Identification and management method for vehicle risk overspeed area

Abstract

The invention discloses an identification and management method for a vehicle risk overspeed area. The method comprises the following steps: obtaining a vehicle risk overspeed area comprehensive index set based on track space-time matching analysis; constructing a multi-dimensional feature vector of the vehicle risk overspeed area; constructing a vehicle risk overspeed area prediction model, and training the vehicle risk overspeed area prediction model; and obtaining a prediction result of the vehicle risk overspeed area comprehensive index based on the real-time data of the vehicle risk overspeed area multi-dimensional feature vector and the trained vehicle risk overspeed area prediction model, and then executing area risk management operation. According to the method, the regional static traffic characteristics, the regional dynamic traffic characteristics, the regional environment influence factors, the regional track attributes and the behavior characteristics are integrated, the vehicle risk overspeed region prediction model is combined to realize accurate prediction of the regional risk, and the management and control strategy is optimized based on prediction result grading, so that the traffic accident rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation management, and particularly relates to a method for identifying and managing vehicle risk over-speed areas. Background Art

[0002] With the acceleration of the urbanization process, traffic congestion and traffic accidents occur frequently, especially traffic accidents caused by vehicle over-speed behavior are increasing day by day. Traditional traffic management methods mainly rely on devices such as traffic monitoring cameras and radar speed measurement, but these methods have problems such as limited coverage and insufficient real-time performance. In recent years, with the development of vehicle networking and big data technologies, the acquisition and analysis of vehicle trajectory data have become possible, providing new ideas for traffic management.

[0003] However, most of the existing technologies only rely on single-dimensional data analysis, lacking comprehensive analysis and processing of multi-dimensional data such as vehicle trajectories, traffic environments, and road characteristics, and it is difficult to achieve real-time prediction and hierarchical control of regional risks, resulting in lagged warnings or insufficient pertinence of optimization measures. Therefore, there is an urgent need for a technical method that can accurately identify risk over-speed areas and achieve dynamic management. Summary of the Invention

[0004] In view of the above deficiencies in the prior art, the present invention provides a method for identifying and managing vehicle risk over-speed areas. By comprehensively integrating multi-dimensional data such as vehicle trajectory data, static traffic characteristics, dynamic traffic characteristics, and environmental factors, and combining with a deep learning model, accurate prediction of regional risks is achieved, and optimization control strategies are implemented hierarchically based on the prediction results, thereby reducing the incidence of traffic accidents.

[0005] In order to achieve the above invention purpose, the technical solution adopted by the present invention is as follows:

[0006] A method for identifying and managing vehicle risk over-speed areas, comprising the following steps:

[0007] S1. Obtain a comprehensive index set of vehicle risk over-speed areas based on trajectory spatio-temporal matching analysis;

[0008] S2. Construct a multi-dimensional feature vector of vehicle risk over-speed areas, including regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental impact factors, regional trajectory attributes, and behavioral characteristics, and obtain historical data of the multi-dimensional feature vector of vehicle risk over-speed areas corresponding to the comprehensive index set of vehicle risk over-speed areas;

[0009] S3. Construct a prediction model for vehicle risk over-speed areas, train the prediction model for vehicle risk over-speed areas based on the comprehensive index set of vehicle risk over-speed areas and the historical data of the multi-dimensional feature vector of vehicle risk over-speed areas corresponding to the comprehensive index set of vehicle risk over-speed areas, and obtain the trained prediction model for vehicle risk over-speed areas;

[0010] S4. Obtain the real-time data of the multi-dimensional feature vector of the vehicle risk over-speed area, and based on the real-time data of the multi-dimensional feature vector of the vehicle risk over-speed area and the trained vehicle risk over-speed area prediction model, obtain the prediction result of the comprehensive index of the vehicle risk over-speed area, and perform regional risk management operations according to the prediction result of the comprehensive index of the vehicle risk over-speed area.

[0011] Further, step S1 includes the following steps:

[0012] S11. Construct a vehicle accident location neighborhood trajectory database based on trajectory spatio-temporal matching analysis;

[0013] S12. Identify vehicle risk over-speed events and calculate vehicle risk over-speed indexes for the vehicle accident location neighborhood trajectory database to extract a vehicle risk over-speed event data set;

[0014] S13. Calculate the comprehensive index of the vehicle risk over-speed area based on the vehicle risk over-speed event data set to obtain a set of comprehensive indexes of the vehicle risk over-speed area.

[0015] Further, step S13 includes the following steps:

[0016] S131. Divide the vehicle risk over-speed area into grids to obtain grid data of the vehicle sub-line over-speed area;

[0017] S132. Based on the vehicle risk over-speed event data set, obtain the risk over-speed time index values of the grid data of the vehicle sub-line over-speed area, and normalize the risk over-speed time index values of the grid data of the vehicle sub-line over-speed area;

[0018] S133. Calculate the comprehensive index of the vehicle risk over-speed area according to the normalized risk over-speed time index values of the grid data of the vehicle sub-line over-speed area to obtain a set of comprehensive indexes of the vehicle risk over-speed area.

[0019] Further, in step S132, the risk over-speed time index values of the grid data of the vehicle sub-line over-speed area include risk over-speed index values for 8 time periods, and the 8 time periods are 7:00 - 9:00, 9:00 - 17:00, 17:00 - 19:00, 19:00 - 23:00 on weekdays, and 7:00 - 9:00, 9:00 - 17:00, 17:00 - 19:00, 19:00 - 23:00 on non-working days; the risk over-speed index values include the risk over-speed occurrence intensity, the risk over-speed incidence rate, and the risk over-speed average index.

[0020] Further, in step S2, the regional static traffic characteristics include a total of 12 variables: the road length attribute feature vector, the road alignment geometric parameter feature vector, and the traffic facility feature vector; the road length attribute feature vector includes 7 variables: road length, bridge type, tunnel type, highway type, road grade, number of lanes, and design speed; the road alignment geometric parameter feature vector includes 2 variables: longitudinal slope and horizontal curve radius; the traffic facility feature vector includes 3 variables: number of signal lights, number of speed limit signs, and number of speed bumps;

[0021] The regional dynamic traffic characteristics include the aggregated traffic flow level feature vector for different time periods and the list vector of vehicle types passing by and their corresponding average vehicle speeds, with a total of 24 variables; the aggregated traffic flow level feature vector for different time periods includes the aggregated traffic volume for 8 time periods, with a total of 8 variables; the list vector of vehicle types passing by and their corresponding average vehicle speeds includes the average vehicle speeds and speed variances for 8 vehicle types, with a total of 16 variables.

[0022] The regional environmental impact factors include the environmental factor feature vector, the road construction condition feature vector, and the date-time period feature vector, with a total of 12 variables; the environmental factor feature vector includes weather, visibility, temperature, and humidity, with a total of 4 variables; the road construction condition feature vector includes the construction status, with a total of 1 variable; the date-time period feature vector includes one-hot variables for 8 time periods, with a total of 7 variables;

[0023] The regional trajectory attributes and behavior characteristics include the list feature vector of the continuous driving duration of passing vehicles for different time periods and the over-speed behavior characteristics, with a total of 10 variables; the list feature vector of the continuous driving duration of passing vehicles for different time periods includes the statistics of the continuous driving duration of vehicles for 8 time periods, with a total of 8 variables; the over-speed behavior characteristics include the historical over-speed behavior incidence rate and the historical over-speed behavior occurrence intensity, with a total of 2 variables.

[0024] Further, in step S3, the vehicle risk over-speed area prediction model includes:

[0025] A structured processing module for structuring the regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental impact factors, and regional trajectory attributes and behavior characteristics to obtain a 58-dimensional structured feature vector;

[0026] An information dimension elevation and compression module for elevating the 58-dimensional structured feature vector through the elu activation function to obtain a 128-dimensional hidden vector containing abstract information; elevating it to 512 dimensions through a multi-layer fully connected neural network, and then gradually compressing the hidden vector containing abstract information to obtain a compressed 64-dimensional feature hidden vector;

[0027] A multi-output inference module is used to respectively pass the 64-dimensional feature hidden vector through three prediction modules to obtain three values as the prediction results of the risk tachycardia occurrence intensity, the risk tachycardia incidence rate, and the risk tachycardia average index;

[0028] A vehicle risk tachycardia area prediction index module is used to perform arithmetic summation on the prediction results of the risk tachycardia occurrence intensity, the risk tachycardia incidence rate, and the risk tachycardia average index to obtain the prediction result of the vehicle risk tachycardia area comprehensive index.

[0029] Further, in the information dimension elevation and compression module, the multi-layer fully connected neural network includes 6 fully connected layers connected in sequence. The number of neurons in the first fully connected layer is set to 128, the number of neurons in the second fully connected layer is set to 256, the number of neurons in the third fully connected layer is set to 512, the number of neurons in the fourth fully connected layer is set to 256, the number of neurons in the fifth fully connected layer is set to 128, and the number of neurons in the sixth fully connected layer is set to 64.

[0030] Further, in step S4, perform area risk management operations according to the prediction result of the vehicle risk tachycardia area comprehensive index. Specifically: judge the size of the prediction result of the vehicle risk tachycardia area comprehensive index; if the prediction result of the vehicle risk tachycardia area comprehensive index is between the top 20% - top 10%, send a yellow warning; if the prediction result of the vehicle risk tachycardia area comprehensive index is between the top 10% - top 5%, send an orange warning and report to the system; if the prediction result of the vehicle risk tachycardia area comprehensive index is in the top 5%, send a red warning, report to the system and conduct dynamic inspections on this area.

[0031] The present invention has the following beneficial effects:

[0032] (1) By comprehensively considering the regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental influencing factors, regional trajectory attributes and behavior characteristics, the present invention can more accurately identify risk tachycardia behaviors;

[0033] (2) The present invention can analyze vehicle trajectory data in real time and perform hierarchical area risk management according to the prediction result of the vehicle risk tachycardia area comprehensive index, effectively improving the level of traffic safety management;

[0034] (3) By constructing a vehicle risk tachycardia area prediction model including a structured processing module, a multi-layer neural network module, an information compression module, a normalization layer, a multi-output inference module and a vehicle risk tachycardia area prediction module, the present invention can automatically extract complex features in vehicle trajectories, improving the accuracy and efficiency of vehicle risk tachycardia area identification. Description of the Drawings

[0035] Figure 1It is a schematic flow chart of a method for identifying and managing the overspeed area of vehicle risks;

[0036] Figure 2 It is a schematic diagram of the structure of a vehicle risk overspeed area prediction model. Specific implementation manners

[0037] The following describes the specific implementation manners of the present invention to facilitate those skilled in the art of this technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation manners. For those of ordinary skill in the art of this technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0038] As Figure 1 shown, a method for identifying and managing the overspeed area of vehicle risks includes steps S1 - S4, specifically as follows:

[0039] S1. Obtain a comprehensive index set of vehicle risk overspeed areas based on trajectory spatio - temporal matching analysis.

[0040] In an alternative embodiment of the present invention, step S1 includes the following steps:

[0041] S11. Construct a trajectory database of the neighborhood of vehicle accident locations based on trajectory spatio - temporal matching analysis.

[0042] Step S11 includes the following steps:

[0043] S111. Extract accident time, accident location, and matching vehicle trajectory data from the vehicle trajectory set of accident records.

[0044] The vehicle trajectory set of accident records is A_Cars(A, C, G), where A is the accident record list, C is the accident - matching vehicle list, and G is the set of matching vehicle trajectory data. In the present invention, the accident records a in the accident record list are traversed. Taking a 5 - minute range before and after the accident record time as the scope, the vehicle c is matched from the accident record a, and then the trajectory data of the matching vehicle c is extracted, denoted as g(a, c).

[0045] S112. Determine the adjacent time period of the accident occurrence and the exact accident occurrence location based on the accident time, accident location, and matching vehicle trajectory data.

[0046] Based on the accident time, accident location, and the matched vehicle trajectory data, sequentially traverse each trajectory point in the trajectory data g(a, c) of the matched vehicle c, and obtain the zero-speed moment t_accident_real closest to the accident record occurrence time as the precise accident occurrence moment. Five minutes before the precise accident occurrence moment is the corresponding accident occurrence adjacent period t_seg, and the longitude and latitude coordinates c_accident_real corresponding to the precise accident occurrence moment are used as the precise accident occurrence location.

[0047] S113. Extract the trajectory points within 100 meters to 300 meters before the accident based on the precise accident occurrence location, and calculate the estimated driving speed before the accident based on the accident occurrence adjacent period and the distance between the first and last trajectory points within 100 meters to 300 meters before the accident.

[0048] Extract the list of trajectory points Tr_pre(a, c) within 100 meters to 300 meters before the trajectory reaches the precise accident occurrence location based on the precise accident occurrence location, and then, based on the accident occurrence adjacent period, calculate the ratio of the distance between the first and last trajectory points in this list to the time to calculate the estimated driving speed v_accident before the accident.

[0049] S114. Traverse the entire vehicle trajectory set, extract the vehicle trajectory data that is spatio-temporally matched with the accident, and calculate the relative estimated driving speed to construct the accident location neighborhood trajectory database.

[0050] For the entire vehicle trajectory set T_Cars(C_total, G_total), where C_total is the list of all vehicles and G_total is the list of corresponding vehicle trajectory data, traverse each vehicle c in C_total, and for its trajectory set g(c), obtain the subset of trajectory points g_S(c, A_m) that is consistent with the time range of A_T_seg, where A_m is the subset of the corresponding accident record list that is consistent with the time range of vehicle c and A_T_seg. For example, if there is an accident a_m in A_T_seg with an occurrence period of t1 - t2, and vehicle c has a trajectory record set g'(c) during this period, then add this set to g_S(c, A_m) and add the accident a_m record to A_m;

[0051] If \(g_S(c, A_m)\) is an empty set, skip to the next vehicle. Otherwise, traverse each accident record \(a_m\) in \(A_m\), search within a range of 500 meters adjacent to the exact occurrence location \(c_{accident\_real\_m}\) of \(a_m\). If there is trajectory data with the same driving direction (the included angle of the forward direction is less than 90 degrees) in the corresponding vehicle trajectory record set \(g'(c)\) within this range, it can be considered that there is a spatio-temporal matching relationship between the trajectory data set \(g(c, a_m)\) of the vehicle within the corresponding range and the accident record \(a_m\). At the same time, by calculating the ratio of the distance between the first and last trajectory points in \(g(c, a_m)\) to the time, the estimated driving speed \(v_{related}(c, a_m)\) before the accident of accident \(a_m\) is calculated.

[0052] After traversing \(T\_Cars(C\_total, G\_total)\), the set \(M\_C\) of spatio-temporally matching vehicles, the set \(M\_G\) of matching trajectories, and the set \(M\_V\) of estimated driving speeds before the accident for all accidents can be obtained. The set \(M\_C\) of spatio-temporally matching vehicles, the set \(M\_G\) of matching trajectories, and the set \(M\_V\) of estimated driving speeds before the accident for all accidents are constructed together into a trajectory database in the neighborhood of the accident location.

[0053] S12. Identify vehicle risk over-speed events and calculate the vehicle risk over-speed index for the trajectory database in the neighborhood of the vehicle accident location to extract the vehicle risk over-speed event data set.

[0054] The present invention calculates the risk over-speed index for the trajectory database in the neighborhood of the accident location, expressed as:

[0055] indi_v(a) = s(a) * avg_v(a) / (v_accident - avg_v(a))

[0056] Where: indi_v(a) is the risk over-speed index of accident a, s(a) is the severity parameter of accident a, avg_v(a) is the average speed in the neighborhood of accident a, and v_accident is the estimated driving speed before the accident.

[0057] Specifically, the present invention divides the severity of accidents into three levels based on historical data, including first-degree accidents, second-degree accidents, and third-degree accidents, and the corresponding severity parameters are 1, 5, and 20 respectively.

[0058] S13. Calculate the comprehensive index of the vehicle risk over-speed area based on the vehicle risk over-speed event data set to obtain the set of comprehensive indexes of the vehicle risk over-speed area.

[0059] Step S13 includes the following steps:

[0060] S131. Divide the vehicle risk over-speed area into grids to obtain the grid data of the vehicle sub-line over-speed area.

[0061] S132. Based on the vehicle risk over-speed event dataset, obtain the risk over-speed time index value of the grid data in the vehicle lane-based over-speed area, and normalize the risk over-speed time index value of the grid data in the vehicle lane-based over-speed area.

[0062] The risk over-speed time index value of the grid data in the vehicle lane-based over-speed area includes the risk over-speed index values for 8 time periods, which are 7:00 - 9:00, 9:00 - 17:00, 17:00 - 19:00, 19:00 - 23:00 on weekdays, and 7:00 - 9:00, 9:00 - 17:00, 17:00 - 19:00, 19:00 - 23:00 on non-working days; the risk over-speed index values include the risk over-speed occurrence intensity, the risk over-speed incidence rate, and the risk over-speed average index.

[0063] S133. According to the normalized risk over-speed time index value of the grid data in the vehicle lane-based over-speed area, calculate the comprehensive index of the vehicle risk over-speed area to obtain the comprehensive index set of the vehicle risk over-speed area.

[0064] S2. Construct a multi-dimensional feature vector for the vehicle risk over-speed area, including regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental impact factors, regional trajectory attributes, and behavior characteristics, and obtain the historical data of the multi-dimensional feature vector of the vehicle risk over-speed area corresponding to the comprehensive index set of the vehicle risk over-speed area.

[0065] In an alternative embodiment of the present invention, the regional static traffic characteristics include a road length attribute feature vector, a road alignment geometric parameter feature vector, and a traffic facility feature vector, with a total of 12 variables; the road length attribute feature vector includes 7 variables: road length, bridge type, tunnel type, highway type, road grade, number of lanes, and design speed; the road alignment geometric parameter feature vector includes 2 variables: longitudinal slope and horizontal curve radius; the traffic facility feature vector includes 3 variables: number of signal lights, number of speed limit signs, and number of speed bumps.

[0066] The regional dynamic traffic characteristics include an aggregated traffic volume level feature vector for different time periods and a list vector of vehicle types passing through and their corresponding average vehicle speeds, with a total of 24 variables; the aggregated traffic volume level feature vector for different time periods includes the aggregated traffic volume for 8 time periods, with a total of 8 variables; the list vector of vehicle types passing through and their corresponding average vehicle speeds includes the average vehicle speeds and speed variances of 8 vehicle types (mini cars, small sedans, SUVs, large buses, small trucks, medium trucks, large trucks, special vehicles), with a total of 16 variables.

[0067] The regional environmental impact factors include the environmental factor feature vector, the road construction situation feature vector, and the date-time period feature vector, with a total of 12 variables; the environmental factor feature vector includes weather, visibility, temperature, and humidity, with a total of 4 variables; the road construction situation feature vector includes the construction status, with a total of 1 variable; the date-time period feature vector includes one-hot variables for 8 time periods, with a total of 7 variables.

[0068] The regional trajectory attributes and behavior characteristics include the continuous driving duration list feature vector of passing vehicles by time period and the overspeed behavior characteristics, with a total of 10 variables; the continuous driving duration list feature vector of passing vehicles by time period includes the statistics of the continuous driving duration of vehicles in 8 time periods, with a total of 8 variables; the overspeed behavior characteristics include the historical overspeed behavior incidence rate and the historical overspeed behavior occurrence intensity, with a total of 2 variables.

[0069] S3. Construct a vehicle risk overspeed area prediction model, and train the vehicle risk overspeed area prediction model based on the historical data of the vehicle risk overspeed area comprehensive index set and the corresponding multi-dimensional feature vector of the vehicle risk overspeed area to obtain the trained vehicle risk overspeed area prediction model.

[0070] In an alternative embodiment of the present invention, the vehicle risk overspeed area prediction model includes a structured processing module, a multi-layer neural network module, a multi-output inference module, and a vehicle risk overspeed area prediction module, as Figure 2 shown.

[0071] The structured processing module is used to structure the regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental impact factors, and regional trajectory attributes and behavior characteristics to obtain a 58-dimensional structured feature vector.

[0072] The information dimension elevation and compression module is used to elevate the 58-dimensional structured feature vector through the elu activation function to obtain a 128-dimensional hidden vector containing abstract information; elevate it to 512 dimensions through a multi-layer fully connected neural network, and then gradually compress the hidden vector containing abstract information to obtain a compressed 64-dimensional feature hidden vector.

[0073] The multi-layer fully connected neural network includes 6 sequentially connected fully connected layers. The number of neurons in the first fully connected layer is set to 128, the number of neurons in the second fully connected layer is set to 256, the number of neurons in the third fully connected layer is set to 512, the number of neurons in the fourth fully connected layer is set to 256, the number of neurons in the fifth fully connected layer is set to 128, and the number of neurons in the sixth fully connected layer is set to 64.

[0074] The multi-output inference module is used to respectively pass the 64-dimensional feature hidden vector through three prediction modules (Dense-32 → dense (sigmoid activation function)) to obtain three values in the range of (0, 1) as the prediction results of the risk tachycardia occurrence intensity, the risk tachycardia incidence rate, and the risk tachycardia average index.

[0075] The vehicle risk tachycardia area prediction index module is used to perform arithmetic summation on the prediction results of the risk tachycardia occurrence intensity, the risk tachycardia incidence rate, and the risk tachycardia average index to obtain the prediction result of the vehicle risk tachycardia area comprehensive index.

[0076] Specifically, considering that the value range is between 0 and 1 and the difference is small, the present invention uses the Huber loss function to train the vehicle risk tachycardia area prediction model. This function is more sensitive to small differences and more robust to outlier differences.

[0077] S4. Obtain the real-time data of the vehicle risk tachycardia area multi-dimensional feature vector, and based on the real-time data of the vehicle risk tachycardia area multi-dimensional feature vector and the trained vehicle risk tachycardia area prediction model, obtain the prediction result of the vehicle risk tachycardia area comprehensive index, and perform area risk management operations according to the prediction result of the vehicle risk tachycardia area comprehensive index.

[0078] In an alternative embodiment of the present invention, the present invention performs area risk management operations according to the prediction result of the vehicle risk tachycardia area comprehensive index, specifically: judge the size of the prediction result of the vehicle risk tachycardia area comprehensive index; if the prediction result of the vehicle risk tachycardia area comprehensive index is between the top 20% and the top 10%, send a yellow warning; if the prediction result of the vehicle risk tachycardia area comprehensive index is between the top 10% and the top 5%, send an orange warning and report it to the system; if the prediction result of the vehicle risk tachycardia area comprehensive index is in the top 5%, send a red warning, report it to the system and conduct dynamic inspections on this area.

[0079] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0080] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction means that implements the functions specified in one or more of the processes and / or blocks Figure 1 of one or more of the processes and / or blocks Figure 1 specified in one or more of the blocks or blocks.

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 of one or more of the processes and / or blocks Figure 1 specified in one or more of the blocks or blocks.

[0082] Specific embodiments are used in the present invention to illustrate the principles and implementation manners of the present invention. The description of the above embodiments is only for helping to understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0083] Those of ordinary skill in the art will realize that the embodiments described herein are for helping the reader understand the principles of the present invention and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations without departing from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A method for identifying and managing vehicle risk overspeeding areas, characterized in that: The following steps are involved: S1. Obtain the comprehensive index set of vehicle risk speeding areas based on trajectory spatiotemporal matching analysis; S2. Construct a multidimensional feature vector of the vehicle risk overspeeding area, including regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental influencing factors, regional trajectory attributes and behavioral characteristics, and obtain historical data of the multidimensional feature vector of the vehicle risk overspeeding area corresponding to the comprehensive index set of the vehicle risk overspeeding area; S3, constructing a vehicle risk overspeeding area prediction model, training the vehicle risk overspeeding area prediction model based on the vehicle risk overspeeding area comprehensive index set and the historical data of the vehicle risk overspeeding area multidimensional feature vector corresponding to the vehicle risk overspeeding area comprehensive index set, and obtaining the trained vehicle risk overspeeding area prediction model; S4. Acquire real-time data of the multidimensional feature vector of the vehicle risk overspeeding area, and obtain the prediction result of the comprehensive index of the vehicle risk overspeeding area based on the real-time data of the multidimensional feature vector of the vehicle risk overspeeding area and the trained vehicle risk overspeeding area prediction model, and perform regional risk management operations according to the prediction result of the comprehensive index of the vehicle risk overspeeding area.

2. The method for identifying and managing vehicle risk overspeeding areas according to claim 1, characterized in that: Step S1 includes the following steps: S11, constructing a neighborhood trajectory database of the vehicle accident location based on trajectory spatiotemporal matching analysis; S12, identifying vehicle risk overspeeding events and calculating vehicle risk overspeeding indexes on the vehicle accident location neighborhood trajectory database to extract a vehicle risk overspeeding event data set; S13. Based on the vehicle risk overspeeding event data set, calculate the vehicle risk overspeeding area comprehensive index to obtain a vehicle risk overspeeding area comprehensive index set.

3. The method for identifying and managing vehicle risk overspeeding areas according to claim 2, characterized in that: Step S13 includes the following steps: S131, dividing the vehicle risk overspeeding area into grids, and obtaining grid data of the vehicle lane overspeeding area; S132, based on the vehicle risk overspeeding event data set, obtaining the risk overspeeding time index value of the grid data of the vehicle lane overspeeding area, and normalizing the risk overspeeding time index value of the grid data of the vehicle lane overspeeding area; S133. Calculate the vehicle risk overspeeding area comprehensive index according to the normalized risk overspeeding time index value of the grid data of the vehicle lane overspeeding area to obtain a vehicle risk overspeeding area comprehensive index set.

4. The method for identifying and managing vehicle risk overspeeding areas according to claim 3, characterized in that: In step S132, the risk speeding time index value of the raster data of the vehicle lane speeding area includes the risk speeding index values ​​of 8 time periods, which are 7:00-9:00, 9:00-17:00, 17:00-19:00, 19:00-23:00 on weekdays, and 7:00-9:00, 9:00-17:00, 17:00-19:00, and 19:00-23:00 on non-working days; the risk speeding index value includes the risk speeding occurrence intensity, the risk speeding occurrence rate and the risk speeding average index.

5. The method for identifying and managing vehicle risk overspeeding areas according to claim 1, characterized in that: In step S2, the regional static traffic characteristics include 12 variables, including road length attribute feature vector, road linear geometry parameter feature vector and traffic facility feature vector; the road length attribute feature vector includes 7 variables: road length, bridge type, tunnel type, highway type, road grade, number of lanes, design speed; the road linear geometry parameter feature vector includes 2 variables: longitudinal slope, horizontal curve radius; the traffic facility feature vector includes 3 variables: number of traffic lights, number of speed limit signs, number of speed bumps; The regional dynamic traffic characteristics include the time-divided traffic flow level characteristic vector and the list vector of passing vehicle models and corresponding average speeds, with a total of 24 variables; the time-divided traffic flow level characteristic vector includes the aggregated traffic volume of 8 time periods, with a total of 8 variables; the passing vehicle model and corresponding average speed list vector includes the average speed and speed variance of 8 vehicle types, with a total of 16 variables. Regional environmental influencing factors include environmental factor feature vectors, road construction condition feature vectors, and date time feature vectors, with a total of 12 variables; the environmental factor feature vector includes weather, visibility, temperature, and humidity, with a total of 4 variables; the road construction condition feature vector includes construction status, with a total of 1 variable; the date time feature vector includes unique hot variables in 8 time periods, with a total of 7 variables; The regional trajectory attributes and behavior characteristics include the feature vector of the continuous driving time list of vehicles passing through in different time periods and the speeding behavior characteristics, a total of 10 variables; the feature vector of the continuous driving time list of vehicles passing through in different time periods includes the statistics of the continuous driving time of vehicles in 8 time periods, a total of 8 variables; the speeding behavior characteristics include the historical speeding behavior incidence rate and the historical speeding behavior intensity, a total of 2 variables.

6. The method for identifying and managing vehicle risk overspeeding areas according to claim 1, characterized in that: In step S3, the vehicle risk overspeeding area prediction model includes: The structured processing module is used to structure the regional static traffic characteristics, regional dynamic traffic characteristics, regional environmental influencing factors, regional trajectory attributes and behavioral characteristics to obtain a 58-dimensional structured feature vector; The information dimension upgrading and compression module is used to upgrade the 58-dimensional structured feature vector through the elu activation function to obtain a 128-dimensional latent vector containing abstract information; the dimension is upgraded to 512 dimensions through a multi-layer fully connected neural network, and then the latent vector containing abstract information is gradually compressed to obtain a compressed 64-dimensional feature latent vector; A multi-output reasoning module is used to pass the 64-dimensional feature latent vector through three prediction modules respectively to obtain three values ​​as the prediction result of the risk overspeeding intensity, the prediction result of the risk overspeeding incidence rate and the prediction result of the risk overspeeding average index; The vehicle risk overspeeding area prediction index module is used to arithmetically add the risk overspeeding intensity prediction results, the risk overspeeding occurrence rate prediction results and the risk overspeeding average index prediction results to obtain the prediction result of the vehicle risk overspeeding area comprehensive index.

7. The method for identifying and managing vehicle risk overspeeding areas according to claim 6, characterized in that: In the information dimension upgrading and compression module, the multi-layer fully connected neural network includes 6 fully connected layers connected in sequence. The number of neurons in the first fully connected layer is set to 128, the number of neurons in the second fully connected layer is set to 256, the number of neurons in the third fully connected layer is set to 512, the number of neurons in the fourth fully connected layer is set to 256, the number of neurons in the fifth fully connected layer is set to 128, and the number of neurons in the sixth fully connected layer is set to 64.

8. The method for identifying and managing vehicle risk overspeeding areas according to claim 1, characterized in that: In step S4, regional risk management operations are performed according to the predicted results of the comprehensive index of the vehicle risk speeding area, specifically: the size of the predicted result of the comprehensive index of the vehicle risk speeding area is judged; if the predicted result of the comprehensive index of the vehicle risk speeding area is between the top 20% and the top 10%, a yellow warning is sent; if the predicted result of the comprehensive index of the vehicle risk speeding area is between the top 10% and the top 5%, an orange warning is sent and reported to the system; if the predicted result of the comprehensive index of the vehicle risk speeding area is in the top 5%, a red warning is sent, reported to the system and a dynamic inspection of the area is carried out.

Images