CN112613786B - Individualized and differentiated driving risk evaluation method based on driving safety event - Google Patents

Individualized and differentiated driving risk evaluation method based on driving safety event Download PDF

Info

Publication number
CN112613786B
CN112613786B CN202011603122.1A CN202011603122A CN112613786B CN 112613786 B CN112613786 B CN 112613786B CN 202011603122 A CN202011603122 A CN 202011603122A CN 112613786 B CN112613786 B CN 112613786B
Authority
CN
China
Prior art keywords
driving
driving safety
risk evaluation
steering wheel
acceleration
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.)
Active
Application number
CN202011603122.1A
Other languages
Chinese (zh)
Other versions
CN112613786A (en
Inventor
张晖
刘永杰
吴超仲
肖逸影
张琦
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.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
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 Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN202011603122.1A priority Critical patent/CN112613786B/en
Publication of CN112613786A publication Critical patent/CN112613786A/en
Application granted granted Critical
Publication of CN112613786B publication Critical patent/CN112613786B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
    • G06F7/58—Random or pseudo-random number generators
    • G06F7/582—Pseudo-random number generators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Educational Administration (AREA)
  • Evolutionary Computation (AREA)
  • Marketing (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Game Theory and Decision Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses an individual differential driving risk evaluation method based on driving safety events. Specifically, historical behavior data of other drivers need to be extracted, driving safety events are extracted according to a preset threshold value, equivalent matching samples are extracted from a database of a normal driving state of the driver according to the extracted sample capacity, the driving behavior data of the driver in the driving safety events and the behavior data of the normal driving state are subjected to statistical inspection, results with significant statistical differences in the inspection are input into a clustering algorithm for classification, risk evaluation indexes are obtained, and driving risk weights of various driving safety events to the driver are obtained according to the risk evaluation indexes. The method fully considers the individual difference of different drivers in various driving safety incident risks.

Description

Individualized and differentiated driving risk evaluation method based on driving safety event
Technical Field
The invention relates to the field of traffic safety, driving behaviors and data mining, in particular to an individual differential driving risk evaluation method based on driving safety events.
Background
In the existing research, quantitative driving risk is generally analyzed for three events, which are divided from high to low in severity in turn: accidents (Crash), imminent accidents (Near-Crash), and security Events (Critical incorporated Events: CIEs). In the driving risk evaluation based on the driving behavior and accident risk association rule, the risk level of the driver is often evaluated in the historical traffic accident. In particular, in the research on the aspect of personalized driving behavior insurance and the like, indexes such as historical accident frequency, severity and amount of out-of-insurance claims are used as indexes for evaluating the behavior risk level of the driver.
Although the historical accident occurrence frequency index can represent the driving risk level of the driver, the method also has the following limitations: 1) the accident rate is too low, and the accident can only occur under extreme working conditions, so that the number of samples is small and the accident rule is difficult to find; 2) the acquisition period is long, resulting in low timeliness and accuracy of driving risk assessment. Therefore, in the existing research, driving safety events (CIEs) are used for judging driving risk states, the judgment standard of the CIEs is lower than that of Near-Crash, and the CIEs are events which have high occurrence frequency and low severity and are marked by high acceleration rate, high deceleration rate or other motion characteristics. The driving control stability and the accident tendency of a driver in different driving safety events have obvious individual difference, and the individual difference between the importance degrees of various driving safety event risks is not considered in the weighting method adopted in the existing research, namely the subjective weighting method and the objective weighting method.
Disclosure of Invention
The invention aims to provide an individualized and differentiated driving risk evaluation method based on driving safety events, which is used for solving the problem that the individual difference of the risks of various driving safety events is not considered in the traditional empowerment methods such as an analytic hierarchy process, an entropy weight method and the like.
In order to solve the technical problem, the invention provides a technical scheme that:
acquiring driving behavior data: collecting real-time driving behavior data of a driver by using a sensor; the driving behavior data includes: the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data, the right steering lamp data, the steering wheel angle and the steering wheel angle acceleration;
judging real-time driving behavior data by using a risk evaluation model to obtain a driving risk evaluation score;
the risk evaluation model is obtained by the following steps:
step 1, extracting driving safety events, selecting historical behavior data of other drivers within a certain time to perform data preprocessing, and extracting the driving safety events by taking m% upper bound of limit operation of the drivers as an extraction threshold value of the driving safety events by using a probability density function of each driving safety event judgment index to obtain the frequency of the driving safety events, wherein m is more than 0 and less than 100; extracting equivalent pairing samples from a database of normal driving states of a driver according to the extracted traffic safety event sample capacity; the database of the normal driving state of the driver is a database containing normal driving behavior data of other drivers;
and 2, performing index statistical test, namely performing significance pairing comparison on the driving behavior data of the driver in each type of driving safety event and the driving behavior data in the normal driving state, and dividing the driving behavior indexes into three types according to the obtained result: no statistical difference, statistical difference and significant statistical difference;
step 3, screening indexes, namely inputting the results with significant statistical differences into a machine learning clustering algorithm for classification, and obtaining risk evaluation indexes corresponding to the results by using a statistical method;
and 4, risk evaluation modeling, namely sequencing results from low to high according to the variation degree of the risk evaluation indexes in corresponding events by utilizing the calculation results of the clustering algorithm, converting the risk tendency grades of the individual drivers in different events to obtain the driving risk weight of various driving safety events to the drivers, and establishing the risk evaluation model.
According to the scheme, the driving behavior data is acquired by adopting a vehicle-mounted CAN and a steering wheel corner sensor; the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data and the right steering lamp data are collected by a vehicle-mounted CAN; and the steering wheel angle acceleration are acquired by a steering wheel angle sensor.
According to the scheme, the driving safety events are specifically divided into four types, namely emergency braking, emergency turning, emergency refueling door and overspeed.
According to the scheme, the driving safety event extraction comprises data preprocessing, and specifically comprises the following steps: firstly, synchronizing data of different sensors, and then carrying out wavelet packet noise reduction processing on the following indexes: the speed average value, the speed standard deviation, the speed maximum value, the speed range difference, the acceleration average value, the deceleration average value, the acceleration standard deviation, the acceleration maximum value, the deceleration maximum value, the steering wheel right corner average value, the steering wheel left corner average value, the steering wheel corner standard deviation, the steering wheel corner maximum value, the steering wheel right corner acceleration average value, the steering wheel left corner acceleration average value, the steering wheel corner acceleration standard deviation and the steering wheel corner acceleration maximum value.
According to the scheme, the driving safety event extraction comprises the calibration of the driving safety event, wherein the judgment index of overspeed is the maximum speed, the judgment index of rapid acceleration is the maximum acceleration, the judgment index of rapid deceleration is the maximum deceleration, and the judgment index of rapid turning is the maximum steering wheel corner acceleration; and taking the m% upper bound of the limit operation of the driver as an event extraction threshold value so as to obtain the frequency of the driving safety events.
According to the scheme, the specific method for extracting the equivalent matched samples from the database of the normal driving state of the driver comprises the following steps: generating a pseudorandom integer by using MATLAB, and reserving the pseudorandom function seed to keep the selected observation value and the sequence thereof unchanged;
setting a pseudo random number generation seed:
rng(seed)
pseudo-random number decimation paired samples Xpair：
Xpair＝randi([imin,imax],m,n)
Where imin is the minimum value of the overall sample index, imax is the maximum value of the overall sample index, and m and n respectively represent the extracted index values and the rows and columns of the composed matrix.
According to the scheme, the three driving behavior indexes obtained by the index statistical test are specifically as follows: no statistical difference exists, and h is equal to 0; there is a statistical difference that h is equal to 1, p is less than 0.05 and greater than 0.01; the statistical difference is significant, h is equal to 1, and p is less than 0.01; where h is the output term, with two results, 0 and 1; h-0 indicates that the null hypothesis is negated at 5% confidence, i.e., the two sets of paired samples are statistically considered to be from the same distribution; h-1 indicates that the null hypothesis is negated, and the two sets of matched samples are statistically considered to be data from different distributions; p represents the set criterion of significant difference, p <0.05 considered the difference, and p <0.01 considered the significant difference.
According to the scheme, in the risk evaluation modeling, a specific conversion formula of the driving risk weight is as follows:
Figure BDA0002871887280000041
wherein, wijRepresenting the driving risk weight of the driver j in the i-type driving safety event; dijAnd n is the total number of the driving safety events.
According to the scheme, the specific acquisition method of the driving risk assessment score comprises the following steps:
Figure BDA0002871887280000042
wherein G represents a driving risk assessment score, i and n represent the category and total number of driving safety events, respectively, and γiA driving risk weight, x, representing a category iiAnd representing the frequency of the traffic safety events of the category i.
The invention has the beneficial effects that: the method fully considers the accident tendency and the difference of driving operation habits of a driver in different driving safety events, can accurately and objectively determine the driving risk weight of various events, and has more reasonable judgment result.
Drawings
FIG. 1 is a general flow diagram;
FIG. 2 is a clustering algorithm classification diagram;
FIG. 3 is a personalized driving risk assessment process;
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present disclosure more apparent, the technical solutions of the embodiments of the present disclosure will be described clearly and completely with reference to the drawings of the embodiments of the present disclosure. It is to be understood that the described embodiments are only a few embodiments of the present disclosure, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the described embodiments of the disclosure without any inventive step, are within the scope of protection of the disclosure.
Referring to fig. 1, the personalized differential driving risk evaluation method based on driving safety events specifically includes the following steps:
acquiring driving behavior data: collecting real-time driving behavior data of a driver by using a sensor; the driving behavior data includes: the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data, the right steering lamp data, the steering wheel angle and the steering wheel angle acceleration;
judging real-time driving behavior data by using a risk evaluation model to obtain a driving risk evaluation score;
the risk evaluation model is obtained by the following steps:
step 1, extracting driving safety events, selecting historical behavior data of other drivers within a certain time to perform data preprocessing, and extracting the driving safety events by taking m% upper bound of limit operation of the drivers as an extraction threshold value of the driving safety events by using a probability density function of each driving safety event judgment index to obtain the frequency of the driving safety events, wherein m is more than 0 and less than 100; extracting equivalent pairing samples from a database of normal driving states of a driver according to the extracted traffic safety event sample capacity; the database of the normal driving state of the driver is a database containing normal driving behavior data of other drivers;
and 2, performing index statistical test, namely performing significance pairing comparison on the driving behavior data of the driver in each type of driving safety event and the driving behavior data in the normal driving state, and dividing the driving behavior indexes into three types according to the obtained result: no statistical difference, statistical difference and significant statistical difference;
step 3, screening indexes, namely inputting the results with significant statistical differences into a machine learning clustering algorithm for classification, and obtaining risk evaluation indexes corresponding to the results by using a statistical method;
and 4, risk evaluation modeling, namely sequencing results from low to high according to the variation degree of the risk evaluation indexes in corresponding events by utilizing the calculation results of the clustering algorithm, converting the risk tendency grades of the individual drivers in different events to obtain the driving risk weight of various driving safety events to the drivers, and establishing the risk evaluation model.
According to the scheme, the driving behavior data is acquired by adopting a vehicle-mounted CAN and a steering wheel corner sensor; the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data and the right steering lamp data are collected by a vehicle-mounted CAN; and the steering wheel angle acceleration are acquired by a steering wheel angle sensor.
Preferably, in the step of acquiring the driving behavior data, the sensors include a vehicle-mounted CAN and a steering wheel angle sensor; the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data and the right steering lamp data are collected by a vehicle-mounted CAN; the steering wheel corner and the steering wheel corner acceleration are collected by a steering wheel corner sensor.
Preferably, the driving safety events are specifically divided into four types, namely emergency braking, emergency cornering, emergency refueling door and overspeed.
Preferably, the driving safety event extraction includes data preprocessing, specifically: firstly, synchronizing data of different sensors, and then carrying out wavelet packet noise reduction treatment on the following indexes: the speed average value, the speed standard deviation, the speed maximum value, the speed range difference, the acceleration average value, the deceleration average value, the acceleration standard deviation, the acceleration maximum value, the deceleration maximum value, the steering wheel right corner average value, the steering wheel left corner average value, the steering wheel corner standard deviation, the steering wheel corner maximum value, the steering wheel right corner acceleration average value, the steering wheel left corner acceleration average value, the steering wheel corner acceleration standard deviation and the steering wheel corner acceleration maximum value.
Preferably, the driving safety event extraction comprises the calibration of the driving safety event, wherein the judgment index of overspeed is the maximum speed, the judgment index of rapid acceleration is the maximum acceleration, the judgment index of rapid deceleration is the maximum deceleration, and the judgment index of rapid turning is the maximum steering wheel angular acceleration; because all the indexes conform to normal distribution, the upper bound of 97.5% of the limit operation of a driver is taken as an event extraction threshold value, so that the driving safety event frequency is obtained, and the method specifically comprises the following steps:
if the index conforms to normal distribution, the probability density function is:
Figure BDA0002871887280000061
in the formula, sigma is the standard deviation of the index, and mu is the mean value of the index;
for each event discrimination index, taking the upper limit of 97.5% of the limit operation of the driver as an extraction threshold of the event:
Figure BDA0002871887280000062
in the formula, xiIs the threshold value, x, of the event discrimination indicatormaxIs the maximum value of the discrimination index.
Preferably, the specific method for extracting the equivalent matched samples from the database of the normal driving state of the driver comprises the following steps: generating a pseudorandom integer by using MATLAB, and reserving the pseudorandom function seed to keep the selected observation value and the sequence thereof unchanged;
setting a pseudo random number generation seed:
rng(seed)
pseudo-random number decimation paired samples Xpair：
Xpair＝randi([imin,imax],m,n)
Where imin is the minimum value of the overall sample index, imax is the maximum value of the overall sample index, and m and n respectively represent the extracted index values and the rows and columns of the composed matrix.
Preferably, the three driving behavior indexes obtained by the index statistical test are specifically: no statistical difference exists, and h is equal to 0; the statistical difference exists, h is equal to 1, and p is less than 0.05 and more than 0.01; the statistical difference is significant, h is equal to 1, and p is less than 0.01; in this embodiment, u test is performed on the paired samples, and the principle is to test the difference between each behavior characteristic of the same driver under different types of risk conditions and the overall behavior characteristic of the same driver without considering the risk conditions, and the method specifically includes the following steps:
the object of the test is the difference between the observed values of the paired samples, if two paired samples X1iAnd X2iThe difference is di＝X1i-X2iIndependent and from normal distribution, then diWhether the maternal expectation value mu is mu or not0The statistic u can be calculated using the following formula:
Figure BDA0002871887280000071
wherein, i is 1 … n, n is the number of the paired driving safety event samples,
Figure BDA0002871887280000072
is the average of the differences of the paired samples, sdStandard deviation for paired sample differences:
Figure BDA0002871887280000073
Figure BDA0002871887280000074
the statistic u is zero hypothesis μ ═ μ0If true, obeying t distribution with the degree of freedom of n-1;
in the embodiment, a matching sample u test function built in MATLAB software is used as a tool to analyze the driving behavior characteristic difference, and the functional form is as follows:
[h,p,ci]＝utest(x,y,alpha)
for the input term, x, y are two sets of paired samples, and alpha is the confidence, where the effect of confidence is to examine the significance level of the difference; the output term h includes both 0 and 1 results: if h is 0, it indicates that the null hypothesis is negated at 5% confidence (inferred from the returned h being 0 when x is set to y), i.e., the x, y two sets of paired samples are statistically considered to be from the same distribution. If h is 1, it indicates that the null hypothesis is negated, i.e., the x, y two sets of matched samples are statistically considered to be data from different distributions, i.e., there is discrimination; p represents a set standard of significant difference, and p <0.05 is generally defined as the difference between the two, and p <0.01 is defined as the significant difference between the two; ci is the (1-alpha) confidence interval in which the actual mean lies.
Preferably, referring to fig. 2, in this embodiment, the K-means clustering algorithm is used to further analyze the results with significant statistical differences, that is, behavior individual differences and risk individual differences in a typical driving safety event, and the principle of the K-means clustering algorithm is as follows:
the K-means cluster object can be any multidimensional, so it clusters a D-dimensional vector of point sets D:
D＝{xi|i＝1…N}
wherein x isi∈RdRepresenting the ith object;
the specific calculation process of the K-means clustering algorithm comprises the following steps:
1) randomly selecting k values from the sample point set D as initial cluster centers (mu)1…μk)；
2) For the a-th iteration, all sample points P in the sample point set D are sampledt(t 1,2 … n) and the center point of each category
Figure BDA0002871887280000081
Sequentially calculating the Euclidean distance d:
Figure BDA0002871887280000082
3) comparing the cluster centers
Figure BDA0002871887280000083
And PbThe distance of (2) is classified into a category group in which the cluster center with the smallest relative distance is located;
4) calculating the clustering distance mean value of the same type of samples, and resetting the clustering center:
Figure BDA0002871887280000091
5) calculating a standard measure function E of all points in DiIf the calculation result satisfies | Ei+1-Ei|<δ stopping iteration, otherwise returning to step 2:
Figure BDA0002871887280000092
the present embodiment performs an experimental guideline by using MATLAB self-contained toolkit, and the functional form is as follows:
[Idx,C,sumD,D]＝kmeans(X,K)
the method comprises the following steps that X represents a clustering sample set matrix, K represents the number of initial clustering centers, Idx is used for storing clustering result labels of all points, C is used for storing the positions of K clustering centroids, sumD is used for storing the sum of distances between all points among clusters and the points of the centroids, and D is used for storing the distance between each point and all centroids;
the category of the K-means is controlled by a parameter K, so that K is the key input of the K-means algorithm and is the initial clustering center of the sample; and in the clustering process, distributing the coordinate points to the category of the clustering center closest to the coordinate points according to the similarity degree of the coordinate points and the initial clustering center, and then calculating the mean value of all respective distances of the k clusters. Iterating until the standard measure function begins to converge; the standard measure function is the mean square error of each cluster under a general condition, so that the samples in the same class are as close as possible, and the samples in different classes are as far away as possible. The number of initial clustering centers has an important influence on the result of the K-means clustering algorithm, and the driving human type is divided into three types by referring to the existing research: and (4) setting the number of the clustering centers to be k to 3 by combining the result distribution characteristics of the matched sample u test for the cautious drivers, the ordinary drivers and the aggressive drivers.
Preferably, in the risk evaluation modeling, a specific conversion formula of the driving risk weight is as follows:
Figure BDA0002871887280000093
wherein, wijRepresenting the driving risk weight of the driver j in the i-type driving safety event; dijRepresenting a risk evaluation index of the driver j in the i-type safety event; and n is the total number of the driving safety events.
Preferably, referring to fig. 3, the specific method for obtaining the driving risk assessment score is as follows:
Figure BDA0002871887280000101
wherein G represents a driving risk assessment score, i and n represent the category and total number of driving safety events, respectively, and γiA driving risk weight, x, representing a category iiAnd representing the frequency of the traffic safety events of the category i.
The above description is only an embodiment of the present invention, and is not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes performed by the present invention or directly or indirectly applied to other related technical fields are included in the scope of the present invention.

Claims (9)

1. A personalized differential driving risk evaluation method based on driving safety events is characterized in that:
acquiring driving behavior data: collecting real-time driving behavior data of a driver by using a sensor; the driving behavior data includes: the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data, the right steering lamp data, the steering wheel angle and the steering wheel angle acceleration;
judging real-time driving behavior data by using a risk evaluation model to obtain a driving risk evaluation score;
the risk evaluation model is obtained by the following steps:
step 1, extracting driving safety events, selecting historical driving behavior data of other drivers within a certain time to perform data preprocessing, and extracting the driving safety events by taking m% upper bound of limit operation of the drivers as an extraction threshold value of the driving safety events by using a probability density function of a judgment index of each driving safety event to obtain the frequency of the driving safety events, wherein m is more than 0 and less than 100; extracting equivalent pairing samples from a database of normal driving states of a driver according to the extracted traffic safety event sample capacity; the database of the normal driving state of the driver is a database containing normal driving behavior data of other drivers;
and 2, performing index statistical test, namely performing significance pairing comparison on the driving behavior data of the driver in each type of driving safety event and the driving behavior data in the normal driving state, and dividing the driving behavior indexes into three types according to the obtained result: no statistical difference, statistical difference and significant statistical difference;
step 3, screening indexes, namely inputting the results with significant statistical differences into a machine learning clustering algorithm for classification, and obtaining risk evaluation indexes corresponding to the results by using a statistical method;
and 4, risk evaluation modeling, namely sequencing results from low to high according to the variation degree of the risk evaluation indexes in corresponding events by utilizing the calculation results of the clustering algorithm, converting the risk tendency grades of the individual drivers in different events to obtain the driving risk weight of various driving safety events to the drivers, and establishing the risk evaluation model.
2. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: acquiring the driving behavior data by adopting a vehicle-mounted CAN and a steering wheel corner sensor; the vehicle-mounted CAN acquires the vehicle speed value, the acceleration, the brake pedal data, the left steering lamp data and the right steering lamp data; and the steering wheel angle acceleration are acquired by a steering wheel angle sensor.
3. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: the driving safety events are specifically divided into four types, namely emergency braking, emergency turning, emergency refueling door and overspeed.
4. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: the driving safety event extraction comprises data preprocessing, and specifically comprises the following steps: firstly, synchronizing data of different sensors, and then carrying out wavelet packet noise reduction treatment on the following indexes: the speed average value, the speed standard deviation, the speed maximum value, the speed range difference, the acceleration average value, the deceleration average value, the acceleration standard deviation, the acceleration maximum value, the deceleration maximum value, the steering wheel right corner average value, the steering wheel left corner average value, the steering wheel corner standard deviation, the steering wheel corner maximum value, the steering wheel right corner acceleration average value, the steering wheel left corner acceleration average value, the steering wheel corner acceleration standard deviation and the steering wheel corner acceleration maximum value.
5. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: the driving safety event extraction comprises the calibration of the driving safety event, wherein the judgment index of overspeed is the maximum speed, the judgment index of rapid acceleration is the maximum acceleration, the judgment index of rapid deceleration is the maximum deceleration, and the judgment index of rapid turning is the maximum steering wheel corner acceleration; and taking the upper m% of the limit operation of the driver as an event extraction threshold value to obtain the frequency of the driving safety events.
6. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: the specific method for extracting equivalent pairing samples from the database of the normal driving state of the driver comprises the following steps: generating a pseudorandom integer by using MATLAB, and reserving the pseudorandom function seed to keep the selected observation value and the sequence thereof unchanged;
setting a pseudo random number generation seed:
rng(seed)
pseudo-random number decimation paired samples Xpair：
Xpair＝randi([imin,imax],m,n)
Where imin is the minimum value of the overall sample index, imax is the maximum value of the overall sample index,
m and n denote the extracted index values and the rows and columns of the composed matrix, respectively.
7. The individualized and differentiated driving risk evaluation method based on driving safety events according to claim 1, characterized in that: the three driving behavior indexes obtained by the index statistical test are specifically as follows: no statistical difference exists, and h is equal to 0; the statistical difference exists, h is equal to 1, and p is less than 0.05 and more than 0.01; the statistical difference is significant, h is equal to 1, and p is less than 0.01; where h is the output term, with two results, 0 and 1; h-0 indicates that the null hypothesis is negated at 5% confidence, i.e., the two sets of paired samples are statistically considered to be from the same distribution; h-1 indicates that the null hypothesis is negated, and the two sets of matched samples are statistically considered to be data from different distributions; p represents the set criterion of significant difference, p <0.05 considered the difference, and p <0.01 considered the significant difference.
8. The individualized and differentiated driving risk assessment based on driving safety events according to claim 1
A method of pricing characterized by: in the risk evaluation modeling, a specific conversion formula of the driving risk weight is as follows:
Figure FDA0002871887270000031
wherein, wijRepresenting the driving risk weight of the driver j in the i-type driving safety event; dijRepresenting a risk evaluation index of the driver j in the i-type safety event; and n is the total number of the driving safety events.
9. The driving safety event-based individualized and differentiated driving risk assessment according to claim 1
A method of pricing characterized by: the specific acquisition method of the driving risk assessment score comprises the following steps:
Figure FDA0002871887270000032
wherein G represents a driving risk assessment score, i and n represent the category and total number of driving safety events, respectively, and γiA driving risk weight, x, representing a category iiAnd representing the frequency of the traffic safety events of the category i.
CN202011603122.1A 2020-12-30 2020-12-30 Individualized and differentiated driving risk evaluation method based on driving safety event Active CN112613786B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011603122.1A CN112613786B (en) 2020-12-30 2020-12-30 Individualized and differentiated driving risk evaluation method based on driving safety event

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011603122.1A CN112613786B (en) 2020-12-30 2020-12-30 Individualized and differentiated driving risk evaluation method based on driving safety event

Publications (2)

Publication Number Publication Date
CN112613786A CN112613786A (en) 2021-04-06
CN112613786B true CN112613786B (en) 2022-05-06

Family

ID=75249333

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011603122.1A Active CN112613786B (en) 2020-12-30 2020-12-30 Individualized and differentiated driving risk evaluation method based on driving safety event

Country Status (1)

Country Link
CN (1) CN112613786B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113221843A (en) * 2021-06-07 2021-08-06 吉林大学 Driving style classification method based on empirical mode decomposition characteristics
CN113689071B (en) * 2021-07-15 2022-11-15 东南大学 Multi-driver risk evaluation-based active intervention method for bad driving behaviors
CN113673826B (en) * 2021-07-20 2023-06-02 中国科学技术大学先进技术研究院 Driving risk assessment method and system based on individual factors of driver
CN113657555B (en) * 2021-09-03 2024-05-07 燕山大学 Improved semi-supervised clustering-based ice and snow environment driving experience evaluation method
CN114291072A (en) * 2021-12-17 2022-04-08 北京中科坤润科技有限公司 Carsickness prevention device and carsickness prevention method

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107909678A (en) * 2017-11-29 2018-04-13 思建科技有限公司 One kind driving risk evaluating method and system
CN110866677A (en) * 2019-10-25 2020-03-06 东南大学 Driver relative risk evaluation method based on benchmark analysis
CN111081374A (en) * 2019-12-16 2020-04-28 华南师范大学 Autism auxiliary diagnosis device based on common attention paradigm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7389178B2 (en) * 2003-12-11 2008-06-17 Greenroad Driving Technologies Ltd. System and method for vehicle driver behavior analysis and evaluation

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107909678A (en) * 2017-11-29 2018-04-13 思建科技有限公司 One kind driving risk evaluating method and system
CN110866677A (en) * 2019-10-25 2020-03-06 东南大学 Driver relative risk evaluation method based on benchmark analysis
CN111081374A (en) * 2019-12-16 2020-04-28 华南师范大学 Autism auxiliary diagnosis device based on common attention paradigm

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Mileage Traveled, Driving Time, and Speeding Behavior as Predictors for Hazmat Transportation Risk Assessment Using Naturalistic Driving Data;Siyao Li;《IEEE》;20191231;全文 *
老年长期照护分级综合评价模型初步构建研究;师亚;《中国全科医学》;20181231;全文 *
驾驶行为安全性多属性评价方法及应用研究;李扬;《中国优秀硕士学位论文全文数据库》;20170331;全文 *

Also Published As

Publication number Publication date
CN112613786A (en) 2021-04-06

Similar Documents

Publication Publication Date Title
CN112613786B (en) Individualized and differentiated driving risk evaluation method based on driving safety event
CN110796859A (en) Real-time traffic state identification and accident risk early warning method based on traffic flow
CN108469806B (en) Driving right transfer method in alternating type man-machine common driving
CN101806229B (en) Mine roof safety evaluation method based on multi-target monitoring data fusion
Shi et al. Real-time driving risk assessment using deep learning with XGBoost
CN110588658B (en) Method for detecting risk level of driver based on comprehensive model
CN101694744A (en) Method and system for evaluating road emergency evacuation capacity and method and system for grading road emergency evacuation capacity
CN108769104B (en) Road condition analysis and early warning method based on vehicle-mounted diagnosis system data
CN104809255A (en) Load shape acquisition method and system
CN110781873A (en) Driver fatigue grade identification method based on bimodal feature fusion
CN111242484A (en) Vehicle risk comprehensive evaluation method based on transition probability
CN111401414B (en) Natural driving data-based dangerous scene extraction and classification method
CN111563555A (en) Driver driving behavior analysis method and system
CN110562261B (en) Method for detecting risk level of driver based on Markov model
CN110901385B (en) Active speed limiting method based on fatigue state of driver
CN110781872A (en) Driver fatigue grade recognition system with bimodal feature fusion
CN110858312A (en) Driver driving style classification method based on fuzzy C-means clustering algorithm
CN108229304A (en) A kind of driving behavior recognition methods based on Clustering of systematization
CN105809193A (en) Illegal operation vehicle recognition method based on Kmeans algorithm
CN114970943A (en) Expressway accident type and severity prediction method
CN113581188A (en) Commercial vehicle driver driving style identification method based on Internet of vehicles data
CN114037015A (en) Driving behavior safety analysis method based on driving mode primitive
CN115774942A (en) Driving style identification model modeling and statistical method based on Internet of vehicles real vehicle data and SVM
CN112927511A (en) Vehicle early warning method based on driver age and gender identification
CN112215044A (en) Driving tendency identification method based on probabilistic neural network

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
GR01 Patent grant
GR01 Patent grant