A traffic risk assessment method and system

Abstract

The application provides a traffic risk assessment method and system, in particular to a traffic risk assessment method and system, the method comprises the following steps: S1) obtaining traffic event data of a road network; S2) constructing a line element network; S3) constructing a risk assessment model based on the line element network and the traffic event data; S4) outputting a spatiotemporal risk distribution matrix; the system comprises: a roadside edge computing unit: generating road traffic events; an intelligent network connection unit: generating collaborative warning information through a vehicle-road collaboration system; a vehicle trajectory platform: analyzing vehicle trajectory data and identifying abnormal driving behavior; a risk assessment unit: a data access module: correlating traffic event data to line elements; a network construction module: dividing road segments by a predetermined length and constructing a line element network; a matrix calculation module: dynamically loading configuration file parameters, setting spatiotemporal radii according to event types, and performing kernel density calculation; a matrix generation module: outputting a spatiotemporal risk distribution matrix.

Description

Technical Field

[0001] This invention relates to a risk assessment method and system, specifically a traffic risk assessment method and system. Background Technology

[0002] With the rapid development of intelligent transportation systems and vehicle-to-everything (V2X) technology, utilizing multi-source data to achieve accurate road risk assessment has become a core direction for improving traffic safety. Current mainstream risk assessment methods have three main limitations: First, traditional risk statistical models based on fixed road segments or grid areas (such as accident frequency statistics and static density clustering) struggle to accurately depict the spatiotemporal propagation characteristics of traffic events within the road network. Due to the connectivity of road networks, the impact of a single event can spread along the topology to adjacent road segments, and the spatiotemporal impact radii of different event types (such as vehicles driving in the wrong direction or collisions involving vulnerable road users) differ significantly. Existing methods generally ignore road network topological constraints and the nonlinear characteristics of event propagation, leading to risk calculations that deviate from reality. Second, existing spatiotemporal analysis methods (such as time-slice statistics and spatial buffer overlay) often treat time and space dimensions separately, failing to capture the dynamic evolution of risks in continuous spatiotemporal space, especially lacking sensitivity to short-term sudden risks in complex urban road networks (such as chain reactions caused by sudden acceleration or deceleration of vehicles). Third, the data utilization is limited in scope. Most systems rely solely on historical accident records or data from fixed detectors, failing to integrate multimodal data generated in real time in a vehicle-road cooperative environment (such as road traffic events generated by roadside edge computing units, cooperative early warning information from connected vehicles, and abnormal driving behavior based on trajectory data mining). This results in limited real-time performance and coverage of risk assessment. These shortcomings make the spatiotemporal granularity of risk heat maps output by existing technologies coarse, making it difficult to support real-time route planning for connected vehicles and precise governance decisions by traffic management departments.

[0003] To overcome these bottlenecks, recent research has attempted to introduce spatiotemporal modeling techniques, but practical applications still face challenges: while graph neural network-based methods can express road network topology, they are insufficient for modeling long-distance spatiotemporal dependencies; classical kernel density estimation, although capable of handling continuous distributions, is not optimized for road network structures, its isotropic kernel function cannot adapt to road directional constraints, and it ignores the curvature propagation and boundary effects of event impacts. Furthermore, existing algorithms generally use fixed-length road segments as analysis units, failing to adequately characterize the risk coupling mechanism between intersections and road segments—intersections have concentrated risks due to multi-directional traffic flow conflicts, while the risk distribution within long road segments varies significantly, and uniform segmentation leads to distorted risk values. At the data level, multi-source heterogeneous event data lacks a dynamic matching mechanism with the road network structure, further limiting model accuracy. Summary of the Invention

[0004] This invention provides a traffic risk assessment method and system that integrates road network topology characteristics, adapts to multi-source spatiotemporal event data, and can accurately quantify risk propagation patterns, providing technical support for connected vehicle active safety systems and intelligent traffic management dynamic governance.

[0005] This invention provides a traffic risk assessment method, comprising the following steps:

[0006] S1) Obtain traffic event data for the road network;

[0007] The road network is a collection of intersections and road segments within the coverage area of ​​the traffic risk assessment system. The data type is a high-precision map of the area, including data such as road segment ID, road segment name, road level, and road coordinates.

[0008] The traffic event data in step S1) includes: road traffic events, collaborative early warning information, and abnormal driving behavior of vehicles identified based on vehicle trajectory data. Data types include event ID, event type, event occurrence time, event location coordinates, and event description information.

[0009] The collaborative early warning information refers to dynamic risk early warning data generated in real time through the vehicle-road cooperative system. It is generated by roadside sensing devices or vehicle-mounted communication units based on the analysis of vehicle, pedestrian and road environment status and transmitted to the cloud control platform.

[0010] S2) Construct a line element network by dividing the road connections between intersections into line elements, treating each line element as a node, and the connection relationship as the connection of line elements, thereby constructing a line element network on the basis of the original road network.

[0011] The line element network is a mathematical model that reconstructs the original road network topology into refined analysis units. Its core is to divide the road connection into several lines by setting a preset length threshold. Each line element is a network node, and the connection relationship between the lines constitutes the network edge, thereby realizing a refined expression of the road network structure and risk calculation.

[0012] Step S2) includes a preset length for the line element. If the line element exceeds the preset length, then the maximum number of equal parts not greater than the preset length is taken.

[0013] S3) Construct a risk assessment model based on line element network and traffic event data. The risk assessment model uses a spatiotemporal kernel density estimation model to calculate the risk density value of each line element in a specified time window and generate a spatiotemporal risk distribution matrix.

[0014] The spatiotemporal kernel density estimation model is a risk density calculation method that integrates temporal and spatial dimensions. It quantifies the spatiotemporal impact range of traffic events on the road network through kernel functions, generating a continuous risk distribution field. Its core is to model the impact of each traffic event as a spatiotemporal diffusion surface, and then superimpose all event surfaces to obtain the risk density matrix for the entire road network.

[0015] The spatiotemporal kernel density estimation model in step S3 calculates the weighted density of events for line elements in time and space using the spatiotemporal search radius.

[0016] The spatiotemporal search radius is a core parameter in the spatiotemporal kernel density estimation model that controls the impact range of traffic events. It includes the spatial search radius and the temporal search radius. The spatial search radius defines the shortest path distance for the event to spread along the road topology, while the temporal search radius defines the duration of the event's impact over time. Its values ​​are dynamically configured through a configuration file, differentiated according to the event type, to adapt to road directionality constraints and the physical characteristics of the event.

[0017] The spatiotemporal kernel density estimation model uses a Gaussian kernel function to process the density estimation of multivariate continuous data. The expression for the Gaussian function is: .

[0018] Step S3) further includes:

[0019] S31) Obtain line element and traffic event data, and construct a spatiotemporal network line element model;

[0020] S32) Obtain the spatiotemporal search radius according to the configuration file;

[0021] The configuration file is a structured data file that predefines the model parameter mapping relationship and is used to dynamically load the running parameters of the spatiotemporal kernel density estimation model according to the event type.

[0022] S33) Based on the principle of spatiotemporal kernel density estimation, calculate the impact range of all events to obtain the risk density on each line element.

[0023] S4) Output the spatiotemporal risk distribution matrix.

[0024] The spatiotemporal risk distribution matrix is ​​represented as follows: .

[0025] in The risk density of line element i in the j-th time window is represented by the value range [0, 100]. The higher the risk density, the higher the risk of a traffic event occurring in road segment i within the j-th time window.

[0026] Step S4) further includes: plotting the three-dimensional risk distribution assessment results and visualizing the traffic risk density matrix.

[0027] The present invention also provides a traffic risk assessment system, comprising:

[0028] Roadside edge computing unit: generates road traffic events;

[0029] Intelligent connected unit: Generates collaborative early warning information through vehicle-road cooperative system;

[0030] Vehicle trajectory platform: Analyzes vehicle trajectory data to identify abnormal driving behavior;

[0031] Risk assessment unit:

[0032] Data access module: Associates traffic incident data with line elements;

[0033] Network construction module: Divide road segments according to preset lengths and construct a line element network;

[0034] Matrix calculation module: dynamically loads configuration file parameters, sets the spatiotemporal radius according to event type, and performs kernel density calculation;

[0035] Matrix generation module: Outputs a spatiotemporal risk distribution matrix.

[0036] Based on the aforementioned spatiotemporal risk distribution matrix, instructions to avoid high-risk road sections are sent to connected vehicles, or suggestions for managing risk points are output to the traffic management platform.

[0037] The present invention has the following beneficial effects:

[0038] 1. This invention analyzes the road topology connectivity based on the line element network model, accurately depicting the spatiotemporal diffusion pattern of traffic events along the road network.

[0039] 2. This invention achieves continuous dynamic risk early warning by simultaneously calculating time and space dimensions through spatiotemporal kernel density estimation.

[0040] 3. This invention integrates roadside event, vehicle collaborative early warning and abnormal behavior data to improve the real-time performance of risk assessment and the completeness of road network coverage.

[0041] 4. This invention dynamically adjusts the scope of event influence, solving the problem of traditional methods ignoring road direction and boundary effects.

[0042] 5. This invention replaces fixed road segments with refined line element units to accurately identify high-risk areas at intersections and risk gradients within road segments. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the steps of a traffic risk assessment method according to the present invention;

[0044] Figure 2 This is a block diagram of a traffic risk assessment system according to the present invention;

[0045] Figure 3 This is a block diagram of the risk assessment unit of a traffic risk assessment system according to the present invention.

[0046] Figure Labels

[0047] Roadside edge calculation unit-1;

[0048] Intelligent Connected Unit-2;

[0049] Vehicle trajectory platform-3;

[0050] Risk Assessment Unit-4;

[0051] Data access module-41;

[0052] Network building block - 42;

[0053] Matrix calculation module -43;

[0054] Matrix generation module - 44. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. The described embodiments are some embodiments of this application, but not all embodiments. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0056] Furthermore, it should be noted that, unless otherwise explicitly specified and limited, terms such as “installation,” “connection,” and “linkage” used in the description of this application should be interpreted broadly. For example, a connection can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; or it can be a connection within two components. Those skilled in the art can understand its specific meaning in this application according to the specific circumstances.

[0057] Figure 1 This is a flowchart illustrating the steps of a traffic risk assessment method according to the present invention, as follows: Figure 1 As shown, a traffic risk assessment method includes the following steps:

[0058] S1) Obtain traffic event data for the road network;

[0059] The road network is a collection of intersections and road segments within the coverage area of ​​the traffic risk assessment system. The data type is a high-precision map of the area, including data such as road segment ID, road segment name, road level, and road coordinates.

[0060] The traffic event data in step S1) includes: road traffic events, collaborative early warning information, and abnormal driving behavior of vehicles identified based on vehicle trajectory data. Data types include event ID, event type, event occurrence time, event location coordinates, and event description information.

[0061] The data types are shown in the table below:

[0062] The collaborative early warning information refers to dynamic risk early warning data generated in real time through the vehicle-road cooperative system. It is generated by roadside sensing devices or vehicle-mounted communication units based on the analysis of vehicle, pedestrian and road environment status and transmitted to the cloud control platform. Typical types include collision warnings for vulnerable traffic participants (such as pedestrian crossing risks), intersection conflict warnings (such as left-turning oncoming collisions), emergency vehicle priority passage warnings (such as ambulance approaching), and abnormal road condition warnings (such as road obstacles). As a key dynamic input source for the spatiotemporal kernel density estimation model, it is used to capture short-term sudden risks, make up for the lag of historical event data, and support the real-time updating and accurate early warning of the risk matrix.

[0063] S2) Construct a line element network by dividing the road connections between intersections into line elements, treating each line element as a node, and the connection relationship as the connection of line elements, thereby constructing a line element network on the basis of the original road network.

[0064] The line element network is a mathematical model that reconstructs the original road network topology into refined analysis units. Its core is to divide the road connection into several lines by setting a preset length threshold. Each line element is a network node, and the connection relationship between the lines constitutes the network edge, thereby realizing a refined expression of the road network structure and risk calculation.

[0065] Step S2) includes a preset length for the line element. If the line element exceeds the preset length, then the maximum number of equal parts not greater than the preset length is taken.

[0066] In one embodiment, the road segment length is 120 meters, and the preset line element length is 50 meters. Since 120 > 50, the maximum number of equal parts is calculated to be 3, and it is divided into 3 sub-line elements of 40 meters each. Each sub-line element is added to the line element network as an independent node, inheriting the connection relationship of the original road segment.

[0067] S3) Construct a risk assessment model based on line element network and traffic event data. The risk assessment model uses a spatiotemporal kernel density estimation model to calculate the risk density value of each line element in a specified time window and generate a spatiotemporal risk distribution matrix.

[0068] The spatiotemporal kernel density estimation model is a risk density calculation method that integrates temporal and spatial dimensions. It quantifies the spatiotemporal impact range of traffic events on the road network through kernel functions, generating a continuous risk distribution field. Its core is to model the impact of each traffic event as a spatiotemporal diffusion surface, and then superimpose all event surfaces to obtain the risk density matrix for the entire road network.

[0069] The spatiotemporal kernel density estimation model in step S3 calculates the weighted density of events for line elements in time and space using the spatiotemporal search radius.

[0070] The spatiotemporal search radius is a core parameter in the spatiotemporal kernel density estimation model that controls the impact range of traffic events. It includes the spatial search radius and the temporal search radius. The spatial search radius defines the shortest path distance for the event to spread along the road topology, while the temporal search radius defines the duration of the event's impact over time. Its values ​​are dynamically configured through a configuration file, differentiated according to the event type, to adapt to road directionality constraints and the physical characteristics of the event.

[0071] The spatiotemporal kernel density estimation model uses a Gaussian kernel function to process the density estimation of multivariate continuous data. The expression for the Gaussian function is: .

[0072] Step S3) further includes:

[0073] S31) Obtain line element and traffic event data, and construct a spatiotemporal network line element model;

[0074] S32) Obtain the spatiotemporal search radius according to the configuration file;

[0075] The configuration file is a structured data file that predefines the model parameter mapping relationship and is used to dynamically load the running parameters of the spatiotemporal kernel density estimation model according to the event type.

[0076] S33) Based on the principle of spatiotemporal kernel density estimation, calculate the impact range of all events to obtain the risk density on each line element.

[0077] S4) Output the spatiotemporal risk distribution matrix.

[0078] The spatiotemporal risk distribution matrix is ​​represented as follows: .

[0079] in The risk density of line element i in the j-th time window is represented by the value range [0, 100]. The higher the risk density, the higher the risk of a traffic event occurring in road segment i within the j-th time window.

[0080] Step S4) further includes: plotting the three-dimensional risk distribution assessment results and visualizing the traffic risk density matrix.

[0081] Figure 2 This is a block diagram of a traffic risk assessment system according to the present invention, such as... Figure 2 As shown, the present invention also provides a traffic risk assessment system, comprising:

[0082] Roadside edge computing unit 1: Generates road traffic events;

[0083] Intelligent Connected Unit 2: Generates collaborative early warning information through the vehicle-road cooperative system;

[0084] Vehicle trajectory platform 3: Analyzes vehicle trajectory data to identify abnormal driving behavior;

[0085] Risk assessment unit 4 Figure 3 This is a block diagram of risk assessment unit 4 of a traffic risk assessment system according to the present invention, as shown below. Figure 3 As shown, the risk assessment unit 4 includes:

[0086] Data access module 41: Associates traffic event data with line elements;

[0087] Network building module 42: Building a line element network model;

[0088] Matrix calculation module 43: Dynamically loads configuration file parameters, sets the spatiotemporal radius according to event type, and performs kernel density calculation;

[0089] Risk matrix generation module 44: Outputs the spatiotemporal risk distribution matrix.

[0090] Based on the aforementioned spatiotemporal risk distribution matrix, instructions to avoid high-risk road sections are sent to connected vehicles, or suggestions for managing risk points are output to the traffic management platform.

[0091] In one embodiment, the traffic risk assessment system provides connected vehicles with diversified vehicle-to-everything (V2X) scenario information services, encompassing both safety and efficiency, based on road traffic events generated by roadside edge computing (MEC) and collaborative early warning information generated based on core intelligent connected vehicle services. Simultaneously, by combining data from mass-produced connected vehicles, testing and demonstration vehicles, it analyzes and calculates abnormal driving behaviors based on vehicle trajectory data. This data is of significant value in road traffic safety risk analysis and other areas.

[0092] Based on road traffic incidents, collaborative early warning information, and abnormal vehicle driving behavior as inputs, and based on the road network topology modeling covered by the risk assessment system, the traffic risk assessment system will mine and identify the risk density on the urban traffic road network, identify high-risk points in road traffic, and conduct road traffic risk analysis with the aim of improving traffic safety.

[0093] In the model algorithm construction process, road connections are divided into line elements no longer than a preset length (in principle, road segments between intersections are marked as one line element; if a line element exceeds the preset length, it is divided into the largest possible number of equal parts not exceeding the preset length). Each line element can be considered as a node, and the connection relationship can be considered as the connection of line elements. Based on this, a line element network is constructed on the original road network, and the risk density of the urban traffic road network is calculated based on the line element network. This provides a spatiotemporal risk distribution matrix for connected vehicle operation, enabling connected vehicles to avoid high-risk road sections in advance and improving vehicle operation safety.

[0094] The spatiotemporal risk matrix generated by the traffic risk assessment system can provide risk alerts to vehicles, including safety risk hotspots in the spatiotemporal dimensions of the urban road network, for use by vehicles unfamiliar with road conditions. It identifies road risk points in actual traffic management operations, allowing for targeted management; traffic management can also combine this data with other data for correlation analysis, and can add or exclude hotspots.

[0095] The impact of traffic incidents on road networks exhibits spatiotemporal diffusion. Due to the connectivity of road networks, the occurrence of one incident will subsequently affect surrounding connected roads, causing temporal and spatial asymmetries and delays. Therefore, the Spatiotemporal Kernel Density Estimation (STKDE) model is chosen to calculate the risk value of road segments. The STKDE model considers the asymmetry, curvature, and boundary effects of the impact of traffic incidents on the road network. It calculates the event-weighted density of road segments in time and space using a search radius, thus allowing the risk trend of road segments to be identified by changes in risk density. For the kernel function, a Gaussian kernel function is used to process the density estimation of multivariate continuous data, which can be expressed as: .

[0096] The model acquires line element data and event data from a spatiotemporal database, treating each line element as a node in the network. It constructs a spatiotemporal network line element model based on the connections between line elements and the matching relationships between events and line elements. On this basis, a spatiotemporal kernel density estimation model is used to calculate the risk density on each spatiotemporal line element, ultimately generating the risk assessment results.

[0097] The spatiotemporal risk distribution matrix R is represented as follows: .

[0098] in The risk density of line element i in the j-th time window is represented by the value range [0, 100]. The higher the risk density, the higher the risk of a traffic event occurring in road segment i within the j-th time window.

[0099] The spatiotemporal risk distribution matrix R reflects the change in the risk value of each line element over time. During calculation, the upper limit of the line element length and the time interval for numerical calculation can be controlled to achieve road network risk identification and analysis at different spatiotemporal granularities. For example, the road network can be divided into 50-meter-long line elements, and a 24-hour day can be divided into 48 30-minute time periods. The algorithm can then provide the risk distribution of different time periods and road sections on urban roads.

[0100] In one specific embodiment, the traffic risk assessment system accesses high-precision map data for a 1.2-kilometer-long road segment connecting three intersections. The system sets the preset length of each line element to 50 meters. If the system detects that the length of the road segment exceeds the preset threshold (50 meters), it automatically divides it into 24 equal segments of 50-meter line elements, generating line element network nodes L001-L024.

[0101] During the morning rush hour that day, the system collected multi-source events in real time:

[0102] 1. Roadside edge calculation unit 1 reported a rear-end collision at L008 (occurred at 7:30);

[0103] 2. Intelligent connected unit 2 captured a collision warning for vulnerable road user L015 (occurred at 7:45);

[0104] 3. The vehicle trajectory platform 3 identifies the sudden deceleration behavior of multiple vehicles in L020 (7:50-8:00).

[0105] Risk assessment unit 4 dynamically loads configuration file parameters:

[0106] The spatial and temporal search radius for a rear-end collision is 200 meters in space and 15 minutes in time.

[0107] The spatiotemporal search radius for weak collisions is a spatial radius of 100 meters and a temporal radius of 5 minutes.

[0108] The spacetime search radius for rapid deceleration is 150 meters in space and 3 minutes in time.

[0109] The calculated risk matrix is ​​displayed as follows:

[0110] 7:30-7:45: The rear-end collision caused the risk value of downstream L009-L012 of L008 to rise to 75 (red alert).

[0111] 7:45-7:50: A weak collision causes the local risk peak of L015 to reach 85, but the range is limited to L014-L016;

[0112] 8:00-8:15: The combined effect of sudden deceleration results in a sustained high risk (>80) for L020-L024.

[0113] The embodiments described above are merely further illustrations of the present invention and are not intended to limit the present invention in any other way. The present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding modifications and changes based on the present invention, but all such modifications and changes should fall within the protection scope of the present invention.

[0114] In the description of this application, it should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. For ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0115] It should be noted that, in this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. It should also be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with reference to certain examples may be combined in other examples.

[0116] The above embodiments are provided for those skilled in the art to implement or use this application. Those skilled in the art can make various modifications or changes to the above embodiments without departing from the spirit of this application. Therefore, the scope of protection of this application is not limited to the above embodiments, but should be the maximum scope that conforms to the innovative features mentioned in the claims.

Claims

1. A traffic risk assessment method, characterized by, Comprising the following steps: S1) obtaining traffic event data of a road network; S2) constructing a line element network, dividing the road connections between intersections into line elements, regarding each line element as a node and the connection relationship as the connection of line elements, to thereby construct a line element network on the basis of the original road network; The step S2) comprises a line element preset length, and if a line element exceeds the preset length, the maximum equal division number not greater than the preset length is taken; S3) constructing a risk assessment model based on the line element network and the traffic event data, the risk assessment model adopting a spatio-temporal kernel density estimation model to calculate a risk density value of each line element in a specified time window to generate a spatio-temporal risk distribution matrix; The step S3) adopts the spatio-temporal kernel density estimation model to calculate the weighted density of events of a line element from time and space through a spatio-temporal search radius; The step S3) further comprises: S31) obtaining line elements and traffic event data to construct a spatio-temporal network line element model; S32) obtaining a spatio-temporal search radius according to a configuration file; S33) calculating the influence range of all events according to the principle of spatio-temporal kernel density estimation to obtain the risk density on each line element; S4) outputting the spatio-temporal risk distribution matrix; The output spatiotemporal risk distribution matrix is represented as follows: , wherein denotes the risk density of link i in the jth time window, which takes a value in the range [0, 100], and the greater the risk density, the higher the risk of link i in the jth time window. Based on the spatio-temporal risk distribution matrix, a high-risk road section avoidance instruction is sent to a connected vehicle, or a risk point management suggestion is output to a traffic management platform.

2. The traffic risk assessment method according to claim 1, characterized in that, The traffic event data in the step S1) comprises road traffic events, cooperative warning information and vehicle abnormal driving behaviors identified based on vehicle trajectory data.

3. The traffic risk assessment method according to claim 1, characterized in that, The spatio-temporal kernel density estimation model uses a Gaussian kernel function to process the density estimation of multivariate continuous data, and the expression of the Gaussian function is: .

4. The traffic risk assessment method according to claim 1, characterized in that, The step S4) further comprises drawing a three-dimensional risk distribution assessment result.

5. A traffic risk assessment system based on the traffic risk assessment method according to any one of claims 1 to 4, characterized in that Comprise: A roadside edge computing unit: generating road traffic events; An intelligent connected unit: generating cooperative warning information through a vehicle-road cooperative system; A vehicle trajectory platform: analyzing vehicle trajectory data to identify abnormal driving behaviors; A risk assessment unit: A data access module: associating traffic event data to line elements;   A network construction module: dividing road sections according to a preset length to construct a line element network; A matrix calculation module: dynamically loading configuration file parameters, setting spatio-temporal radii according to event types and performing kernel density calculation;   A matrix generation module: outputting a spatio-temporal risk distribution matrix.

Images