CN115860461A - Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection - Google Patents
Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection Download PDFInfo
- Publication number
- CN115860461A CN115860461A CN202211453450.7A CN202211453450A CN115860461A CN 115860461 A CN115860461 A CN 115860461A CN 202211453450 A CN202211453450 A CN 202211453450A CN 115860461 A CN115860461 A CN 115860461A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- variable
- motor vehicle
- traffic
- variables
- 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.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 15
- 238000000034 method Methods 0.000 claims abstract description 39
- 238000004458 analytical method Methods 0.000 claims abstract description 13
- 230000006399 behavior Effects 0.000 claims description 20
- 238000005259 measurement Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 9
- 230000000717 retained effect Effects 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 238000013459 approach Methods 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 230000000694 effects Effects 0.000 claims description 5
- 238000000546 chi-square test Methods 0.000 claims description 4
- 238000007476 Maximum Likelihood Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 239000011541 reaction mixture Substances 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 238000012353 t test Methods 0.000 claims description 3
- 206010039203 Road traffic accident Diseases 0.000 abstract description 8
- 238000007619 statistical method Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005315 distribution function Methods 0.000 description 2
- 238000009827 uniform distribution Methods 0.000 description 2
- 235000011293 Brassica napus Nutrition 0.000 description 1
- 240000008100 Brassica rapa Species 0.000 description 1
- 235000000540 Brassica rapa subsp rapa Nutrition 0.000 description 1
- 240000002834 Paulownia tomentosa Species 0.000 description 1
- 235000010678 Paulownia tomentosa Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012502 risk assessment Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a method for evaluating risk factors of traffic conflicts of non-motor vehicles at plane intersections, which comprises the following steps: 1. collecting vehicle running track data in a plane intersection by using an unmanned aerial vehicle aerial video, constructing a safety substitute evaluation index with quantified traffic conflict risk, identifying the traffic conflict of the non-motor vehicles, and dividing the traffic conflict into slight conflict and serious conflict according to the severity of the conflict approaching accident; 2. acquiring risk driving behavior data related to non-motor vehicle traffic conflicts by using a road fixed camera which is shot synchronously with the aerial photography of the unmanned aerial vehicle; 3. and establishing a random parameter binomial Logit model with heterogeneity of mean and variance, and quantitatively analyzing the influence of the risk driving behavior of the rider on the traffic conflict severity of the non-motor vehicle. The invention can realize the high-efficiency and objective evaluation of the traffic safety of the non-motor vehicles at the plane intersection so as to improve the accuracy of the analysis of the traffic accident risk factors.
Description
Technical Field
The invention relates to a risk factor evaluation method for a non-motor vehicle traffic conflict under a plane intersection, and belongs to the field of traffic safety risk analysis based on traffic conflicts.
Background
With the rise of our country's shared economy and takeaway culture, the number of non-motor vehicles is growing rapidly. The non-motor vehicles are poor in driving stability and few in safety protection facilities, drivers do not need driving licenses and safety training, driving skills and safety consciousness are weak, dangerous driving behaviors such as violating traffic lights, occupying motor vehicle lanes, reversely driving, speeding and violating manned persons are frequent, and road traffic accidents caused by the non-motor vehicles are increased. Particularly, at plane intersections, the phenomenon of mixed movement of machines and non-machines is prominent, and traffic conflicts and traffic accidents are high. Therefore, it is urgently needed to analyze the traffic safety of non-motor vehicles at the plane intersection.
The traditional method for evaluating the traffic safety risk factors of the non-motor vehicles at the plane intersection mainly comprises a traffic accident statistical analysis method and a questionnaire survey method. The traffic accident statistical analysis method relies on the report data of police, researchers can extract accurate and reliable information, mainly focuses on external factors such as vehicles, roads, environments and the like, and driving behavior information before an accident happens can be rarely obtained from official records. And the defects of insufficient report, after-the-fact evaluation, long observation period and the like exist. Compared with the statistical analysis of traffic accidents, the questionnaire survey method can more comprehensively and deeply analyze the risky driving behaviors and relevant motivational factors thereof. However, the questionnaire survey method adopts a self-reporting mode, the data effectiveness is easily influenced by the memory deviation and social expectation, inaccurate estimation of risk factors is caused, and strong causal statement on the result is difficult to make.
Aiming at the limitations of the method, the evaluation of the traffic safety risk factors of the non-motor vehicles at the plane intersection based on the traffic conflict theory is widely applied. However, the existing method for evaluating the traffic safety risk factor of the non-motor vehicles at the plane intersection based on the traffic conflict theory has the following problems: 1. the analysis of the influence factors of the severity of the traffic conflict of the non-motor vehicles mainly focuses on the factors such as road characteristics and the like, and the systematic exploration on the risk driving behaviors of riders is lacked; 2. the selection of the traffic conflict index does not consider the vehicle motion conditions and conflict characteristics under different conflict angles, so that the traffic conflict risk before and after the conflict vehicle reaches the conflict point is difficult to accurately reflect; 3. the construction of statistical analysis models ignores the heterogeneity problem not observed in the data, and deviations in the estimated parameters can occur, leading to erroneous inferences and predictions.
Disclosure of Invention
The invention aims to overcome the problems in the prior art, and provides a method for evaluating risk factors of non-motor vehicle traffic conflicts at a plane intersection, so that the high-efficiency and objective evaluation on the non-motor vehicle traffic safety at the plane intersection can be realized, and the individual heterogeneity characteristics are fully excavated to improve the accuracy of the analysis of the traffic accident risk factors, thereby improving the traffic safety level of the non-motor vehicles.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention relates to a method for evaluating risk factors of traffic conflicts of non-motor vehicles at plane intersections, which is characterized by comprising the following steps:
step 1, collecting vehicle running track data in a current plane intersection from an unmanned aerial vehicle aerial video, and extracting motion state data of a non-motor vehicle A and a vehicle B from the vehicle running track data, wherein the motion state data comprises the following steps: vehicle position and speed;
step 2, if the current non-motor vehicle A and the vehicle B approach each other on a horizontal straight line, executing step 2.1-step 2.3, and if the current non-motor vehicle A and the vehicle B approach each other in a staggered manner, executing step 2.4-step 2.7;
step 2.1, if the current non-motor vehicle A and the current vehicle B run on the same horizontal straight line in the same direction, calculating the time TTC of collision between the non-motor vehicle A and the current vehicle B at the current time t by using the formula (1) A,B (t)：
In the formula (1), l A,B (t) represents the spatial distance from the front of the rear vehicle B to the rear of the front vehicle A at the current moment t, v B (t) speed of the vehicle of the following vehicle B at the present time t, v A (t) represents the speed of the preceding vehicle a at the current time t;
if the current non-motor vehicle A and the current vehicle B oppositely run on a horizontal straight line, calculating the time TTC of the collision between the non-motor vehicle A and the current vehicle B at the current time t by using an equation (2) A ′ ,B (t)：
L 'in the formula (2)' A,B (t) represents the spatial distance between the heads of two vehicles at the current moment t, v A (t)、v B (t) respectively representing the speeds of the non-motor vehicle A and the vehicle B at the current moment t;
step 2.2, if the speed of the rear vehicle B is greater than that of the front vehicle A at the moment of t +1 or the speed of any one of the non-motor vehicle A and the vehicle B at the moment of t +1 is not 0 in the opposite driving process in the same-direction driving process, indicating that the two vehicles are expected to collide, assigning t +1 to t, returning to the step 2.1 to execute, otherwise, indicating that the two vehicles are not expected to collide, and selecting from all calculated collision time of the non-motor vehicle A and the vehicle BObtaining the minimum time to collide (minTTC) by taking the minimum value A,B ；
Step 2.3, obtaining the risk grade y of traffic conflict between the non-motor vehicle A and the vehicle B by using the formula (3) A,B If y is A,B If not, executing the step 3; otherwise, returning to the step 1;
in the formula (3), θ 1 Representing traffic conflict discrimination threshold, θ, based on TTC index 2 Representing a traffic conflict risk rating threshold based on the TTC index; y is A,B =0 for slight collision, y A,B =1 denotes a serious collision, y A,B = ∞ represents no collision;
step 2.4, calculating the traffic conflict risk measurement index of the non-motor vehicle A and the vehicle B at the current moment t by using the formula (4)
In the formula (4), d A (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the non-motor vehicle A when the non-motor vehicle A and the vehicle B keep the current driving direction unchanged at the current moment t, d B (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the vehicle B when the current driving direction of the non-off motor vehicle A and the vehicle B is kept unchanged at the current moment t;
step 2.5, if the tail of the non-motor vehicle A or the vehicle B passes through the intersection point of the driving tracks of the two vehicles, selecting the minimum value from all the calculated traffic conflict risk measurement indexes and recording the minimum value as the minimum traffic conflict risk measurement index
Otherwise, after assigning t +1 to t, returning to the step 2.4 for execution;
step 2.6, utilize formula (5)Calculating post-intrusion time index PET of non-motor vehicle A and vehicle B A,B ：
PET A,B ＝t A -t B (5)
In the formula (5), t A Indicating the moment, t, at which the head of the non-motor vehicle A passes the intersection B Indicating the moment at which the rear of vehicle B that first passed the intersection leaves the intersection;
step 2.7, obtaining the risk level y 'of traffic conflict between the non-motor vehicle A and the vehicle B by utilizing the formula (6)' A,B (ii) a If y' A,B If not, executing the step 3; otherwise, returning to the step 1;
in the formula (6), τ 1 、λ 1 Respectively representing risk measurement indexes T based on traffic conflicts 2 Traffic conflict discrimination threshold, τ, for post-intrusion time index PET 2 、λ 2 Respectively representing risk measurement indexes T based on traffic conflicts 2 The traffic conflict risk grade of the post-intrusion time index PET is divided into threshold values; y' A,B =0 denotes a slight conflict, y' A,B =1 indicates severe collision; y' A,B = ∞ represents no collision;
step 3, extracting K risky driving behaviors of the non-motor vehicle A driver from the road camera shooting data shot synchronously with the unmanned aerial vehicle aerial photography and forming an interpretation variable set X influencing any ith starting traffic conflict risk level i ＝(X i1 ,X i2 ,…,X ik ,…,X iK ) Wherein X is ik Indicating the k risky driving behavior of the ith traffic conflict; set X of interpretation variables for making ith traffic conflict risk level i Traffic conflict risk level y A,B Or y' A,B Is marked as y i (ii) a From X i And y i Forming an ith sample; thus obtaining N samples according to the processes of the step 1 to the step 3;
step 4, establishing a random parameter binomial Logit model with heterogeneity of mean and variance, and quantitatively analyzing influence factors influencing the severity of the traffic conflict of the non-motor vehicles;
step 4.1, when X ik When the variable is a continuous variable, the t test is adopted to X ik Performing single factor significance analysis when X is ik When classifying variables, chi-square test is used for X ik Performing single factor significance analysis, comparing the tested P value with the set significance level alpha, and adding a first interpretation variable set if the P value is less than the significance level alpha
If not, continuing to judge the next risk driving behavior; thus resulting in a final first interpretation variable set->
Step 4.2, calculating the final first interpretation variable set
And deleting the interpretation variables with the variance inflation factor larger than or equal to theta to obtain a second interpretation variable set ^ and ^ r>
Step 4.3, constructing a stepwise regression model, and integrating the second interpretation variables
Introducing the explained variables into a regression equation one by one, respectively performing F test on all the explained variables in the regression equation during each introduction, if the P value obtained by the test is more than or equal to the set significant level alpha, rejecting the corresponding explained variables, and otherwise, reserving the corresponding explained variables to obtain a third explained variable set (ion-value-based analysis method or mean-value-based analysis method) based on the third explained variable set>
Step 4.4, establishing a Logit regression model by using the formula (7):
in the formula (7), p i Indicates the probability of the ith sample being a severe collision, 1-p i Representing the probability that the ith sample is a slight conflict; beta is a beta i A parameter vector to be estimated for the ith sample is obtained by the formula (8); epsilon i A random error term for the ith sample;
β i ＝β+δ i M i +σ i exp(ω i s i )v i (8)
in the formula (8), β represents a parameter vector β i Mean value of (D), M i Is the influence parameter vector beta i Interpretation of the mean variable vector, δ i Is M i Corresponding estimated parameter vector, s i To capture beta i Standard deviation of (a) i Interpretive variable vector of medium heterogeneity, omega i Is s i A corresponding estimated parameter vector; v. of i Is the perturbation term for the ith sample;
step 4.4, assume third set of interpretation variables
Parameter vector beta of i Are all random parameters, and specify a probability density function for the distribution of the random parameters, a variable vector M will be explained i And interpreting the variable vector s i Are initialized to a set of explanatory variables X i ；
And 4.5, performing parameter estimation on the Logit regression model by adopting a simulation-based maximum likelihood method, and solving an estimated parameter vector beta 'when the likelihood function shown as the formula (9) is maximum' i 、δ′ i 、ω′ i And as a model regression result;
step (ii) of4.6, according to the regression result of the model, if the third interpretation variable set
Of an interpretive variable in the estimated parameter vector β' i If the P values of the standard deviation and the mean value of the corresponding estimated parameters in the set are less than the set significance level alpha, setting the corresponding explanatory variables as random parameters, and determining the corresponding explanatory variables in the estimated parameter vector beta 'based on the goodness of fit of the Logit regression model' i Otherwise, setting the estimation parameters of the corresponding interpretation variables as fixed parameters;
if the variable vector M is explained i Is in the estimated parameter vector delta' i The P value of the corresponding estimated parameter is less than the set significance level alpha, then the variable vector M is explained i The corresponding interpretation variable is retained, otherwise, from M i Eliminating corresponding explanation variables;
if the variable vector sigma is explained i Of an interpretive variable in the estimated parameter vector ω' i The P value of the corresponding estimated parameter is less than the set significance level alpha, then the variable vector sigma is explained i In which the corresponding interpretation variable is retained, otherwise, from σ i Removing corresponding explanation variables;
step 4.7, processing according to the processes of the step 4.5 to the step 4.6 until the P values of the estimated parameters in the model regression result are all smaller than the set significance level alpha, so as to obtain a final model regression result;
step 4.8, according to the final model regression result, when the third interpretation variable set
Mth interpretation variable>
For successive variables, the value of the elasticity coefficient->
When interpreting the set of variables +>
Is explained in>
For categorizing a variable, the value of the pseudo-elasticity coefficient of the variable->
In the formula (11), the reaction mixture is,
indicates when an interpretation variable>
The probability of a severe collision occurring when equal to 1,
indicates when an interpretation variable>
Probability of severe collision when equal to 0;
step 4.9, obtaining a third interpretation variable set according to the process of the step 4.8
The marginal effects of all the explained variables in the (A) and the (B) are sorted in descending order, and a third explained variable set is judged to be greater than or equal to>
And the influence of each risky driving behavior on the traffic conflict risk degree is used as a risk factor evaluation result of the traffic conflict of the non-motor vehicles at the plane intersection.
The electronic device comprises a memory and a processor, and is characterized in that the memory is used for storing programs for supporting the processor to execute the risk factor evaluation method, and the processor is configured to execute the programs stored in the memory.
The invention relates to a computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, performs the steps of the risk factor assessment method.
Compared with the prior art, the invention has the beneficial effects that:
1. compared with the traditional method for evaluating the traffic safety risk factors of the non-motor vehicles at the plane intersection, the method for evaluating the traffic safety risk factors of the non-motor vehicles at the plane intersection has the advantages that the severity of the conflict approach accident is measured by constructing the traffic conflict index, the risk driving behaviors of the non-motor vehicles at the plane intersection are fully considered, the influence rule of the risk driving behaviors on the traffic conflict risk of the non-motor vehicles is analyzed, and the statistical advantage characteristics of large sample, short period, small area, high reliability and the like are realized, so that the high efficiency and the objectivity of the traffic safety analysis of the non-motor vehicles at the plane intersection are improved.
2. In the invention, the uncertainty of the mobility of the conflict point under the linear traffic conflict and the expected collision under the non-linear traffic conflict is considered, the safety substitute evaluation index reflecting the traffic conflict risk before and after the conflict vehicle reaches the conflict point is accurately constructed, and the accuracy of the traffic conflict risk evaluation is improved.
3. The invention establishes a random parameter binomial Logit model with heterogeneity of mean and variance, and quantitatively analyzes the influence of the risk driving behavior of the rider on the non-motorized traffic conflict. By specifying the distribution function of the model estimation parameters and taking the mean value and the variance of the distribution function as the estimation functions of the observed interpretation variables, the problem of unobserved heterogeneity in data is captured, and the accuracy of the model parameter estimation and the inference results is improved.
Drawings
FIG. 1 is a flow chart of a risk factor assessment method of the present invention.
Detailed Description
In this embodiment, as shown in fig. 1, a method for evaluating risk factors of non-motor vehicle traffic conflicts at a plane intersection is performed according to the following steps, taking a cross-shaped signal plane intersection of a tung city road and a turnip lake road in a central urban area of Hefei city, anhui province as an example:
step 1, collecting vehicle running track data in a current plane intersection from an unmanned aerial vehicle aerial video, and extracting motion state data of a non-motor vehicle A and a vehicle B from the vehicle running track data, wherein the data comprises the following steps: vehicle position and speed;
step 2, if the current non-motor vehicle A and the current non-motor vehicle B approach each other on a horizontal straight line, and the mobility of the conflict point under the straight line traffic conflict is considered, executing the step 2.1-the step 2.3, and measuring the risk of the straight line traffic conflict; if the current non-motor vehicle A and the current non-motor vehicle B approach each other in a staggered mode, and uncertainty of expected collision of non-linear traffic conflict is considered, executing the step 2.4-the step 2.7, and comprehensively reflecting the approaching degree of the vehicles before and after reaching the conflict point;
step 2.1, if the current non-motor vehicle A and the current vehicle B run on the same horizontal straight line in the same direction, calculating the time TTC of collision between the non-motor vehicle A and the current vehicle B at the current time t by using the formula (1) A,B (t), identifying a rear-end collision:
in the formula (1), l A,B (t) represents the spatial distance from the front of the rear vehicle B to the rear of the front vehicle A at the current moment t, v B (t) represents the speed of the vehicle of the following vehicle B at the present time t, v A (t) represents the speed of the preceding vehicle a at the current time t;
if the current non-motor vehicle A and the current non-motor vehicle B oppositely run on a horizontal straight line, calculating the time TTC of the collision between the non-motor vehicle A and the current non-motor vehicle B at the current time t by using the formula (2) A ′ ,B (t), identifying the forward directionConflict:
in the formula (2), l A ′ ,B (t) represents the spatial distance between the heads of two vehicles at the current moment t, v A (t)、v B (t) respectively representing the speeds of the non-motor vehicle A and the vehicle B at the current moment t;
step 2.2, if the speed of the rear vehicle B is greater than that of the front vehicle A at the moment of t +1 or the speed of any one of the non-motor vehicle A and the vehicle B at the moment of t +1 is not 0 in the opposite driving process, indicating that the two vehicles are expected to collide, assigning t +1 to t, returning to the step 2.1 to execute, otherwise, indicating that the two vehicles are not expected to collide, and selecting the minimum value from all calculated collision time of the non-motor vehicle A and the vehicle B to obtain the minimum collision time minTTC A,B ，；
Step 2.3, selecting theta 1 Using =3s as collision time TTC index to judge threshold of collision, and selecting theta 2 And (4) taking The Time of Collision (TTC) as an index for dividing threshold values of slight conflict and serious conflict to obtain the risk grade y of traffic conflict between the non-motor vehicle A and the vehicle B by using the formula (3) A,B (ii) a If y A,B If not, executing the step 3; otherwise, returning to the step 1;
in the formula (3), θ 1 Representing traffic conflict discrimination thresholds, theta, based on TTC index 2 Representing a traffic conflict risk rating threshold based on the TTC index; y is A,B =0 for slight collision, y A,B =1 denotes a serious collision, y A,B = ∞ represents no collision;
step 2.4, calculating the traffic conflict risk measurement index of the non-motor vehicle A and the vehicle B at the current moment t by using the formula (4)
Measurement ofRisk of traffic collision before vehicle reaches conflict point under non-linear traffic collision:
in the formula (4), d A (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the non-motor vehicle A when the non-motor vehicle A and the vehicle B keep the current driving direction unchanged at the current moment t, d B (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the vehicle B when the current driving direction of the non-motor vehicle A and the vehicle B is kept unchanged at the current moment t;
step 2.5, if the tail of the non-motor vehicle A or the vehicle B passes through the intersection point of the driving tracks of the two vehicles, selecting the minimum value from all the calculated traffic conflict risk measurement indexes and recording the minimum value as the minimum traffic conflict risk measurement index
Otherwise, after assigning t +1 to t, returning to the step 2.4 for execution;
step 2.6, calculating the post-invasion time index PET of the non-motor vehicle A and the vehicle B by using the formula (5) A,B And (3) measuring the risk of the vehicle after reaching the conflict point in the non-linear traffic conflict:
PET A,B ＝t A -t B (5)
in the formula (5), t A Indicating the moment, t, at which the head of the non-motor vehicle A passes the intersection B Indicating the moment at which the rear of vehicle B that first passed the intersection leaves the intersection;
step 2.7, select tau separately 1 ＝3s、λ 1 =3s as traffic conflict risk metric T 2 Judging conflict threshold value by post-invasion time index PET index, selecting tau 2 ＝1.5s、λ 2 =1.5s as T 2 The PET index is divided into threshold values of slight conflict and serious conflict, and the risk level y 'of traffic conflict between the non-motor vehicle A and the vehicle B is obtained by utilizing the formula (6)' A,B (ii) a If y' A,B If not, executing the step 3; otherwiseReturning to the step 1;
in the formula (6), τ 1 、λ 1 Respectively represent a radical based on T 2 Traffic conflict discrimination threshold value of PET index, tau 2 、λ 2 Respectively represent a radical based on T 2 The traffic conflict risk grade of the PET index is divided into a threshold value; y' A,B =0 denotes slight conflict, y' A,B =1 represents a serious collision;
step 3, extracting 18 (K = 18) risky driving behaviors of the non-motor vehicle A driver from the road camera shooting data shot synchronously with the unmanned aerial vehicle aerial photography and forming an interpretation variable set X influencing any ith initial traffic conflict risk level i ＝(X i1 ,X i2 ,…,X ik ,…,X i18 ) Wherein X is ik The k-th risky driving behavior of the ith traffic conflict is shown, and in order to facilitate the fitting of the discrete result model, the explanatory variables are recoded and assigned as shown in the table 1; set X of interpretation variables for making ith traffic conflict risk level i Traffic conflict risk level y A,B Or y' A,B Is marked as y i (ii) a From X i And y i Forming an ith sample; thus, N =438 samples were obtained according to the procedure of step 1-step 3;
TABLE 1 explanatory variables for analysis of impact factors for non-motor traffic conflict risk levels
Step 4, establishing a random parameter binomial Logit model with heterogeneity of mean value and variance, and quantitatively analyzing influence factors influencing the severity of traffic conflicts of the non-motor vehicles;
step 4.1, when X ik When the variable is a continuous variable, the t test is adopted to X ik Performing single factor significance analysis when X is ik For classifying variables, chi-square test is used for X ik Performing single factor significance analysis, comparing the tested P value with the set significance level alpha, and adding a first interpretation variable set if the P value is less than the significance level alpha
If not, continuing to judge the next risky driving behavior; thus resulting in a final first interpretation variable set->
In an embodiment, a set of interpretation variables X i The interpretation variables in (1) are all classification variables, so the chi-square test is adopted to carry out single-factor significance test, when the significance evaluation index P value is less than 0.05 (alpha = 0.05), the alternative interpretation variables are considered to have significant influence on the severity of the traffic conflict of the non-motor vehicle, otherwise, the alternative interpretation variables are deleted, and the result is shown in table 2:
TABLE 2 Single factor analysis results
As can be seen from Table 2, 8 explanatory variables, including gender, takeaway deliverer, running a red light, occupying motorway, not observing the coming vehicle at the rear while turning, not decelerating while driving into an intersection, violating road right-of-way, and not maintaining safe lateral distance, have statistical significance (P)<= 0.05). Thus, the above explained variables are retained, resulting in a first set of explained variables
Step 4.2, calculating the final first interpretation variable set
And the variance inflation factor of each interpretation variable is deleted, and the interpretation variables with the variance inflation factor of more than or equal to 10 (theta = 10) are deleted, so that a second set of interpretation variables ≥ is obtained>
The method eliminates the influence of multiple collinearity existing among the explained variables, meets the requirement that the variables are mutually independent when the Logit regression model carries out parameter estimation, and the multiple collinearity test is shown in a table 3:
TABLE 3 multiple collinearity diagnostic results
As can be seen from Table 3, all the VIF values of the 8 explanatory variables selected in step 4.2 with statistical significance are less than 2, and the multiple collinearity problem does not exist. Thus, the 8 explanatory variables described above are retained, resulting in a second set of explanatory variables
Step 4.3, constructing a stepwise regression model, and integrating the second interpretation variables
The explained variables are introduced into the regression equation one by one, all the explained variables in the regression equation are subjected to F test during each introduction, if the P value obtained by the test is more than or equal to the set significant level alpha, the corresponding explained variables are removed, otherwise, the corresponding explained variables are retained, and therefore a third explained variable set with significant prediction effect is further screened out
The stepwise regression results are shown in table 4:
TABLE 4 stepwise regression results
As can be seen from Table 4, the set of explanatory variables obtained by stepwise regression results screening
Step 4.4, establishing a Logit regression model by using the formula (7):
in formula (7), p i Indicates the probability of the ith sample being a severe collision, 1-p i Representing the probability that the ith sample is a slight conflict; beta is a i A parameter vector to be estimated for the ith sample is obtained by the formula (8); epsilon i A random error term for the ith sample;
β i ＝β+δ i M i +σ i exp(ω i s i )v i (8)
in the formula (8), β represents a parameter vector β i Mean value of M i Is the influence parameter vector beta i Interpretation of the mean variable vector, δ i Is M i Corresponding estimated parameter vector, s i To catch beta i Standard deviation of (a) i Interpretive variable vector of medium heterogeneity, omega i Is s i A corresponding estimated parameter vector; v. of i Is the perturbation term for the ith sample;
step 4.4, assume third set of interpretation variables
Parameter vector beta of i Are random parameters, and are assigned to obey normal distribution or log-normal distribution or uniform distribution, and the variable vector M will be explained i And interpreting the variable vector s i Are initialized to a set of interpretation variables X i ；
And 4.5, performing parameter estimation on the Logit regression model by adopting a simulation-based maximum likelihood method, and solving an estimated parameter vector beta 'when the likelihood function shown as the formula (9) is maximum' i 、δ′ i 、ω′ i And as a model regression result;
step 4.6, according to the model regression result, if the third interpretation variable set
Of an interpretive variable in the estimated parameter vector β' i If the standard deviation and the mean value P of the corresponding estimation parameters are less than the set significance level alpha, setting the corresponding interpretation variables as random parameters, and determining the estimated parameter vector beta of the corresponding interpretation variables based on the goodness of fit of the Logit regression model i ' the probability density function of the corresponding estimation parameter, otherwise, the estimation parameter of the corresponding interpretation variable is set as a fixed parameter;
if the variable vector M is explained i Of an interpretive variable in the estimated parameter vector δ' i The P value of the corresponding estimated parameter is less than the set significance level alpha, then the variable vector M is explained i The corresponding interpretation variable is retained, otherwise, from M i Removing corresponding explanation variables;
if the variable vector sigma is explained i Of an interpretive variable in the estimated parameter vector ω' i The P value of the corresponding estimated parameter is less than the set significance level alpha, then the variable vector sigma is explained i Middle retention phaseVariables should be interpreted, otherwise, from σ i Removing corresponding explanation variables;
step 4.7, processing according to the processes of the step 4.5 to the step 4.6 until the P values of the estimated parameters in the model regression result are all smaller than the set significance level α, so as to obtain a final model regression result, as shown in table 5:
TABLE 5 random parameter Lomit model parameter estimation results of mean and variance heterogeneity
As can be seen from table 5, the interpretation variables do not slow down when driving into the intersection, run a red light, take out deliverers, violate road traffic, do not maintain safe lateral distance, and occupy motorways have a significant impact on the severity of the non-motor traffic conflict. The explanation variable does not decelerate when entering the intersection, has random parameter characteristics, and has variance heterogeneity in the 'gender' explanation variable. Compared with a triangular distribution (-239.00217) and a uniform distribution (-239.06153), the random parameter obeys normal distribution when the model converges, and has a better log-likelihood value (-238.96525), and the fitting goodness of the model is optimal. Thus, a normal distribution with a mean of 1.84469 and a variance of 2.31186 is determined as a probability density function of random parameters, i.e., 78.75% of the riders are prone to severe collisions and 21.25% of the riders are less likely to have severe collisions for non-motor riders that do not slow down when entering the intersection. The remaining explanatory variable estimation parameters are fixed parameters.
Step 4.8, according to the final model regression result, when the third interpretation variable set
The mth interpretation variable->
For continuous variables, the value of the elasticity coefficient->
Represents a variable->
The change in (b) corresponds to a change value of the probability of the non-motor vehicle experiencing a severe collision; when interpreting the set of variables->
Is explained in>
For categorizing a variable, the value of the pseudo-elasticity coefficient ≥ of the variable is calculated using equation (11)>
Indicates that the interpretation variable->
When the reference value is changed into the current value, the non-motor vehicle has a change value of the serious conflict probability; therefore, the influence of the risky driving behaviors on the severity of the traffic conflict of the non-motor vehicle is quantitatively analyzed;
in the formula (11), the reaction mixture is,
indicates when the interpretation variable->
Probability of a severe conflict occurring equal to 1, <' >>
Indicates when the interpretation variable->
Probability of a severe collision occurring when equal to 0;
in the embodiment, the selected interpretation variables are all classification variables, so the influence of the interpretation variables on the severity of the non-motor vehicle traffic conflict is quantified through a calculation formula (11).
Step 4.9, obtaining a third interpretation variable set according to the process of the step 4.8
The marginal effects of all the explained variables in the (A) and the (B) are sorted in descending order, and a third explained variable set is judged to be greater than or equal to>
The influence of each risky driving behavior on the traffic conflict risk degree is used as a risk factor evaluation result of the traffic conflict of the non-motor vehicles at the intersection, and the result is shown in table 6:
TABLE 6 average marginal effect of significant variables on severity of non-motor vehicle traffic conflicts
As can be seen from table 6, interpreting variables that run red light, do not maintain lateral safe distance, violate road right of way, take delivery, occupy motorways, and drive into intersections without slowing down significantly increases the probability of a serious conflict with a non-motor vehicle. Wherein, the probability of serious conflict of non-motor vehicles caused by red light running is the maximum, and is 3.51 percent; secondly, the transverse safety distance is not kept, and is 2.78%; the probability of serious conflict of the non-motor vehicles caused by violation of road right of way, the fact that the riders are takeaway distributors and the fact that the riders occupy the motor vehicle lanes is equivalent to 1.71%, 1.69% and 1.68% respectively; the speed reduction is lowest when the automobile enters the intersection and is 0.52 percent.
In this embodiment, an electronic device includes a memory for storing a program that supports a processor to execute the risk factor assessment method, and a processor configured to execute the program stored in the memory.
In this embodiment, a computer-readable storage medium stores a computer program, and the computer program is executed by a processor to perform the steps of the risk factor assessment method.
In conclusion, the method for evaluating the risk factors of the traffic conflicts of the non-motor vehicles at the plane intersection has feasibility, can objectively and efficiently analyze the factors influencing the severity of the traffic conflicts of the plane intersection, fully excavates the heterogeneous characteristics in the data and improves the accuracy of analyzing the risk factors of the traffic accidents.
Claims (3)
1. A method for evaluating risk factors of non-motor vehicle traffic conflicts at plane intersections is characterized by comprising the following steps:
step 1, collecting vehicle running track data in a current plane intersection from an unmanned aerial vehicle aerial video, and extracting motion state data of a non-motor vehicle A and a vehicle B from the vehicle running track data, wherein the motion state data comprises the following steps: vehicle position and speed;
step 2, if the current non-motor vehicle A and the vehicle B approach each other on a horizontal straight line, executing step 2.1-step 2.3, and if the current non-motor vehicle A and the vehicle B approach each other in a staggered manner, executing step 2.4-step 2.7;
step 2.1, if the current non-motor vehicle A and the current vehicle B run on the same horizontal straight line in the same direction, calculating the time TTC of collision between the non-motor vehicle A and the current vehicle B at the current time t by using the formula (1) A,B (t)：
In the formula (1), l A,B (t) represents the spatial distance from the front of the rear vehicle B to the rear of the front vehicle A at the current moment t, v B (t) speed of the vehicle of the following vehicle B at the present time t, v A (t) shows the preceding vehicle A at the present time tSpeed;
if the current non-motor vehicle A and the current non-motor vehicle B oppositely run on a horizontal straight line, calculating the time TTC of the collision between the non-motor vehicle A and the current non-motor vehicle B at the current time t by using the formula (2) A ′ ,B (t)：
L 'in the formula (2)' A,B (t) represents the spatial distance between the heads of two vehicles at the current moment t, v A (t)、v B (t) respectively representing the speeds of the non-motor vehicle A and the vehicle B at the current moment t;
step 2.2, if the speed of the rear vehicle B is greater than that of the front vehicle A at the moment of t +1 or the speed of any one of the non-motor vehicle A and the vehicle B at the moment of t +1 is not 0 in the opposite driving process, indicating that the two vehicles are expected to collide, assigning t +1 to t, returning to the step 2.1 to execute, otherwise, indicating that the two vehicles are not expected to collide, and selecting the minimum value from all calculated collision time of the non-motor vehicle A and the vehicle B to obtain the minimum collision time minTTC A,B ；
Step 2.3, obtaining the risk grade y of traffic conflict between the non-motor vehicle A and the vehicle B by using the formula (3) A,B If y is A,B If not, executing the step 3; otherwise, returning to the step 1;
in the formula (3), θ 1 Representing traffic conflict discrimination thresholds, theta, based on TTC index 2 Representing a traffic conflict risk rating threshold based on the TTC index; y is A,B =0 for slight collision, y A,B =1 denotes a serious collision, y A,B = ∞ denotes no collision;
step 2.4, calculating the traffic conflict risk measurement index of the non-motor vehicle A and the vehicle B at the current moment t by using the formula (4)
In the formula (4), d A (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the non-motor vehicle A when the non-motor vehicle A and the vehicle B keep the current driving direction unchanged at the current moment t, d B (t) represents the distance between the intersection point of the expected tracks of the two vehicles and the vehicle B when the current driving direction of the non-off motor vehicle A and the vehicle B is kept unchanged at the current moment t;
step 2.5, if the tail of the non-motor vehicle A or the vehicle B passes through the intersection point of the driving tracks of the two vehicles, selecting the minimum value from all the calculated traffic conflict risk measurement indexes and recording the minimum value as the minimum traffic conflict risk measurement index
Otherwise, after assigning t +1 to t, returning to the step 2.4 for execution;
step 2.6, calculating the post-invasion time index PET of the non-motor vehicle A and the vehicle B by using the formula (5) A,B ：
PET A,B ＝t A -t B (5)
In the formula (5), t A Indicating the moment, t, at which the head of the non-motor vehicle A passes the intersection B Indicating the moment at which the rear of vehicle B that first passed the intersection leaves the intersection;
step 2.7, obtaining the risk level y 'of traffic conflict between the non-motor vehicle A and the vehicle B by utilizing the formula (6)' A,B (ii) a If y' A,B If not, executing the step 3; otherwise, returning to the step 1;
in the formula (6), τ 1 、λ 1 Respectively representing risk measurement indexes T based on traffic conflicts 2 Traffic conflict discrimination threshold, τ, for post-intrusion time index PET 2 、λ 2 Respectively representing risk measurement indexes T based on traffic conflicts 2 The traffic conflict risk level of the post-invasion time index PET is divided into threshold values; y' A,B =0 denotes a slight conflict, y' A,B =1 represents a serious collision; y' A,B = ∞ represents no collision;
step 3, extracting K risky driving behaviors of the non-motor vehicle A driver from road camera shooting data shot synchronously with unmanned aerial vehicle aerial photography and forming an interpretation variable set X influencing any ith traffic conflict risk level i ＝(X i1 ,X i2 ,…,X ik ,…,X iK ) Wherein X is ik The k risky driving behavior of the ith traffic conflict is shown; set X of interpretation variables for making ith traffic conflict risk level i Traffic conflict risk rating y A,B Or y' A,B Is marked as y i (ii) a From X i And y i Forming an ith sample; thus obtaining N samples according to the processes of the step 1 to the step 3;
step 4, establishing a random parameter binomial Logit model with heterogeneity of mean and variance, and quantitatively analyzing influence factors influencing the severity of the traffic conflict of the non-motor vehicles;
step 4.1, when X ik When the variable is a continuous variable, the t test is adopted to X ik Performing single factor significance analysis when X is ik When classifying variables, chi-square test is used for X ik Performing single factor significance analysis, comparing the tested P value with the set significance level alpha, and if the P value is less than the significance level alpha, adding a first interpretation variable set
If not, continuing to judge the next risky driving behavior; thus resulting in a final first interpretation variable set->
Step 4.2, calculate the final first interpretationVariable set
And deleting the interpretation variables with the variance inflation factor larger than or equal to theta to obtain a second interpretation variable set ^ and ^ r>
Step 4.3, constructing a stepwise regression model, and integrating the second interpretation variables
The explained variables are introduced into the regression equation one by one, all the explained variables in the regression equation are subjected to F test each time the explained variables are introduced, if the P value obtained by the test is more than or equal to the set significant level alpha, the corresponding explained variables are removed, otherwise, the corresponding explained variables are retained, and the third explained variable set (or the third explained variable set) is obtained>
Step 4.4, establishing a Logit regression model by using the formula (7):
in formula (7), p i Indicates the probability of the ith sample being a severe collision, 1-p i Representing the probability that the ith sample is a slight conflict; beta is a i A parameter vector to be estimated for the ith sample is obtained by the formula (8); epsilon i A random error term for the ith sample;
β i ＝β+δ i M i +σ i exp(ω i s i )v i (8)
in the formula (8), β represents a parameter vector β i Mean value of (D), M i Is an influence parameter vector beta i Interpretation of the mean variable vector, δ i Is M i Corresponding estimated parameter vector, s i To catch beta i Standard deviation of (a) i Interpretive variable vector of medium heterogeneity, omega i Is s i A corresponding estimated parameter vector; v. of i Is the perturbation term for the ith sample;
step 4.4, assume third set of interpretation variables
Parameter vector beta of i Are all random parameters, and specify a probability density function for the distribution of the random parameters, a variable vector M will be explained i And interpreting the variable vector s i Are initialized to a set of interpretation variables X i ；
Step 4.5, carrying out parameter estimation on the Logit regression model by adopting a simulation-based maximum likelihood method, and solving an estimated parameter vector beta when the likelihood function shown as the formula (9) is maximized i ′、δ′ i 、ω′ i And as a model regression result;
step 4.6, according to the model regression result, if the third interpretation variable set
In the estimation of the parameter vector beta i If the P values of the standard deviation and the mean value of the corresponding estimated parameters are less than the set significance level alpha, the corresponding interpretation variables are set as random parameters, and the goodness of fit of the Logit regression model is used for determining the corresponding interpretation variables in the estimated parameter vector beta i ' the probability density function of the corresponding estimation parameter, otherwise, the estimation parameter of the corresponding interpretation variable is set as a fixed parameter;
if the variable vector M is explained i Of an interpretive variable in the estimated parameter vector δ' i The P value of the corresponding estimation parameter is less than the set significance level alpha, then the solution is carried outVector of variables M i The corresponding interpretation variable is retained, otherwise, from M i Eliminating corresponding explanation variables;
if the variable vector sigma is explained i Of an interpretive variable in the estimated parameter vector ω' i The P value of the corresponding estimated parameter is less than the set significance level alpha, then the variable vector sigma is explained i In which the corresponding interpretation variable is retained, otherwise, from σ i Eliminating corresponding explanation variables;
step 4.7, processing according to the processes of the step 4.5 to the step 4.6 until the P values of the estimated parameters in the model regression result are all smaller than the set significance level alpha, so as to obtain a final model regression result;
step 4.8, according to the final model regression result, when the third interpretation variable set
The mth interpretation variable->
For continuous variables, the value of the elasticity coefficient->
When interpreting the set of variables->
An interpretation variable in->
For categorizing a variable, the value of the pseudo-elasticity coefficient ≥ of the variable is calculated using equation (11)>
In the formula (11), the reaction mixture is,
indicates when the interpretation variable->
Probability of a severe conflict occurring equal to 1, <' >>
Indicates when the interpretation variable->
Probability of a severe collision occurring when equal to 0;
step 4.9, obtaining a third interpretation variable set according to the process of the step 4.8
The marginal effects of all the explained variables in the (A) and the (B) are sorted in descending order, and a third explained variable set is judged to be greater than or equal to>
The influence of each risky driving behavior on the traffic conflict risk degree is used as a risk factor evaluation result of the non-motor vehicle traffic conflict at the plane intersection.
2. An electronic device comprising a memory and a processor, wherein the memory is configured to store a program that enables the processor to perform the risk factor assessment method of claim 1, and wherein the processor is configured to execute the program stored in the memory.
3. A computer-readable storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the steps of the risk factor assessment method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211453450.7A CN115860461A (en) | 2022-11-21 | 2022-11-21 | Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211453450.7A CN115860461A (en) | 2022-11-21 | 2022-11-21 | Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115860461A true CN115860461A (en) | 2023-03-28 |
Family
ID=85664286
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211453450.7A Pending CN115860461A (en) | 2022-11-21 | 2022-11-21 | Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115860461A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648539A (en) * | 2024-01-30 | 2024-03-05 | 山东理工大学 | Rural highway traffic accident severity evaluation method |
-
2022
- 2022-11-21 CN CN202211453450.7A patent/CN115860461A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117648539A (en) * | 2024-01-30 | 2024-03-05 | 山东理工大学 | Rural highway traffic accident severity evaluation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Goyani et al. | Investigation of traffic conflicts at unsignalized intersection for reckoning crash probability under mixed traffic conditions | |
| Yu et al. | Trajectory data based freeway high-risk events prediction and its influencing factors analyses | |
| Lyu et al. | Using naturalistic driving data to identify driving style based on longitudinal driving operation conditions | |
| CN110197588A (en) | A kind of truck driving behavior appraisal procedure and device based on GPS track data | |
| Kovaceva et al. | Identification of aggressive driving from naturalistic data in car-following situations | |
| CN112668172B (en) | Following behavior modeling method considering heterogeneity of vehicle type and driving style and model thereof | |
| CN110705582B (en) | Ship bridge-crossing behavior feature extraction method based on improved K-Means clustering algorithm | |
| Li et al. | Crash Risk Prediction Model of Lane‐Change Behavior on Approaching Intersections | |
| CN113436432A (en) | Method for predicting short-term traffic risk of road section by using road side observation data | |
| CN114822044B (en) | Driving safety early warning method and device based on tunnel | |
| Kidando et al. | Evaluating traffic congestion using the traffic occupancy and speed distribution relationship: An application of Bayesian Dirichlet process mixtures of generalized linear model | |
| CN115860461A (en) | Risk factor evaluation method for traffic conflict of non-motor vehicles at plane intersection | |
| CN112308136B (en) | Driving distraction detection method based on SVM-Adaboost | |
| Katrakazas et al. | A simulation study of predicting conflict-prone traffic conditions in real-time | |
| CN115114786B (en) | Assessment method, system and storage medium for traffic flow simulation model | |
| CN116361175A (en) | Method for creating test scenes of automatic driving vehicles in different safety domains | |
| CN113192340B (en) | Method, device, equipment and storage medium for identifying highway construction vehicles | |
| CN115027497A (en) | Target vehicle cut-in intention prediction method and readable storage medium | |
| CN114820216A (en) | Driving risk quantification method, classification method and prediction method | |
| CN116997890A (en) | Generating an unknown unsafe scenario, improving an automated vehicle, and a computer system | |
| Ma et al. | Lane change analysis and prediction using mean impact value method and logistic regression model | |
| Mahajan | Real-time driving intention prediction and crash risk estimation from naturalistic driving data using machine learning | |
| Wang et al. | Improved Time-to-collision Considering Vehicle Speed Adaptation based on Trajectory Data | |
| Tian et al. | Understanding scenarios for cooperative V2P safety applications using connected vehicle datasets | |
| CN117719524B (en) | Driving safety risk identification early warning method, device, terminal and storage medium |
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 |