square method. Wherein a is a development parameter and u is a gray effect amount.

Let a be (a, u)^TThen, the least squares method yields:

a＝(B^TB)^-1B^TY_n

wherein B and Y_nAnd are respectively:

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

from the above equations, the first order gray differential equation is solved as:

thus, the predicted values are obtained:

further, the difference between the predicted value and the true value of the power spectral density with significant characteristic indicators is calculated:

wherein,

represents a pair t_NPrediction of the power spectral density of a characteristic index with significance at a time, x (t)_N) Represents t_NThe true measurement of the power spectral density extraction with significant characteristic indicators at the time.

Further, calculating a difference between the predicted value and the true value of the power spectral density having the significant characteristic index, further includes: the model was tested for relative error (t) and posterior difference C, P:

the aim is to remove data whose power spectral density prediction values with significant characteristic indicators do not comply with the relative error check and the posterior error check.

Further, identifying the degree of danger of the traffic scene according to the difference value includes:

calculating the mean value mu and the standard deviation sigma between the predicted value and the true value of the power spectral density with the characteristic index of significance according to the difference value²；

Normal distribution characteristic (mu, sigma) of difference between predicted value and actual value of power spectral density of characteristic index with significance obtained by gray prediction method²) The indexes of the traffic factor risk degree are as follows: complexity_scene＝{μ_{Characteristic index with significance},σ² _{Characteristic index with significance}And reflecting the danger degree of the traffic scene through the index.

According to another aspect of the present invention, there is provided a dangerous traffic scene identification system based on electroencephalogram data, including:

the first data acquisition module is configured to acquire a stimulus object and an experimental variable based on a traffic scene;

the second data acquisition module is configured for acquiring EEG data containing electroencephalogram signals of a driver within a preset time length when the driving test of the driver is completed through a stimulus object and an experimental variable;

the preprocessing module is configured for preprocessing the acquired electroencephalogram signals and removing signal noise in the electroencephalogram signals;

the characteristic index extraction and analysis module is configured for extracting characteristic indexes of the preprocessed electroencephalogram signals and analyzing the characteristic indexes by adopting a mathematical statistics method to obtain characteristic indexes with significance;

a calculating module configured to calculate a difference between a predicted value and a true value of the power spectral density of the characteristic indicator having significance;

and the danger degree identification module is used for identifying the danger degree of the traffic scene according to the difference value according to the configuration.

In addition, the present invention also provides an apparatus comprising:

one or more processors;

a memory for storing one or more programs,

the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of the above.

According to another aspect of the invention, there is provided a computer readable storage medium storing a computer program which, when executed by a processor, implements a method as defined in any one of the above.

Compared with the prior art, the invention has the following beneficial effects:

1. according to the method for identifying the dangerous traffic scene based on the electroencephalogram data, disclosed by the invention, the scene danger degree is researched from the view angle of a driver, the traffic scene is established as a research object in different traffic environments, the electroencephalogram data in the driving process of the driver is used as an index for quantifying and evaluating the danger degree of the traffic scene, the internal association between the electroencephalogram of the driver and the traffic environment factors is explored, and the traffic environment danger degree identification model based on the electroencephalogram of the driver is established.

The method quantifies the danger degree of the traffic system from the visual angle of human drivers, analyzes the danger degree of a traffic scene more objectively, analyzes traffic factors (stimulus objects and experimental variables) in the traffic environment in a targeted manner, judges the danger degree of the traffic environment, can scientifically measure the danger degree of the traffic scene, accurately evaluates the mental load of the drivers, carries out effective safety early warning on the drivers, is beneficial to controlling the mental load of the drivers in a reasonable interval, and effectively reduces the driving risk.

2. According to the dangerous traffic scene identification system based on the electroencephalogram data, the electroencephalogram signal of the driver is an index for quantitatively evaluating the risk degree of the traffic scene according to the stimulus object and the experimental variable of the traffic scene, the internal association between the electroencephalogram of the driver and the stimulus object and the experimental variable is explored, a traffic scene risk degree identification model based on the electroencephalogram signal is established, the index capable of reflecting the mental load of the driver is selected as the index for evaluating the risk degree of the traffic scene, the mental load of the driver is favorably controlled in a reasonable interval, and the driving risk is effectively reduced.

3. According to the equipment disclosed by the invention, the processor executes the method for identifying the dangerous traffic scene based on the electroencephalogram data, so that effective safety early warning can be carried out on the driver, the mental load of the driver can be controlled in a reasonable interval, and the driving risk can be effectively reduced.

4. The readable storage medium stores the electroencephalogram data-based dangerous traffic scene identification method which is realized by the processor when being executed, so that effective safety early warning is carried out on a driver, the mental load of the driver is favorably controlled in a reasonable interval, and the driving risk is effectively reduced.

Drawings

FIG. 1 is a schematic view of a pedestrian crossing scene with different dangers in experimental example 1;

FIG. 2 is a flowchart showing pretreatment of electroencephalogram signals in Experimental example 1;

FIG. 3 is a graph showing the correlation between the electroencephalogram 32 IC components and the electro-ocular signals of a driver in Experimental example 1;

FIG. 4 is a diagram of EEG signals before and after a channel of a driver's electroencephalogram is subjected to eye-drop in Experimental example 1;

FIG. 5 is a comparison graph before and after baseline wandering of a driver EEG in Experimental example 1;

FIG. 6 is a electroencephalogram comparison graph before and after a pretreatment of a certain driver in Experimental example 1;

FIG. 7 is a full brain power spectral density profile of a driver in Experimental example 1;

FIG. 8 is a predicted curve of the power spectral density of alpha wave in the top lobe area of a driver in experimental example 1;

FIG. 9 is a diagram showing the calculation of the difference value based on the gray prediction method in Experimental example 1;

FIG. 10 is a diagram illustrating the probability density distribution of the difference between the predicted value and the actual value under the influence of different risk factors in Experimental example 1;

FIG. 11 is a difference distribution diagram of the power spectral density of the top lobe area alpha wave of the driver in the pedestrian crossing dangerous scene in example 2;

FIG. 12 is a difference distribution diagram of the power spectral density of the top lobe region alpha wave of the driver in the dangerous scene of vehicle speed change in example 2;

FIG. 13 is a difference distribution diagram of the power spectral density of the top lobe area alpha wave of the driver in the dangerous scene of the vehicle lane change in embodiment 2;

FIG. 14 is a probability density distribution graph of the difference between the predicted value and the actual value of the three risk factors in example 2;

FIG. 15 is a flow chart of the present invention;

FIG. 16 is a schematic structural diagram of an apparatus according to embodiment 1 of the present invention.

Detailed Description

In order to better understand the technical scheme of the invention, the invention is further explained by combining the drawings and the specific embodiments in the specification.

Example 1:

the embodiment provides a dangerous traffic scene identification system based on brain electrical data, includes:

the first data acquisition module is configured to acquire a stimulus object and an experimental variable based on a traffic scene;

the second data acquisition module is configured to acquire EEG data including electroencephalogram signals of a driver within a preset time length when a driving test of the driver is completed through a stimulus object and an experimental variable, and the second data acquisition module divides a brain channel into 32 independent channels.

The preprocessing module is configured for preprocessing the acquired electroencephalogram signals and removing signal noise in the electroencephalogram signals, and specifically, the preprocessing module comprises,

the device comprises a trap wave and band-pass filtering denoising unit, a band-pass filter and a band-pass filtering denoising unit, wherein the trap wave and band-pass filtering denoising unit is configured to remove high-frequency noise by adopting a band-pass FIR filter with the bandwidth of 1-50Hz and reserve a frequency band signal of 1-50 HZ;

an ICA method-based electro-oculogram denoising unit is configured for decomposing the electroencephalogram data into 32 independent components, then calculating correlation coefficients between the independent components and electro-oculogram, removing a plurality of independent components with larger correlation coefficients, and then carrying out ICA inverse operation reconstruction on the rest independent components to remove the electroencephalogram signals of the electro-oculogram noise;

the experimental variable stimulation section data extraction unit is configured for dividing the electroencephalogram signals according to experimental variable stimulation sections and non-experimental variable stimulation sections, extracting electroencephalogram data induced by similar stimulation from continuous electroencephalogram data, dividing the electroencephalogram data into a plurality of sections of experimental variable stimulation sections with equal length, and intercepting data sections by taking the time starting point of stimulation as zero time and a time window as length.

And the baseline correction unit is configured to subtract all electroencephalogram signal values of the experimental variable stimulation section from the basic amplitude value at the zero moment one by one to obtain a new steady potential value.

The characteristic index extraction and analysis module is configured for extracting characteristic indexes from the preprocessed electroencephalogram signals and analyzing the characteristic indexes by adopting a mathematical statistics method to obtain characteristic indexes with significance, and specifically, the characteristic index extraction and analysis module comprises,

the characteristic index extraction unit is configured to extract power spectral density of a plurality of rhythm waves from the EEG data as a characteristic index, and the extraction process comprises the following steps: extracting alpha wave, beta wave, theta wave and wave from the EEG signals respectively, wherein the power spectral densities of the four rhythm waves are used as characteristic indexes of the whole brain average power spectral density:

wherein f is_upRepresenting the upper frequency bound, f, corresponding to the rhythm wave_downRepresenting correspondence of rhythmic wavesThe lower frequency bound is used as an alternative, and alpha waves, beta waves, theta waves and waves of the whole brain are extracted to obtain 4 characteristic indexes;

dividing a brain channel into four brain areas according to frontal lobe, parietal lobe, occipital lobe and temporal lobe, and extracting power spectral densities of brain waves, theta waves, alpha waves and beta waves from the four brain areas respectively to obtain 16 characteristic indexes;

the characteristic index analysis unit of the significance is configured to analyze and obtain the characteristic index with the significance by adopting a mathematical statistics method, and the analysis process comprises the following steps:

analyzing the electroencephalogram signal of a driver in the driving process by adopting a power spectrum estimation method,

wherein the AR power spectrum estimation model is as follows:

the order p in the AR power spectrum estimation model is estimated by using a cost function, -jtw is a complex exponential signal, and the coefficient c of the AR power spectrum estimation model_piAnd σ²Solving by using a Burg algorithm to minimize the sum of powers of the prediction errors of the items before and after the AR model;

checking and analyzing the P (w) difference of the collected characteristic indexes by adopting a normality check;

carrying out one-way anova on the characteristic indexes which obey the requirements of normal distribution and homogeneity of variance, and carrying out nonparametric anova on the characteristic indexes which do not obey the normal distribution or do not meet the requirements of homogeneity of variance;

and determining the characteristic index with the P (w) smaller than the preset value as the characteristic index with significance.

A calculating module configured to calculate a difference between a predicted value and a true value of the power spectral density having the significant characteristic indicator, specifically, the calculating module includes:

the prediction value calculation unit is configured for calculating the prediction value of the power spectral density of the characteristic index with significance, and comprises the following steps:

dividing the preset time into n time windows, wherein the EEG signal data with the power spectral density of the significant characteristic indexes corresponding to the time windows are as follows:

x⁽⁰⁾＝(x⁽⁰⁾(1),x⁽⁰⁾(2),…,x⁽⁰⁾(n)), where n is the number of electroencephalogram signal data.

The primary accumulation generation sequence of the original data is as follows:

x⁽¹⁾＝(x⁽¹⁾(1),x⁽¹⁾(2),…,x⁽¹⁾(n))

wherein:

based on x⁽¹⁾Establishing a first-order gray differential equation of a GM (1,1) model:

and calculating the parameters a and u to be identified of the equation by using a least square method. Wherein a is a development parameter and u is a gray effect amount.

Let a be (a, u)^TThen, the least squares method yields:

a＝(B^TB)^-1B^TY_n

wherein B and Y_nAnd are respectively:

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

from the above equations, the first order gray differential equation is solved as:

thus, the predicted values are obtained:

a difference value calculating unit configured to calculate a difference value between the predicted value and the true value of the power spectral density having the significant characteristic indicator, the steps being as follows:

calculating the difference e (t) between the predicted value and the true value of the power spectral density with the characteristic index of significance_N)：

Wherein,

represents a pair t_NPrediction of the power spectral density of a characteristic index with significance at a time, x (t)_N) Represents t_NThe true measurement of the power spectral density extraction with significant characteristic indicators at the time.

The inspection module is configured to inspect the power spectral density predicted value of the characteristic index of significance, and comprises the following steps:

the model was tested for relative error (t) and posterior difference C, P:

the aim is to remove data that the power spectral density predicted value with significant characteristic indexes does not meet relative error test and posterior difference test, specifically, according to table 1, the experiment is based on a gray color prediction model, removes unqualified relative error (t) test and posterior difference C, P,

TABLE 1

	Value of epsilon (t)	P value	C value
				Good effect	0.05	0.95≤P	C≤0.35
Qualified	0.01	0.80≤P＜0.95	0.35＜P≤0.50
				Basic qualification of	0.10	0.70≤P＜0.80	0.50＜P≤0.65
Fail to be qualified	0.20	P＜0.70	0.65＜C

The danger degree identification module is used for identifying the danger degree of the traffic scene according to the difference value, and comprises the following steps:

calculating the mean value mu and the standard deviation sigma between the predicted value and the true value of the power spectral density with the characteristic index of significance according to the difference value²；

Normal distribution characteristic (mu, sigma) of difference between predicted value and actual value of power spectral density of characteristic index with significance obtained by gray prediction method²) The indexes of the traffic factor risk degree are as follows: complexity_scene＝{μ_{Characteristic index with significance},σ² _{Characteristic index with significance}}；

And identifying the danger degree of the traffic scene according to the index.

The embodiment provides a dangerous traffic scene identification method applying the system based on electroencephalogram data, which comprises the following steps:

step 1: acquiring a stimulus object and an experimental variable based on a traffic scene;

step 2: completing driving test of a driver through a stimulating object and an experimental variable, and acquiring EEG data of the driver containing an electroencephalogram signal within a preset time;

and step 3: preprocessing the acquired electroencephalogram signals, and removing signal noise in the electroencephalogram signals; preprocessing the acquired electroencephalogram signals, including removing signal noise in the electroencephalogram signals, including notch and band-pass filtering denoising, eye electrical denoising based on an ICA method, extracting experimental variable stimulation segment data, and correcting a base line.

And 4, step 4: extracting characteristic indexes from the preprocessed electroencephalogram signals, and analyzing by adopting a mathematical statistics method to obtain the characteristic indexes with significance, wherein the method specifically comprises the following steps:

step 4-1: extracting the power spectral density of a plurality of rhythm waves from EEG data as a characteristic index, including: extracting alpha wave, beta wave, theta wave and wave from the EEG signals respectively, wherein the power spectral densities of the four rhythm waves are used as characteristic indexes of the whole brain average power spectral density:

wherein f is_upRepresenting the upper frequency bound, f, corresponding to the rhythm wave_downRepresenting the lower frequency bound corresponding to the rhythm wave; as an alternative, extracting alpha waves, beta waves, theta waves and waves of the whole brain to obtain 4 characteristic indexes;

dividing a brain channel into four brain areas according to frontal lobe, parietal lobe, occipital lobe and temporal lobe, and extracting power spectral densities of brain waves, theta waves, alpha waves and beta waves from the four brain areas respectively to obtain 16 characteristic indexes;

step 4-2: the method for analyzing and obtaining the characteristic indexes with significance by adopting a mathematical statistics method comprises the following steps: adopting a power spectrum estimation method to analyze the brain electrical signals of a driver in the driving process,

wherein the AR power spectrum estimation model is as follows:

the order p in the AR power spectrum estimation model is estimated by using a cost function, -jtw is a complex exponential signal, and the coefficient c of the AR power spectrum estimation model_piAnd σ²Solving by using a Burg algorithm to minimize the sum of powers of the prediction errors of the items before and after the AR model;

checking and analyzing the P (w) difference of the collected characteristic indexes by adopting a normality check;

carrying out one-way anova on the characteristic indexes which obey the requirements of normal distribution and homogeneity of variance, and carrying out nonparametric anova on the characteristic indexes which do not obey the normal distribution or do not meet the requirements of homogeneity of variance;

and determining the characteristic index with the P (w) smaller than the preset value as the characteristic index with significance.

And 5: calculating a difference between a predicted value and a true value of the power spectral density of the characteristic indicator with significance, including;

step 5-1: calculating a predicted value of the power spectral density with significant characteristic indicators:

dividing the preset time into n time windows, wherein the EEG signal data with the power spectral density of the significant characteristic indexes corresponding to the time windows are as follows:

x⁽⁰⁾＝(x⁽⁰⁾(1),x⁽⁰⁾(2),…,x⁽⁰⁾(n)), where n is the number of electroencephalogram signal data.

The primary accumulation generation sequence of the original data is as follows:

x⁽¹⁾＝(x⁽¹⁾(1),x⁽¹⁾(2),…,x⁽¹⁾(n))

wherein:

based on x⁽¹⁾Establishing a first-order gray differential equation of a GM (1,1) model:

and calculating the parameters a and u to be identified of the equation by using a least square method. Wherein a is a development parameter and u is a gray effect amount.

Let a be (a, u)^TThen, the least squares method yields:

a＝(B^TB)^-1B^TY_n

wherein B and Y_nAnd are respectively:

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

from the above equations, the first order gray differential equation is solved as:

thus, the predicted values are obtained:

and identifying the danger degree of the traffic scene according to the difference value.

Step 5-2: the model was tested for relative error (t) and posterior difference C, P:

the aim is to remove data of the power spectral density predicted value with significant characteristic indexes which do not accord with the relative error test and the posterior difference test, specifically, according to the table 1, the experiment is based on a gray color prediction model, and unqualified relative error (t) test and the posterior difference C, P are removed.

Step 5-3: calculating the difference between the predicted value and the true value of the power spectral density with the significant characteristic indicators:

wherein,

represents a pair t_NPrediction of the power spectral density of a characteristic index with significance at a time, x (t)_N) Represents t_NThe true measurement of the power spectral density extraction with significant characteristic indicators at the time.

Step 6: identifying the danger degree of the traffic scene according to the difference value, comprising the following steps:

step 6-1: calculating the mean value mu and the standard deviation sigma between the predicted value and the true value of the power spectral density of the characteristic index with significance according to the difference value²；

Step 6-2: passing through ashNormal distribution characteristic (mu, sigma) of difference value between predicted value and actual value of power spectral density of characteristic index with significance obtained by color prediction method²) The indexes of the traffic factor danger degree are as follows: complexity_scene＝{μ_{Characteristic index with significance},σ² _{Characteristic index with significance}And reflecting the danger degree of the traffic scene through the index.

And carrying out correlation analysis on the screened significant characteristic indexes, knowing the correlation degree among the characteristic indexes, then reducing the dimension of the indexes with higher correlation degree, constructing an optimal index set of traffic risk influence factors, and constructing a traffic scene risk identification model. Therefore, before and after the driver is influenced by the traffic factor, the degree of smoothness of the change of the electroencephalogram signals can reflect the risk degree of the current traffic factor.

An apparatus of this embodiment, the apparatus comprising: one or more processors;

a memory for storing one or more programs,

when the one or more programs are executed by the one or more processors, the one or more processors execute the method, the processor executes the method for identifying the dangerous traffic scene based on the electroencephalogram data, effective safety early warning can be performed on a driver, the mental load of the driver can be controlled in a reasonable interval, and the driving risk is effectively reduced.

In the embodiment, the computer readable storage medium stores a computer program, and when the computer program is executed by a processor, the computer readable storage medium implements any one of the methods described above, and stores the method for identifying dangerous traffic scenes based on electroencephalogram data, which is implemented when the computer program is executed by the processor, so that driving risks are effectively reduced. Further introduction is as follows:

the computer system includes a Central Processing Unit (CPU)101, which can perform various appropriate actions and processes in accordance with a program stored in a read-only memory (ROM)102 or a program loaded from a storage section into a Random Access Memory (RAM) 103. In the RAM103, various programs and data necessary for system operation are also stored. The CPU 101, ROM 102, and RAM103 are connected to each other via a bus 104. An input/output (I/O) interface 105 is also connected to bus 104.

The following components are connected to the I/O interface 105: an input portion 106 including a keyboard, a mouse, and the like; an output section including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 108 including a hard disk and the like; and a communication section 109 including a network interface card such as a LAN card, a modem, or the like. The communication section 109 performs communication processing via a network such as the internet. The drives are also connected to the I/O interface 105 as needed. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 108 as necessary.

In particular, the process described above with reference to the flowchart of fig. 15 may be implemented as a computer software program according to an embodiment of the present invention. For example, embodiment 1 of the invention comprises a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section, and/or installed from a removable medium. The above-described functions defined in the system of the present application are executed when the computer program is executed by the Central Processing Unit (CPU) 101.

It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

Block diagram 16 in the drawings illustrates the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments 1 of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logic function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules or units described in the embodiments of the present invention may be implemented by software, or may be implemented by hardware, and the described units may also be disposed in a processor. Where the names of these elements do not in some way constitute a limitation on the elements themselves. The described units or modules may also be provided in a processor, and may be described as: a dangerous traffic scene identification system based on electroencephalogram data comprises: the device comprises a first data acquisition module, a second data acquisition module, a preprocessing module, a characteristic index extraction and analysis module, a calculation module and a danger degree identification module, wherein the names of the modules do not limit the unit per se under certain conditions, for example, the preprocessing module can also be described as a preprocessing module for preprocessing the acquired electroencephalogram signals and removing signal noise in the electroencephalogram signals.

As another aspect, the present application also provides a computer-readable medium that may be contained in the electronic device described in the above embodiments; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs, and when the one or more programs are executed by an electronic device, the electronic device is enabled to implement the method for recognizing the dangerous traffic scene based on the electroencephalogram data as described in the above embodiment.

For example, the electronic device may implement the following as shown in fig. 15: step S1: acquiring a stimulation object and an experimental variable based on a traffic scene; step S2: the driver EEG data acquisition device is used for acquiring EEG data of a driver including a brain electricity signal in a preset time length when the driver is tested by stimulating objects and experimental variables; step S3: preprocessing the acquired electroencephalogram signals, and removing signal noise in the electroencephalogram signals; step S4: extracting characteristic indexes from the preprocessed electroencephalogram signals, and analyzing by adopting a mathematical statistics method to obtain characteristic indexes with significance; step S5: calculating a difference value between a predicted value and a true value of the power spectral density with the characteristic indexes of significance; step S6: and identifying the danger degree of the traffic scene according to the difference value.

It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functionality of two or more modules or units described above may be embodied in one module or unit, according to embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.

Moreover, although the steps of the methods of the present disclosure are depicted in the drawings in a particular order, this does not require or imply that the steps must be performed in this particular order, or that all of the depicted steps must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions, etc.

Experimental example 1:

this experimental example uses pedestrian's traffic factor of crossing as amazing object to different distances between driver's vehicle and pedestrian's amazing trigger point of crossing are experimental variables, have designed the different experimental scenes of two kinds of danger degrees, and the driver brain electricity characteristic under the traffic factor influence of research different danger degrees has established the identification model, and concrete step is as follows:

the first step is as follows: in the experiment, the distances of pedestrian crossing street stimulation trigger points in two scenes are respectively 120m and 60m, and the specific scene setting is shown in fig. 1. In the experiment, 30 drivers participate, and each driver is required to complete the driving tests of 120m and 60m scenes respectively. In the experiment, the appearance sequence of two scenes is set to be random for balancing the sequence factors of the scenes.

When the distance between the driver vehicle and the street pedestrian stimulus is 120m, the driver can discover the street pedestrian earlier and make a decision on driving behavior with a longer reaction time, so that the stimulus occurring at the 120m position is considered as a traffic factor with a smaller danger degree. When the distance between the driver's vehicle and the street pedestrian stimulus is set to 60m, the driver must make an emergency action immediately after discovering the stimulus in order to avoid a traffic accident. The pedestrian crossing the street at 60m has greater stimulation to the driver, so the stimulation at 60m is considered as a traffic factor with greater danger degree.

The second step is that: the method comprises the steps of obtaining EEG data of a driver containing EEG signals within a preset time length, and collecting the EEG data of the driver from a 32-lead EEG channel by a non-invasive collection method by adopting Neurone 32-lead EEG collection equipment as an alternative. A10-20 international standard lead system of the International society for electroencephalography (EEG) is taken as a placement standard of an EEG, a reference electrode is positioned at a Cz point, and the sampling frequency is 500 Hz.

The experimental example respectively extracts 24s electroencephalogram data before the pedestrian cross the street and 3s electroencephalogram data when the pedestrian cross the street, and takes 3s as a time window to preprocess the acquired electroencephalogram signals and remove signal noise. The preprocessing process of the electroencephalogram signal can be divided into four main steps, and the specific flow is shown in fig. 2.

Step1, notch and band-pass filtering denoising:

the trap filter can remove specific frequency components, firstly, a trap wave is used to remove 50HZ power frequency noise, and influence on other useful frequency components is avoided as much as possible:

wherein, ω is₀The cutoff frequency of the wave limiter is the power frequency signal of 50Hz in the study. Then a band-pass FIR filter with the bandwidth of 1-50Hz is adopted to remove the high-frequency noise and reserve the frequency band signal of 1-50 Hz.

Step2. removing electrooculogram noise based on ICA method:

in this study, ocular artifacts were removed using Independent Component Analysis (ICA). The main principle is that the electroencephalogram data are decomposed into 32 independent components, then the correlation coefficient between each independent component and the ocular is calculated, the independent component with the larger correlation coefficient is removed, and the rest independent components are subjected to ICA inverse operation to be reconstructed into pure electroencephalogram signals. As shown in fig. 3, the analysis process of the eye-drop analysis for a driver ICA method. The correlation between the component IC18 and IC27 and the eye electrical signals is large, and the components are eye electrical artifact components causing original signal interference, so that the IC18 component and the IC27 component are removed, and the electroencephalogram signals after the eye motion artifacts are separated can be obtained by reconstructing the electroencephalogram signals. As shown in fig. 4, for removing the EEG signals before and after the eye electrical noise, the waveform glitch reduction can be observed from the time domain, and the eye electrical artifact is basically removed, so as to obtain pure electroencephalogram data.

Step3. stimulation segment data extraction:

after pure electroencephalogram data are obtained, the electroencephalogram signals are divided according to an event stimulation section and a non-event stimulation section, electroencephalogram data induced by the similar stimulation are extracted from continuous electroencephalogram data, and the electroencephalogram data are divided into a plurality of stimulation event data sections with equal length. This example sets a 3s time window length and sets three homogeneous stimulation events in one scene.

Step4. baseline correction:

the baseline correction causes the brain electrical data to fluctuate around the horizontal axis from being skewed to one side of the horizontal axis. As shown in FIG. 5, a contrast graph of electroencephalogram data before and after baseline wander for a certain channel is shown.

Finally, the electroencephalogram waveform obtained through the four preprocessing steps is normal and has an analyzed state. As shown in fig. 6.

The third step: the power spectrum estimation is carried out by adopting an AR model method, and the method can obtain a higher power spectrum and can be conveniently converted into a characteristic vector only by short-range data:

the AR power spectrum estimation model is as follows:

the order p in the AR power spectrum estimation model typically uses a cost functionThe number is estimated, the experimental example p is 20, the coefficient c_piAnd σ²The Burg algorithm is used herein for the evaluation. As shown in fig. 7, is a full brain power spectral density profile of a driver during driving. The principle is that the sum of the power of the prediction error of the front term and the rear term of the AR model is minimum, and the known electroencephalogram signal data is directly obtained

In order to obtain the reflection coefficient K_pThen, the reflection coefficient is used to obtain the AR parameter, K, by using the Levinson recursion algorithm_pAnd the AR parameters are used for calibrating the AR power spectrum estimation model.

In order to extract useful information from brain waves, the experimental example extracts power spectral densities (α wave, β wave, θ wave, wave) of four rhythm waves from all brain electrical signals as characteristic indexes:

wherein f is_upRepresenting the upper frequency bound, f, corresponding to the rhythm wave_downRepresenting the lower frequency bound for the rhythmic wave.

The fourth step: 4 indicators on the whole brain mean power spectral density were analyzed.

Under the 3s time window, the whole-brain average alpha wave power spectral density and the whole-brain average beta wave power spectral density are both subjected to normal distribution, but only the whole-brain average beta wave power spectral density is subjected to homogeneity (p is 0.258>0.05) through testing, so that variance analysis (F test) is carried out on electroencephalogram signal data of the whole-brain average beta wave power spectral density, and nonparametric test (H test) is carried out on electroencephalogram signal data of other three indexes. Through analysis of variance, under a 3s time window, the whole brain average power spectral density has no significant difference in four frequency bands, so that the whole brain average power spectral density index of each frequency band cannot be used as a representation index of the mental load of a driver.

Analyzing power spectral density indexes of waves, theta waves, alpha waves and beta waves extracted from four brain areas of frontal lobe, parietal lobe, occipital lobe and temporal lobe. Under the time window of 3s, 4 indexes of the mean power spectral density of the apical lobe area, the mean beta power spectral density of the apical lobe area, the mean power spectral density of the occipital lobe area and the mean alpha power spectral density of the temporal lobe area meet normal distribution and variance homogeneity, so that single-factor analysis is carried out on the indexes, and non-parameter test is carried out on the other indexes which do not meet the normal distribution or the variance homogeneity. The result shows that the power spectral density of the average alpha wave in the apical lobe area has significant difference (p is 0.011<0.05), and the power spectral density can be used as a unique characterization index of mental load of a driver, and the rest indexes have no significance.

In the experiment, the power spectral density of the alpha wave in the top lobe area of the driver is used as a characteristic index with significance to analyze and quantify the danger degree of the traffic factor.

The fifth step: electroencephalogram signal data are obtained based on experiments, the power spectral densities of top leaf area alpha waves of a driver in two experimental scenes are respectively analyzed, and the obtained analysis results are shown in fig. 7.

Under the influence of different traffic hazard factors, the power spectral density of alpha waves in the top lobe area of a driver tends to become smaller, but under the influence of different hazard factor stimuli (120m and 60m), the power spectral density of the alpha waves in the top lobe area becomes smaller by different amplitudes. When a driver drives in a pedestrian street crossing scene at a distance of 120m, the danger degree of influence factors in the traffic scene is low, the response of the driver to stimulation is small, the change of the brain wave waveform before and after the stimulation is relatively smooth, and the change of the power spectral density of alpha waves in the apical lobe area is relatively small. When a driver drives in a pedestrian street crossing scene at a distance of 60m, the danger degree of traffic factors is high, and the power spectral density changes of the brain wave waveform and the alpha wave of the apical lobe area of the driver are also greatly changed. Therefore, before and after the driver is influenced by the traffic complexity factor, the degree of change and smoothness of the power spectral density of the alpha wave in the top leaf area can reflect the danger degree of the current factor.

And a sixth step: the experiment predicts the stimulation time based on the power spectral density of the alpha wave in the apical lobe area at the stimulation preorder time. And dividing the preamble time of 24s into 8 prediction samples by taking 3s as a time window to predict the power spectral density value of the top lobe area alpha wave at the stimulation moment. The experiment carries out grey prediction of the stimulation time based on a grey prediction method to obtain a local fitting function of the power spectral density of alpha waves in the top leaf area, and the method comprises the following specific steps:

the electroencephalogram signal data of the power spectral density of alpha waves in the top lobe area of the driver are as follows:

x⁽⁰⁾＝(x⁽⁰⁾(1),x⁽⁰⁾(2),…,x⁽⁰⁾(n))

wherein n is the number of electroencephalogram signal data.

The primary accumulation generation sequence of the original data is as follows:

x⁽¹⁾＝(x⁽¹⁾(1),x⁽¹⁾(2),…,x⁽¹⁾(n))

wherein:

based on x⁽¹⁾Establishing a first-order gray differential equation of a GM (1,1) model:

and calculating the parameters a and u to be identified of the equation by using a least square method. Wherein a is a development parameter and u is a gray effect amount.

Let a be (a, u)^TThen, the least squares method yields:

a＝(B^TB)^-1B^TY_n

wherein B and Y_nAnd are respectively:

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

from the above, the first order gray differential equation is solved as:

obtaining a predicted value:

the seventh step: the model was tested for relative error (t) and posterior difference C, P:

gray prediction and inspection are carried out on all the electroencephalogram data of the driver, the predicted values of the alpha wave power spectral density of the top leaf area meet error inspection, and the curve is shown in figure 8 and is an electroencephalogram signal gray prediction curve of the driver.

Finally, the following is obtained:

wherein, the difference e (t) between the predicted value and the true value of the power spectral density of the alpha wave in the top leaf area of the driver_N)。

Indicates that t is predicted by gray prediction_NPredicted value of power spectral density of alpha wave in top leaf area at time, x (t)_N) Represents t_NThe power spectral density value of the alpha wave in the top leaf region measured actually at the time is shown in fig. 9, where J ═ t_N-3,t_N-2,t_N-1Respectively representing the sampling time corresponding to the three preamble sampling points, and utilizing the obtained local fitting function,and predicting the power spectrum density of the top leaf area alpha wave of the electroencephalogram signal at the corresponding moment of the stimulation point.

Eighth step: mean μ and standard deviation σ in a normal distribution formed by the difference between the predicted value and the actual value²The intensity of the change of the electroencephalogram signals is reflected, and the complex degree of traffic factors is indirectly reflected. The results of a comparative analysis of normal distribution curves (after fitting) of the power spectral density difference of alpha waves in the top lobe area of the driver under the influence of the pedestrian street-crossing stimulation with smaller complexity (120m) and larger complexity (60m) are shown in fig. 10. Traffic complexity factors with larger complexity, and the absolute value of the mean value mu and the standard deviation sigma in the normal distribution characteristics corresponding to the difference value²Are all larger.

Experimental example 2:

the experimental example establishes an identification model aiming at the risk degrees of three traffic risk factors, namely pedestrian crossing, vehicle speed change and vehicle lane change, and contrasts and analyzes the risk degrees of the three traffic risk factors, and comprises the following specific steps:

the electroencephalogram data preprocessing process and the characteristic index determining process which are the same as those in the experimental example 1 are not repeated, and the experimental example 1 is different from the experimental example 1 in that scene risk degrees containing different elements are contrastively analyzed:

in the experiment, based on the obtained significant electroencephalogram index as the power spectral density of the top leaf area alpha wave, the difference e (t) between the predicted value and the true value of the power spectral density of the top leaf area alpha wave of the driver is calculated_N) And carrying out normal test on the distribution condition of the difference value. The power spectral density difference values of the top lobe area alpha wave of three traffic risk factors of pedestrian crossing, vehicle speed change and vehicle lane change are obtained, and all have good normal distribution characteristics, as shown in fig. 11 to 13.

And (3) recording the normal analysis results of the power spectral density difference values of the top leaf area alpha waves of the three traffic risk factors as the risk comparison indexes of the top leaf area alpha waves:

Complexity_{street crossing pedestrian}＝{μ_{Top zone alpha wave},σ² _{Top zone alpha wave}}＝{1.084,1.980}

Complexity_{Variable speed vehicle}＝{μ_{Top zone alpha wave},σ² _{Top zone alpha wave}}＝{0.736,1.431}

Complexity_{Lane-changing vehicle}＝{μ_{Top zone alpha wave},σ² _{Top zone alpha wave}}＝{0.833,1.693}

The normal distribution curve of the difference (after fitting) is shown in fig. 14. The traffic factor with higher risk degree is the absolute value of the mean value mu and the standard deviation sigma in the corresponding normal distribution characteristics²Are all larger. Therefore, according to the risk index, the risk of the three traffic factors is as follows from small to large: pedestrian crossing street>Vehicle lane change>The vehicle is shifted.

Pedestrian crossing suddenly appears on a road without a mark line, so that uncertainty and abruptness are high for a driver, the response time given to the driver is short, the caused mental load is maximum, and driving regulations tell the driver that the severity of an accident caused by pedestrian collision is larger compared with the severity of vehicle rear-end collision and vehicle scratching accidents, and the pedestrian stimulation causes the mental load of the driver to be larger. The front vehicle speed change behavior has great abruptness, but before the vehicle speed change, the change of the lamp can play a certain warning role for other vehicles, so that the reaction decision time of a driver is increased. Compared with the factors of the pedestrian crossing the street without early warning, the influence is small and the complexity is low. The lane changing behavior of the vehicle is a driving process behavior with a time length, and the sudden behavior is much smaller than that of pedestrians and vehicle speed change. The stimulus is given to the driver for a longer time to react when it occurs, and the influence on the mental load of the driver is the smallest among the three influencing factors.

The above description is only a preferred embodiment of the application and is illustrative of the principles of the technology employed. It will be appreciated by a person skilled in the art that the scope of the invention as referred to in the present application is not limited to the specific combination of features described above, but also covers other embodiments where any combination of the features described above or their equivalents is used without departing from the inventive concept described above. For example, the above features may be replaced with (but not limited to) features having similar functions disclosed in the present application.

Claims (10)

1. A dangerous traffic scene identification method based on electroencephalogram data is characterized by comprising the following steps:

acquiring a stimulus object and an experimental variable based on a traffic scene;

completing driving test of a driver through a stimulating object and an experimental variable, and acquiring EEG data of the driver containing an electroencephalogram signal within a preset time;

preprocessing the acquired electroencephalogram signals, and removing signal noise in the electroencephalogram signals;

extracting characteristic indexes from the preprocessed electroencephalogram signals, and analyzing by adopting a mathematical statistics method to obtain characteristic indexes with significance;

calculating the difference between the predicted value and the true value of the power spectral density with the significant characteristic indexes;

and identifying the danger degree of the traffic scene according to the difference value.

2. The method for identifying the dangerous traffic scene based on the electroencephalogram data as claimed in claim 1, wherein the acquired electroencephalogram signals are preprocessed, and the preprocessing comprises signal noise removal in the electroencephalogram signals, including notch and band-pass filtering denoising, electrooculogram denoising based on an I CA method, experimental variable stimulation segment data extraction and baseline correction.

3. The method for identifying the dangerous traffic scene based on the electroencephalogram data as claimed in claim 1, wherein the step of extracting characteristic indexes from the preprocessed electroencephalogram signals comprises the following steps: the power spectral density of several rhythmic waves is extracted from the EEG data as a characteristic index.

4. The method for identifying dangerous traffic scenes based on electroencephalogram data according to claim 3, wherein the step of extracting the power spectral density of a plurality of rhythm waves from EEG data as a characteristic index comprises the following steps: extracting alpha wave, beta wave, theta wave and wave from the EEG signals respectively, wherein the power spectral densities of the four rhythm waves are used as characteristic indexes of the whole brain average power spectral density:

wherein f is_upRepresenting the upper frequency bound, f, corresponding to the rhythm wave_downRepresenting the lower bound of the frequency to which the rhythmic wave corresponds.

5. The method for identifying dangerous traffic scenes based on electroencephalogram data according to claim 4, wherein characteristic indexes with significance are obtained by adopting a mathematical statistics method, and the method comprises the following steps:

extracting alpha waves, beta waves, theta waves and waves of the whole brain to obtain 4 characteristic indexes;

dividing a brain channel into four brain areas according to frontal lobe, parietal lobe, occipital lobe and temporal lobe, and extracting power spectral densities of brain waves, theta waves, alpha waves and beta waves from the four brain areas respectively to obtain 16 characteristic indexes;

analyzing the electroencephalogram signal of a driver in the driving process by adopting a power spectrum estimation method,

wherein the AR power spectrum estimation model is as follows:

the order p in the AR power spectrum estimation model is estimated by using a cost function, -jtw is a complex exponential signal, and the coefficient c of the AR power spectrum estimation model_piAnd σ²Solving by using a Burg algorithm to minimize the sum of powers of the prediction errors of the items before and after the AR model;

checking and analyzing the P (w) difference of the collected characteristic indexes by adopting a normality check;

carrying out one-way anova on the characteristic indexes which obey the requirements of normal distribution and homogeneity of variance, and carrying out nonparametric anova on the characteristic indexes which do not obey the normal distribution or do not meet the requirements of homogeneity of variance;

and determining the characteristic index with the P (w) smaller than the preset value as the characteristic index with significance.

6. The method for identifying the dangerous traffic scene based on the electroencephalogram data as claimed in claim 1, wherein the step of calculating the predicted value of the power spectral density of the characteristic index with significance comprises the following steps:

dividing the preset time into n time windows, wherein the EEG signal data with the power spectral density of the significant characteristic indexes corresponding to the time windows are as follows:

x⁽⁰⁾＝(x⁽⁰⁾(1),x⁽⁰⁾(2),…,x⁽⁰⁾(n)), where n is the number of electroencephalogram signal data.

The primary accumulation generation sequence of the original data is as follows:

x⁽¹⁾＝(x⁽¹⁾(1),x⁽¹⁾(2),…,x⁽¹⁾(n))

wherein:

based on x⁽¹⁾Establishing a first-order gray differential equation of a GM (1,1) model:

and calculating the parameters a and u to be identified of the equation by using a least square method. Wherein a is a development parameter and u is a gray effect amount.

Let a be (a, u)^TThen, the least squares method yields:

a＝(B^TB)^-1B^TY_n

wherein B and Y_nAnd are respectively:

Y_n＝(x⁽⁰⁾(2),x⁽⁰⁾(3),…,x⁽⁰⁾(n))^T

from the above equations, the first order gray differential equation is solved as:

thus, the predicted values are obtained:

7. the method for identifying dangerous traffic scenes based on electroencephalogram data according to claim 6, wherein the difference between the predicted value and the true value of the power spectral density with the characteristic indicator of significance is calculated as follows:

wherein,

represents a pair t_NPrediction of the power spectral density of a characteristic index with significance at a time, x (t)_N) Represents t_NThe true measurement of the power spectral density extraction with significant characteristic indicators at the time.

8. The method for identifying a dangerous traffic scene based on electroencephalogram data according to claim 7, wherein calculating the difference between the predicted value and the true value of the power spectral density having the characteristic indicator of significance further comprises: the model was tested for relative error (t) and posterior difference C, P:

9. the method for recognizing the dangerous traffic scene based on the electroencephalogram data as claimed in claim 1, wherein the step of recognizing the degree of danger of the traffic scene according to the difference comprises the following steps:

calculating the mean value mu and the standard deviation sigma between the predicted value and the true value of the power spectral density with the characteristic index of significance according to the difference value²；

Normal distribution characteristic (mu, sigma) of difference value between predicted value and actual value of power spectral density of characteristic index with significance obtained by gray prediction method²) The indexes of the traffic factor risk degree are as follows: complexity_scene＝{μ_{Characteristic index with significance},σ² _{Characteristic index with significance}And reflecting the danger degree of the traffic scene through the index.

10. The utility model provides a dangerous traffic scene identification system based on brain electrical data which characterized in that includes:

the first data acquisition module is configured to acquire a stimulus object and an experimental variable based on a traffic scene;

the second data acquisition module is configured for acquiring EEG data of the driver containing electroencephalogram signals within a preset time length when the driving test of the driver is completed through the stimulating object and the experimental variable;

the preprocessing module is configured for preprocessing the acquired electroencephalogram signals and removing signal noise in the electroencephalogram signals;

the characteristic index extraction and analysis module is configured for extracting characteristic indexes from the preprocessed electroencephalogram signals and analyzing the characteristic indexes by adopting a mathematical statistics method to obtain characteristic indexes with significance;

a calculating module configured to calculate a difference between a predicted value and a true value of the power spectral density of the characteristic indicator having significance;

and the danger degree identification module is used for identifying the danger degree of the traffic scene according to the difference value according to the configuration.

Images