A method for measuring driver's risk prediction ability based on beta band electroencephalogram signal

Abstract

The application discloses a kind of based on the measurement method of driver dangerous pre-judgment ability of beta band electroencephalogram signal, characterized in that setting dangerous scene, dangerous pre-scene, three kinds of dangerous signal perception test tasks of safe scene, the electroencephalogram signal of measured personnel in three kinds of tasks is collected;The power value corresponding to the baseline normalization of beta band in the time window 450~550ms in Fz and Pz electrode channel under different tasks of measured personnel is calculated;The index value under different tasks is calculated respectively, and the statistical test p value and mean value of difference between each other.If and and and and, it indicates that measurement is effective, and the dangerous pre-judgment ability level of the measured personnel is, and the greater the value indicates that the pre-judgment ability is stronger;The method has the advantages that the pre-judgment perception ability of measured personnel before danger appears is objectively and directly evaluated.

Description

Technical Field

[0001] This invention relates to the field of electroencephalography (EEG) neurofunctional measurement technology, and in particular to a method for measuring a driver's ability to predict danger based on beta-band EEG signals. Background Technology

[0002] Danger prediction ability refers to the capacity to anticipate and prepare for potential hazards before clear warning signs appear. In driving operations, a driver's danger prediction ability is crucial for safety. An excellent driver not only handles various scenarios proficiently but, more importantly, can anticipate various hazardous situations. Therefore, objectively and accurately measuring an individual's danger prediction ability is vital for evaluating a driver's hazard handling capabilities, improving the scientific and personalized nature of driver safety training, identifying low-risk drivers, and rationally assigning driving tasks. However, currently, there is no objective and accurate assessment method to measure a driver's danger prediction ability. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method for measuring a driver’s risk prediction ability based on beta-band EEG signals. This method can accurately and efficiently assess the degree of an individual’s risk prediction ability by analyzing the conditional differences in EEG signals.

[0004] The technical solution adopted by this invention to solve the above problems is as follows: a method for measuring a driver's hazard prediction ability based on beta-band EEG signals, comprising the following specific steps: (1) The subjects were fitted with multi-channel EEG caps and EEG acquisition devices were used to record EEG signals; (2) In the driving road environment of the driving simulator, select one type of danger: set up three types of tasks: danger perception test in dangerous scenario, danger perception test in pre-danger scenario, and danger perception test in safe scenario; in each of the three types of tasks, collect the EEG signals of the test subjects for 10 minutes to obtain the raw EEG data of the test subjects. (3) Preprocess the raw EEG data, including signal amplification, segmentation, signal noise reduction, bandpass filtering and artifact removal. (4) Based on the preprocessed EEG data, the baseline normalized frequency band power (unit: dB) of the β band (14~20Hz part) of the Fz and Pz electrodes within a time window of 450~550ms after the end of the stimulation video was calculated for the subjects under different tasks; the baseline normalized frequency band power value was averaged between the Fz and Pz electrodes to obtain the baseline normalized power value P_dB of the β band when danger was perceived in dangerous scenarios. 危险The baseline normalized power value P_dB of the β band when danger is perceived in a pre-danger scenario. 危险前 The baseline normalized power value P_dB corresponding to the β band in a safe scenario. 安全 ; (5) Calculate P_dB 危险 With P_dB 安全 p-value of the statistical test of the difference between them P_dB was calculated. 危险 With P_dB 危险前 p-value of the statistical test of the difference between them P_dB was calculated. 危险前 With P_dB 安全 p-value of the statistical test of the difference between them The mean was calculated. mean mean Then, calculate the test subject's ability to predict the risk of this type of hazard: if and and and and This indicates that the measurement was valid, the person being tested has the ability to predict this type of hazard, and the level of hazard prediction ability is [value missing]. The larger the value, the stronger the predictive ability; if the conditions and and and and If the result is not valid, then the measurement is invalid.

[0005] If multiple hazard types need to be measured, repeat steps (1) to (5) above for different hazard types, and use the total number of hazard types with predictive ability. N and the mean level of risk prediction ability The higher the value, the stronger the overall predictive ability.

[0006] Furthermore, in step (2), the test task is a danger assessment task. The test extracts the EEG data of the test subject within 2000ms after the end of the stimulus video. The number of test trials for each task is ensured to be 20 or more.

[0007] Furthermore, in step (3), during the preprocessing of the raw EEG data, the raw EEG data is subjected to bandpass filtering of 0.5 to 45 Hz, and independent component analysis and data reconstruction are performed using the FastICA algorithm based on the principle of maximum negative entropy, so as to effectively remove artifact interference; and the preprocessed EEG data retains the β band as the target data for subsequent analysis.

[0008] Furthermore, in step (4), the method for calculating the baseline-normalized frequency band power value is as follows: (4-1) For each type of task, the test time is 10 minutes. Wavelet transform is performed on the preprocessed EEG data to obtain the frequency power value of the β band (14~20Hz) in the Fz and Pz electrode channels in the interval of 450~550ms after the end of the stimulation video. (4-2) For three scenarios—dangerous scene perception, danger perception before danger, and safe scene—frequency band power is calculated separately. Using -500ms to -100ms before the end of the stimulus video as the baseline, the baseline-normalized frequency band power values ​​are obtained. The calculation formula is as follows: , , , , , , , , , , , , , , , Where: the symbol P represents the β-band power value under a single stimulus trial, and k is the frequency value, which ranges from 14 to 20 Hz; This represents the β-band power value (in dB) obtained after baseline normalization for a single stimulus trial. The symbol is... This represents the β-band power value during the baseline period from -500ms to -100ms before stimulation.

[0009] Furthermore, in step (5), the mean mean mean , N , The calculation formula is as follows: , , , , , Where: symbol This represents the average β-band power value obtained after baseline normalization of multiple stimulus trials. i, j, and m respectively represent the sequence numbers of the trials in the dangerous scenario, the pre-danger scenario, and the safe scenario. h represents the sequence number of the danger type with predictive capability. This represents the level of hazard prediction capability for the h-th hazard type.

[0010] Compared with the prior art, the advantages of the present invention are: (1) This method objectively and directly assesses the risk prediction ability of the test subjects by analyzing the task-specific changes of the β-band power index in the EEG signal data. It has important reference value for evaluating the driver's risk handling ability, improving the scientific and personalized level of driver safety training, screening low-risk drivers, and rationally assigning driving tasks. (2) This method uses baseline-normalized β-band power to extract neural activity features in the critical time window of 450-550ms in the Fz and Pz electrode channels. It is simple and efficient for recognizing the brain's ability to predict and perceive danger when danger is about to occur. Attached Figure Description

[0011] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0012] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0013] As shown in the figure, a method for measuring a driver's hazard prediction ability based on beta-band EEG signals includes the following specific steps: (1) The subject wears a multi-channel EEG cap and uses an EEG acquisition device to record EEG signals; the multi-channel EEG measurement device can be an existing EEG acquisition system, such as the EMOTIV EPOC FlexSaline Sensor Kit EEG acquisition system. (2) In the driving road environment of the driving simulator, select a type of danger, such as a pedestrian suddenly crossing the road, and then set up three types of tasks: danger perception test in dangerous scenario, danger perception test in pre-danger scenario, and danger perception test in safe scenario. Danger scenario is to play a dangerous video of a pedestrian crossing the road. When the pedestrian turns left and lifts his leg to cross the road, the video stops playing. Pre-danger scenario is to play a dangerous video of a pedestrian crossing the road. The video stops a moment before the pedestrian turns left and crosses the road. Safe scenario is to play a safe video of a pedestrian walking normally. The video stops at random times. When the video stops, the test subject should quickly judge whether it is dangerous and press the button to provide feedback. Two judgment results will be obtained: danger perceived and danger not perceived. In the three types of tasks, the test subject's EEG signal is collected for 10 minutes. The test subject's EEG data within 2000ms after the end of the stimulation video is extracted. The number of tests under each task should be 20 or more. (3) Preprocessing of raw EEG data includes signal amplification, segmentation, signal denoising, bandpass filtering, and artifact removal. Specifically, bandpass filtering and artifact removal are performed on the raw EEG data using a bandpass filter of 0.5 to 45 Hz, and independent component analysis and data reconstruction are performed using the FastICA algorithm based on the principle of maximum negative entropy to effectively remove interference from electrooculography, electromyography, and other artifacts. The preprocessed EEG data retains the β (14 to 20 Hz) frequency band as the target data for subsequent analysis, providing a clear and reliable basic signal for danger perception analysis. (4) Based on the preprocessed EEG data, the baseline normalized frequency band power (unit: dB) of the β band (14~20Hz part) of the Fz and Pz electrodes within a time window of 450~550ms after the end of the stimulation video was calculated for the subjects under different tasks; the baseline normalized frequency band power value was averaged between the Fz and Pz electrodes to obtain the baseline normalized power value P_dB of the β band when danger was perceived in dangerous scenarios. 危险 The baseline normalized power value P_dB of the β band when danger is perceived in a pre-danger scenario. 危险前 The baseline normalized power value P_dB corresponding to the β band in a safe scenario. 安全 Specifically: (4-1) For each type of task, the test time is 10 minutes. Wavelet transform is performed on the preprocessed EEG data to obtain the frequency power value of the β band (14~20Hz) in the Fz and Pz electrode channels in the interval of 450~550ms after the end of the stimulation video. (4-2) For three scenarios—dangerous scene perception, danger perception before danger, and safe scene—frequency band power is calculated separately. Using -500ms to -100ms before the end of the stimulus video as the baseline, the baseline-normalized frequency band power values ​​are obtained. The calculation formula is as follows: , , , , , , , , , , , , , , , Where: the symbol P represents the β-band power value under a single stimulus trial, and k is the frequency value, which ranges from 14 to 20 Hz; This represents the β-band power value (in dB) obtained after baseline normalization for a single stimulus trial. The symbol is... This represents the β-band power value during the baseline period from -500ms to -100ms before stimulation; (5) Calculate P_dB 危险 With P_dB 安全 p-value of the statistical test of the difference between them P_dB was calculated. 危险 With P_dB 危险前 p-value of the statistical test of the difference between them P_dB was calculated. 危险前 With P_dB 安全 p-value of the statistical test of the difference between them The mean was calculated. mean mean Then, calculate the test subject's ability to predict the risk of this type of hazard: if and and and and This indicates that the measurement was valid, the person being tested has the ability to predict this type of hazard, and the level of hazard prediction ability is [value missing]. The larger the value, the stronger the predictive ability; if the conditions and and and and If the condition is not met, then the measurement is invalid. mean mean mean The calculation formula is as follows: , , , Wherein: Wherein: symbol This represents the average β-band power value obtained after baseline normalization of multiple stimulus trials, where i, j, and m represent the sequence numbers of the trials in the dangerous scenario, the trials in the pre-danger scenario, and the trials in the safe scenario, respectively.

[0014] If multiple hazard types need to be measured, repeat steps (1) to (5) above for different hazard types, and use the total number of hazard types with predictive ability. N and the mean level of risk prediction ability This is used for measurement, and the larger the value, the stronger the overall predictive ability. N , The calculation formula is as follows: , , Where: h represents the sequence number of the hazard type with predictive capability. This represents the level of hazard prediction capability for the h-th hazard type.

[0015] The scope of protection of this invention includes, but is not limited to, the above embodiments. The scope of protection is defined by the claims. Any substitutions, modifications, or improvements to this technology that are easily conceived by those skilled in the art fall within the scope of protection of this invention.

Claims

1. A method for measuring a driver's hazard prediction ability based on beta-band electroencephalogram (EEG) signals, characterized in that... The specific steps include the following: (1) The subjects were fitted with multi-channel EEG caps and EEG acquisition devices were used to record EEG signals; (2) In the driving road environment of the driving simulator, select one type of danger: set up three types of tasks: danger perception test in dangerous scenario, danger perception test in pre-danger scenario, and danger perception test in safe scenario; in each of the three types of tasks, collect the EEG signals of the test subjects for 10 minutes to obtain the raw EEG data of the test subjects. (3) Preprocess the raw EEG data, including signal amplification, segmentation, signal noise reduction, bandpass filtering and artifact removal. (4) Based on the preprocessed EEG data, the baseline-normalized frequency band power values ​​of the β band of the Fz and Pz electrodes within a time window of 450–550 ms after the end of the stimulation video were calculated for the subjects under different tasks. The baseline-normalized frequency band power values ​​were averaged between the Fz and Pz electrodes to obtain the baseline-normalized power value P_dB of the β band corresponding to the perception of danger in dangerous scenarios. 危险 The baseline normalized power value P_dB of the β band when danger is perceived in a pre-danger scenario. 危险前 The baseline normalized power value P_dB corresponding to the β band in a safe scenario. 安全 ; (5) Calculate P_dB 危险 With P_dB 安全 p-value of the statistical test of the difference between them P_dB was calculated. 危险 With P_dB 危险前 p-value of the statistical test of the difference between them P_dB was calculated. 危险前 With P_dB 安全 p-value of the statistical test of the difference between them The mean was calculated. mean mean Then, calculate the test subject's ability to predict the risk of this type of hazard: if and and and and This indicates that the measurement was valid, the person being tested has the ability to predict this type of hazard, and the level of hazard prediction ability is [value missing]. The larger the value, the stronger the predictive ability; if the conditions and and and and If the result is not valid, then the measurement is invalid.

2. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 1, characterized in that: In step (2), the test task is a danger assessment task. The test extracts the EEG data of the test subject within 2000ms after the end of the stimulus video. The number of tests under each task is ensured to be 20 or more.

3. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 1, characterized in that: In step (3), during the preprocessing of the raw EEG data, the raw EEG data is subjected to bandpass filtering of 0.5 to 45 Hz, and independent component analysis and data reconstruction are performed using the FastICA algorithm based on the principle of maximum negative entropy to effectively remove artifact interference; and the preprocessed EEG data retains the β band as the target data for subsequent analysis.

4. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 1, characterized in that: In step (4), the method for calculating the baseline-normalized frequency band power value is as follows: (4-1) For each type of task, the test time is 10 minutes. Wavelet transform is performed on the preprocessed EEG data to obtain the frequency power value of the β band in the Fz and Pz electrode channels in the interval of 450-550ms after the end of the stimulation video. (4-2) For three scenarios—dangerous scene perception, danger perception before danger, and safe scene—frequency band power is calculated separately. Using -500ms to -100ms before the end of the stimulus video as the baseline, the baseline-normalized frequency band power values ​​are obtained. The calculation formula is as follows: ， ， ， ， ， ， ， ， ， ， ， ， ， ， ， Where: the symbol P represents the β-band power value under a single stimulus trial, and k is the frequency value, which ranges from 14 to 20 Hz; This represents the β-band power value obtained after baseline normalization in a single stimulus trial, in dB. This represents the β-band power value during the baseline period from -500ms to -100ms before stimulation.

5. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 4, characterized in that: In step (5), the mean mean mean The calculation formula is as follows: ， ， ， Where: symbol This represents the average β-band power value obtained after baseline normalization of multiple stimulus trials, where i, j, and m represent the sequence numbers of the trials in the dangerous scenario, the trials in the pre-danger scenario, and the trials in the safe scenario, respectively.

6. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 1, characterized in that: If multiple hazard types need to be measured, repeat steps (1) to (5) for different hazard types, and use the total number of hazard types with predictive ability. N and the mean level of risk prediction ability The higher the value, the stronger the overall predictive ability.

7. The method for measuring a driver's hazard prediction ability based on beta-band EEG signals as described in claim 6, characterized in that: N and The calculation formula is as follows: ， ， Where: h represents the sequence number of the hazard type with predictive capability. This represents the level of hazard prediction capability for the h-th hazard type.

Images