WO2024060588A1 - 高速公路隧道事故实时风险预测方法、装置、设备及介质 - Google Patents

高速公路隧道事故实时风险预测方法、装置、设备及介质 Download PDF

Info

Publication number
WO2024060588A1
WO2024060588A1 PCT/CN2023/087326 CN2023087326W WO2024060588A1 WO 2024060588 A1 WO2024060588 A1 WO 2024060588A1 CN 2023087326 W CN2023087326 W CN 2023087326W WO 2024060588 A1 WO2024060588 A1 WO 2024060588A1
Authority
WO
WIPO (PCT)
Prior art keywords
accident
data
tunnel
import
traffic flow
Prior art date
Application number
PCT/CN2023/087326
Other languages
English (en)
French (fr)
Inventor
黄合来
金杰灵
邹国庆
姚广
李烨
周波
许星伟
李永汉
戴剑军
Original Assignee
中南大学
湖南纽狐科技有限公司
湖南省交通科学研究院有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中南大学, 湖南纽狐科技有限公司, 湖南省交通科学研究院有限公司 filed Critical 中南大学
Publication of WO2024060588A1 publication Critical patent/WO2024060588A1/zh

Links

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
    • 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
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • 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
    • G08G1/048—Detecting movement of traffic to be counted or controlled with provision for compensation of environmental or other condition, e.g. snow, vehicle stopped at detector
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count

Definitions

  • This application belongs to the field of traffic safety technology, and particularly relates to a real-time risk prediction method, device, equipment and medium for highway tunnel accidents.
  • Tunnels are regarded as important sections of highways because of their important role in reducing travel time, protecting the ecological environment, and improving highway operations. With the gradual development of infrastructure construction, the mileage of highways is increasing. The number of tunnels has increased significantly over the past few years and continues to grow.
  • tunnels since tunnels usually have less lighting than open roads and have a semi-enclosed spatial structure, they may affect the behavior of drivers, making it easier to cause traffic conflicts and further lead to collision accidents. Therefore, research aimed at improving tunnel traffic safety is of certain significance to improving the safety level of highway systems.
  • the embodiments of the present application provide a method, device, equipment and medium for real-time risk prediction of highway tunnel accidents, which can solve the problem of being unable to predict the risk of highway tunnel accidents in real time.
  • embodiments of this application provide a real-time risk prediction method for highway tunnel accidents, including:
  • n is an integer, and 5 ⁇ n ⁇ 15;
  • the target import and export traffic flow data and weather data corresponding to the accident are fused and processed to obtain the accident fusion data of the accident, and the comparison data of the accident fusion data is obtained;
  • the tunnel accident real-time risk prediction model was trained, and the trained tunnel accident real-time risk prediction model was obtained;
  • the import and export traffic flow data of the target tunnel and the weather data of the area where the target tunnel is located are collected every n minutes;
  • the import and export traffic flow data collected each time are the import and export traffic flow data within n minutes before the collection time, and the weather data collected each time
  • the data is weather data within n minutes before the collection time;
  • the accident precursor characteristic data is extracted from the collected import and export traffic flow data and weather data, and based on the accident precursor characteristics
  • the real-time risk prediction model of tunnel accidents after data and training predicts the risk of tunnel accidents.
  • the accident precursor characteristics that affect the risk of tunnel accidents are determined based on the calculated impact values.
  • select the target variable from the defined variables based on the calculated correlation value including:
  • the correlation coefficient between the two variables corresponding to the correlation value and whether the accident occurred is calculated separately, and the calculated correlation coefficient is The variable corresponding to the smaller one is deleted;
  • determine the accident precursor characteristics that affect the risk of tunnel accidents based on the calculated impact values including:
  • Target variables with impact values less than or equal to the preset impact value and target variables corresponding to weather data are used as accident precursor features that affect tunnel accident risks.
  • import and export traffic flow data include: import flow, average import occupancy, average import speed, export flow, average export occupancy, and average export speed.
  • comparison data of accident fusion data including:
  • embodiments of the present application provide a real-time risk prediction device for highway tunnel accidents, including:
  • the first acquisition module is used to acquire accident data and import and export traffic flow data of the target tunnel within a preset historical time period
  • the data matching module is used to match the target entrance and exit traffic flow data within n minutes before the accident from the entrance and exit traffic flow data for each accident in the accident data, and determine the area where the target tunnel is located n before the accident.
  • the second acquisition module is used to fuse the target import and export traffic flow data and weather data corresponding to the accident for each accident, obtain the accident fusion data of the accident, and obtain the comparison data of the accident fusion data;
  • the first determination module is used to use the accident fusion data and control data corresponding to all accidents in the accident data as experimental data sets;
  • the second determination module is used to determine the accident precursor characteristics that affect the risk of tunnel accidents based on the experimental data set;
  • the model training module is used to train the tunnel accident real-time risk prediction model based on the experimental data set, with the accident precursor characteristics as the independent variable and whether the accident occurred as the dependent variable, and obtain the trained tunnel accident real-time risk prediction model;
  • the data collection module is used to collect the import and export traffic flow data of the target tunnel every n minutes and the weather data of the area where the target tunnel is located; the import and export traffic flow data collected each time is the import and export traffic flow data within n minutes before the collection time. , the weather data collected each time is the weather data within n minutes before the collection time;
  • the risk prediction module is used to extract accident precursor characteristic data from the collected import and export traffic flow data and weather data each time based on the independent variables of the tunnel accident real-time risk prediction model. , and predict tunnel accident risks based on accident precursor characteristic data and the trained tunnel accident real-time risk prediction model.
  • embodiments of the present application provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor.
  • the processor executes the computer program, the above-mentioned real-time detection of highway tunnel accidents is realized. Risk prediction methods.
  • embodiments of the present application provide a computer-readable storage medium.
  • the computer-readable storage medium stores a computer program.
  • the computer program is executed by a processor, the above-mentioned real-time risk prediction method for highway tunnel accidents is implemented.
  • the corresponding information for each accident that occurred in the tunnel within the preset historical time period is obtained.
  • Import and export traffic flow data and weather data and use the import and export traffic flow data and weather data corresponding to each accident as the accident fusion data of the accident; then for each accident that occurred in the tunnel within the preset historical time period, obtain the The control data corresponding to the accident fusion data of the accident, and the accident fusion data and control data corresponding to all accidents that occurred within the preset historical time period are used as the experimental data set; then the experimental data set is analyzed to determine the risk of tunnel accidents Accident precursor characteristics that have a significant impact, and based on this experimental data set, with accident precursor characteristics as independent variables and whether the accident occurred as a dependent variable, a real-time risk prediction model for tunnel accidents was trained, and the trained real-time risk prediction model for tunnel accidents was obtained; finally Based on the real-time collected import and export traffic flow
  • Figure 1 is a flow chart of a real-time risk prediction method for highway tunnel accidents provided by an embodiment of the present application
  • FIG2 is a schematic diagram of a ROC curve in a test result of the present application.
  • Figure 3 is a schematic structural diagram of a real-time risk prediction device for highway tunnel accidents provided by an embodiment of the present application
  • Figure 4 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.
  • the term “if” may be interpreted as “when” or “once” or “in response to determining” or “in response to detecting” depending on the context. ". Similarly, the phrase “if determined” or “if [the described condition or event] is detected” may be interpreted, depending on the context, to mean “once determined” or “in response to a determination” or “once the [described condition or event] is detected ]” or “in response to detection of [the described condition or event]”.
  • references to "one embodiment” or “some embodiments” etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements “in one embodiment”, “in some embodiments”, “in some other embodiments”, “in some other embodiments”, etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean “one or more but not all embodiments", unless otherwise specifically emphasized in other ways.
  • the terms “including”, “comprising”, “having” and their variations all mean “including but not limited to”, unless otherwise specifically emphasized in other ways.
  • embodiments of this application provide a real-time risk prediction method for highway tunnel accidents. This method uses the accident data, entrances and exits of the tunnel within a preset historical time period.
  • the accident precursor characteristics are used as independent variables, accident Whether it occurs is used as the dependent variable to train the real-time risk prediction model of tunnel accidents and obtain the trained real-time risk prediction model of tunnel accidents; finally, real-time prediction is based on the real-time collected import and export traffic flow data, weather data and the trained real-time risk prediction model of tunnel accidents.
  • Tunnel accident risk realize real-time prediction of highway tunnel accident risk, and then provide reference for tunnel accident prevention measures, so as to intervene in advance to improve the driving safety of vehicles in tunnels.
  • the embodiments of the present application provide a real-time risk prediction method for highway tunnel accidents.
  • This method can be executed by a terminal device or by a device (such as a chip) used in the terminal device.
  • the following embodiments use this method by Take terminal device execution as an example.
  • the terminal device may be a tablet, a server, a laptop, etc., which is not limited in the embodiments of the present application.
  • the real-time risk prediction method for highway tunnel accidents includes the following steps:
  • Step 11 Obtain the accident data and import and export traffic flow data of the target tunnel within the preset historical time period.
  • the above-mentioned target tunnel is a tunnel that requires real-time risk prediction of accidents;
  • the above-mentioned preset historical time period can be set according to the actual situation, for example, set to the past year;
  • the above-mentioned accident data includes the accidents that occurred in the target tunnel within the preset historical time period.
  • the above import and export traffic flow data includes: import flow, average import occupancy, average import speed, export flow, average export occupancy, and average export Vehicle speed.
  • the inlet flow refers to the number of vehicles passing the detector (the detector is installed at the entrance of the target tunnel) per unit time, the unit is: vehicles/hour, for example, the number of vehicles passing the detector within 5 minutes is N, the import flow
  • the average entrance occupancy rate refers to the ratio of the time T v when a vehicle exists in a certain detection section or detection area at the entrance of the target tunnel to the total statistical time T (such as 5 minutes);
  • the average entrance speed refers to the vehicle speed passing through the detector per unit time (The detector is installed at the entrance of the target tunnel)
  • the average vehicle speed, the unit is kilometers/hour;
  • the exit flow refers to the detector (the detector is installed at the target tunnel) per unit time (detector at the exit), the unit is: vehicles/hour, for example, the number of vehicles passing the detector within 5 minutes is N, and the exit flow
  • the average exit occupancy rate refers to the ratio of the time T v when a vehicle exists in a certain detection section or detection area at the target tunnel exit to the
  • the above accident data and import and export traffic flow data can be obtained from the website of the relevant transportation department. It is understandable that after obtaining accident data and import and export traffic flow data from relevant websites, recorded data with missing information and misjudged accidents can be deleted to ensure the integrity of the accident data in step 11.
  • Step 12 for each accident in the accident data, respectively, match the target import and export traffic flow data within n minutes before the accident from the import and export traffic flow data, and determine the weather data of the area where the target tunnel is located within n minutes before the accident.
  • n is an integer, and 5 ⁇ n ⁇ 15. It should be noted that since traffic flow monitoring is generally done at intervals of 5 minutes, the value of n can generally be 5, 10 or 15.
  • the above target entrance and exit traffic flow data is the entry and exit in step 11.
  • Part of the oral traffic flow data specifically the import and export traffic flow data within n minutes before the accident.
  • the import and export traffic flow data obtained in step 11 can be matched to the import and export traffic flow data within n minutes before the accident occurred.
  • weather data of the target tunnel area within n minutes before the accident can be obtained from the relevant website of the meteorological department, and the weather data can be sunny, cloudy, rainy, snowy, etc. Specifically, the weather data of the target tunnel area within n minutes before the accident can be found according to the time and place of the accident in the recorded data.
  • Step 13 For each accident, the target import and export traffic flow data and weather data corresponding to the accident are fused to obtain the accident fusion data of the accident, and the comparison data of the accident fusion data is obtained.
  • the target entrance and exit traffic flow data and weather data corresponding to the accident can be fused according to the time and location of the accident. Processing to obtain accident fusion data, which facilitates the determination of control data and enriches subsequent experimental data sets, which in turn helps to subsequently improve the training accuracy of the real-time risk prediction model for tunnel accidents.
  • the target import and export traffic flow data and weather data corresponding to the accident are fused.
  • the obtained accident fusion data includes: import flow, average import occupancy, average import speed, export flow, average export occupancy, average export speed,
  • the difference in import and export flow i.e. the difference between import flow and export flow
  • the difference in import and export occupancy i.e. the difference between the average import occupancy and the average export occupancy
  • the import and export speed difference i.e. the average import speed and the average export speed
  • weather data i.e. the average import speed and the average export speed
  • the control data of the accident fusion data can be obtained through a case control method.
  • the comparison data can be obtained as follows: first, use the case comparison method to obtain the entrance and exit traffic flow data of the target tunnel at the preset relevant time, and obtain the weather data of the area where the target tunnel is located at the preset relevant time; and then use The import and export traffic flow data and weather data obtained by the case comparison method are fused, and the fused data are used as control data for the accident fusion data.
  • the comparison data includes the inlet flow, average inlet occupancy, average inlet speed, outlet flow, average outlet occupancy, average outlet speed, and inlet and outlet flow difference (i.e., the difference between inlet flow and outlet flow) of the target tunnel at the preset relevant time. value), import and export share rate difference (that is, the difference between the average import share and the average export share), the import and export speed difference (that is, the difference between the average import speed and the average export speed) and the corresponding weather data.
  • the above-mentioned preset relevant time is related to the time of occurrence of the accident.
  • the data of the first 7 days, 14 days, last 7 days and 14 days of the accident location and period can be selected as control data.
  • the recorded data of the accident and the target entrance and exit traffic flow data and weather data corresponding to the accident can be fused. That is, the above-mentioned accident fusion data also includes recorded data.
  • the obtained comparison data also includes the recorded data of the accident.
  • Step 14 Use the accident fusion data and control data corresponding to all accidents in the accident data as the experimental data set.
  • data may be fused according to the time and location of the accident, and the accident fused data and control data may be combined into an experimental data set.
  • Step 15 Based on the experimental data set, determine the accident precursor characteristics that affect the risk of tunnel accidents.
  • the Spearman correlation coefficient and the conditional logistic regression model can be used to select accident precursor characteristics that have a significant impact on the risk of tunnel accidents, thereby facilitating subsequent real-time and accurate tunnel accident detection. Risk prediction.
  • Step 16 Based on the experimental data set, using the accident precursor characteristics as the independent variable and whether the accident occurred as the dependent variable, train the tunnel accident real-time risk prediction model to obtain the trained tunnel accident real-time risk prediction model.
  • the above experimental data set can be divided into training set data and test set data.
  • Set data then use the training set data to train the tunnel accident real-time risk prediction model, use the test set data to test the tunnel accident real-time risk prediction model, and obtain the trained tunnel accident real-time risk prediction model.
  • Step 17 Collect the entrance and exit traffic flow data of the target tunnel and the weather data of the area where the target tunnel is located every n minutes.
  • the import and export traffic flow data collected each time is the import and export traffic flow data within n minutes before the collection time (the import and export traffic flow data includes: import flow, average import occupancy, average import speed, export flow, average export occupancy) and the average vehicle speed at the exit).
  • the weather data collected each time is the weather data within n minutes before the collection time (the weather data can be sunny, cloudy, rainy, snowy, etc.).
  • Step 18 For each collected import and export traffic flow data and weather data, according to the independent variables of the tunnel accident real-time risk prediction model, extract the accident precursor characteristic data from the collected import and export traffic flow data and weather data, and based on The accident precursor characteristic data and the trained tunnel accident real-time risk prediction model predict the tunnel accident risk.
  • the above-mentioned real-time risk prediction model for tunnel accidents may be a commonly used decision function based on support vector machines.
  • the expression of the decision function can be:
  • f(x) represents the dependent variable, that is, whether the accident occurred, is the support vector
  • b * is the offset
  • b * can be calculated by the Karush-Kuhn-Tucker condition
  • K is the kernel function
  • x i is the data of the accident precursor characteristics of the i-th sample in the training set data
  • x is the need to predict
  • the data of accident precursor characteristics of the tunnel i.e., the accident precursor characteristic data in step 18
  • n is the number of samples in the training set data.
  • the method provided by the embodiment of the present application analyzes the accident data, import and export traffic flow data and related weather data of the tunnel within a preset historical time period to obtain the import and export traffic within n minutes before the accident.
  • flow data and weather data and use the import and export traffic flow data and weather data corresponding to each accident as the accident fusion data of the accident; then for each accident that occurred in the tunnel within the preset historical time period, obtain the accident data of the accident.
  • the control data corresponding to the fusion data, and the accident fusion data and control data corresponding to all accidents that occurred within the preset historical time period are used as experimental data sets; Then the experimental data set was analyzed to determine the accident precursor characteristics that have a significant impact on tunnel accident risks.
  • the accident precursor characteristics were used as independent variables and whether the accident occurred as the dependent variable to train the real-time risk of tunnel accidents.
  • Prediction model to obtain the trained real-time risk prediction model of tunnel accidents; finally, based on the real-time collected import and export traffic flow data, weather data and the trained real-time risk prediction model of tunnel accidents, real-time prediction of whether a traffic accident occurs in the tunnel is realized, so as to realize the prediction of highway tunnels.
  • Real-time prediction of accident risks can provide reference for tunnel accident prevention measures (such as traffic flow, speed control, etc.), so as to intervene in advance to improve the driving safety of vehicles in tunnels.
  • step 15 The specific implementation process of the above step 15 will be exemplified below with reference to specific embodiments.
  • step 15 above based on the experimental data set, to determine the accident precursor characteristics that affect the risk of tunnel accidents includes the following steps:
  • Step 1 For each type of data in the experimental data set, define a variable for this type of data, and use this type of data as the data of the variable.
  • the above experimental data set includes ten categories: import flow, average import occupancy, average import speed, export flow, average export occupancy, average export speed, import and export flow difference, import and export occupancy difference, import and export speed difference, and weather data.
  • a variable needs to be defined for each type of data in the ten types of data, and this type of data is used as the data of the corresponding variable.
  • the inlet flow as an example. Assume that the variable defined for the inlet flow is t, then all the inlet flow data in the experimental data set is the value of the variable t.
  • Step 2 Use Spearman correlation coefficient to calculate the correlation value between the defined variables.
  • the data of the two variables can be combined and the Spearman correlation coefficient can be used to calculate the correlation value of the two variables.
  • the expression of the above Spearman correlation coefficient is: S represents the calculated correlation value, n represents the number of observations (that is, the total number of data contained in the accident fusion data and the control data), and di represents the difference in the position of the paired variables after sorting the two variables. It should be noted that since the Spearman correlation coefficient is a commonly used correlation analysis method, the principle of the Spearman correlation coefficient will not be described in too much detail here.
  • Step 3 Select the target variable from the defined variables based on the calculated correlation value.
  • the accuracy of features requires deleting features that are weakly correlated with whether an accident occurs after calculating the correlation value between each two variables.
  • the correlation value i.e., Spearman correlation coefficient value
  • the preset correlation value such as 0.7
  • the correlation coefficient between the two variables corresponding to the correlation value and whether the accident occurred i.e., the Spearman correlation coefficient value
  • delete the variable corresponding to the smaller of the calculated correlation coefficients and then define All variables except the deleted variable are used as target variables.
  • the correlation coefficient between inlet flow and outlet flow is 0.85
  • the correlation coefficient between inlet flow and accident factors ie, whether the accident occurs
  • the correlation coefficient between outlet flow and accident factors ie, whether the accident occurs
  • the correlation coefficient is 0.3, then the variables corresponding to the export flow are deleted and the variables corresponding to the import flow are retained.
  • the correlation coefficient between the above variables and whether the accident occurs can also be calculated through the Spearman correlation coefficient. It can be understood that in order to facilitate the calculation of the correlation coefficient between the variable and whether the accident occurred, in actual processing, a variable can be defined for whether the accident occurred (the value of this variable is 0 or 1, 0 means no accident occurred, 1 Indicates that an accident occurred), the value of this variable in the accident fusion data is 1, and the value of this variable in the control data is 1 or 0.
  • Step 4 Use the conditional logistic regression model to calculate the impact value of each target variable on the risk of tunnel accidents.
  • Step 5 Determine the accident precursor characteristics that affect the tunnel accident risk based on the calculated impact value.
  • target variables with impact values less than or equal to a preset impact value (such as 0.05) and target variables corresponding to weather data can be used as accident precursor features that affect tunnel accident risks.
  • the import and export traffic flow data and weather data need to be analyzed to extract the information for each accident.
  • the data corresponding to the precursor features i.e., the accident precursor feature data
  • the trained tunnel accident real-time risk prediction model based on the extracted accident precursor feature data to perform tunnel accident risk prediction.
  • 187 tunnel traffic records of 15 tunnels in California in 2018 are first collected. accidents, deleting some missing information and misjudged accidents, and obtained a total of 180 valid accident data.
  • the average import speed, average export occupancy, average export speed, import and export flow difference, import and export occupancy difference, The import and export speed difference and weather are used as target variables; finally, the conditional logistic regression model is used to calculate the impact value of each target variable on the tunnel accident risk, and the impact value is less than or equal to the preset impact value (the preset impact value here is 0.05)
  • the target variables here, average import speed, average export occupancy, import and export flow difference, import and export occupancy difference
  • the influence values calculated using the conditional logistic regression model are shown in Table 4. Estimate represents the correlation strength between the target variable and the tunnel accident risk, and Pr(>
  • the test results are shown in Table 5, and the ROC curve is shown in Figure 2.
  • the F1 score is an indicator used in statistics to measure the accuracy of the binary classification model. It takes into account both the precision and recall rate of the classification model.
  • the ROC curve is the receiver operating characteristic curve. It should be noted that the "-" in Table 5 means that the item does not exist.
  • the total accuracy of the real-time risk prediction model for tunnel accidents is around 91%, indicating that the prediction accuracy of the real-time risk prediction model for tunnel accidents is high; as can be seen from Figure 2, the AUC value (the AUC value refers to the area under the ROC curve) is 0.97, which is close to 1, indicating that the real-time risk prediction model of tunnel accidents is highly realistic.
  • an embodiment of the present application provides a real-time risk prediction device for highway tunnel accidents.
  • the real-time risk prediction device 300 for highway tunnel accidents includes:
  • the first acquisition module 301 is used to acquire accident data and import and export traffic flow data of the target tunnel within a preset historical time period;
  • the data matching module 302 is used to match the target import and export traffic flow data within n minutes before the accident from the import and export traffic flow data for each accident in the accident data, and determine the weather data of the area where the target tunnel is located within n minutes before the accident; n is an integer, and 5 ⁇ n ⁇ 15;
  • the second acquisition module 303 is used to fuse the target import and export traffic flow data and weather data corresponding to the accident for each accident, obtain the accident fusion data of the accident, and obtain the comparison data of the accident fusion data;
  • a first determination module 304 is used to use the accident fusion data and control data corresponding to all accidents in the accident data as an experimental data set;
  • the second determination module 305 is used to determine the accident precursor characteristics that affect the risk of tunnel accidents based on the experimental data set;
  • the model training module 306 is used to train the real-time risk prediction model of tunnel accidents based on the experimental data set, using the characteristics of accident precursors as independent variables and whether the accident occurs as dependent variables, and obtain the trained real-time risk prediction model of tunnel accidents;
  • the data collection module 307 is used to collect the import and export traffic flow data of the target tunnel and the weather data of the area where the target tunnel is located every n minutes; the import and export traffic flow data collected each time is the import and export traffic flow within n minutes before the collection time. Data, the weather data collected each time is the weather data within n minutes before the collection time;
  • the risk prediction module 308 is used for each collected import and export traffic flow data and weather data, According to the independent variables of the tunnel accident real-time risk prediction model, the accident precursor characteristic data is extracted from the collected import and export traffic flow data and weather data, and the tunnel accident risk is predicted based on the accident precursor characteristic data and the trained tunnel accident real-time risk prediction model. .
  • Module completion means dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above.
  • Each functional unit and module in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.
  • the above-mentioned integrated unit can be hardware-based. It can also be implemented in the form of software functional units.
  • the specific names of each functional unit and module are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present application.
  • For the specific working processes of the units and modules in the above system please refer to the corresponding processes in the foregoing method embodiments, and will not be described again here.
  • an embodiment of the present application provides a terminal device.
  • the terminal device D10 of this embodiment includes: at least one processor D100 (only one processor is shown in Figure 4), a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100.
  • the processor D100 executes the computer program D102, the steps in any of the above-mentioned method embodiments are implemented.
  • the processor D100 executes the computer program D102, it obtains the tunnel's accident data, import and export traffic flow data and related weather data within the preset historical time period by analyzing the tunnel's accident data during the preset historical time period. Import and export traffic flow data and weather data corresponding to each accident that occurred within the period, and use the import and export traffic flow data and weather data corresponding to each accident as the accident fusion data of the accident; then for the tunnel within the preset historical time period For each accident that occurs, the control data corresponding to the accident fusion data of the accident is obtained, and the accident fusion data and control data corresponding to all accidents that occurred within the preset historical time period are used as the experimental data set; then the experimental data set is analysis to determine the accident precursor characteristics that have a significant impact on the risk of tunnel accidents, and based on this experimental data set, Therefore, the precursor characteristics are the independent variables and whether the accident occurs is the dependent variable.
  • the real-time risk prediction model of tunnel accidents is trained to obtain the trained real-time risk prediction model of tunnel accidents. Finally, it is based on the real-time collected import and export traffic flow data, weather data and the trained The tunnel accident real-time risk prediction model predicts the tunnel accident risk in real time, realizes the real-time prediction of the highway tunnel accident risk, and then provides a reference for tunnel accident prevention measures, so as to intervene in advance to improve the driving safety of vehicles in the tunnel.
  • the so-called processor D100 can be a central processing unit (CPU, Central Processing Unit).
  • the processor D100 can also be other general-purpose processors, digital signal processors (DSP, Digital Signal Processor), application specific integrated circuits (ASIC, Application Specific Integrated Circuit), off-the-shelf programmable gate array (FPGA, Field-Programmable GateArray) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • DSP digital signal processors
  • ASIC Application Specific Integrated Circuit
  • FPGA Field-Programmable GateArray
  • a general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
  • the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may also be an external storage device of the terminal device D10, such as a plug-in hard disk, a smart memory card (SMC, Smart Media Card), or a secure digital device equipped on the terminal device D10. (SD, Secure Digital) card, Flash Card, etc. Further, the memory D101 may also include both an internal storage unit of the terminal device D10 and an external storage device.
  • the memory D101 is used to store operating systems, application programs, boot loaders (Boot Loaders), data and other programs, such as program codes of the computer programs.
  • the memory D101 can also be used to temporarily store data that has been output or will be output.
  • Module completion means dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above.
  • Each functional unit and module in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.
  • the above-mentioned integrated unit can either use hardware. It can be implemented in the form of software or in the form of software functional units.
  • the specific names of each functional unit and module are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present application.
  • Embodiments of the present application also provide a computer-readable storage medium.
  • the computer-readable storage medium stores a computer program.
  • the steps in each of the above method embodiments can be implemented.
  • Embodiments of the present application provide a computer program product.
  • the steps in each of the above method embodiments can be implemented when the terminal device executes it.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium.
  • this application can implement all or part of the processes in the methods of the above embodiments by instructing relevant hardware through a computer program.
  • the computer program can be stored in a computer-readable storage medium.
  • the computer program When executed by a processor, the steps of each of the above method embodiments may be implemented.
  • the computer program includes computer program code, which may be in the form of source code, object code, executable file or some intermediate form.
  • the computer-readable medium may at least include: any entity or device capable of carrying computer program code to a highway tunnel accident real-time risk prediction device/terminal equipment, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory) ), random access memory (RAM, Random Access Memory), electrical carrier signals, telecommunications signals, and software distribution media.
  • ROM Read-only memory
  • RAM Random Access Memory
  • electrical carrier signals telecommunications signals
  • software distribution media for example, U disk, mobile hard disk, magnetic disk or CD, etc.
  • computer-readable media may not be electrical carrier signals and telecommunications signals.
  • the disclosed devices/network devices and methods can be implemented in other ways.
  • the apparatus/network equipment embodiments described above are only illustrative.
  • the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units. Or components can be combined or can be integrated into another system, or some features can be omitted, or not implemented.
  • the coupling or direct coupling or communication connection between each other shown or discussed may be through some interfaces, indirect coupling or communication connection of devices or units, which may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种高速公路隧道事故实时风险预测方法、装置、设备及介质，其中方法包括：获取目标隧道在预设历史时间段内的事故数据、进出口交通流数据以及天气数据；获取目标隧道在预设历史时间段内发生的每次事故对应的进出口交通流数据和天气数据，并获取进出口交通流数据和天气数据的对照数据；将进出口交通流数据、天气数据以及对照数据作为实验数据集；基于实验数据集提取影响隧道事故风险的事故先兆特征；基于实验数据集、事故先兆特征训练隧道事故实时风险预测模型；基于训练后的隧道事故实时风险预测模型实时预测隧道事故风险。能够对高速公路隧道事故风险进行实时预测。

Description

高速公路隧道事故实时风险预测方法、装置、设备及介质 技术领域
本申请属于交通安全技术领域，尤其涉及一种高速公路隧道事故实时风险预测方法、装置、设备及介质。
背景技术
隧道因其在减少通行时间、保护生态环境、改善高速公路运营等方面的重要作用而被视为高速公路的重要路段。随着基础设施建设的逐步发展，高速公路通车里程不断增加。在过去的几年中，隧道数量显著增加，并且仍在持续增长。
然而，由于隧道通常比开放道路的照明更少并且具有半封闭的空间结构，它们可能会影响驾驶员的行为，从而更容易造成交通冲突，进一步导致碰撞事故的发生。因此，以提升隧道交通安全为目标进行研究对提升高速公路系统的安全水平具有一定意义。
目前关于隧道交通安全的研究大多集中在基于历史的事故记录数据的碰撞数据的碰撞特征、碰撞频率和影响伤害严重程度的因素的统计分析上，缺乏实时性，在实际的应用中有一定的局限性。近几年来，事故实时风险分析预测成为热点，然而，目前高速公路隧道交通事故实时风险预测目前尚处于空白阶段。
发明内容
本申请实施例提供了一种高速公路隧道事故实时风险预测方法、装置、设备及介质，可以解决无法对高速公路隧道事故风险进行实时预测的问题。
第一方面，本申请实施例提供了一种高速公路隧道事故实时风险预测方法，包括：
获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据；
分别针对事故数据中的每个事故，从进出口交通流数据中匹配出事故发生前n分钟内的目标进出口交通流数据，并确定目标隧道所在区域在事故发生前n分钟内的天气数据；n为整数，且5≤n≤15；
分别针对每个事故，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到事故的事故融合数据，并获取事故融合数据的对照数据；
将事故数据中所有事故对应的事故融合数据和对照数据作为实验数据集；
基于实验数据集，确定影响隧道事故风险的事故先兆特征；
基于实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；
每隔n分钟采集目标隧道的进出口交通流数据以及目标隧道所在区域的天气数据；每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据，每次采集的天气数据为采集时刻前n分钟内的天气数据；
针对每次采集到的进出口交通流数据和天气数据，根据隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
可选的，基于实验数据集，确定影响隧道事故风险的事故先兆特征，包括：
分别针对实验数据集中的每类数据，为该类数据定义一变量，并将该类数据作为该变量的数据；
使用斯皮尔曼相关系数计算定义的各变量之间的相关性值；
根据计算得到的相关性值，从定义的各变量中选取目标变量；
利用条件Logistic回归模型计算每个目标变量对隧道事故风险的影响值；
根据计算得到的影响值确定影响隧道事故风险的事故先兆特征。
可选的，根据计算得到的相关性值，从定义的各变量中选取目标变量，包括：
若计算得到的相关性值中存在大于预设相关性值的相关性值，则分别计算该相关性值对应的两个变量与事故是否发生之间的相关系数，并将计算得到的相关系数中较小者对应的变量删除；
将定义的所有变量中除被删除变量以外的其他变量作为目标变量。
可选的，根据计算得到的影响值确定影响隧道事故风险的事故先兆特征，包括：
将影响值小于等于预设影响值的目标变量、以及天气数据对应的目标变量作为影响隧道事故风险的事故先兆特征。
可选的，进出口交通流数据包括：进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率以及出口平均车速。
可选的，获取事故融合数据的对照数据，包括：
利用病例对照法，获取事故融合数据的对照数据。
第二方面，本申请实施例提供了一种高速公路隧道事故实时风险预测装置，包括：
第一获取模块，用于获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据；
数据匹配模块，用于分别针对事故数据中的每个事故，从进出口交通流数据中匹配出事故发生前n分钟内的目标进出口交通流数据，并确定目标隧道所在区域在事故发生前n分钟内的天气数据；n为整数，且5≤n≤15；
第二获取模块，用于分别针对每个事故，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到事故的事故融合数据，并获取事故融合数据的对照数据；
第一确定模块，用于将事故数据中所有事故对应的事故融合数据和对照数据作为实验数据集；
第二确定模块，用于基于实验数据集，确定影响隧道事故风险的事故先兆特征；
模型训练模块，用于基于实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；
数据采集模块，用于每隔n分钟采集目标隧道的进出口交通流数据以及目标隧道所在区域的天气数据；每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据，每次采集的天气数据为采集时刻前n分钟内的天气数据；
风险预测模块，用于针对每次采集到的进出口交通流数据和天气数据，根据隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
第三方面，本申请实施例提供了一种终端设备，包括存储器、处理器以及存储在存储器中并可在处理器上运行的计算机程序，处理器执行计算机程序时实现上述的高速公路隧道事故实时风险预测方法。
第四方面，本申请实施例提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序，计算机程序被处理器执行时实现上述的高速公路隧道事故实时风险预测方法。
本申请的上述方案有如下的有益效果：
在本申请的实施例中，通过对隧道在预设历史时间段内的事故数据、进出口交通流数据以及相关天气数据进行分析，得到隧道在预设历史时间段内发生的每次事故对应的进出口交通流数据和天气数据，并将每次事故对应的进出口交通流数据和天气数据作为该事故的事故融合数据；然后针对隧道在预设历史时间段内发生的每次事故，获取该事故的事故融合数据对应的对照数据，并将预设历史时间段内发生的所有事故对应的事故融合数据和对照数据作为实验数据集；接着对该实验数据集进行分析，确定出对隧道事故风险有显著影响的事故先兆特征，并基于该实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；最终基于实时采集的进出口交通流数据、天气数据以及训练后的隧道事故实时风险预测模型实时预测隧道事故风险，实现对高速公路隧道事故风险的实时预测，进而为隧道事故预防措施提供参考，以便提前干预提高隧道内车辆的行车安全。
本申请的其它有益效果将在随后的具体实施方式部分予以详细说明。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅 仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请一实施例提供的高速公路隧道事故实时风险预测方法的流程图；
图2为本申请一测试结果中ROC曲线的示意图；
图3为本申请一实施例提供的高速公路隧道事故实时风险预测装置的结构示意图；
图4为本申请一实施例提供的终端设备的结构示意图。
具体实施方式
以下描述中，为了说明而不是为了限定，提出了诸如特定系统结构、技术之类的具体细节，以便透彻理解本申请实施例。然而，本领域的技术人员应当清楚，在没有这些具体细节的其它实施例中也可以实现本申请。在其它情况中，省略对众所周知的系统、装置、电路以及方法的详细说明，以免不必要的细节妨碍本申请的描述。
应当理解，当在本申请说明书和所附权利要求书中使用时，术语“包括”指示所描述特征、整体、步骤、操作、元素和/或组件的存在，但并不排除一个或多个其它特征、整体、步骤、操作、元素、组件和/或其集合的存在或添加。
还应当理解，在本申请说明书和所附权利要求书中使用的术语“和/或”是指相关联列出的项中的一个或多个的任何组合以及所有可能组合，并且包括这些组合。
如在本申请说明书和所附权利要求书中所使用的那样，术语“如果”可以依据上下文被解释为“当...时”或“一旦”或“响应于确定”或“响应于检测到”。类似地，短语“如果确定”或“如果检测到[所描述条件或事件]”可以依据上下文被解释为意指“一旦确定”或“响应于确定”或“一旦检测到[所描述条件或事件]”或“响应于检测到[所描述条件或事件]”。
另外，在本申请说明书和所附权利要求书的描述中，术语“第一”、“第二”、“第三”等仅用于区分描述，而不能理解为指示或暗示相对重要性。
在本申请说明书中描述的参考“一个实施例”或“一些实施例”等意味着在本申请的一个或多个实施例中包括结合该实施例描述的特定特征、结构或特点。由此，在本说明书中的不同之处出现的语句“在一个实施例中”、“在一些实施例中”、“在其他一些实施例中”、“在另外一些实施例中”等不是必然都参考相同的实施例，而是意味着“一个或多个但不是所有的实施例”，除非是以其他方式另外特别强调。术语“包括”、“包含”、“具有”及它们的变形都意味着“包括但不限于”，除非是以其他方式另外特别强调。
针对目前无法对高速公路隧道事故风险进行实时预测的问题，本申请实施例提供了一种高速公路隧道事故实时风险预测方法，该方法通过对隧道在预设历史时间段内的事故数据、进出口交通流数据以及相关天气数据进行分析，得到隧道在预设历史时间段内发生的每次事故对应的进出口交通流数据和天气数据，并将每次事故对应的进出口交通流数据和天气数据作为该事故的事故融合数据；然后针对隧道在预设历史时间段内发生的每次事故，获取该事故的事故融合数据对应的对照数据，并将预设历史时间段内发生的所有事故对应的事故融合数据和对照数据作为实验数据集；接着对该实验数据集进行分析，确定出对隧道事故风险有显著影响的事故先兆特征，并基于该实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；最终基于实时采集的进出口交通流数据、天气数据以及训练后的隧道事故实时风险预测模型实时预测隧道事故风险，实现对高速公路隧道事故风险的实时预测，进而为隧道事故预防措施提供参考，以便提前干预提高隧道内车辆的行车安全。
下面结合具体实施例对本申请提供的高速公路隧道事故实时风险预测方法进行示例性的说明。
本申请实施例提供了一种高速公路隧道事故实时风险预测方法，该方法可以由终端设备执行，也可以由应用于终端设备中的装置(比如芯片)来执行，下述实施例以该方法由终端设备执行为例。作为一种示例，该终端设备可以是平板，服务器或者笔记本电脑等，本申请实施例对此不做限定。
如图1所示，本申请实施例提供的高速公路隧道事故实时风险预测方法包括如下步骤：
步骤11，获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据。
上述目标隧道为需要进行事故实时风险预测的隧道；上述预设历史时间段可根据实际情况进行设定，例如设定为近一年；上述事故数据包括目标隧道在预设历史时间段内发生的所有事故的记录数据，该记录数据包括事故发生的地点、时间等描述信息；上述进出口交通流数据包括：进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率以及出口平均车速。其中，进口流量指的是单位时间内通过检测器(该检测器为设置于目标隧道进口处的检测器)的车辆数，单位为：辆/小时，例如5分钟内通过该检测器的车辆数为N，进口流量进口平均占有率指的是目标隧道进口处某检测截面或检测区内有车存在的时间Tv与统计总时间T(如5分钟)之比；进口平均车速指的是单位时间内通过检测器(该检测器为设置于目标隧道进口处的检测器)车辆车速的平均值，单位为公里/小时；类似的，出口流量指的是单位时间内通过检测器(该检测器为设置于目标隧道出口处的检测器)的车辆数，单位为：辆/小时，例如5分钟内通过该检测器的车辆数为N，出口流量出口平均占有率指的是目标隧道出口处某检测截面或检测区内有车存在的时间Tv与统计总时间T(如5分钟)之比；出口平均车速指的是单位时间内通过检测器(该检测器为设置于目标隧道出口处的检测器)车辆车速的平均值，单位为公里/小时。
具体的，在本申请的一些实施例中，可以从相关交通部门的网站上获得上述事故数据和进出口交通流数据。可以理解的是，在从相关网站获得事故数据和进出口交通流数据后，可将存在信息缺失以及误判事故的记录数据删除，以确保步骤11中事故数据的完整性。
步骤12，分别针对事故数据中的每个事故，从进出口交通流数据中匹配出事故发生前n分钟内的目标进出口交通流数据，并确定目标隧道所在区域在事故发生前n分钟内的天气数据。
其中，n为整数，且5≤n≤15。需要说明的是，由于交通流监测一般都是5分钟的间隔，因此n的取值一般可以为5、10或者15。
在本申请的一些实施例中，上述目标进出口交通流数据为步骤11中进出 口交通流数据的一部分，具体为事故发生前n分钟内的进出口交通流数据。具体的，可以根据记录数据中事故发生的时间，从步骤11得到的进出口交通流数据中匹配到事故发生前n分钟内的进出口交通流数据。
在本申请的一些实施例中，可通过从气象部门的相关网站获取目标隧道所在区域在事故发生前n分钟内的天气数据，该天气数据可以为晴天、阴天、雨天、雪天等天气数据。具体的，可以根据记录数据中事故发生的时间和地点，查询找到目标隧道所在区域在事故发生前n分钟内的天气数据。
步骤13，分别针对每个事故，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到事故的事故融合数据，并获取事故融合数据的对照数据。
在本申请的一些实施例中，在确定出事故对应的目标进出口交通流数据和天气数据后，可根据事故发生的时间和地点，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到事故融合数据，从而便于确定对照数据，丰富后续的实验数据集，进而有助于后续提升隧道事故实时风险预测模型的训练精度。
其中，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到的事故融合数据包括：进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率、出口平均车速、进出口流量差(即进口流量与出口流量的差值)、进出口占有率差(即进口平均占有率与出口平均占有率的差值)、进出口速度差(即进口平均车速与出口平均车速的差值)以及天气数据。
在本申请的一些实施例中，可以通过病例对照法获取事故融合数据的对照数据。具体的，对照数据的获取方式可以为：首先利用病例对照法，获取目标隧道在预设相关时间的进出口交通流数据，并获取目标隧道所在区域在预设相关时间的天气数据；然后将利用病例对照法获取到的进出口交通流数据和天气数据进行融合处理，并将融合处理后的数据作为事故融合数据的对照数据。
需要说明的是，对照数据与事故融合数据所包含的数据类型是一样的，只是二者所包含数据对应的时间不同。即，对照数据包括目标隧道在预设相关时间的进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率、出口平均车速、进出口流量差(即进口流量与出口流量的差值)、进出口占有 率差(即进口平均占有率与出口平均占有率的差值)、进出口速度差(即进口平均车速与出口平均车速的差值)以及相应的天气数据。
其中，上述预设相关时间与事故的发生时间相关。一般的，可选取事故发生地点、时段的前7天、14天、后7天以及14天的数据为对照数据。
示例性的，假设事故发生的时间为2018年1月1日上午9:03，地点为A隧道，则提取A隧道在2017年12月18日、12月25日及2018年1月8日、2018年1月15日上午9:00的数据(包括进出口交通流数据和天气数据)作为对照数据。
在本申请的一些可能的实施例中，在确定出事故对应的目标进出口交通流数据和天气数据后，可对事故的记录数据以及事故对应的目标进出口交通流数据和天气数据进行融合。即，上述事故融合数据还包括记录数据。
需要说明的是，若事故融合数据包括记录数据，则获取的对照数据也包括事故的记录数据。
步骤14，将事故数据中所有事故对应的事故融合数据和对照数据作为实验数据集。
在本申请的一些实施例中，可根据事故发生的时间和地点将数据融合，并将事故融合数据和对照数据组合成实验数据集。
步骤15，基于实验数据集，确定影响隧道事故风险的事故先兆特征。
在本申请的一些实施例中，可通过斯皮尔曼(Spearman)相关系数和条件Logistic回归模型，帅选出对隧道事故风险有显著影响的事故先兆特征，从而便于后续实时、准确地进行隧道事故风险预测。
需要说明的是，由于根据高速公路事故实时风险相关研究及隧道的半封闭特性，推测隧道事故风险与进出口交通特性及天气条件相关，因此基于上述实验数据集进行事故先兆特征的确定，能提升事故先兆特征的准确性，进而有助于提升事故风险预测的准确性。
步骤16，基于实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型。
在本申请的一些实施例中，可将上述实验数据集划分为训练集数据和测试 集数据，然后利用训练集数据对隧道事故实时风险预测模型进行训练，利用测试集数据对隧道事故实时风险预测模型进行测试，得到训练后的隧道事故实时风险预测模型。
步骤17，每隔n分钟采集目标隧道的进出口交通流数据以及目标隧道所在区域的天气数据。
每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据(该进出口交通流数据包括：进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率以及出口平均车速)，每次采集的天气数据为采集时刻前n分钟内的天气数据(该天气数据可以为晴天、阴天、雨天、雪天等天气数据)。
步骤18，针对每次采集到的进出口交通流数据和天气数据，根据隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
在本申请的一些实施例中，上述隧道事故实时风险预测模型可以为常用的基于支持向量机的决策函数。具体的，该决策函数的表达式可以为：
其中，f(x)表示因变量，即事故是否发生，为支持向量，b*为偏置量，与b*可由卡罗需-库恩-塔克(Karush-Kuhn-Tucker)条件算出，K为核函数，xi为训练集数据中第i个样本的事故先兆特征的数据，x为需要预测的隧道的事故先兆特征的数据(即步骤18中的事故先兆特征数据)，yi∈γ＝{-1,1}为事故是否发生的输出特征值，n为训练集数据中样本的数量。
值得一提的是，本申请实施例提供的方法通过对隧道在预设历史时间段内的事故数据、进出口交通流数据以及相关天气数据进行分析，得到事故发生前n分钟内的进出口交通流数据和天气数据，并将每次事故对应的进出口交通流数据和天气数据作为该事故的事故融合数据；然后针对隧道在预设历史时间段内发生的每次事故，获取该事故的事故融合数据对应的对照数据，并将预设历史时间段内发生的所有事故对应的事故融合数据和对照数据作为实验数据集； 接着对该实验数据集进行分析，确定出对隧道事故风险有显著影响的事故先兆特征，并基于该实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；最终基于实时采集的进出口交通流数据、天气数据以及训练后的隧道事故实时风险预测模型实时预测隧道是否发生交通事故，实现对高速公路隧道事故风险的实时预测，进而为隧道事故预防措施(如流量、车速控制等措施)提供参考，以便提前干预提高隧道内车辆的行车安全。
下面结合具体实施例对上述步骤15得具体实现过程进行示例性说明。
在本申请的一些实施例中，上述步骤15，基于实验数据集，确定影响隧道事故风险的事故先兆特征得具体实现方式包括如下步骤：
步骤一，分别针对实验数据集中的每类数据，为该类数据定义一变量，并将该类数据作为该变量的数据。
上述实验数据集中包括进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率、出口平均车速、进出口流量差、进出口占有率差、进出口速度差以及天气数据这十类数据。在本申请的一些实施例中，需为这十类数据中的每类数据定义一变量，并将该类数据作为对应变量的数据。在此以进口流量进行示例说明，假设为进口流量定义的变量为t，则实验数据集中所有进口流量的数据为变量t的取值。
步骤二，使用斯皮尔曼相关系数计算定义的各变量之间的相关性值。
在本申请的一些实施例中，对于定义的任意两变量，可结合这两个变量的数据，使用斯皮尔曼相关系数计算这两个变量的相关性值。
其中，上述斯皮尔曼相关系数得表达式为：S表示计算得到的相关性值，n表示观测值数量(即事故融合数据和对照数据所包含的数据总条数)，di表示对应两个变量分别排序后成对的变量位置之差。需要说明的是，由于斯皮尔曼相关系数为常用的相关性分析方法，因此在此，不对斯皮尔曼相关系数的原理进行过多赘述。
步骤三，根据计算得到的相关性值，从定义的各变量中选取目标变量。
在本申请的一些实施例中，为了剔除具有强相关性的特征，提高事故先兆 特征的准确性，需要在计算得到每两变量之间的相关性值之后，删除与事故是否发生相关性较弱的特征。
具体的，在计算得到任意两变量之间的相关性值(即斯皮尔曼相关系数值)后，若计算得到的相关性值中存在大于预设相关性值(如0.7)的相关性值，则分别计算该相关性值对应的两个变量与事故是否发生之间的相关系数(即斯皮尔曼相关系数值)，并将计算得到的相关系数中较小者对应的变量删除，然后将定义的所有变量中除被删除变量以外的其他变量作为目标变量。
示例性的，若进口流量与出口流量的相关性值为0.85，其中进口流量与事故因素(即事故是否发生)之间的相关系数为0.45，出口流量与事故因素(即事故是否发生)之间的相关性系数为0.3，则删除出口流量对应的变量，保留进口流量对应的变量。
在本申请的一些实施例中，上述变量与事故是否发生之间的相关系数也可以通过斯皮尔曼相关系数进行计算。可以理解的是，为便于计算变量与事故是否发生之间的相关系数，在实际处理时，可以为事故是否发生定义一变量(该变量的取值为0或者1，0表示未发生事故，1表示发生事故)，其中事故融合数据中该变量的值为1，对照数据中该变量的值为1或者0。
步骤四，利用条件Logistic回归模型计算每个目标变量对隧道事故风险的影响值。
步骤五，根据计算得到的影响值确定影响隧道事故风险的事故先兆特征。
在本申请的一些实施例中，可将影响值小于等于预设影响值(如0.05)的目标变量、以及天气数据对应的目标变量作为影响隧道事故风险的事故先兆特征。
需要说明的是，在实际预测过程中，在每采集到一次进出口交通流数据以及目标隧道所在区域的天气数据后，需要对该进出口交通流数据和天气数据进行分析，提取到每个事故先兆特征对应的数据(即事故先兆特征数据)，然后在基于提取到的事故先兆特征数据输入训练后的隧道事故实时风险预测模型进行隧道事故风险预测。
在此以一具体实例对事故先兆特征的确定过程进行示例性说明。
在该实例中，首先收集加利福尼亚洲15个隧道2018年的187条隧道交通 事故，删除部分信息缺失以及误判事故，共得到180条有效事故数据，同时收集2018年15个隧道的进出口交通流数据，以及2018年15个隧道所在区域的天气数据(例如，2018年1月1日马林郡的天气为晴天)；然后从180条有效事故数据中筛选出隧道A的15条有效事故数据，并将这15条有效事故数据作为隧道A的事故数据，分别针对15条有效事故数据中的每个事故，查询匹配出该事故发生前n(此处n的取值为5)分钟内的进出口交通流数据和天气数据，再基于匹配出的进出口交通流数据和天气数据，得到隧道A的实验数据集；接着将该实验数据集中的每类数据定义为一变量，这些变量的描述统计情况如表1和表2所示。其中，在实际处理过程中，天气数据对应的变量可以为一个，在此处为便于统计情况的描述，对晴天、阴天、雨天、雪天进行分开说明。
表1
表2
在定义好变量后，使用斯皮尔曼相关系数计算每两个变量之间的斯皮尔曼相关系数值，并计算定义的每个变量与事故是否发生这个变量之间的斯皮尔曼相关系数值。其中，计算得到的斯皮尔曼相关系数值如表3所示。
表3
结合计算得到斯皮尔曼相关系数值以及目标变量的筛选原则，选定进口平均车速、出口平均占有率、出口平均车速、进出口流量差、进出口占有率差、 进出口速度差、天气为目标变量；最终利用条件Logistic回归模型计算每个目标变量对隧道事故风险的影响值，并将影响值小于等于预设影响值(此处预设影响值为0.05)的目标变量(此处为进口平均车速、出口平均占有率、进出口流量差、进出口占有率差)，以及天气数据对应的目标变量作为影响隧道事故风险的事故先兆特征。其中，利用条件Logistic回归模型计算得到的影响值如表4所示，Estimate表示目标变量与隧道事故风险的相关强度，Pr(>|z|)表示目标变量对隧道事故风险的影响值。需要说明的是，表1至表4中的“-”表示该项不存在。
表4
下面结合具体测试结果对上述隧道事故实时风险预测模型的效果进行示例性的说明。
通过利用上述测试集数据对上述隧道事故实时风险预测模型进行测试，得到的测试结果如表5所示，ROC曲线如图2所示。其中，F1分数是统计学中用来衡量二分类模型精确度的一种指标，它同时兼顾了分类模型的精确率和召回率，ROC曲线为接受者操作特征曲线。需要说明的是，表5中的“-”表示该项不存在。
表5
从表5可知，隧道事故实时风险预测模型的总精度在91％左右，表明隧道事故实时风险预测模型的预测精度高；从图2可知，AUC值(AUC值指的是ROC曲线下的面积)为0.97，接近于1，表明隧道事故实时风险预测模型的真实性高。
下面结合具体实施例对本申请提供的高速公路隧道事故实时风险预测装置进行示例性的说明。
如图3所示，本申请的实施例提供了一种高速公路隧道事故实时风险预测装置，该高速公路隧道事故实时风险预测装置300包括：
第一获取模块301，用于获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据；
数据匹配模块302，用于分别针对事故数据中的每个事故，从进出口交通流数据中匹配出事故发生前n分钟内的目标进出口交通流数据，并确定目标隧道所在区域在事故发生前n分钟内的天气数据；n为整数，且5≤n≤15；
第二获取模块303，用于分别针对每个事故，对事故对应的目标进出口交通流数据和天气数据进行融合处理，得到事故的事故融合数据，并获取事故融合数据的对照数据；
第一确定模块304，用于将事故数据中所有事故对应的事故融合数据和对照数据作为实验数据集；
第二确定模块305，用于基于实验数据集，确定影响隧道事故风险的事故先兆特征；
模型训练模块306，用于基于实验数据集，以事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；
数据采集模块307，用于每隔n分钟采集目标隧道的进出口交通流数据以及目标隧道所在区域的天气数据；每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据，每次采集的天气数据为采集时刻前n分钟内的天气数据；
风险预测模块308，用于针对每次采集到的进出口交通流数据和天气数据， 根据隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
需要说明的是，上述装置/单元之间的信息交互、执行过程等内容，由于与本申请方法实施例基于同一构思，其具体功能及带来的技术效果，具体可参见方法实施例部分，此处不再赘述。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。实施例中的各功能单元、模块可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中，上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。另外，各功能单元、模块的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
如图4所示，本申请的实施例提供了一种终端设备，如图4所示，该实施例的终端设备D10包括：至少一个处理器D100(图4中仅示出一个处理器)、存储器D101以及存储在所述存储器D101中并可在所述至少一个处理器D100上运行的计算机程序D102，所述处理器D100执行所述计算机程序D102时实现上述任意各个方法实施例中的步骤。
具体的，所述处理器D100执行所述计算机程序D102时，通过对隧道在预设历史时间段内的事故数据、进出口交通流数据以及相关天气数据进行分析，得到隧道在预设历史时间段内发生的每次事故对应的进出口交通流数据和天气数据，并将每次事故对应的进出口交通流数据和天气数据作为该事故的事故融合数据；然后针对隧道在预设历史时间段内发生的每次事故，获取该事故的事故融合数据对应的对照数据，并将预设历史时间段内发生的所有事故对应的事故融合数据和对照数据作为实验数据集；接着对该实验数据集进行分析，确定出对隧道事故风险有显著影响的事故先兆特征，并基于该实验数据集，以事 故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；最终基于实时采集的进出口交通流数据、天气数据以及训练后的隧道事故实时风险预测模型实时预测隧道事故风险，实现对高速公路隧道事故风险的实时预测，进而为隧道事故预防措施提供参考，以便提前干预提高隧道内车辆的行车安全。
所称处理器D100可以是中央处理单元(CPU，Central Processing Unit)，该处理器D100还可以是其他通用处理器、数字信号处理器(DSP，Digital Signal Processor)、专用集成电路(ASIC，Application Specific Integrated Circuit)、现成可编程门阵列(FPGA，Field-Programmable GateArray)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
所述存储器D101在一些实施例中可以是所述终端设备D10的内部存储单元，例如终端设备D10的硬盘或内存。所述存储器D101在另一些实施例中也可以是所述终端设备D10的外部存储设备，例如所述终端设备D10上配备的插接式硬盘，智能存储卡(SMC，Smart Media Card)，安全数字(SD，Secure Digital)卡，闪存卡(Flash Card)等。进一步地，所述存储器D101还可以既包括所述终端设备D10的内部存储单元也包括外部存储设备。所述存储器D101用于存储操作系统、应用程序、引导装载程序(BootLoader)、数据以及其他程序等，例如所述计算机程序的程序代码等。所述存储器D101还可以用于暂时地存储已经输出或者将要输出的数据。
需要说明的是，上述装置/单元之间的信息交互、执行过程等内容，由于与本申请方法实施例基于同一构思，其具体功能及带来的技术效果，具体可参见方法实施例部分，此处不再赘述。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。实施例中的各功能单元、模块可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中，上述集成的单元既可以采用硬 件的形式实现，也可以采用软件功能单元的形式实现。另外，各功能单元、模块的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现可实现上述各个方法实施例中的步骤。
本申请实施例提供了一种计算机程序产品，当计算机程序产品在终端设备上运行时，使得终端设备执行时实现可实现上述各个方法实施例中的步骤。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请实现上述实施例方法中的全部或部分流程，可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读介质至少可以包括：能够将计算机程序代码携带到高速公路隧道事故实时风险预测装置/终端设备的任何实体或装置、记录介质、计算机存储器、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、电载波信号、电信信号以及软件分发介质。例如U盘、移动硬盘、磁碟或者光盘等。在某些司法管辖区，根据立法和专利实践，计算机可读介质不可以是电载波信号和电信信号。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述或记载的部分，可以参见其它实施例的相关描述。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的实施例中，应该理解到，所揭露的装置/网络设备和方法，可以通过其它的方式实现。例如，以上所描述的装置/网络设备实施例仅仅是示意性的，例如，所述模块或单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通讯连接可以是通过一些接口，装置或单元的间接耦合或通讯连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
以上所述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。

Claims (9)

  1. 一种高速公路隧道事故实时风险预测方法，其特征在于，包括：
    获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据；
    分别针对所述事故数据中的每个事故，从所述进出口交通流数据中匹配出所述事故发生前n分钟内的目标进出口交通流数据，并确定所述目标隧道所在区域在所述事故发生前n分钟内的天气数据；n为整数，且5≤n≤15；
    分别针对每个所述事故，对所述事故对应的目标进出口交通流数据和天气数据进行融合处理，得到所述事故的事故融合数据，并获取所述事故融合数据的对照数据；
    将所述事故数据中所有事故对应的事故融合数据和对照数据作为实验数据集；
    基于所述实验数据集，确定影响隧道事故风险的事故先兆特征；
    基于所述实验数据集，以所述事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；
    每隔n分钟采集所述目标隧道的进出口交通流数据以及所述目标隧道所在区域的天气数据；每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据，每次采集的天气数据为采集时刻前n分钟内的天气数据；
    针对每次采集到的进出口交通流数据和天气数据，根据所述隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于所述事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
  2. 根据权利要求1所述的方法，其特征在于，所述基于所述实验数据集，确定影响隧道事故风险的事故先兆特征，包括：
    分别针对所述实验数据集中的每类数据，为该类数据定义一变量，并将该类数据作为该变量的数据；
    使用斯皮尔曼相关系数计算定义的各变量之间的相关性值；
    根据计算得到的相关性值，从定义的各变量中选取目标变量；
    利用条件Logistic回归模型计算每个目标变量对隧道事故风险的影响值；
    根据计算得到的影响值确定影响隧道事故风险的事故先兆特征。
  3. 根据权利要求2所述的方法，其特征在于，所述根据计算得到的相关性值，从定义的各变量中选取目标变量，包括：
    若计算得到的相关性值中存在大于预设相关性值的相关性值，则分别计算该相关性值对应的两个变量与事故是否发生之间的相关系数，并将计算得到的相关系数中较小者对应的变量删除；
    将定义的所有变量中除被删除变量以外的其他变量作为目标变量。
  4. 根据权利要求2所述的方法，其特征在于，所述根据计算得到的影响值确定影响隧道事故风险的事故先兆特征，包括：
    将影响值小于等于预设影响值的目标变量、以及天气数据对应的目标变量作为影响隧道事故风险的事故先兆特征。
  5. 根据权利要求1所述的方法，其特征在于，所述进出口交通流数据包括：进口流量、进口平均占有率、进口平均车速、出口流量、出口平均占有率以及出口平均车速。
  6. 根据权利要求1所述的方法，其特征在于，所述获取所述事故融合数据的对照数据，包括：
    利用病例对照法，获取所述事故融合数据的对照数据。
  7. 一种高速公路隧道事故实时风险预测装置，其特征在于，包括：
    第一获取模块，用于获取目标隧道在预设历史时间段内的事故数据和进出口交通流数据；
    数据匹配模块，用于分别针对所述事故数据中的每个事故，从所述进出口交通流数据中匹配出所述事故发生前n分钟内的目标进出口交通流数据，并确定所述目标隧道所在区域在所述事故发生前n分钟内的天气数据；n为整数，且5≤n≤15；
    第二获取模块，用于分别针对每个所述事故，对所述事故对应的目标进出口交通流数据和天气数据进行融合处理，得到所述事故的事故融合数据，并获取所述事故融合数据的对照数据；
    第一确定模块，用于将所述事故数据中所有事故对应的事故融合数据和对 照数据作为实验数据集；
    第二确定模块，用于基于所述实验数据集，确定影响隧道事故风险的事故先兆特征；
    模型训练模块，用于基于所述实验数据集，以所述事故先兆特征为自变量、事故是否发生作为因变量，训练隧道事故实时风险预测模型，得到训练后的隧道事故实时风险预测模型；
    数据采集模块，用于每隔n分钟采集所述目标隧道的进出口交通流数据以及所述目标隧道所在区域的天气数据；每次采集的进出口交通流数据为采集时刻前n分钟内的进出口交通流数据，每次采集的天气数据为采集时刻前n分钟内的天气数据；
    风险预测模块，用于针对每次采集到的进出口交通流数据和天气数据，根据所述隧道事故实时风险预测模型的自变量，从采集到的进出口交通流数据和天气数据中提取事故先兆特征数据，并基于所述事故先兆特征数据和训练后的隧道事故实时风险预测模型预测隧道事故风险。
  8. 一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现如权利要求1至6任一项所述的高速公路隧道事故实时风险预测方法。
  9. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至6任一项所述的高速公路隧道事故实时风险预测方法。
PCT/CN2023/087326 2022-09-22 2023-04-10 高速公路隧道事故实时风险预测方法、装置、设备及介质 WO2024060588A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211170362.6 2022-09-22
CN202211170362.6A CN115565373B (zh) 2022-09-22 2022-09-22 高速公路隧道事故实时风险预测方法、装置、设备及介质

Publications (1)

Publication Number Publication Date
WO2024060588A1 true WO2024060588A1 (zh) 2024-03-28

Family

ID=84743550

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/087326 WO2024060588A1 (zh) 2022-09-22 2023-04-10 高速公路隧道事故实时风险预测方法、装置、设备及介质

Country Status (2)

Country Link
CN (1) CN115565373B (zh)
WO (1) WO2024060588A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115565373B (zh) * 2022-09-22 2024-04-05 中南大学 高速公路隧道事故实时风险预测方法、装置、设备及介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103198713A (zh) * 2013-03-21 2013-07-10 东南大学 基于交通数据和天气数据的减少交通事故的车辆调控方法
CN112101759A (zh) * 2020-09-03 2020-12-18 交通运输部科学研究院 高速公路隧道的风险评估模型构建、风险评估方法及装置
CN112149922A (zh) * 2020-11-03 2020-12-29 南京信息职业技术学院 高速公路隧道下行线出入口区域事故严重程度预测方法
KR20210032758A (ko) * 2019-09-17 2021-03-25 주식회사 퀀텀게이트 터널 내 교통 흐름 정보에 기초한 교통 안전 제공 시스템 및 방법
CN112990545A (zh) * 2021-02-08 2021-06-18 东南大学 一种高速公路交织区交通安全状态预测方法
JP2021182189A (ja) * 2020-05-18 2021-11-25 株式会社オリエンタルコンサルタンツ 事故予測方法、コンピュータプログラム、事故予測装置及び学習モデル生成方法
CN114970943A (zh) * 2022-03-29 2022-08-30 浙江省交通投资集团有限公司智慧交通研究分公司 一种高速公路事故类型与严重程度预测方法
CN115565373A (zh) * 2022-09-22 2023-01-03 中南大学 高速公路隧道事故实时风险预测方法、装置、设备及介质

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20170065898A (ko) * 2015-12-04 2017-06-14 한국교통연구원 도로 상에서의 기상악화로 인한 교통사고 발생 가능성 예측 방법 및 이를 위한 장치
KR101890497B1 (ko) * 2016-12-29 2018-09-28 (주) 더아이엠씨 교통 위험도 분석방법 및 이를 포함하는 저장매체
CN111126868B (zh) * 2019-12-30 2023-07-04 中南大学 一种道路交通事故发生风险确定方法及系统
CN112419709B (zh) * 2020-10-16 2021-11-09 同济大学 一种基于路段异质性的快速路事故风险实时预测方法
CN113095584B (zh) * 2021-04-23 2023-08-15 中南大学 一种基于交通流特征的交叉口安全水平短时预测方法、系统、终端及可读存储介质
CN113269960A (zh) * 2021-05-17 2021-08-17 郑州航空工业管理学院 一种高速公路隧道群交通事故风险的分析方法
CN113744526B (zh) * 2021-08-25 2022-12-23 贵州黔通智联科技股份有限公司 一种基于lstm和bf的高速公路风险预测方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103198713A (zh) * 2013-03-21 2013-07-10 东南大学 基于交通数据和天气数据的减少交通事故的车辆调控方法
KR20210032758A (ko) * 2019-09-17 2021-03-25 주식회사 퀀텀게이트 터널 내 교통 흐름 정보에 기초한 교통 안전 제공 시스템 및 방법
JP2021182189A (ja) * 2020-05-18 2021-11-25 株式会社オリエンタルコンサルタンツ 事故予測方法、コンピュータプログラム、事故予測装置及び学習モデル生成方法
CN112101759A (zh) * 2020-09-03 2020-12-18 交通运输部科学研究院 高速公路隧道的风险评估模型构建、风险评估方法及装置
CN112149922A (zh) * 2020-11-03 2020-12-29 南京信息职业技术学院 高速公路隧道下行线出入口区域事故严重程度预测方法
CN112990545A (zh) * 2021-02-08 2021-06-18 东南大学 一种高速公路交织区交通安全状态预测方法
CN114970943A (zh) * 2022-03-29 2022-08-30 浙江省交通投资集团有限公司智慧交通研究分公司 一种高速公路事故类型与严重程度预测方法
CN115565373A (zh) * 2022-09-22 2023-01-03 中南大学 高速公路隧道事故实时风险预测方法、装置、设备及介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ZHUANG-LIN MA, SHAO CHUN-FU, LI XIA: "Analysis of factors affecting accident severity in highway tunnels based on logistic model ", JOURNAL OF JILIN UNIVERSITY(ENGINEERING AND TECHNOLOGY EDITION), vol. 40, no. 2, 15 March 2020 (2020-03-15), pages 423 - 426, XP093150073 *

Also Published As

Publication number Publication date
CN115565373B (zh) 2024-04-05
CN115565373A (zh) 2023-01-03

Similar Documents

Publication Publication Date Title
Yang et al. A Bayesian dynamic updating approach for urban expressway real-time crash risk evaluation
CN111462488B (zh) 一种基于深度卷积神经网络和交叉口行为特征模型的交叉口安全风险评估方法
CN110428091B (zh) 基于数据分析的风险识别方法及相关设备
WO2024060588A1 (zh) 高速公路隧道事故实时风险预测方法、装置、设备及介质
CN108847022B (zh) 一种微波交通数据采集设备的异常值检测方法
CN113870564B (zh) 一种封闭路段交通拥堵分类方法、系统、电子设备及存储介质
Das et al. Association of reduced visibility with crash outcomes
CN110562261B (zh) 一种基于马尔可夫模型检测驾驶员风险等级的方法
CN115080638B (zh) 微观仿真的多源数据融合分析方法、电子设备及存储介质
CN112634614B (zh) 一种长下坡交通事件实时检测方法、装置及存储介质
Meng et al. Expressway crash prediction based on traffic big data
CN110620760A (zh) 一种SVM和贝叶斯网络的FlexRay总线融合入侵检测方法和检测装置
JP2023524623A (ja) 交通異常を検出するための方法、装置、機器、記憶媒体およびプログラム
CN114879192A (zh) 基于路侧毫米波雷达的决策树车型分类方法、电子设备
CN106910334B (zh) 一种基于大数据预测路段状况的方法及装置
CN114970943A (zh) 一种高速公路事故类型与严重程度预测方法
Nihan et al. Evaluation of dual-loop data accuracy using video ground truth data
CN110705628B (zh) 一种基于隐马尔可夫模型检测驾驶员风险等级的方法
Ding et al. Clustering framework to identify traffic conflicts and determine thresholds based on trajectory data
Howlader et al. A bivariate, non-stationary extreme value model for estimating opposing-through crash frequency by severity by applying artificial intelligence-based video analytics
CN113192340B (zh) 高速公路施工车辆的识别方法、装置、设备及存储介质
CN104361349A (zh) 基于车检器和收费数据融合的异常交通状态识别方法及系统
Torres Investigating traffic crashes involving autonomous vehicles
CN110543827B (zh) 一种基于高斯混合模型与深度学习的多类别车辆检测方法
CN113012430A (zh) 一种车辆排队长度检测方法、装置、设备及可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23866885

Country of ref document: EP

Kind code of ref document: A1