CN116959248B - Expressway accident risk optimization method based on interactive self-interpretation model - Google Patents
Expressway accident risk optimization method based on interactive self-interpretation model Download PDFInfo
- Publication number
- CN116959248B CN116959248B CN202310903318.XA CN202310903318A CN116959248B CN 116959248 B CN116959248 B CN 116959248B CN 202310903318 A CN202310903318 A CN 202310903318A CN 116959248 B CN116959248 B CN 116959248B
- Authority
- CN
- China
- Prior art keywords
- risk
- accident
- traffic
- accident risk
- traffic flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 40
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000005457 optimization Methods 0.000 title claims abstract description 34
- 230000003993 interaction Effects 0.000 claims abstract description 36
- 238000010586 diagram Methods 0.000 claims abstract description 29
- 238000013058 risk prediction model Methods 0.000 claims abstract description 26
- 206010039203 Road traffic accident Diseases 0.000 claims abstract description 24
- 238000004088 simulation Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 7
- 238000011144 upstream manufacturing Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000012954 risk control Methods 0.000 abstract description 6
- 238000010801 machine learning Methods 0.000 description 7
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 239000000654 additive Substances 0.000 description 2
- 230000000996 additive effect Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Geometry (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides a highway accident risk optimization method based on an interactive self-interpretation model, which comprises the following steps: constructing a traffic risk accident risk prediction model; based on the traffic risk accident influencing factors, according to the SHAP model, obtaining interaction relations among the traffic risk accident influencing factors; constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor; and optimizing the traffic accident risk probability of the expressway according to the accident risk thermodynamic diagram. The method solves the problems that in the prior art, the analysis of the traffic accident causes is difficult to sort the importance of influencing factors and the black box of the accident prediction model is difficult to search the optimization direction in the traffic accident risk control.
Description
Technical Field
The invention relates to the technical field of traffic intelligent control, in particular to a highway accident risk optimization method based on an interactive self-interpretation model.
Background
With the rapid development of digital technology and the popularization and application of the digital technology in various industries, the public expects and demands on the intellectualization and the intellectualization of expressway service, how to solve the problems of traffic safety, congestion relief, traffic efficiency improvement, traffic flow regulation and control, automatic driving adaptation and intelligent networking by using modern means is more and more urgent, and industry response and implementation are needed, so that ubiquitous highway users can obtain the intellectualized and intelligent high-level service.
It can be explained that machine learning has been an important research direction for machine learning over the years. There is a need for data scientists to prevent model bias and help decision makers understand how to use our model correctly. The more stringent the scene, the more the model is required to provide evidence that proves how they are functioning and avoid errors. Regarding model interpretation, besides models that are naturally very well interpreted, such as linear models and decision trees, many models in sklean have an interface of importance, and the importance of features can be checked. In fact this has embodied the model explanatory idea in sand projection. However, the traditional method of computing importance is in fact quite controversial and not always consistent. SHAP belongs to a model post-interpretation method, and the core idea is to calculate marginal contributions of features to model output and interpret 'black box model' from both global and local layers. SHAP builds an additive interpretation model, with all features considered "contributors".
In summary, the prior art has the problems that the analysis of the traffic accident causes is difficult to sort the importance of influencing factors and the black box of the accident prediction model in the traffic accident risk control is difficult to search the optimization direction.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide an expressway accident risk optimization method based on an interactive self-interpretation model, and solves the problems that in the prior art, the importance of influencing factors is difficult to sort through traffic accident cause analysis, and the black box of an accident prediction model in traffic accident risk control is difficult to search for an optimization direction.
In order to achieve the above object, the present invention provides the following solutions:
a highway accident risk optimization method based on an interactive self-interpretation model comprises the following steps:
constructing a traffic risk accident risk prediction model;
based on the traffic risk accident influencing factors, according to the SHAP model, obtaining interaction relations among the traffic risk accident influencing factors;
constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor;
and optimizing the traffic accident risk probability of the expressway according to the accident risk thermodynamic diagram.
Preferably, the construction method of the traffic risk accident risk prediction model comprises the following steps:
acquiring traffic flow data;
calculating the current traffic flow accident risk by using a traffic flow accident risk detection algorithm;
constructing a self-variable data set according to the traffic flow data;
performing dimension reduction processing on the self-variable data set to obtain a training data set;
and constructing the traffic risk accident risk prediction model according to the training data set and the traffic flow accident risk based on the XGBoost model.
Preferably, the acquiring traffic flow data includes:
acquiring microscopic traffic simulation software and a group driving simulation platform;
constructing an interactive simulation experiment platform according to the microscopic traffic simulation software and the group driving simulation platform;
and acquiring traffic flow data by using the interactive simulation experiment platform.
Preferably, the accident occurrence definition formula in the traffic flow accident risk detection algorithm is as follows:
wherein TTC is i I the collision time of the vehicle relative to the front vehicle at the time t, because the vehicle position acquired by data is the position of the vehicle head, X is the position of the vehicle head i (t) is the position of the head of the vehicle at the moment i, X h (t) is the position of the front h head of the front car at the moment i, l h Is the length of the body of the h car, V i (t) is the instantaneous speed of the i-car at time t, V h And (t) is the instantaneous speed of the vehicle at time t.
Preferably, the acquiring traffic flow data by using the interactive simulation experiment platform includes:
acquiring the section flow, the speed and the occupancy of the upstream of the specific position and the section flow, the speed and the occupancy of the downstream of the specific position by using the interactive simulation experiment platform;
and acquiring traffic flow data according to the upstream section flow, the vehicle speed and the occupancy rate and the downstream section flow, the vehicle speed and the occupancy rate.
Preferably, the method further comprises the following steps of:
preprocessing the traffic flow data, and filtering invalid data in the traffic flow data to obtain first traffic flow data;
and correcting the first traffic flow data, and normalizing the parameter characteristic indexes in the first traffic flow data.
Preferably, the method further comprises:
analyzing the traffic risk accident influence factors by using the traffic risk accident risk prediction model to obtain importance ranking of the influence factors;
and carrying out real-time reconstruction updating on the traffic risk accident risk prediction model according to the importance sorting of the influencing factors.
Preferably, the constructing an accident risk thermodynamic diagram of each traffic risk accident influence factor interaction according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor comprises:
the traffic risk accident influencing factors are processed in pairs, and the traffic risk accident influencing factors of every two pairs are obtained;
the traffic risk accident influencing factors of the pairs are amplified or reduced in the same proportion, and corresponding accident risk values are calculated by using a traffic risk accident risk prediction model;
and constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the accident risk values.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides an expressway accident risk optimization method based on an interactive self-interpretation model, which is used for carrying out importance analysis on important influence factors of traffic flow accident risk by utilizing an XGBoost machine learning model, solving the problem that the importance of the influence factors is difficult to sort by the traffic accident cause analysis, analyzing the interaction relationship between the important influence factors of the traffic flow accident risk based on a SHAP model, optimizing the accident risk probability of the expressway by utilizing an accident risk thermodynamic diagram of pairwise interaction, solving the problem that the black box of an accident prediction model is difficult to search for the optimization direction in the traffic accident risk control, and simultaneously providing an interaction optimization method.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a highway accident risk optimization method based on an interactive self-explanatory model according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a highway accident risk optimization method based on an interactive self-explanatory model according to an embodiment of the present invention;
fig. 3 is a schematic diagram of analysis of a variable SHAP value in an expressway accident risk optimization method based on an interactive self-explanatory model according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an interactive micro-traffic simulation platform in a highway accident risk optimization method based on an interactive self-explanatory model according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a traffic flow parameter acquisition position in an expressway accident risk optimization method based on an interactive self-interpretation model according to an embodiment of the present invention;
fig. 6 is a schematic diagram of sorting importance of traffic accident risk influencing factors based on XGBoost in the expressway accident risk optimization method based on the interactive self-explanatory model according to the embodiment of the present invention;
fig. 7 is an accident risk thermodynamic diagram under the influence of variable interaction in the expressway accident risk optimization method based on the interactive self-explanatory model according to the embodiment of the invention.
Fig. 8 is a schematic diagram illustrating analysis of an interaction variable SHAP value in an expressway accident risk optimization method based on an interaction self-explanatory model according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention aims to provide an expressway accident risk optimization method based on an interactive self-interpretation model, which solves the problems that in the prior art, the importance of influencing factors is difficult to sort by traffic accident cause analysis, and the black box of an accident prediction model in traffic accident risk control is difficult to search for an optimization direction.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
As shown in fig. 1, the present invention provides a highway accident risk optimization method based on an interactive self-interpretation model, comprising:
step 100: constructing a traffic risk accident risk prediction model;
step 200: based on the traffic risk accident influencing factors, according to the SHAP model, obtaining interaction relations among the traffic risk accident influencing factors;
step 300: constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor;
step 400: and optimizing the traffic accident risk probability of the expressway according to the accident risk thermodynamic diagram.
Further, as shown in fig. 2, the construction method of the traffic risk accident risk prediction model is as follows:
acquiring traffic flow data;
calculating the current traffic flow accident risk by using a traffic flow accident risk detection algorithm;
constructing a self-variable data set according to the traffic flow data;
performing dimension reduction processing on the self-variable data set to obtain a training data set;
and constructing the traffic risk accident risk prediction model according to the training data set and the traffic flow accident risk based on the XGBoost model.
Specifically, taking the traffic flow accident risk as a dependent variable, taking the traffic flow section parameter as an independent variable, training a machine learning model based on XGBoost, taking the traffic flow section parameter as an input variable, performing data standardization processing, establishing an original data set by using the traffic flow accident risk, and performing dimension reduction processing on the independent variable by adopting a variable importance ordering method carried by the XGBoost to obtain a training data set. The variable importance ranking is shown in figure 3.
Further, the acquiring traffic flow data includes:
acquiring microscopic traffic simulation software and a group driving simulation platform;
constructing an interactive simulation experiment platform according to the microscopic traffic simulation software and the group driving simulation platform;
and acquiring traffic flow data by using the interactive simulation experiment platform.
Specifically, a three-dimensional road scene of the expressway is designed, the three-dimensional road scene of the expressway is built based on a group driving simulation platform, and meanwhile, the scene is also imported into micro traffic simulation software;
establishing an interactive simulation experiment platform based on the microscopic traffic simulation software and the group driving simulation platform, wherein the microscopic traffic simulation software provides microscopic traffic flow and vehicles around driving vehicles for the experiment platform, and the group driving simulation platform provides a plurality of simulated vehicles controlled by different drivers for the experiment platform; loading background traffic and simulating vehicles to the three-dimensional road scene based on the interactive traffic simulation platform; the interactive simulation platform is shown in fig. 4. And after the experimental platform is started, the simulated traffic flow data is exported in real time, and the current traffic flow accident risk is calculated by using a traffic flow accident risk detection algorithm.
Specifically, the accident occurrence definition formula in the traffic flow accident risk detection algorithm is as follows:
wherein TTC is i I the collision time of the vehicle relative to the front vehicle at the time t, because the vehicle position acquired by data is the position of the vehicle head, X is the position of the vehicle head i (t) is the position of the head of the vehicle at the moment i, X h (t) is the position of the front h head of the front car at the moment i, l h Is the length of the body of the h car, V i (t) is the instantaneous speed of the i-car at time t, V h And (t) is the instantaneous speed of the vehicle at time t.
Further, as shown in fig. 5, the obtaining traffic flow data by using the interactive simulation experiment platform includes:
acquiring the section flow, the speed and the occupancy of the upstream of the specific position and the section flow, the speed and the occupancy of the downstream of the specific position by using the interactive simulation experiment platform;
and acquiring traffic flow data according to the upstream section flow, the vehicle speed and the occupancy rate and the downstream section flow, the vehicle speed and the occupancy rate.
Specifically, the traffic flow data comprises 200m, 800m, 1400m and 2000m upstream of a specific position, and 400m and 1000m downstream of the specific position, wherein the statistical time granularity is 5min, and the statistical time is 5min, 10min, 15min, 20min, 25min and 30min before each dangerous driving behavior, and the characteristic parameters comprise an average value and variance, namely 216 traffic flow characteristic parameters. The traffic flow parameter acquisition location is shown in fig. 6.
Further, after obtaining the traffic flow data, the method further comprises:
preprocessing the traffic flow data, and filtering invalid data in the traffic flow data to obtain first traffic flow data;
and correcting the first traffic flow data, and normalizing the parameter characteristic indexes in the first traffic flow data.
Specifically, invalid data in the simulated traffic flow data is filtered to obtain first traffic flow data, the first traffic flow data is corrected, normalization processing is carried out on each characteristic index, and all the characteristic indexes are normalized to be in the interval of [0,1 ].
Further, the method further comprises the following steps:
analyzing the traffic risk accident influence factors by using the traffic risk accident risk prediction model to obtain importance ranking of the influence factors;
and carrying out real-time reconstruction updating on the traffic risk accident risk prediction model according to the importance sorting of the influencing factors.
Specifically, the traffic flow accident risk is taken as a dependent variable, the traffic flow section parameter after the dimension reduction treatment is taken as the independent variable, and the XGBoost-based machine learning model is retrained. The traffic flow accident independent variables after dimension reduction comprise:
smaanL 4T6, svarL4T6, fvarL4T4, fmeanL4T4, smaanL 4T5, svarL3T5, svarL5T6, omeanL6T1, ovarL6T1, fvarL4T3, fvarL4T5, omeanL5T6, svarL1T4, fmeanL4T1, ovar L5T6, variable descriptions are shown in Table 1. The prediction model confusion matrix is shown in table 2. Tables 1 and 2 are as follows:
table 1 variable description table
Table 2 prediction model confusion matrix table
Further, the constructing an accident risk thermodynamic diagram of each traffic risk accident influence factor interaction according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor comprises:
the traffic risk accident influencing factors are processed in pairs, and the traffic risk accident influencing factors of every two pairs are obtained;
the traffic risk accident influencing factors of the pairs are amplified or reduced in the same proportion, and corresponding accident risk values are calculated by using a traffic risk accident risk prediction model;
and constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the accident risk values.
Specifically, based on the SHAP model, the interaction relation among important influence factors of traffic flow accident risks is analyzed, an accident risk thermodynamic diagram of the interaction of the variables is constructed, and the expressway traffic accident risk probability is optimized based on the accident risk thermodynamic diagram. Taking traffic flow section parameters smeanL4T6, svarL4T6, fvarL4T4, fmeanL4T4, smeanL4T5, svarL3T5, svarL5T6, omeanL6T1, ovarL6T1, fvarL4T3, fvarL4T5, omeanL5T6, svarL1T4, fmeanL4T1, ovarL5T6 after the dimension reduction treatment as input variables of the SHAP model;
and comparing the SHAP value of each variable with the SHAP values of all the variables in the data set, and analyzing the interaction relation between important influence factors of traffic flow accident risk to obtain the dependency between the interaction variables, wherein the interaction relation between the two variables of svarL4T6 and fvarL4T4 is shown in fig. 6.
The interactive variables are processed in pairs respectively, and the two variables of the svarL4T6 and the fvarL4T4 are respectively increased and decreased, wherein the increasing proportion of the svarL4T6 is 0.05, 0.10, 0.15, 0.20, 0.25, 0.30, 0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75, 0.80, 0.85, 0.90, 0.95 and 1.00, and the decreasing proportion of the fvarL4T4 is 0.05, 0.10, 0.15, 0.20, 0.25, 0.30, 0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75, 0.80, 0.85, 0.90, 0.95 and 1.00.
And calculating the real-time accident risk value of the expressway after variable adjustment to be 0.20 by using the trained XGBoost accident prediction model after the variable is increased by 0.05 times by using the variable which is increased and the variable is reduced by 0.10 times by using the svarL4T6 for example. Whereas the accident risk value of svarL4T6 corresponds to the original value of fvarL4T4 is 0.22. In this way, a thermodynamic diagram of the traffic accident risk values at all adjustment scales is plotted as shown in fig. 7.
Based on real-time accident risk values of the expressway after adjustment of all relevant variables, drawing accident risk thermodynamic diagrams of interaction of all relevant two variables, comparing the accident risk of each adjustment proportion of the two variables with the base point value by taking the accident risk of the two variables before adjustment as the base point, and obtaining the traffic accident risk optimization direction of the expressway according to the variable trend of the variable adjustment proportion of the accident risk values. Taking two variables shown in fig. 7-8 as examples, the bluer the color is, the lower the accident risk probability is, the lower right corner direction of the thermodynamic diagram is the adjustment optimization direction of svarL4T6 and fvarL4T4, so that the value of fvarL4T4 is reduced while the value of svarL4T6 is required to be increased, and the accident risk optimization direction facing the two variables is obtained, so that the accident risk optimization direction under the condition of interaction relation of all related variables can be obtained by the method.
The beneficial effects of the invention are as follows:
according to the method, an XGBoost machine learning model is utilized, and a variable importance analysis function in the XGBoost machine learning model is adopted to conduct importance analysis on important influence factors of traffic flow accident risks, so that the importance scores of the influence factors affecting the traffic accident risk probability are obtained, the problem that the importance of the influence factors is difficult to sort through traffic accident cause analysis is solved, and therefore the problem that a traditional research means does not have a cut-in point in the traffic accident risk optimization direction is solved.
The invention is based on SHAP model, analyzes the interactive relation between important influencing factors of traffic flow accident risk, SHAP belongs to the method of model post interpretation, its core idea is to calculate the marginal contribution of the feature to model output, then explain the 'black box model' from the global and local two layers, through constructing an additive interpretation model, all features are regarded as 'contributor'. The accident risk thermodynamic diagram of two-by-two interaction is utilized to obtain the optimization direction of the traffic accident risk of the expressway, the problem that the black box of the accident prediction model is difficult to search the optimization direction in the traffic accident risk control is solved, and meanwhile, the interaction optimization method is provided. The invention has the characteristics of replicable popularization and strong robustness. In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (6)
1. The highway accident risk optimization method based on the interactive self-interpretation model is characterized by comprising the following steps of:
constructing a traffic risk accident risk prediction model;
based on the traffic risk accident influencing factors, according to the SHAP model, obtaining interaction relations among the traffic risk accident influencing factors;
constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor;
optimizing the traffic accident risk probability of the expressway according to the accident risk thermodynamic diagram;
the construction method of the traffic risk accident risk prediction model comprises the following steps:
acquiring traffic flow data;
calculating the current traffic flow accident risk by using a traffic flow accident risk detection algorithm;
constructing a self-variable data set according to the traffic flow data;
performing dimension reduction processing on the self-variable data set to obtain a training data set;
based on an XGBoost model, constructing a traffic risk accident risk prediction model according to the training data set and the traffic flow accident risk;
constructing an accident risk thermodynamic diagram of each traffic risk accident influence factor interaction according to the traffic risk accident risk prediction model and the interaction relation between each traffic risk accident influence factor, wherein the accident risk thermodynamic diagram comprises the following components:
the traffic risk accident influencing factors are processed in pairs, and the traffic risk accident influencing factors of every two pairs are obtained;
the traffic risk accident influencing factors of the pairs are amplified or reduced in the same proportion, and corresponding accident risk values are calculated by using a traffic risk accident risk prediction model;
constructing accident risk thermodynamic diagrams of each traffic risk accident influence factor in a pairwise interaction mode according to the accident risk values;
and comparing the accident risk of the two variables in the accident risk thermodynamic diagram before adjustment with the base point value according to the accident risk value and the variable trend of the variable adjustment proportion according to the accident risk value to obtain the expressway traffic accident risk optimization direction.
2. The method for optimizing highway accident risk based on the interactive self-explanatory model according to claim 1, wherein the acquiring traffic flow data comprises:
acquiring microscopic traffic simulation software and a group driving simulation platform;
constructing an interactive simulation experiment platform according to the microscopic traffic simulation software and the group driving simulation platform;
and acquiring traffic flow data by using the interactive simulation experiment platform.
3. The expressway accident risk optimization method based on the interactive self-explanatory model according to claim 1, wherein the accident occurrence definition formula in the traffic flow accident risk detection algorithm is as follows:
；
the collision time of the i vehicle relative to the front vehicle at the time t is the position of the vehicle head at the time t, and the position of the vehicle head at the time t is the position of the vehicle head at the time t, the vehicle body length of the h vehicle, the instantaneous speed of the vehicle at the time t, and the instantaneous speed of the h vehicle at the time t.
4. The method for optimizing the risk of an accident on an expressway based on an interactive self-explanatory model according to claim 2, wherein said obtaining traffic flow data using the interactive simulation experiment platform comprises:
acquiring the section flow, the speed and the occupancy of the upstream of the specific position and the section flow, the speed and the occupancy of the downstream of the specific position by using the interactive simulation experiment platform;
and acquiring traffic flow data according to the upstream section flow, the vehicle speed and the occupancy rate and the downstream section flow, the vehicle speed and the occupancy rate.
5. The method for optimizing highway accident risk based on the interactive self-explanatory model according to claim 1, wherein the step of obtaining traffic flow data further comprises the steps of:
preprocessing the traffic flow data, and filtering invalid data in the traffic flow data to obtain first traffic flow data;
and correcting the first traffic flow data, and normalizing the parameter characteristic indexes in the first traffic flow data.
6. The method for optimizing highway accident risk based on the interactive self-explanatory model according to claim 1, further comprising:
analyzing the traffic risk accident influence factors by using the traffic risk accident risk prediction model to obtain importance ranking of the influence factors;
and carrying out real-time reconstruction updating on the traffic risk accident risk prediction model according to the importance sorting of the influencing factors.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310903318.XA CN116959248B (en) | 2023-07-22 | 2023-07-22 | Expressway accident risk optimization method based on interactive self-interpretation model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310903318.XA CN116959248B (en) | 2023-07-22 | 2023-07-22 | Expressway accident risk optimization method based on interactive self-interpretation model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116959248A CN116959248A (en) | 2023-10-27 |
| CN116959248B true CN116959248B (en) | 2024-04-05 |
Family
ID=88454248
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310903318.XA Active CN116959248B (en) | 2023-07-22 | 2023-07-22 | Expressway accident risk optimization method based on interactive self-interpretation model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116959248B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101826258A (en) * | 2010-04-09 | 2010-09-08 | 北京工业大学 | Method for predicting simple accidents on freeways |
| CN102360525A (en) * | 2011-09-28 | 2012-02-22 | 东南大学 | Discriminant analysis-based high road real-time traffic accident risk forecasting method |
| CN106355883A (en) * | 2016-10-20 | 2017-01-25 | 同济大学 | Risk evaluation model-based traffic accident happening probability acquiring method and system |
| CN109118773A (en) * | 2018-09-30 | 2019-01-01 | 中交第公路勘察设计研究院有限公司 | A kind of traffic accidents methods of risk assessment |
| CN110120151A (en) * | 2019-04-24 | 2019-08-13 | 华南理工大学 | A kind of On-Ramp on Freeway merging area vehicle collision probability forecasting method |
| KR20210086381A (en) * | 2019-12-30 | 2021-07-08 | 계명대학교 산학협력단 | System and Method for Predicting Traffic Accident Risk |
-
2023
- 2023-07-22 CN CN202310903318.XA patent/CN116959248B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101826258A (en) * | 2010-04-09 | 2010-09-08 | 北京工业大学 | Method for predicting simple accidents on freeways |
| CN102360525A (en) * | 2011-09-28 | 2012-02-22 | 东南大学 | Discriminant analysis-based high road real-time traffic accident risk forecasting method |
| CN106355883A (en) * | 2016-10-20 | 2017-01-25 | 同济大学 | Risk evaluation model-based traffic accident happening probability acquiring method and system |
| CN109118773A (en) * | 2018-09-30 | 2019-01-01 | 中交第公路勘察设计研究院有限公司 | A kind of traffic accidents methods of risk assessment |
| CN110120151A (en) * | 2019-04-24 | 2019-08-13 | 华南理工大学 | A kind of On-Ramp on Freeway merging area vehicle collision probability forecasting method |
| KR20210086381A (en) * | 2019-12-30 | 2021-07-08 | 계명대학교 산학협력단 | System and Method for Predicting Traffic Accident Risk |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116959248A (en) | 2023-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106895975B (en) | Bearing fault diagnosis method based on Stacked SAE deep neural network | |
| CN108764137A (en) | Vehicle traveling lane localization method based on semantic segmentation | |
| CN106203330A (en) | A kind of vehicle classification method based on convolutional neural networks | |
| DE102018108947A1 (en) | Apparatus for correcting an utterance error of a user and method thereof | |
| EP4145345A1 (en) | Data annotation method and device, and fine granularity identification method and device | |
| CN111767790B (en) | Chromatographic peak identification method based on convolutional neural network | |
| CN110929762B (en) | Limb language detection and behavior analysis method and system based on deep learning | |
| CN101964063A (en) | Method for constructing improved AdaBoost classifier | |
| DE112013002654T5 (en) | Method for classifying text | |
| CN105825241A (en) | Driver braking intention identification method based on fuzzy neural network | |
| CN111353505A (en) | Network model capable of realizing semantic segmentation and depth of field estimation jointly and training method | |
| CN112966709B (en) | Deep learning-based fine vehicle type identification method and system | |
| DE102020207537A1 (en) | State diagnostic apparatus and method of a movable system component | |
| CN114913493A (en) | Lane line detection method based on deep learning | |
| CN116630932A (en) | Road shielding target detection method based on improved YOLOV5 | |
| CN114528646B (en) | Vehicle running condition construction method based on online map data | |
| CN113886562A (en) | AI resume screening method, system, equipment and storage medium | |
| CN115761834A (en) | Multi-task mixed model for face recognition and face recognition method | |
| CN116959248B (en) | Expressway accident risk optimization method based on interactive self-interpretation model | |
| CN117765480A (en) | Method and system for early warning migration of wild animals along road | |
| CN113642114A (en) | Modeling method for humanoid random car following driving behavior capable of making mistakes | |
| DE102023130154A1 (en) | Test procedure of actual roads for automated driving system, device and storage medium | |
| CN111414511A (en) | Automatic voiceprint modeling and warehousing method, device and equipment | |
| CN115774942A (en) | Driving style identification model modeling and statistical method based on Internet of vehicles real vehicle data and SVM | |
| CN115049991A (en) | Shared device pose uniformity judging method, device, terminal and 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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Song Hao Inventor after: Wang Junhua Inventor after: ShangGuan Qiangqiang Inventor before: Song Hao Inventor before: Wang Junhua |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |