CN111657923A - Method and system for testing danger perception capability of driver - Google Patents

Method and system for testing danger perception capability of driver Download PDF

Info

Publication number
CN111657923A
CN111657923A CN202010626935.6A CN202010626935A CN111657923A CN 111657923 A CN111657923 A CN 111657923A CN 202010626935 A CN202010626935 A CN 202010626935A CN 111657923 A CN111657923 A CN 111657923A
Authority
CN
China
Prior art keywords
driver
danger
dangerous
physiological characteristic
physiological
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202010626935.6A
Other languages
Chinese (zh)
Inventor
冯树民
吴迪
孙雅丽
盛彬
潘晨龙
赵琥
宋子龙
黄秋菊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Institute of Technology
Original Assignee
Harbin Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Institute of Technology filed Critical Harbin Institute of Technology
Priority to CN202010626935.6A priority Critical patent/CN111657923A/en
Publication of CN111657923A publication Critical patent/CN111657923A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/18—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state for vehicle drivers or machine operators
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2503/00—Evaluating a particular growth phase or type of persons or animals
    • A61B2503/20—Workers
    • A61B2503/22—Motor vehicles operators, e.g. drivers, pilots, captains
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Public Health (AREA)
  • Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Pathology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Psychiatry (AREA)
  • Biophysics (AREA)
  • Veterinary Medicine (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • Physiology (AREA)
  • Mathematical Physics (AREA)
  • Child & Adolescent Psychology (AREA)
  • Social Psychology (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychology (AREA)
  • Hospice & Palliative Care (AREA)
  • Educational Technology (AREA)
  • Developmental Disabilities (AREA)
  • Data Mining & Analysis (AREA)
  • Pulmonology (AREA)
  • Primary Health Care (AREA)
  • Epidemiology (AREA)
  • Databases & Information Systems (AREA)
  • Fuzzy Systems (AREA)
  • Evolutionary Computation (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention relates to a method for testing danger perception capability of a driver, which comprises the following steps: classifying and combining the running conditions of the automobile to construct a dangerous scene library required in driving simulation; designing a physiological characteristic index capable of representing the danger perception capability of a driver; selecting an experienced driver to carry out driving simulation on the dangerous scenes in the dangerous scene library to obtain physiological characteristic index data of the experienced driver in different dangerous scenes; establishing a driver danger perception model and performing model training by using empirical driver physiological characteristic index data; and step five, testing the danger perception capability of the tested driver by using the driver danger perception model. The invention also provides a test system based on the method, which can accurately and pertinently test the danger perception capability of the driver so as to directionally improve the defensive driving skill of the driver, increase the perception capability of the driver and reduce the traffic accident rate.

Description

Method and system for testing danger perception capability of driver
Technical Field
The invention belongs to the technical field of driving behavior recognition, and particularly relates to a driver danger perception test method and system based on multi-source physiological characteristics.
Background
The reform is open to the present, and the motor vehicle reservation in China is increased sharply. Along with the continuous improvement of the industrialization level, the automobile manufacturing cost is lower and lower, and the quantity of motor vehicles in China still has growing potential. At present, the road safety problem in China only becomes more severe in the face of the rapid growth trend of the quantity of the huge motor vehicles. The road traffic system is a dynamic system consisting of people, vehicles and roads. Domestic and foreign researches show that compared with other factors, the proportion of the driver factors in the causes of the traffic accidents is 93%, and the proportion of the driver factors in the traffic accidents caused by the perception errors of the driver is the largest. Therefore, the danger perception capability of the driver is improved on the basis of testing the danger perception capability of the driver, and the important functions of actively preventing and controlling traffic accidents and reducing the traffic accident rate can be achieved.
The danger perception capability of the driver is comprehensively and effectively tested, so that the driving defense skill of the driver can be improved in a targeted manner, and the traffic accident rate is reduced. The current common test methods include questionnaires, static pictures, scene experiments, simulation and real vehicle tests. Among the methods, the driving simulation is a convenient system. However, the existing dangerous scene library of the driving simulation has no targeted design method, in addition, the danger perception of the driver belongs to the physiological characteristic range of the driver, and the existing test indexes of the danger perception capability of the driver do not contain evaluation indexes in the physiological characteristic aspect, so that the danger perception capability of the driver under various different working conditions cannot be comprehensively reflected by the existing driving simulation.
Disclosure of Invention
The invention provides a driver danger perception test method and system based on multi-source physiological characteristics, and aims to solve the technical problem that the danger perception capability of a driver cannot be comprehensively reflected under different working conditions in the prior art.
The invention relates to a method for testing danger perception capability of a driver, which comprises the following steps:
classifying and combining the running conditions of the automobile to construct a dangerous scene library required in driving simulation;
secondly, designing a physiological characteristic index capable of representing the dangerous sensing capability of the driver according to the electrocardio-signal and electroencephalogram signal characteristics of the driver after being stimulated by the external environment;
selecting an experienced driver to carry out driving simulation on the dangerous scene in the dangerous scene library, extracting physiological signals of the driver, preprocessing the physiological signals, calculating physiological characteristic index data of the experienced driver in different dangerous scenes, and storing the physiological characteristic index data in the database;
establishing a danger perception model, training the model by using physiological characteristic indexes of experienced drivers in different dangerous scenes in a database, obtaining standard intervals representing the physiological characteristic indexes of the different dangerous scenes in the model, and embedding the standard intervals into a data analysis platform after the model training is finished;
and fifthly, testing the danger sensing capability of the tested driver, wearing data acquisition equipment for the tested driver, selecting a dangerous scene from a dangerous scene library, performing driving simulation test by combining a driving simulator, extracting physiological signal data, and judging the danger sensing capability of the tested driver in different dangerous scenes through a data processing platform and a data analysis platform.
Further, in the second step, in the multi-source physiological characteristic indexes for representing the dangerous perception capability of the driver, the electrocardio characteristic indexes comprise a heart rate index and a heart rate variability index, the heart rate index is a heart rate increase rate, the heart rate variability time domain index is R-R interval standard deviation SDNN, the root mean square RMSSD of the difference values of two adjacent R-R intervals and NN50 account for the percentage PNN50 of all R-R intervals, and the heart rate variability frequency domain index HRV is the spectral density power value LF of the low frequency band; the EEG characteristic index is the power spectral density mean value of alpha wave and beta wave.
Further, in the second step, the heart rate index is a heart rate increase rate HGR＝(Hf-Hb)/HbIn which H isGRIs the heart rate increase rate, H, of the tested driver in the dangerous scenefFor the heart rate influence value, H, of the driver in a dangerous scenebThe time domain index of the heart rate variability is the reference value before the dangerous scene appears
Figure BDA0002566861480000021
Where N is the total number of heart beats in a given time period, RRiIs the ith R-R interval in a period of time,
Figure BDA0002566861480000022
is the average of N R-R intervals over a period of time;
Figure BDA0002566861480000023
where N is the total number of heart beats in a given time period, RRi、RRi-1Is the length of two adjacent R-R intervals;
Figure BDA0002566861480000024
wherein NN is the total number of R-R intervals in a certain time period, and NN50 isThe difference between adjacent R-R intervals is greater than the R-R interval of 50 ms.
Furthermore, in the third step, the physiological signals to be extracted in the driving simulation test are electrocardiosignals and electroencephalogram signals within 3 seconds after the driver appears in a dangerous scene.
Further, in the third step, when the data processing platform preprocesses the physiological signals, for the electrocardiosignals, filtering and denoising original signals, then carrying out fast Fourier transform, finally obtaining a heart rate time domain graph by resampling, and deriving heart rate and R-R interval data; for electroencephalogram signals, channel positioning is carried out on the obtained electroencephalogram signals, useless electrodes are deleted, then re-reference is carried out by using a whole brain averaging method, independent component analysis is carried out, and the frequency ranges of alpha waves and beta waves are obtained after wavelet denoising and fast Fourier transform.
Further, in the fourth step, when a driver perception model in the data analysis platform is constructed, a Support Vector Machine (SVM) classification model needs to be established, electroencephalogram characteristic indexes and electrocardio signal characteristic indexes of experienced drivers in different dangerous scenes are used for training the model, penalty parameters and kernel function parameters in the SVM classification model are optimized by using a Particle Swarm Optimization (PSO) algorithm to obtain a driver danger perception model, and physiological characteristic index standard intervals in different dangerous scenes are formed in the model to judge whether the tested driver has danger perception capability in the dangerous scenes.
Further, in the fifth step, aiming at the test result of the tested driver, the tested driver is subjected to m times of driving simulation tests, the physiological characteristic index of the tested driver in a specific dangerous scene is calculated, the physiological characteristic index is input into the driver danger perception model of the data analysis platform, the physiological characteristic index is compared with the physiological characteristic index standard interval in the model, and the danger perception hit rate S of the tested driver is output (S ═ S)1，s2，...，sn) N is the number of dangerous scenes in the driving simulation, si(i-1, 2, …, n) is the hit rate of the ith dangerous scene in the driving simulation, specifically,
Figure BDA0002566861480000031
wherein m is the number of driving simulation times; gjAnd the judgment result is 0 or 1, and the judgment result indicates whether the physiological characteristic change of the tested driver in the jth driving simulator in the ith dangerous scene is within a standard interval or not.
After the test is finished, the driver is determined not to have corresponding danger perception capability for the dangerous scene with perception hit rate lower than 60%, and targeted training is enhanced.
The invention also relates to a system for testing the driver danger perception capability, which comprises a data acquisition device, a data processing platform and an analysis platform.
Further, the data acquisition device comprises a physiological sensor and an electroencephalograph, the collected electrocardio signals and the collected electroencephalograms are transmitted to the data processing platform to be preprocessed, all physiological characteristic indexes of the driver are calculated, the driver danger perception model is embedded into the data analysis platform, the physiological characteristic indexes calculated by the data processing platform are received, the physiological characteristic indexes are compared with physiological characteristic index standard intervals in the model, and whether the driver has corresponding danger perception capability under different dangerous scenes or not is evaluated.
Furthermore, physiological signals of electrocardio, respiration, skin electricity, myoelectricity and the like of the driver are collected by adopting a physiological sensor, and the sampling frequency is 64 Hz; electroencephalograms of the driver are collected by an electroencephalograph, and the sampling frequency is 256 Hz.
Has the advantages that:
the danger perception of the driver belongs to the category of the physiological characteristics of the driver, and after research, after receiving external environment stimulation, the change of the physiological characteristics of the driver can reflect whether the driver perceives the environmental danger or not and also determine whether the subsequent operation of the driver can avoid the accident occurrence, so the change of the physiological characteristics is exactly the substantial representation of the danger perception of the driver.
The driver danger perception capability test method provided by the invention has the advantages that the essential characteristics of driver danger perception are returned, the physiological characteristic indexes representing the driver danger perception capability are designed, the physiological characteristic signals of the driver are collected, the driver danger perception model is established, and whether the driver has the corresponding danger perception capability in a specific scene is judged. In addition, because the danger perception abilities of the drivers are different under different working conditions, the invention provides a classification of the driving conditions of the automobile according to the influence factors influencing the perception abilities of the drivers, and the classification can be used as elements for designing dangerous scenes to enrich a dangerous scene library in driving simulation and measure the danger perception abilities of the drivers more comprehensively.
The method and the device can accurately and pertinently test the danger perception capability of the driver, so that the defensive driving skill of the driver can be directionally improved, the perception capability of the driver is improved, and the traffic accident rate is reduced.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a classification chart of the driving condition of the vehicle according to the present invention;
FIG. 3 is a system diagram of the physiological characteristic indicators of the driver's danger perception capability in accordance with the present invention;
FIG. 4 is an exemplary diagram of a first hazardous scenario in accordance with the present invention;
FIG. 5 is an exemplary diagram of a second hazardous scenario in accordance with the present invention;
FIG. 6 is an exemplary diagram of a third hazardous scenario in accordance with the present invention;
FIG. 7 is an exemplary graphical illustration of a hazardous scenario of the present invention.
Detailed Description
The invention discloses a method for testing danger perception capability of a driver, which comprises the following steps:
firstly, the danger sensing abilities of drivers under different operation conditions are different, in order to comprehensively test the danger sensing abilities of the drivers, the operation conditions of an automobile are divided into road conditions, traffic conditions, control conditions and meteorological conditions, the concrete classification is shown in figure 2, and a plurality of elements required for constructing a danger scene library in a driving simulation test are provided; then according to the concrete classification, selecting different working condition elements for combination, and constructing a dangerous scene library required in driving simulation;
and secondly, designing an electrocardio characteristic index and an electroencephalogram characteristic index capable of representing the danger perception capability of the driver according to the electrocardio signal and the electroencephalogram signal after the driver is stimulated by the external environment, and providing an index basis for establishing a driver danger perception model, wherein the electrocardio characteristic index mainly comprises a heart rate index and a heart rate variability index, and the electroencephalogram index mainly is the power spectral density mean value of alpha waves and beta waves. The index system is shown in figure 3.
The heart rate increase rate H is the heart rate index of the electrocardio characteristic indexGR＝(Hf-Hb)/HbIn which H isGRIs the heart rate increase rate, H, of the tested driver in the dangerous scenefFor the heart rate influence value, H, of the driver in a dangerous scenebThe reference value before the dangerous scene appears is the driver.
The heart rate variability index (HRV) in the electrocardio characteristic indexes is divided into a time domain index and a frequency domain index, wherein the heart rate variability time domain index has R-R interval standard deviation SDNN in a certain time, root mean square RMSSD of difference values of two adjacent R-R intervals, and percentage PNN50 of NN50 in all R-R intervals. In particular to
Figure BDA0002566861480000041
Where N is the total number of heart beats in a given time period, RRiIs the ith R-R interval in a period of time,
Figure BDA0002566861480000044
is the average of N R-R intervals over a period of time.
Figure BDA0002566861480000042
Where N is the total number of heart beats in a given time period, RRi、RRi-1Is the length of two adjacent R-R intervals.
Figure BDA0002566861480000043
Wherein NN is the total number of R-R intervals in a certain time period, and NN50 is the number of R-R intervals with the difference value of adjacent R-R intervals being more than 50 ms.
The heart rate variability frequency domain index is a spectral density power value LF of an HRV low frequency range, represents the activity of sympathetic nerves of a driver, and is an important index for researching the psychology and the workload of the driver.
The EEG index is the power spectral density mean value of alpha wave and beta wave, taking alpha wave as an example, the waveform power spectral density mean value is
Figure BDA0002566861480000051
Wherein f isuUpper limit value f of α frequency banddAnd p (f) is the power spectral density of α waves of the electroencephalogram signal, which is the lower limit value of the α frequency band.
Selecting an experienced driver, wearing data acquisition equipment such as a physiological sensor and an electroencephalograph, performing driving simulation on dangerous scenes in a dangerous scene library by combining a driving simulator, extracting electroencephalogram signals and electrocardiosignals, and transmitting the electroencephalogram signals and the electrocardiosignals to a data processing platform; and (4) preprocessing the physiological signals by the data processing platform, calculating the physiological characteristic indexes in the step two, and storing the physiological characteristic indexes in the database.
Selecting a driver with the license holding time of more than 10 years, the driving mileage of more than 25000 kilometers and the driving time of more than 15 hours per week in the last month as an experience driver, carrying out driving simulation test on each dangerous scene in the dangerous scene library, and enabling the experience driver to be familiar with a driving simulator and wear related test instruments before the test so as to eliminate the influence of an experimental device on experimental data.
The relevant data extracted by testing mainly comprises electrocardiosignals and electroencephalogram signals within 3 seconds after the occurrence of dangerous scenes of experienced drivers.
And transmitting the extracted electrocardio-signals and electroencephalogram signals into a data processing platform for preprocessing. For electrocardiosignals, original data are required to be filtered and denoised, then fast Fourier transform is carried out, finally, a heart rate time domain graph is obtained by resampling, and heart rate and R-R interval data are derived; for electroencephalogram signals, channel positioning is carried out on the obtained electroencephalogram signals, useless electrodes are deleted, then re-reference is carried out by using a whole brain averaging method, independent component analysis is carried out, and the frequency ranges of alpha waves and beta waves are obtained after wavelet denoising and fast Fourier transform. And after the physiological signal preprocessing is finished, the data processing platform calculates the physiological characteristic indexes of the experienced drivers according to the second step and stores the physiological characteristic indexes into the database.
And step four, establishing a danger perception model, training the model according to the physiological characteristic indexes of the experienced driver in different dangerous scenes in the database, and obtaining a physiological characteristic standard interval (a physiological characteristic standard library) of the driver in the model. And after the training is finished, embedding the danger perception model into a data analysis platform.
A driver danger perception model is established based on a support vector machine multi-classifier, an electroencephalogram characteristic index and an electrocardiosignal characteristic index (7 items in total) training model of an experienced driver under different dangerous scenes and calculated by a data processing platform are used, and a penalty parameter and a kernel function parameter in the support vector machine model are optimized by utilizing a particle swarm algorithm. After model training is completed, physiological characteristic standard intervals (physiological characteristic standard libraries) representing different dangerous scenes can be automatically formed in the driver danger perception model to serve as comparison libraries of physiological characteristic indexes of the detected driver in different dangerous scenes, and whether the detected driver has danger perception capability in the dangerous scenes or not is judged.
And fifthly, testing the danger sensing capability of the tested driver, wearing data acquisition equipment for the tested driver, selecting a dangerous scene from a dangerous scene library, performing driving simulation test by combining a driving simulator, extracting physiological signals, and judging the danger sensing capability of the tested driver in different dangerous scenes through a data processing platform and a data analysis platform.
And carrying out driving simulation test on the tested driver, and eliminating the influence of the experimental device on the normal driving of the tested driver before the test. And extracting physiological signals in the test, preprocessing the physiological signals by a data processing platform, calculating electrocardio characteristic indexes and electroencephalogram characteristic indexes of the tested driver, inputting the electrocardio characteristic indexes and the electroencephalogram characteristic indexes into a driver danger perception model of the data analysis platform, comparing the physiological characteristic indexes of the tested driver with a physiological characteristic standard library by the perception model, and outputting whether the tested driver has corresponding danger perception capability in a specific dangerous scene or not.
The first embodiment is as follows: in this embodiment, a dangerous scene is constructed for driving simulation to comprehensively test and evaluate the driver's danger perception capability, and the specific method is as follows:
according to the influence factors of driver danger perception, the operation conditions of the automobile are divided into road conditions, traffic conditions, control conditions and meteorological conditions, and the specific classification is shown in an attached figure 2. According to the classification of the operation conditions of the automobile, extracting different contents from the classification as elements of a dangerous scene library, and constructing a dangerous scene:
in the example of the dangerous scene in fig. 4, the road conditions are selected according to the condition classification: urban roads, roundabouts; traffic conditions are as follows: the traffic volume is large, and the traffic components are all automobiles; the control working condition is as follows: marking lines; meteorological conditions: in the daytime. And constructing a traffic danger scene as a roundabout merging road. The purpose is to test whether the driver notices the roundabout fork and enters the vehicle.
In the example of the dangerous scene, fig. 5, the road conditions are selected according to the condition classification: highway, two-way two lane; traffic conditions are as follows: the traffic volume is small, and the traffic comprises trucks and automobiles; the control working condition is as follows: marking lines; meteorological conditions: in the daytime. The construction of the traffic dangerous scene is that the vehicle overtaking at the curve of the opposite vehicle cannot be foreseen by the driver to be tested because the truck and the curve block the sight. The problem of whether the driver can consider bend and freight train blind area is aim at testing.
In the example of the dangerous scene in fig. 6, the road conditions are selected according to the condition classification: suburb roads, two-way lanes, no central dividing zone; traffic conditions are as follows: the traffic volume is large, and the traffic components are all automobiles and pedestrians; the control working condition is as follows: a signal lamp; meteorological conditions: and 4, fog days. And constructing a traffic danger scene as crossing vehicle meeting. The purpose is to test whether the driver notices the front left-turning vehicle and the passerby which may flee in the right.
The above is an example of constructing a traffic hazard scene according to the classification content of the working conditions. The three-dimensional dangerous scene in the driving simulation test can be designed according to the design, and states of traffic components in the scene, such as speed and acceleration of pedestrians and vehicles, can be debugged automatically. Different dangerous scenes can be selected from the dangerous scene library in the driving simulation test and fused into a complete driving road section.
The second embodiment is as follows: the embodiment is a method for acquiring the electrocardio-electroencephalogram signals and electroencephalogram signals of a driver by a data processing platform and preprocessing data of initial signals, and the specific method is as follows:
the acquired data are all physiological signal data of the driver within 3s after the dangerous scene appears in the driving simulation. The physiological sensor can be used for collecting electrocardio, respiration, skin electricity, myoelectricity and other physiological signals of a driver, and the sampling frequency is 64 Hz. The electrocardio is clamped at the earlobe through the clamp for data acquisition, and the skin electricity and the myoelectricity are connected with the fingers, the palm, the skin surface, the inner side of the forearm of the arm, the inner side of the calf and the like of the driver through the electrode plates and the leads. Electroencephalograms of the driver are collected by an electroencephalograph, and the sampling frequency is 256 Hz.
For the pretreatment of the electrocardiosignals, the collected electrocardiosignals are subjected to filtering, denoising and smoothness treatment, then fast Fourier transform is carried out, finally resampling is carried out, a time domain graph of the heart rate can be obtained, and data of the heart rate and the R-R interval are derived.
For preprocessing of electroencephalogram signals, firstly, channel positioning is carried out on the obtained electroencephalogram signals, useless electrodes are deleted, then, re-reference is carried out by using a whole brain averaging method, and independent component analysis is carried out; then, because a large amount of redundancy can be generated by using a continuous wavelet transform method to process signals in a computer, a discrete wavelet function is used to perform wavelet transform, the expansion factor is discretized according to power series, the translation factor is uniformly discretized, and the signals are further processed and noise removed; and finally, converting the electroencephalogram signals from a time domain to a frequency domain by using fast Fourier transform, carrying out power spectrum analysis, extracting waveforms, obtaining a power spectrum of the waveform of each frequency band in each time period, and calculating a power spectral density mean value.
The third concrete implementation mode: in the embodiment, a driver danger perception model is established, the driver danger perception model is established based on a support vector machine, and a physiological characteristic index training model of an experienced driver is used, and the method specifically comprises the following steps:
according to the physiological characteristic indexes of an experienced driver in a driving simulation test, a driver danger perception model is established by adopting a support vector machine multi-classifier, so that whether the tested driver has danger perception capability in a specific danger scene is evaluated.
Selecting a driver with the license holding time of more than 10 years, the driving mileage of more than 25000 kilometers and the driving time of more than 15 hours per week in the last month as an experience driver, carrying out driving simulation test on each dangerous scene in the dangerous scene library, and enabling the experience driver to be familiar with a driving simulator and wear related test instruments before the test so as to eliminate the influence of an experimental device on experimental data. In the simulation process, physiological signals of a driver in 3s after dangerous scenes appear are extracted, data preprocessing is carried out, physiological characteristic indexes (7 items in total) of the driver in each dangerous scene are calculated and counted, a model is input, punishment parameters and kernel function parameters in a classification model are optimized by utilizing a particle swarm algorithm, model training is completed, and physiological characteristic index standard intervals (physiological characteristic standard libraries) representing different dangerous scenes are obtained in the model.
The fourth concrete implementation mode: in this embodiment, whether the driver to be tested has the corresponding danger sensing capability in a specific dangerous scene is tested, and the test method is as follows:
and selecting a dangerous scene from the dangerous scene library to construct a driving simulation road section in the driving simulator, carrying out m driving simulation tests on the tested driver, and eliminating the influence of the experimental device on the normal driving of the tested driver before the tests. Extracting electroencephalogram signals and electrocardiosignals within 3 seconds after each dangerous scene appears, processing according to the second implementation mode, then calculating a perception model, inputting physiological characteristic indexes of a detected driver into the driver danger perception model, comparing the driver danger perception model with a physiological characteristic standard library, and judging whether the detected driver has corresponding danger perception capability in a specific dangerous scene.
The data processing platform outputs the hit rate S ═ S (S) of the detected driver' S danger perception1，s2，...，sn) N is the number of dangerous scenes in the driving simulation,si(i-1, 2, …, n) is the hit rate of the ith dangerous scene in the driving simulation, specifically the hit rate
Figure BDA0002566861480000081
Wherein m is the number of driving simulation times; gjAnd the judgment result is 0 or 1, and the judgment result indicates whether the physiological characteristic change of the tested driver in the jth driving simulator in the ith dangerous scene is within a standard interval or not.
After the test, the driver is determined not to have corresponding danger perception capability for the dangerous scene with perception hit rate lower than 60%, and the targeted training is strengthened.
The fifth concrete implementation mode: in this embodiment, a system for testing driver's danger awareness is provided, which includes a data acquisition device, a data processing platform, and an analysis platform.
The data acquisition device comprises acquisition instruments such as a physiological sensor and an electroencephalograph, and transmits the collected electrocardio and electroencephalogram signals to the data processing platform for preprocessing, and finally evaluates the danger perception capability of the driver based on the analysis platform.
The physiological sensors are adopted to collect electrocardio, respiration, skin electricity, myoelectricity and other physiological signals of a driver, and the sampling frequency is 64 Hz. Electroencephalograms of the driver are collected by an electroencephalograph, and the sampling frequency is 256 Hz.
The data processing platform integrates physiological signals collected by the physiological sensor, information data are synchronized on the processing platform, the two signals are analyzed at the same time, the electrocardio-electroencephalogram signals and the electroencephalogram signals are respectively preprocessed in specific time, and all physiological characteristic indexes of a driver are calculated.
And embedding a driver danger perception model in the data analysis platform, receiving the physiological characteristic indexes calculated by the data processing platform, comparing the physiological characteristic indexes with a physiological characteristic standard library in the model, and evaluating whether the driver has corresponding danger perception capability in different dangerous scenes.
The method constructs a driver danger perception model by using a support vector machine multi-classifier, and tests the danger perception capability of a driver; in addition, seven physiological characteristic indexes capable of representing dangerous perception capability of the driver are designed by adopting the complementary advantages of the electrocardiosignals and the electroencephalogram signals, and the test accuracy of the model is improved. Based on the above, the invention also provides a set of test system which can carry out comprehensive test, evaluation and targeted training on the danger perception capability of the driver by combining the driving simulator.
The present invention is not limited to the above embodiments, and those skilled in the art can easily make various changes and modifications according to the main concept and spirit of the present invention, so the protection scope of the present invention shall be subject to the protection scope claimed in the claims.

Claims (10)

1. A method of testing the perceived risk of a driver, comprising the steps of:
classifying and combining the running conditions of the automobile to construct a dangerous scene library required in driving simulation;
secondly, designing a physiological characteristic index capable of representing the dangerous sensing capability of the driver according to the electrocardio-signal and electroencephalogram signal characteristics of the driver after being stimulated by the external environment;
selecting an experienced driver to carry out driving simulation on the dangerous scene in the dangerous scene library, extracting physiological signals of the driver, preprocessing the physiological signals, calculating physiological characteristic index data of the experienced driver in different dangerous scenes, and storing the physiological characteristic index data in the database;
establishing a danger perception model, training the model by using physiological characteristic indexes of experienced drivers in different dangerous scenes in a database, obtaining standard intervals representing the physiological characteristic indexes of the different dangerous scenes in the model, and embedding the standard intervals into a data analysis platform after the model training is finished;
and fifthly, testing the danger sensing capability of the tested driver, wearing data acquisition equipment for the tested driver, selecting a dangerous scene from a dangerous scene library, performing driving simulation test by combining a driving simulator, extracting physiological signal data, and judging the danger sensing capability of the tested driver in different dangerous scenes through a data processing platform and a data analysis platform.
2. The method for testing the driver danger perceptibility according to claim 1, wherein in the second step, when designing the physiological characteristic index capable of characterizing the driver danger perceptibility, the electrocardiogram characteristic index includes a heart rate index and a heart rate variability index, the heart rate index is a heart rate increase rate, the heart rate variability time domain index is a standard deviation SDNN of R-R intervals, a root mean square RMSSD of difference values of two adjacent R-R intervals, a percentage PNN50 of NN50 in all R-R intervals, and a spectral density power value LF of a low frequency band of a heart rate variability frequency domain index HRV; the EEG characteristic index is the power spectral density mean value of alpha wave and beta wave.
3. The method for testing the driver's danger perception according to claim 2, wherein in the second step, the heart rate index is a heart rate increase rate HGR＝(Hf-Hb)/HbIn which H isGRIs the heart rate increase rate, H, of the tested driver in the dangerous scenefFor the heart rate influence value, H, of the driver in a dangerous scenebThe time domain index of the heart rate variability is the reference value before the dangerous scene appears
Figure FDA0002566861470000011
Where N is the total number of heart beats in a given time period, RRiIs the ith R-R interval in a period of time,
Figure FDA0002566861470000012
is the average of N R-R intervals over a period of time;
Figure FDA0002566861470000013
Figure FDA0002566861470000014
where N is the total number of heart beats in a given time period, RRi、RRi-1Is the length of two adjacent R-R intervals;
Figure FDA0002566861470000015
wherein NN is the total number of R-R intervals in a certain time period, and NN50 is the number of R-R intervals with the difference value of adjacent R-R intervals being more than 50 ms.
4. The method for testing the driver's danger perceptibility according to claim 1, wherein in the third step, the physiological signal to be extracted in the driving simulation test is an electrocardiosignal and an electroencephalogram signal within 3 seconds after the driver appears in the dangerous scene.
5. The method for testing the driver's danger perception capability of claim 1, wherein in the third step, when the data processing platform preprocesses the physiological signals, the electrocardiosignals need to be filtered and denoised, then fast Fourier transform is carried out, finally, a heart rate time domain graph is obtained by resampling, and heart rate and R-R interval data are derived; for electroencephalogram signals, channel positioning is carried out on the obtained electroencephalogram signals, useless electrodes are deleted, then re-reference is carried out by using a whole brain averaging method, independent component analysis is carried out, and the frequency ranges of alpha waves and beta waves are obtained after wavelet denoising and fast Fourier transform.
6. The method for testing the driver danger perception capability of claim 1, wherein in the fourth step, when a driver perception model in the data analysis platform is constructed, a Support Vector Machine (SVM) classification model needs to be established, an electroencephalogram characteristic index and an electrocardio signal characteristic index of an experienced driver under different dangerous scenes are used for training the model, a Particle Swarm Optimization (PSO) is used for optimizing punishment parameters and kernel function parameters in the SVM classification model to obtain the driver danger perception model, and physiological characteristic index standard intervals under different dangerous scenes are formed in the model to judge whether the tested driver has the danger perception capability in the dangerous scenes.
7. The test ride of claim 1And fifthly, aiming at the test result of the tested driver, performing m times of driving simulation tests on the tested driver, calculating the physiological characteristic index of the tested driver in a specific dangerous scene, inputting the physiological characteristic index into a driver danger perception model of the data analysis platform, comparing the physiological characteristic index with a physiological characteristic index standard interval in the model, and outputting the danger perception hit rate S (S) of the tested driver1，s2，...，sn) N is the number of dangerous scenes in the driving simulation, si(i is 1,2, …, n) is the hit rate of the ith dangerous scene in the driving simulation,
Figure FDA0002566861470000021
wherein m is the number of driving simulation times; gjAnd the judgment result is 0 or 1, and the judgment result indicates whether the physiological characteristic change of the tested driver in the jth driving simulator in the ith dangerous scene is within a standard interval or not. After the test is finished, the driver is determined not to have corresponding danger perception capability for the dangerous scene with perception hit rate lower than 60%, and targeted training is enhanced.
8. A system for testing the danger perception capability of a driver is characterized by comprising a data acquisition device, a data processing platform and an analysis platform.
9. The system for testing the driver's danger perception capability of claim 8, wherein the data acquisition device comprises a physiological sensor and an electroencephalograph, the data acquisition device transmits the collected electrocardio-and electroencephalograph signals to a data processing platform for preprocessing and calculating each physiological characteristic index of the driver, a driver danger perception model is embedded in the data analysis platform, the physiological characteristic indexes calculated by the data processing platform are received and compared with the physiological characteristic index standard interval in the model, and whether the driver has corresponding danger perception capability in different danger scenes or not is evaluated.
10. The system for testing the danger perception capability of the driver according to claim 9, wherein physiological signals of electrocardio, respiration, skin electricity, myoelectricity and the like of the driver are collected by a physiological sensor, and the sampling frequency is 64 Hz; electroencephalograms of the driver are collected by an electroencephalograph, and the sampling frequency is 256 Hz.
CN202010626935.6A 2020-07-02 2020-07-02 Method and system for testing danger perception capability of driver Withdrawn CN111657923A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010626935.6A CN111657923A (en) 2020-07-02 2020-07-02 Method and system for testing danger perception capability of driver

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010626935.6A CN111657923A (en) 2020-07-02 2020-07-02 Method and system for testing danger perception capability of driver

Publications (1)

Publication Number Publication Date
CN111657923A true CN111657923A (en) 2020-09-15

Family

ID=72391233

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010626935.6A Withdrawn CN111657923A (en) 2020-07-02 2020-07-02 Method and system for testing danger perception capability of driver

Country Status (1)

Country Link
CN (1) CN111657923A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112349171A (en) * 2020-11-11 2021-02-09 济南北方交通工程咨询监理有限公司 Driving safety simulation system and method based on virtual reality technology
CN113077129A (en) * 2021-03-19 2021-07-06 东南大学 Driver danger perception capability evaluation method suitable for high-altitude area
CN113449790A (en) * 2021-06-25 2021-09-28 贵州省都匀公路管理局 Mountain trunk highway high-risk road section identification method based on SVM
CN115089182A (en) * 2022-05-23 2022-09-23 长安大学 Method for evaluating risk perception capability of multidimensional driver

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101530329A (en) * 2009-03-06 2009-09-16 北京理工大学 Semi-physical driving fatigue vision simulation system platform
US20120252709A1 (en) * 2011-04-01 2012-10-04 John Felts Vessels, contact surfaces, and coating and inspection apparatus and methods
CN107822623A (en) * 2017-10-11 2018-03-23 燕山大学 A kind of driver fatigue and Expression and Action method based on multi-source physiologic information
CN108615110A (en) * 2018-04-23 2018-10-02 吉林大学 The driving risk evaluating system of view-based access control model characteristic
CN109765060A (en) * 2018-12-29 2019-05-17 同济大学 A kind of automatic driving vehicle traffic coordinating virtual test system and method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101530329A (en) * 2009-03-06 2009-09-16 北京理工大学 Semi-physical driving fatigue vision simulation system platform
US20120252709A1 (en) * 2011-04-01 2012-10-04 John Felts Vessels, contact surfaces, and coating and inspection apparatus and methods
CN107822623A (en) * 2017-10-11 2018-03-23 燕山大学 A kind of driver fatigue and Expression and Action method based on multi-source physiologic information
CN108615110A (en) * 2018-04-23 2018-10-02 吉林大学 The driving risk evaluating system of view-based access control model characteristic
CN109765060A (en) * 2018-12-29 2019-05-17 同济大学 A kind of automatic driving vehicle traffic coordinating virtual test system and method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
盛彬: "基于信息融合的驾驶危险场景分析识别研究", 《中国优秀硕士学位论文全文数据库》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112349171A (en) * 2020-11-11 2021-02-09 济南北方交通工程咨询监理有限公司 Driving safety simulation system and method based on virtual reality technology
CN113077129A (en) * 2021-03-19 2021-07-06 东南大学 Driver danger perception capability evaluation method suitable for high-altitude area
CN113449790A (en) * 2021-06-25 2021-09-28 贵州省都匀公路管理局 Mountain trunk highway high-risk road section identification method based on SVM
CN115089182A (en) * 2022-05-23 2022-09-23 长安大学 Method for evaluating risk perception capability of multidimensional driver

Similar Documents

Publication Publication Date Title
CN111657923A (en) Method and system for testing danger perception capability of driver
Zhenhai et al. Driver drowsiness detection based on time series analysis of steering wheel angular velocity
CN111803065B (en) Dangerous traffic scene identification method and system based on electroencephalogram data
CN101491443A (en) Relational model of driver fatigue and vehicle riding trail
CN112419719B (en) Method and system for evaluating traffic operation safety of highway
CN113743471B (en) Driving evaluation method and system
Wan et al. On‐road experimental study on driving anger identification model based on physiological features by ROC curve analysis
CN111643077B (en) Identification method for complexity of traffic dynamic factors based on electroencephalogram data
Kang et al. Analysis of vehicle maneuverability and driving characteristics on a curved road condition
CN113753059B (en) Method for predicting takeover capacity of driver under automatic driving system
CN106913333B (en) Method for selecting sensitivity characteristic index of continuous attention level
Koh et al. Driver drowsiness detection via PPG biosignals by using multimodal head support
CN108711204B (en) Driving abnormity detection system and method integrating human-vehicle-road multi-source information
CN105405253A (en) Method and apparatus for monitoring fatigue state of driver
CN111311093B (en) Road intersection risk assessment and early warning method based on driver physiological data
Bittner et al. Detecting of fatigue states of a car driver
Perello-March et al. Physiological measures of risk perception in highly automated driving
Qi et al. Model of driver’s eye movement and ECG index under tunnel environment based on spatiotemporal data
Wei et al. Driver's mental workload classification using physiological, traffic flow and environmental factors
Hu et al. Analysis on biosignal characteristics to evaluate road rage of younger drivers: A driving simulator study
Distefano et al. Mining of the association rules between driver electrodermal activity and speed variation in different road intersections
Hoogendoorn et al. Mental workload, longitudinal driving behavior, and adequacy of car-following models for incidents in other driving lane
Fu et al. Effects of in-vehicle navigation on perceptual responses and driving behaviours of drivers at tunnel entrances: a naturalistic driving study
CN110742602B (en) Method for recognizing aggressive driving state based on electroencephalogram and vehicle driving data
CN213883185U (en) Information acquisition system for vehicle running process

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20200915