CN112937592A - Method and system for identifying driving style based on headway - Google Patents

Method and system for identifying driving style based on headway Download PDF

Info

Publication number
CN112937592A
CN112937592A CN202110308239.5A CN202110308239A CN112937592A CN 112937592 A CN112937592 A CN 112937592A CN 202110308239 A CN202110308239 A CN 202110308239A CN 112937592 A CN112937592 A CN 112937592A
Authority
CN
China
Prior art keywords
headway
driving
drivers
driving style
identifying
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.)
Granted
Application number
CN202110308239.5A
Other languages
Chinese (zh)
Other versions
CN112937592B (en
Inventor
吕能超
高谨谨
王玉刚
吴超仲
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 CN202110308239.5A priority Critical patent/CN112937592B/en
Publication of CN112937592A publication Critical patent/CN112937592A/en
Application granted granted Critical
Publication of CN112937592B publication Critical patent/CN112937592B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/08—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to drivers or passengers
    • B60W40/09—Driving style or behaviour
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • B60W2050/0001—Details of the control system
    • B60W2050/0043—Signal treatments, identification of variables or parameters, parameter estimation or state estimation

Landscapes

  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Human Computer Interaction (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method for identifying a driving style based on headway, which comprises the following steps: s1, collecting driving videos and headway data of a plurality of drivers in natural driving environment, and classifying the driving styles of all the drivers; s2, carrying out segmented clustering on the headway, and converting the headway into codes according to the preset grade range to which the headway value belongs; s3, detecting a car following mode by using a sliding time window method; s4, dividing the typical mode into different car following states according to the distribution situation of the headway time in the typical mode, and sequentially giving different scores; s5, randomly selecting the headway data of a part of drivers as a model training set, and obtaining the threshold value of each driving style by combining the known driving style results; s6, taking the data of the other drivers as a test set, and identifying the driving style of the drivers in the test set; and S7, comparing and cross-verifying, and judging the accuracy of result identification.

Description

Method and system for identifying driving style based on headway
Technical Field
The invention relates to the field of driving behavior analysis, in particular to a method and a system for identifying a driving style based on headway.
Background
With the rapid increase of the holding capacity of motor vehicles, the form of road traffic safety is more severe, and drivers are the most unstable main factors in a 'person-vehicle-road' closed loop system, and are often the main causes of traffic accidents due to the reasons of poor observation of driving environment, decision errors, misoperation and the like.
The driving style is an integral index representing the inherent driving mode of a driver, researches show that the style of the driver is closely related to traffic safety, the driver is generally classified into a cautious type, a normal type and an aggressive type according to the driving behavior of the driver in China, and the safety of the aggressive type driver is relatively low in the driving process. The intelligent network connection field provides the driver with an auxiliary driving mode according with the driving characteristics of the driver according to different driving styles of the driver, and improves the safety and humanization of automobile service.
Summarizing, the current classification method for the driving style of the driver can be roughly divided into two categories, one is a subjective questionnaire method, namely, the driving style of the driver is subjectively judged by filling in the questionnaire by the driver; the other is an objective driving data evaluation method, namely, the driving style of the driver is objectively judged by analyzing the driving behavior data of the driver; and in the objective evaluation, a clustering algorithm is mostly adopted to perform clustering analysis on a plurality of index data of the driver, so that the driving style of the driver is obtained. The existing driving style classification method is too simple, and has the problems of unreasonable classification, low recognition accuracy and the like, so that the precision of driving style recognition is greatly reduced.
Disclosure of Invention
The invention mainly aims to provide a method and a system for identifying a following mode by using a headway so as to classify driving styles, which can improve the accuracy of driving style identification.
The technical scheme adopted by the invention is as follows:
the method for identifying the driving style based on the headway is characterized by comprising the following specific steps of:
s1, collecting driving videos and headway data of a plurality of drivers in the natural driving environment, and classifying the driving styles of all the drivers by adopting a three-component scale method;
s2, carrying out segmented clustering on the headway, and converting the headway into codes according to the preset grade range to which the headway value belongs;
s3, detecting a following mode by using a sliding time window method, wherein the code of each m continuous time lengths is one following mode, screening out the following modes which frequently appear by drivers with different driving styles by combining the classification results of the driving styles, and taking the following modes as typical modes corresponding to each driving style, wherein m is a natural number;
s4, dividing the typical mode into different car following states according to the distribution situation of the headway time in the typical mode, and sequentially giving different scores;
s5, randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style results;
s6, taking the data of the other drivers as a test set, calculating the scores of the drivers, and identifying the driving styles of the drivers in the test set according to the driving style threshold set in the step S5;
and S7, comparing the result identified in the step S6 with the divided driving style result, performing cross validation, and judging the accuracy of result identification.
According to the technical scheme, the typical mode is divided into 5 types of following states of aggressive type, common type, conservative type and conservative type.
According to the technical scheme, scores of 5, 3, 0, -3 and-5 are sequentially given to the five types of car following states.
According to the technical scheme, the preset grade range is divided into [0, t ] according to the distribution condition of the distance between the vehicle heads from small to large1)，[t1,t2)，[t2,t3)，…，[tn-1,tn]N levels in total, and setting codes as 1, 2, 3, 4, … in sequence, wherein n is a natural number.
In the above technical solution, step S2 specifically includes:
converting the original data into a segmentation and aggregation approximate representation form, wherein the frame duration of the segmentation and aggregation approximation is t seconds, and the average value of the t seconds represents the headway value in the t seconds; and converting the segmented aggregation approximate data into character strings according to the grade of the headway and the grade coding rule.
Following the above technical solution, in step S3, if each m code combinations is identified as a driving mode and n levels are total, n is generatedmAnd (5) a car following mode is adopted.
In step S5, the Score of the driver in the training set is calculated as:
Figure BDA0002988511880000031
in the formula: peScore, time in mode e as a percentage of total driving time for the driver's car following modeeAnd the score is the score corresponding to the car following mode.
In connection with the above technical solution, the determination of the threshold of each driving style in step S5 is specifically: the maximum value of the driver score of each driving style is set as the maximum value of the driving style, and the minimum value of the score is set as the minimum value of the driving style.
In step S1, the driving styles of all drivers are classified by a three-point scale method in a manner that three experts watch videos.
The invention also provides a system for identifying the driving style based on the headway, which comprises the following steps:
the data acquisition module is used for acquiring driving videos and headway data of a plurality of drivers in natural driving environments;
the classification module is used for classifying the driving styles of all drivers by adopting a three-component scale method;
the code conversion module is used for carrying out segmented clustering on the headway and converting the headway into codes according to the preset grade range to which the headway value belongs;
the following mode determining module is used for detecting a following mode by using a sliding time window method, codes of every m continuous time lengths are one following mode, frequent following modes of drivers with different driving styles are screened out by combining the classification results of the driving styles and are used as typical modes corresponding to each driving style, wherein m is a natural number;
the assignment module is used for dividing the typical mode into different car following states according to the distribution situation of the head time distances in the typical mode and sequentially giving different scores;
the model training module is used for randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style result;
the model testing module is used for taking the data of the rest drivers as a testing set, calculating the scores of the drivers and identifying the driving styles of the drivers in the testing set according to a certain driving style threshold value;
and the cross validation module is used for comparing the result identified by the model test module with the divided driving style result for cross validation and judging the accuracy of result identification.
The invention has the following beneficial effects: the method classifies the driving style of the driver by utilizing the distribution characteristics of the headway, and identifies the driving style by analyzing the headway of the driver and adopting a quantitative method, so that the method is simple and has strong operability; the classification condition of the locomotive time interval is adjusted according to different locomotive interval distribution conditions in different environments, and the method is flexible and high in applicability. The data types collected by some current radars or vehicle-mounted data collection equipment are limited, only data with less types such as speed and headway can be collected, the driving style can be classified by using the headway, and the method is easy to realize and has stronger operability.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method for identifying driving style based on headway in an embodiment of the present invention;
FIG. 2 is a flow chart of a method of identifying driving style based on headway in accordance with another embodiment of the present invention;
fig. 3 is a schematic structural diagram of a system for identifying a driving style based on headway in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, the method for identifying a driving style based on headway in the embodiment of the present invention includes the following steps:
s1, collecting driving videos and headway data of a plurality of drivers in the natural driving environment, and classifying the driving styles of all the drivers by adopting a three-component scale method;
s2, carrying out segmented clustering on the headway, and converting the headway into codes according to the preset grade range to which the headway value belongs;
s3, detecting a following mode by using a sliding time window method, wherein the code of each m continuous time lengths is one following mode, screening out the following modes which frequently appear by drivers with different driving styles by combining the classification results of the driving styles, and taking the following modes as typical modes corresponding to each driving style, wherein m is a natural number;
s4, dividing the typical mode into different car following states according to the distribution situation of the headway time in the typical mode, and sequentially giving different scores;
s5, randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style results;
s6, taking the data of the other drivers as a test set, calculating the scores of the drivers, and identifying the driving styles of the drivers in the test set according to the driving style threshold set in the step S5;
and S7, comparing the result identified in the step S6 with the divided driving style result, performing cross validation, and judging the accuracy of result identification.
Another embodiment of the present invention provides a driving style classification method for identifying a following mode based on headway, which specifically includes the following steps:
s1, collecting driving videos and natural driving data (including headway data) of a driver in a natural driving environment by using a laser radar and a camera device, and dividing the driving style of the driver by a method of watching the videos by three experts in the embodiment;
s2, processing the headway data of the driver by using a segmented clustering approximate method, if the segmented clustering time length is t seconds, calculating the headway average value in each t second, representing the headway in the t seconds by the average value, and then converting the data into codes 1, 2, 3, …, n corresponding to the grades of the headway data; sequentially dividing the locomotive headway into n levels from small to large according to the distribution condition of the headway, and sequentially setting codes as 1, 2, 3, …, n;
s3, detecting the driving modes by using a sliding time window method, identifying the code of each m continuous time lengths as a following mode, screening out the following mode which occurs more frequently in each driving style by combining the driving style results divided by experts, and taking the following mode as a typical mode corresponding to the driving style;
s4, dividing the typical mode into 5 types of following states such as aggressive, common, conservative and conservative according to the headway distribution condition of the driver in the typical mode, and sequentially giving scores of 5, 3, 0, -3 and-5;
s5, randomly selecting a part of drivers as a model training set, calculating scores of the drivers according to the scores and time percentages corresponding to the 5 following states in the S4 of each driver, and combining the scores of the drivers in the training set with the corresponding driving styles to obtain the threshold values of the driving styles;
s6, using the rest drivers as a model test set, namely analyzing the data of the drivers in the test set by using the method to obtain the scores of the drivers, and identifying the driving styles of the drivers in the test set according to the set driving style thresholds;
and S7, comparing and cross-verifying the driving styles of the drivers in the test set obtained in the S6 by taking the result of subjective division of experts as a standard, and judging the accuracy of result identification.
Further, the specific method of step S1 of the present invention is:
the method comprises the steps of collecting vehicle head time distance data and driving videos of a certain number of drivers in a natural driving environment by using a laser radar and a video collecting device, watching the videos by three experts, scoring the driving style of each driver by using a three-component scale (1 is conservative, 2 is common, and 3 is aggressive), and obtaining the driving style of each driver, wherein the scoring rule is as follows:
Figure BDA0002988511880000071
in the formula: eAA score for the first expert, EBA score for a second expert, ECA score value for the third driver.
Further, the specific method of step S2 of the present invention is:
making frequency distribution map of headway data of all drivers, and dividing headway into n grades from small to large according to the distribution condition of headway, namely [0, t1)，[t1,t2)，[t2,t3)，…，[tn-1,tn]One for each level, and the headway time belonging to the range of each level is set to the code 1, 2, 3, 4, …, n in turn.
The method comprises the steps of performing symbol aggregation approximate processing on original data of the headway of a driver, converting the original data into a segmentation aggregation approximate representation form, namely averaging the acquired original data every t seconds, representing the headway within t seconds by the average value, and converting the segmentation approximate data into character strings according to headway grade division conditions and coding rules, namely converting the character strings into codes 1, 2, 3, 4, …, n.
Further, the specific method of step S3 of the present invention is:
detecting the following mode of each driver by using a sliding time window method, wherein the code of each m continuous times is marked as one following mode, and the following modes are divided into n grades according to the time interval of the vehicle head to obtain n gradesmA vehicle following mode is planted; when the typical driving mode is selected, the same codes of drivers of each driving style are automatically divided into one group, namely, each following mode is divided into one group, the frequency of each mode is calculated, the driving mode with frequent occurrence of the driver of each driving style is screened out according to the known driving style of the driver, and then the previous e following modes with high occurrence frequency of each driving style are taken as the typical modes of the driving style, so that 3e typical modes are selected from the 3 driving styles.
Further, the specific method of step S4 of the present invention is:
dividing the typical mode into 5 types of following states such as aggressive, common, conservative and conservative according to the head time distance distribution condition of the typical mode, and sequentially giving 5 types of following states with scores of 5, 3, 0, -3 and-5;
further, the specific method of step S5 of the present invention is:
calculating the score of each driver according to the score corresponding to the 5 types of following states of each driver in the test set and the percentage of the total time, namely:
Figure BDA0002988511880000091
in the formula:
score is the driver's Score value;
Pethe percentage of the time that the following mode of the driver is the mode e to the total driving time;
scoreeis the score of the car following mode e.
Randomly selecting test data of a part of drivers as a model training set, and setting a threshold value of a driving style according to the score of the drivers in the training set and the known driving style, namely taking the maximum value of the score of each driving style subjectively evaluated by an expert as the maximum value of the driving style, and taking the minimum value of the score as the minimum value of the driving style; when a fault occurs, the fault part is averagely divided into two critical driving style thresholds, and when an overlapping phenomenon occurs, the driving style with high overlapping range frequency is extracted.
Further, the specific method of step S6 of the present invention is:
and (4) testing the models, namely analyzing the headway data of the drivers in the test set by using the same method, calculating the score of the drivers in the test set, and identifying the driving styles of the drivers in the test set according to the threshold values of the driving styles set in the S5.
Further, the specific method of step S7 of the present invention is:
and (3) verifying the driving style accuracy of the driver in the test set identified in the S7 by taking the subjective division result of the expert as a standard, wherein the accuracy P is as follows:
Figure BDA0002988511880000092
in the formula:
p is the driving style identification accuracy;
a is the number of drivers in the test set;
aithe number of the drivers with the recognition results consistent with the subjective division results of the experts in the a drivers is obtained.
In a third embodiment of the present invention, as shown in fig. 2, a method for identifying a driving style based on headway includes:
s1, collecting driving data (including headway) and driving videos of 44 drivers in a natural driving state by using devices such as a laser radar and a camera, and classifying the driving styles of the 44 drivers by three experts through a video watching method;
it should be noted that the specific method for the expert to divide the driving style of the driver is to score the driving style by watching the driving video and adopting a three-component scale, and the scoring rule is the following formula:
Figure BDA0002988511880000101
in the formula: eAA score for the first expert, EBA score for a second expert, ECA score value for the third driver.
S2, dividing the frequency distribution of the headway time of the 44 drivers into 6 grades according to the following range, namely 6 grades including 0-0.6S, 0.6S-1.0S, 1.0S-1.5S, 1.5S-2.0S, 2.0S-2.5S and 2.5S- ∞; the grade codes are sequentially set to be 1, 2, 3, 4, 5 and 6;
it should be noted that the grade division is divided according to the distribution of headway, and is not a fixed division principle, and the grade division is not necessarily equal-distance division;
s3, preprocessing the original data of the headway, namely, carrying out sectional clustering processing on the headway with the acquisition frequency of 10HZ by using a sectional clustering algorithm, wherein the sectional time length is 1S, and then converting the data per second into codes;
it should be noted that, when preprocessing the headway, firstly, the original data is converted into the simplified code by using a symbolic aggregation approximation method, firstly, the original data is converted into a segmentation aggregation approximation representation form, that is, headway data within 1S is averaged, and the headway value within 1S is represented by an average value, and then, the segmentation clustering data is converted into a character string, that is, the code in S2.
S4, detecting the following mode of the driver, wherein the code generated every 3S is one following mode, namely, a 3S sliding time window is taken to detect the following mode of each driver, the following mode with higher occurrence frequency of the driver of each driving style is screened out by combining with the known driving style, and the following mode with higher occurrence frequency of each driving style is taken as the typical mode of the corresponding driving style;
it should be noted that every 3s is a driving mode, and each mode has 6 levels, so that 6 levels are generated in total3216 driving modes. When the typical mode of each driving style is selected, the accumulated frequency of the following modes of each driver is detected and calculated, then the accumulated frequency of the following modes of each driving style driver is obtained according to the driving style classification result divided by experts, then the first 10 driving modes with high accumulated frequency of each driving style are selected as the typical following modes, and 30 typical modes are selected in total.
And S5, dividing the typical mode into 5 types of following states such as aggressive, common, conservative and conservative according to the headway distribution condition of the typical mode, and sequentially giving scores of 5, 3, 0, -3 and-5 to serve as the basis for scoring the driver in the next step.
S6, randomly selecting test data of 30 drivers as a model training set, scoring the drivers according to scores and time percentages corresponding to the 5 following states of each driver, and setting thresholds of all driving styles according to the scoring conditions of the drivers in the training set and the subjective division results of experts;
it should be noted that, the score calculation formula of the driver is:
Figure BDA0002988511880000111
in the formula:
score is the driver's Score value;
Pithe time of the following mode of the driver is the percentage of the total driving time of the mode i;
scoreiis the score of the car following mode i.
When setting each driving style threshold value, the highest score of the driver of each driving style divided by expert subjective evaluation is used as the maximum value of the driving style, and the lowest score is used as the minimum value of the driving style.
S7, testing the extracted models, namely analyzing the test data of the rest 14 drivers by using the same method and calculating to obtain scores, and then identifying the driving styles of the drivers in the test set according to the driving style thresholds set in S6;
s9, verifying the extracted model, namely comparing the driving style of the driver in the test set identified in S7 with the result of expert division in S1, and obtaining the identification accuracy of the extracted method by taking the result of expert division as a standard;
Figure BDA0002988511880000121
it should be noted that the value range of P is: p is more than or equal to 0 and less than or equal to 1, aiThe value range is as follows: b is not less than 0i≤14。
As shown in fig. 3, the system for identifying a driving style based on headway in the embodiment of the present invention includes:
the data acquisition module is used for acquiring driving videos and headway data of a plurality of drivers in natural driving environments;
the classification module is used for classifying the driving styles of all drivers by adopting a three-component scale method;
the code conversion module is used for carrying out segmented clustering on the headway and converting the headway into codes according to the preset grade range to which the headway value belongs;
the following mode determining module is used for detecting a following mode by using a sliding time window method, codes of every m continuous time lengths are one following mode, frequent following modes of drivers with different driving styles are screened out by combining the classification results of the driving styles and are used as typical modes corresponding to each driving style, wherein m is a natural number;
the assignment module is used for dividing the typical mode into different car following states according to the distribution situation of the head time distances in the typical mode and sequentially giving different scores;
the model training module is used for randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style result;
the model testing module is used for taking the data of the rest drivers as a testing set, calculating the scores of the drivers and identifying the driving styles of the drivers in the testing set according to a certain driving style threshold value;
and the cross validation module is used for comparing the result identified by the model test module with the divided driving style result for cross validation and judging the accuracy of result identification.
In conclusion, the natural driving data of a part of drivers are randomly selected to serve as a model training set, and the rest are taken as a test set; the method comprises the steps of screening out typical modes with high occurrence frequency of drivers of all driving styles by analyzing the following modes of drivers of different styles in a training set, dividing the typical modes into different following modes and giving corresponding scores, then scoring the drivers according to the scores corresponding to the typical modes appearing by the drivers and time percentages, setting the threshold value of each driving style by combining the scores of the drivers of all driving styles in the training set and the known driving style types, and identifying the driving style of the drivers serving as a test set according to the set threshold value. The driving style is quantitatively divided by utilizing a single headway index, the data acquisition aspect is strong in implementability, and the accuracy of the method can be verified through an example.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims (10)

1. A method for identifying a driving style based on headway is characterized by comprising the following specific steps:
s1, collecting driving videos and headway data of a plurality of drivers in the natural driving environment, and classifying the driving styles of all the drivers by adopting a three-component scale method;
s2, carrying out segmented clustering on the headway, and converting the headway into codes according to the preset grade range to which the headway value belongs;
s3, detecting a following mode by using a sliding time window method, wherein the code of each m continuous time lengths is one following mode, screening out the following modes which frequently appear by drivers with different driving styles by combining the classification results of the driving styles, and taking the following modes as typical modes corresponding to each driving style, wherein m is a natural number;
s4, dividing the typical mode into different car following states according to the distribution situation of the headway time in the typical mode, and sequentially giving different scores;
s5, randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style results;
s6, taking the data of the other drivers as a test set, calculating the scores of the drivers, and identifying the driving styles of the drivers in the test set according to the driving style threshold set in the step S5;
and S7, comparing the result identified in the step S6 with the divided driving style result, performing cross validation, and judging the accuracy of result identification.
2. The method for identifying the driving style based on headway as claimed in claim 1, wherein the typical pattern is divided into 5 types of following states of aggressive, more aggressive, common, more conservative, and conservative.
3. The method for identifying the driving style based on the headway distance as claimed in claim 2, wherein the scores of 5, 3, 0, -3 and-5 are sequentially given to the five types of following states.
4. According to claim 1The method for identifying the driving style based on the headway is characterized in that the preset grade range is divided into [0, t ] according to the headway distribution condition and from small to large1)，[t1,t2)，[t2,t3)，…，[tn-1,tn]N levels in total, and setting codes as 1, 2, 3, 4, … in sequence, wherein n is a natural number.
5. The method for identifying the driving style based on the headway distance as recited in claim 4, wherein the step S2 specifically comprises:
converting the original data into a segmentation and aggregation approximate representation form, wherein the frame duration of the segmentation and aggregation approximation is t seconds, and the average value of the t seconds represents the headway value in the t seconds; and converting the segmented aggregation approximate data into character strings according to the grade of the headway and the grade coding rule.
6. The method for identifying driving style based on headway as claimed in claim 4, wherein in step S3, every m code combinations are identified as a driving mode, and n levels are generatedmAnd (5) a car following mode is adopted.
7. The method for identifying driving style based on headway as claimed in claim 1, wherein in step S5, the Score of the driver in the training set is calculated by the formula:
Figure FDA0002988511870000021
in the formula: peScore, time in mode e as a percentage of total driving time for the driver's car following modeeAnd the score is the score corresponding to the car following mode.
8. The method for identifying the driving style based on the headway distance as recited in claim 1, wherein the threshold determination of each driving style in the step S5 is specifically as follows: the maximum value of the driver score of each driving style is set as the maximum value of the driving style, and the minimum value of the score is set as the minimum value of the driving style.
9. The method for identifying driving styles based on headway as claimed in claim 1, wherein in step S1, the driving styles of all drivers are classified by a three-point scale method in a manner that three experts watch videos.
10. The utility model provides a system for discerning driving style based on headway which characterized in that includes:
the data acquisition module is used for acquiring driving videos and headway data of a plurality of drivers in natural driving environments;
the classification module is used for classifying the driving styles of all drivers by adopting a three-component scale method;
the code conversion module is used for carrying out segmented clustering on the headway and converting the headway into codes according to the preset grade range to which the headway value belongs;
the following mode determining module is used for detecting a following mode by using a sliding time window method, codes of every m continuous time lengths are one following mode, frequent following modes of drivers with different driving styles are screened out by combining the classification results of the driving styles and are used as typical modes corresponding to each driving style, wherein m is a natural number;
the assignment module is used for dividing the typical mode into different car following states according to the distribution situation of the head time distances in the typical mode and sequentially giving different scores;
the model training module is used for randomly selecting the headway data of a part of drivers as a model training set, calculating the score of each driver according to the scores of different following states of each driver in the model training set and the corresponding time percentage, and obtaining the threshold value of each driving style by combining the known driving style result;
the model testing module is used for taking the data of the rest drivers as a testing set, calculating the scores of the drivers and identifying the driving styles of the drivers in the testing set according to a certain driving style threshold value;
and the cross validation module is used for comparing the result identified by the model test module with the divided driving style result for cross validation and judging the accuracy of result identification.
CN202110308239.5A 2021-03-23 2021-03-23 Method and system for identifying driving style based on headway Active CN112937592B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110308239.5A CN112937592B (en) 2021-03-23 2021-03-23 Method and system for identifying driving style based on headway

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110308239.5A CN112937592B (en) 2021-03-23 2021-03-23 Method and system for identifying driving style based on headway

Publications (2)

Publication Number Publication Date
CN112937592A true CN112937592A (en) 2021-06-11
CN112937592B CN112937592B (en) 2022-04-15

Family

ID=76228226

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110308239.5A Active CN112937592B (en) 2021-03-23 2021-03-23 Method and system for identifying driving style based on headway

Country Status (1)

Country Link
CN (1) CN112937592B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109263639A (en) * 2018-08-24 2019-01-25 武汉理工大学 Driving path planing method based on state Grid Method
CN109927725A (en) * 2019-01-28 2019-06-25 吉林大学 A kind of self-adaption cruise system and implementation method with driving style learning ability
CN110516746A (en) * 2019-08-29 2019-11-29 吉林大学 A kind of driver's follow the bus behavior genre classification method based on no label data
CN110688729A (en) * 2019-08-26 2020-01-14 南京航空航天大学 LSTM-IDM (least squares inverse discrete cosine transform) -following characteristic fusion method based on adaptive Kalman filtering, storage medium and equipment
CN110858312A (en) * 2018-08-23 2020-03-03 重庆大学 Driver driving style classification method based on fuzzy C-means clustering algorithm
CN111994084A (en) * 2020-09-21 2020-11-27 华南理工大学 Method and system for identifying driving style of driver and storage medium

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110858312A (en) * 2018-08-23 2020-03-03 重庆大学 Driver driving style classification method based on fuzzy C-means clustering algorithm
CN109263639A (en) * 2018-08-24 2019-01-25 武汉理工大学 Driving path planing method based on state Grid Method
CN109927725A (en) * 2019-01-28 2019-06-25 吉林大学 A kind of self-adaption cruise system and implementation method with driving style learning ability
CN110688729A (en) * 2019-08-26 2020-01-14 南京航空航天大学 LSTM-IDM (least squares inverse discrete cosine transform) -following characteristic fusion method based on adaptive Kalman filtering, storage medium and equipment
CN110516746A (en) * 2019-08-29 2019-11-29 吉林大学 A kind of driver's follow the bus behavior genre classification method based on no label data
CN111994084A (en) * 2020-09-21 2020-11-27 华南理工大学 Method and system for identifying driving style of driver and storage medium

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
侯海晶: "高速公路驾驶人换道意图识别方法研究", 《中国博士学位论文全文数据库》 *

Also Published As

Publication number Publication date
CN112937592B (en) 2022-04-15

Similar Documents

Publication Publication Date Title
WO2020244288A1 (en) Method and apparatus for evaluating truck driving behaviour based on gps trajectory data
CN113486822B (en) Surrounding vehicle track prediction method and system based on driving intention
CN110188797B (en) Intelligent automobile rapid test method based on Bayesian optimization
CN110588658B (en) Method for detecting risk level of driver based on comprehensive model
CN110562261B (en) Method for detecting risk level of driver based on Markov model
CN109635852B (en) User portrait construction and clustering method based on multi-dimensional attributes
CN103455820A (en) Method and system for detecting and tracking vehicle based on machine vision technology
CN109191828B (en) Traffic participant accident risk prediction method based on ensemble learning
CN111931810A (en) Energy-saving driving behavior analysis system based on multiple vehicles
CN111144446B (en) Driver identity recognition method and system based on space-time grid
CN111563555A (en) Driver driving behavior analysis method and system
CN113689594A (en) Novel automobile driving condition construction method
CN109993966A (en) A kind of method and device of building user portrait
CN112767684A (en) Highway traffic jam detection method based on charging data
CN111627132B (en) Driving inclination-based automobile driving characteristic evaluation and early warning method
CN113581188A (en) Commercial vehicle driver driving style identification method based on Internet of vehicles data
CN117021959B (en) Method for acquiring coasting recovery torque based on driving style identification
CN113673304B (en) Vehicle-mounted expected functional safety hazard analysis and evaluation method based on scene semantic driving
CN112937592B (en) Method and system for identifying driving style based on headway
CN110705628B (en) Method for detecting risk level of driver based on hidden Markov model
CN116257651B (en) Intelligent monitoring system for abnormal sound of through channel cab apron
CN114707573A (en) Unsupervised driving style analysis method based on basic driving operation event
CN113591780A (en) Method and system for identifying driving risk of driver
CN113221843A (en) Driving style classification method based on empirical mode decomposition characteristics
CN113505955A (en) User driving behavior scoring method based on TSP system

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