A highway risk map construction method and device integrating multi-source data

Abstract

The present invention relates to the field of traffic safety technology, in particular to a method and device for constructing a highway risk map that integrates multi-source data, the method comprising the following steps: S1: selecting traffic conflict risk, driving behavior risk and vehicle safety risk as risk dimensions, selecting risk indicators according to traffic conflict risk, driving behavior risk and vehicle safety risk, and calculating risk indicator data and risk indicator weights; S2: calculating the total risk of a road section according to the weight of the risk dimension, the weight of the risk indicator and the risk indicator data; S3: obtaining a risk map according to the total risk of the road section and the weight of the risk dimension. The method of the present invention strengthens the application of multi-source data fusion in highway risk assessment by constructing a risk map and identifying high-risk areas, can provide a scientific basis and technical support for traffic safety management, and improve the safety of highways.

Description

Technical Field

[0001] The present invention relates to the field of traffic safety technology, and in particular to a method and device for constructing a highway risk map by integrating multi-source data. Background Art

[0002] With the rapid development of the economy and the sharp increase in the number of motor vehicles, highways are facing severe safety challenges. In-depth analysis of the risk mechanism of highways, effective identification and prediction of potential risk areas, and targeted preventive management are of great significance to improving highway safety.

[0003] For example, a Chinese invention patent with announcement number CN118504965A discloses a multi-source data-driven deep prediction method for urban traffic accident risk maps, including the following steps: Step 1, collect satellite images, traffic accident data, and GPS trajectory data of a specified area, and divide the specified area into multiple grid areas; Step 2, calculate the traffic flow data of each grid area based on the GPS trajectory data of the specified area; Step 3, analyze the traffic accident data in each area based on the traffic accident data; and divide the traffic accident data of each area into historical accident data and future accident data; Step 4, use the traffic accident data, GPS trajectory data, and historical accident data as input to train a deep learning model; Step 5, use the trained deep learning model to calculate the traffic accident risk prediction situation at the next moment.

[0004] Existing technologies combine traffic flow data, accident data, monitoring data, etc., and establish highway risk assessment and prediction methods based on data analysis, artificial intelligence, visualization and other technical means. However, existing risk analysis methods still have the following limitations:

[0005] ① Existing risk analysis mainly relies on macro data and ignores detailed analysis at the micro level. These methods fail to comprehensively and accurately identify and quantify the risks in specific traffic behaviors and environments.

[0006] ② When conducting risk assessment on highways, the impact of specific risk factors on highway safety was not determined, and an accurate highway risk assessment model could not be established.

[0007] ③ The existing technology is insufficient in terms of integration and applicability, and cannot fully reflect the risks in different sections and dimensions of the entire highway. Summary of the invention

[0008] The present invention proposes a method and device for constructing a highway risk map by integrating multi-source data. By combining macro and micro perspectives, vehicle, road and driver-related data are acquired through micro traffic flow simulation and driving simulation, and a highway risk map is constructed based on traffic conflict risk, driving behavior risk and vehicle safety risk, so as to realize the monitoring of traffic risks.

[0009] In order to achieve the above objectives, the following technical solutions are proposed:

[0010] A method for constructing a highway risk map by integrating multi-source data includes the following steps:

[0011] S1: Select traffic conflict risk, driving behavior risk and vehicle safety risk as risk dimensions, select risk indicators based on traffic conflict risk, driving behavior risk and vehicle safety risk, and calculate risk indicator data and risk indicator weights;

[0012] S2: Calculate the total risk of the road section based on the weight of the risk dimension, the weight of the risk indicator and the risk indicator data;

[0013] S3: Obtain a risk map based on the total risk of the road section and the weights of the risk dimensions.

[0014] This method integrates microscopic traffic simulation and driving simulation data to assess highway risks, including vehicle status, driver physiological signals, etc., and construct a risk map. It integrates multi-source data and comprehensively assesses risks. It combines macro and micro perspectives to monitor traffic risks, dynamically update risk maps, improve risk identification accuracy, provide a scientific basis for safety management, and enhance highway safety.

[0015] Preferably, the traffic conflict risk selects TTC index and PET index as risk index;

[0016] The TTC index uses the TTC formula to calculate the rear-end collision of straight traffic on the main line, and the PET index uses the PET formula to calculate the lateral conflict in the merge zone.

[0017] Preferably, the data acquisition of the TTC index and the PET index comprises the following steps:

[0018] The road section is simulated according to the SUMO software to obtain the vehicle position, lane ID, vehicle length, speed, acceleration, headway and distance from the conflict point in the road section simulation, which are used to calculate the TTC formula and PET formula.

[0019] Preferably, the driving behavior risk selects the driver's electrocardiogram index and electromyography index as risk indicators;

[0020] The ECG index extracts the R wave in the original ECG signal and calculates the RR wave interval, and calculates the RR wave interval standard deviation based on the RR wave interval to characterize the physiological fluctuation law and state stability of the driver in different driving situations.

[0021] The electromyographic index uses the RMS formula to calculate the surface electromyographic signals of the main force-generating muscles of the steering operation during the simulated driving of the car.

[0022] Preferably, the vehicle safety risk selects stability index and comfort index as risk index;

[0023] Using the driving simulation experiment platform, we conduct BIM modeling and real vehicle driving analysis on the road, obtain the longitudinal stability and lateral stability data of the stability index based on the API interface, and calculate the stability of the vehicle under the condition that there is no lateral or longitudinal slippage;

[0024] The data of the comfort index is obtained through the data interface of the physical engine, and then the comfort of the vehicle is determined by calculating the acceleration interference values ​​in the horizontal and vertical directions.

[0025] Preferably, the weights of the risk dimensions and the weights of the risk indicators are calculated by rough set theory, and the specific steps include:

[0026] S11: Discretize the risk indicators and risk dimension data, convert the continuous values ​​in the risk indicators and risk dimension data into limited intervals or categories, use each risk dimension and risk indicator as a conditional attribute, and use risk assessment as a decision attribute;

[0027] S12: Use the positive domain in rough set theory to evaluate the importance of conditional attributes and calculate the positive domain contribution of conditional attributes relative to decision attributes;

[0028] S13: Calculate the weight of the condition attribute, where the weight of the condition attribute is determined by the relative importance of the condition attribute.

[0029] Preferably, the formula for the importance of the conditional attribute is:

[0030]

[0031] Among them, γ A (D) represents the importance of condition attribute A to decision attribute D; POS A (D) indicates the positive domain, is the complete set of objects;

[0032] The formula for relative importance is:

[0033]

[0034] Among them, ωi is the weight of the i-th attribute; n is the number of risk indicators.

[0035] Preferably, the formula for the total risk of the road section is:

[0036]

[0037] Among them, TSR is the total risk of the road section; TCR is the traffic conflict risk, DBR is the driving behavior risk, and VSR is the vehicle safety risk; ω T ,ω D ,ω V are the weights of traffic conflict risk, driving behavior risk and vehicle safety risk respectively.

[0038] Preferably, the calculation formula of the risk dimension in the total risk of the road section is:

[0039]

[0040] Among them, e 1 、e 2 、e 3 、e 4 、e 5 、e 6 are the standardized risk values ​​of TTC index, PET index, ECG index, EMG index, stability index and comfort index; ω 1 ,ω 2 ,ω 3 ,ω 4 ,ω 5 ,ω 6 They are the weights of TTC, PET, ECG, EMG, stability and comfort in their corresponding risk dimensions.

[0041] Based on the same concept, a highway risk map construction device that integrates multi-source data is also proposed, including at least one processor and a memory that is communicatively connected to the at least one processor; the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any of the above-mentioned methods for constructing a highway risk map that integrates multi-source data.

[0042] Compared with the prior art, the present invention has the following beneficial effects: the method of the present invention combines macro and micro perspectives, obtains vehicle, road and driver-related data through micro traffic flow simulation and driving simulation, constructs a highway risk map based on traffic conflict risks, driving behavior risks and vehicle safety risks, realizes the monitoring of traffic risks, and can intuitively display the risk distribution of different sections of the highway. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flow chart of a highway risk map construction method and device integrating multi-source data in Example 1;

[0044] Figure 2 is a schematic diagram of the highway risk calculation model in Example 1;

[0045] Figure 3 is a flow chart of the SUMO simulation process in Example 1;

[0046] Figure 4 is a flow chart of the driving simulation process in Example 1;

[0047] Figure 5 It is a general flow chart of constructing the highway risk map in Example 1;

[0048] Figure 6 is a schematic diagram of the mode layer of the highway risk map in Example 1;

[0049] Figure 7 This is an example diagram of the highway risk map in Example 1;

[0050] Figure 8 This is a schematic diagram of highway risk coupling in Example 1. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

[0052] Example 1

[0053] A method and device for constructing a highway risk map integrating multi-source data, the flow chart is as follows Figure 1 As shown, the following steps are included:

[0054] S1: Select traffic conflict risk, driving behavior risk and vehicle safety risk as risk dimensions, select risk indicators based on traffic conflict risk, driving behavior risk and vehicle safety risk, and calculate risk indicator data and risk indicator weights;

[0055] S2: Calculate the total risk of the road section based on the weight of the risk dimension, the weight of the risk indicator and the risk indicator data;

[0056] S3: Obtain a risk map based on the total risk of the road section and the weights of the risk dimensions.

[0057] This method combines microscopic traffic flow simulation with driving simulation to obtain information such as the position, speed, acceleration of highway vehicles and the physiological signals of drivers, determine the risk indicator thresholds and weights and calculate the classification, and construct a highway risk map based on traffic conflict risks, driving behavior risks and vehicle safety risks. The innovation of the present invention lies in the proposal of a complete highway risk map construction method that integrates multi-source data to comprehensively evaluate the risks under specific traffic behaviors and environments.

[0058] As a specific embodiment, the main steps include:

[0059] 1) If Figure 2 As shown in the figure, the establishment of risk indicators and the division of thresholds: indicators in three dimensions, namely traffic conflict, driving behavior and vehicle safety, are selected to evaluate highway risks:

[0060] The total risk of a road section is calculated as follows:

[0061]

[0062] in: ; ;

[0063] Where, TSR is the total segment risk (Total Segment Risk); TCR is the traffic conflict risk (Traffic Conflict Risk); DBR is the driving behavior risk (Driving Behavior Risk); VSR is the vehicle safety risk (Vehicle Safety Risk); ω T ,ω D ,ω V is the weight of each dimension of risk; 1 、e 2 、e 3 、e 4 、e 5 、e 6 are the standardized risk values ​​of TTC, PET, ECG, EMG, stability, and comfort risk indicators; ω 1 ,ω 2 ,ω 3 ,ω 4 ,ω 5 ,ω 6 is the weight of each risk indicator in its corresponding risk dimension.

[0064] The threshold division of each risk index is shown in Table 1, where the stability index is the proportion of unstable vehicles in a certain road section to all vehicles, and the other indicators are the average values ​​of all vehicles.

[0065] Table 1 Risk indicator threshold classification

[0066]

[0067] (1) Traffic conflict risk based on TTC and PET: TTC is used to calculate rear-end conflicts in straight-ahead traffic on the main line, and PET is used to calculate lateral conflicts in merge zones:

[0068]

[0069]

[0070] In the formula, v i (t) is the location speed of the following vehicle at the position coordinate of the road at time t; v i-1 (t) is the location speed of the preceding vehicle at the coordinates of the road at time t; h i is the headway l i-1 is the length of the vehicle in front; the distance between vehicle A on the main road and the starting point of the merging conflict area is x A , the speed is v A ; The distance between ramp vehicle B and the starting point of the merging area is x B , the speed is v B , the acceleration is a B .

[0071] (2) Driving behavior risks based on electrocardiogram and electromyography: Analyze the driver's stress level based on electrocardiogram and electromyography signals to verify the safety of the road section.

[0072] ① ECG signal: By extracting the R wave from the original ECG signal and calculating the RR wave interval, the statistical discrete trend analysis method is used to study the standard deviation of the RR wave interval (SDNN):

[0073]

[0074] Where, N is the total number of heart beats; RR i is the i-th RR interval; RR mean is the average value of the RR interval.

[0075] ② Electromyographic signal: RMS is used to process the surface electromyographic signal (EMG) of the main muscles that exert force in steering operation during simulated driving of the car:

[0076]

[0077] In the formula, T is the duration of the selected time period (s), and EMG is the value of the electromyographic signal.

[0078] (3) Vehicle safety risks based on stability and comfort:

[0079] ① Stability: From different directions, vehicle driving stability is divided into longitudinal stability and lateral stability. When a car is driving on a slope, longitudinal slip occurs before longitudinal overturning; when a car is driving on a circular curve, lateral slip occurs before lateral overturning. Therefore, the conditions for lateral and longitudinal stability are that there is no lateral or longitudinal slip:

[0080]

[0081] In the formula, i ψ is the road limit slope; α ψ is the road limit slope angle; ψ is the adhesion coefficient; G d is the axle weight of the driving wheel; G α is the total weight of the vehicle; v is the speed of the vehicle; ψ h is the lateral friction coefficient, generally (0.6~0.7) ψ; i h It is the transverse superelevation slope.

[0082] ② Comfortability: Taking acceleration disturbance as a quantitative evaluation index, the acceleration disturbance values ​​in the horizontal direction (lateral direction) and vertical direction (longitudinal direction) are as follows:

[0083]

[0084]

[0085] In the formula, σ x is the lateral acceleration disturbance value; σ y is the longitudinal acceleration disturbance value; T is the total observation time; α x Lateral acceleration; α y Longitudinal acceleration; is the average lateral acceleration, is the longitudinal average acceleration, and its expression is: .

[0086] Since the human body has different sensitivities to different vibration directions, different weighting coefficients are given to the vibration of the left and right and front and back directions of the human body, which are 0.6 in the left and right (lateral) direction and 0.4 in the front and back (longitudinal) direction. The total acceleration interference value is as follows:

[0087]

[0088] 2) Data acquisition:

[0089] At present, the mainstream microscopic traffic flow simulation software at home and abroad focuses on the simulation of the operation status of traffic flow (group vehicles), and the driving simulation software focuses on the simulation of single vehicle dynamics and driver driving behavior. In order to more objectively simulate the real traffic flow operation status of highways, this invention explores the combination of microscopic traffic flow simulation and driving simulation to obtain multi-source highway risk data. In addition, road design documents, traffic accident data and intersection point data are collected to provide a basis for the division of risk assessment sections.

[0090] ① Microscopic traffic flow simulation data acquisition Figure 3 As shown, the present invention uses SUMO (Simulation of Urban Mobility) software to model and analyze the microscopic traffic flow simulation scene of the road section. In order to reduce the error between the operation of the simulation software and the actual traffic flow operation state and ensure the authenticity of the simulation results, before the simulation, according to the survey of the actual road, the key parameters are calibrated from the perspectives of representative vehicle models, traffic composition, following model, traffic volume, expected speed, lane change behavior, long-distance and short-distance ratio distribution, etc. The simulation time is set to be no less than 4200s, and the first 600s is the traffic flow disorder stability time. Therefore, the recording time of the simulation data is the data after 600s, and the effective recording time is no less than 3600s (i.e. 1h).

[0091] Through the TraCI interface and the simulation output file (.tripinfo.xml), the vehicle position, lane ID, vehicle length, speed, acceleration, headway, and distance from the conflict point in the simulation are obtained, and TTC and PET are calculated. If there is no vehicle in front of the vehicle, no calculation is required. The TTC indicator is the safety level, and the standardized indicator is 0. If the vehicle is not at a merge or diverge point or there is no conflicting vehicle, no calculation is required. The PET indicator is the safety level, and the standardized indicator is 0.

[0092] ② Driving simulation system data acquisition Figure 4 As shown:

[0093] The driving simulation system uses a simulation experiment platform of Unreal Engine + Compact. Through BIM modeling of basic road data and real vehicle driving analysis, and the use of corresponding cardiovascular and physiological monitoring equipment, it can effectively conduct comprehensive analysis from multiple angles and factors such as overall traffic safety, driver cardiovascular and physiological, environmental impact, vehicle motion indicators, and road linear indicators.

[0094] A physiological multi-channel instrument (with surface electromyography, electrocardiography and other main human physiological signal detection functions) is used to connect the physiological signal data acquisition end with the data receiving end using wireless data transmission to obtain electrocardiogram signals (R waves), total number of heart beats and electromyography signals (EMG), and calculate SDNN and RMS respectively.

[0095] Based on the API interface, road parameters (curvature, slope, etc.), road characteristics (adhesion coefficient), vehicle configuration (gross weight, axle weight) are obtained in real time to calculate the vehicle's lateral and longitudinal stability. If the vehicle is not on a ramp or circular curve, no calculation is required. The stability index is a safety level, and the standardized index is 0.

[0096] Through the data interface of the physics engine Unreal Engine, the lateral and longitudinal accelerations of the vehicle are obtained and the comfort index σ is calculated.

[0097] 3) Risk calculation and classification:

[0098] In highway safety risk analysis, it is necessary to evaluate the importance of traffic conflict risk, driving behavior risk, vehicle safety risk and various risk indicators. The present invention uses rough set theory to calculate weights. The core idea is to determine the weights by the importance of attributes without relying on any prior information. First, the risk indicator data needs to be discretized, the continuous values ​​are converted into finite intervals or categories, and a decision table is established (that is, each indicator is used as a conditional attribute, and risk assessment is used as a decision attribute). Secondly, the positive region in the rough set theory is used to evaluate the importance of each attribute, and the positive region contribution of each conditional attribute to the decision attribute is calculated. The specific formula is:

[0099]

[0100] Among them, γ A (D) represents the importance of condition attribute A to decision attribute D; POS A (D) indicates the positive domain, The complete set of objects.

[0101] Finally, the weight of each attribute is calculated. The weight is determined by the relative importance of the attribute. The formula is:

[0102]

[0103] Among them, ω i is the weight of the i-th attribute; n is the number of risk indicators.

[0104] Through the above steps, the traffic conflict, driving behavior, vehicle safety and the weights of each risk index are obtained as shown in Table 2.

[0105] Table 2 Highway risk weights

[0106]

[0107] To facilitate comparison and analysis, the risk indicators are standardized and each risk indicator is mapped to the range of 0 to 1.

[0108] For TTC, PET, and ECG, the standardized formula is:

[0109]

[0110] In the formula, e is the risk index value; e _max is the maximum value of the risk index of this level; e _min is the minimum value of the risk index of this level; the standardized TTC, PET, and ECG risk index values ​​are e 1 、e 2 、e 3 .

[0111] For myoelectricity, stability, and comfort, the standardized formula is:

[0112]

[0113] As shown in Table 3, the standardized risk index values ​​of myoelectricity, stability, and comfort are e 4 、e 5 、e 6 .

[0114] Table 3 Standardized risk level classification table

[0115]

[0116] 4) Risk map drawing: including the construction of the road section ontology, road section knowledge extraction, knowledge fusion and knowledge storage. First, the road section ontology is constructed, and the road sections are divided according to the actual road characteristics (road curvature, slope, bridges, tunnels, entrances and exits and other special areas), traffic flow characteristics, historical traffic accidents, etc. The road section data includes the road section ID, starting point pile number, etc., to form a unified knowledge framework; secondly, the road section knowledge is extracted, based on micro traffic flow simulation and driving simulation, the risk data corresponding to the three dimensions of traffic conflict, driving behavior, and vehicle safety are obtained, and converted into nodes and relationships in the map; then, knowledge fusion is performed to integrate information from different data sources to solve data heterogeneity and redundancy problems, and the risks of each dimension and the total risk of the road section are calculated; finally, knowledge storage is performed, and data is updated regularly to ensure the accuracy and timeliness of the map. The construction process of the highway risk map is as follows: Figure 5 shown.

[0117] By integrating multi-source highway risk data, a top-down approach is adopted to construct a knowledge graph, such as Figure 6 As shown in the figure, the road section is the most basic element in the highway. The risk map starts from the road section and the highway risk map ontology model (H) can be defined as the following structure:

[0118]

[0119] Among them, Seg represents the collection of road sections, which refers to several road sections after the highway is divided; Dim represents the collection of risk dimensions; Met represents the collection of risk indicators; Rel represents the relationship, including the relationship between road sections and risk dimensions, and the relationship between risk dimensions and risk indicators; Val represents the risk value at each level.

[0120] When constructing a highway risk map, we describe the road sections, risk dimensions, risk indicators and their corresponding risk values ​​based on the entity-relationship-entity and entity-attribute-value relationships, and build a complete risk assessment framework. The risk map model consists of four core elements: entity, relationship, attribute and value. Entities represent objects in the map, relationships connect entities, attributes are used to describe the characteristics of entities or the connection between entities and values, and values ​​are used to represent the specific values ​​of attributes.

[0121] In the highway risk map, the segment (Segment) is the basic entity with multiple risk dimensions (Dim), such as traffic conflict risk (TCR), driving behavior risk (DBR) and vehicle safety risk (VSR). These risk dimensions further contain multiple risk indicators (Metric), such as TTC and PET in traffic conflict risk. The relationships in the map are used to connect these entities and describe the semantic connection between them. For example, the relationship between the segment and traffic conflict can be expressed as (Segment, has_risk_dimension, TCR); the relationship between traffic conflict and risk indicator TTC can be expressed as (TCR, has_metric, TTC). Each entity has a specific risk value (Value) to quantify the degree of risk, for example, (Segment, has_value, 0.3 (safer)), (TCR, has_value, 0.6 (more dangerous)), (TTC, has_value, 0.9 (dangerous)). Through these risk values, the risk of each segment can be quantitatively evaluated, thereby providing data support for decision-making.

[0122] Through the iteration and superposition of multiple triples, the risk map forms a multi-level topological structure. In this topology, each risk factor and its relationship chain constitutes a recursive risk propagation path in turn. It not only realizes the interaction modeling of risk factors between single nodes, but also constructs an integrated representation of the overall risk factors at the macro level, revealing the complex interdependence between factors at different levels.

[0123] The risk map drawing includes three levels: the first level is the road section level, the second level is the risk dimension level, and the third level is the risk indicator level. The size of the entity circle represents the importance of the risk. The larger the circle, the higher the risk importance of the entity. The depth of the color represents the level of risk. The darker the color, the higher the risk level. Among them, the risk importance represented by the size of the circle is determined by the risk level and risk weight. The calculation formula for the importance of different levels is as follows:

[0124] Third layer: ,in ,

[0125] Second layer: ,in ,

[0126] First layer:

[0127] Among them, Q T Corresponding to TSR; Q D Corresponding DBR; Q V Corresponding to VSR.

[0128] In the visualization design of the risk map, the higher the color saturation, the higher the risk level. The larger the relative area of ​​the circle under the same type of risk, the higher the importance. Figure 7 shown.

[0129] Based on the risk map, a risk coupling method for the same risk level in different road sections is proposed, such as Figure 8 As shown in the figure, more effective risk integration and management can be achieved. Different sections with the same risk level in the risk dimension layer and risk indicator layer are merged to form a unified risk node. Define section S i The risk set of the risk indicator layer is , the risk set of the risk dimension layer is ; Each risk index e, d corresponds to a risk level. If two sections S i and S jIf there are indicators with the same risk level in the graph, the two can be coupled. To simplify the graph structure, duplicate edges can be deleted to obtain a coupled risk graph. This process aims to eliminate redundant information and enhance the understanding of the common risk characteristics of multiple road sections. Through risk coupling, the overall performance of different road sections under specific risk indicators can be better analyzed and evaluated, allowing traffic management departments to more effectively identify and respond to these risks and provide a clear basis for decision-making.

[0130] Finally, it should be noted that the embodiments described in detail above are only the best practices of the invention and cannot be used to limit the scope of rights of the invention. Equivalent replacement of the technical solutions recorded in the aforementioned embodiments does not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the invention, and they should all be included in the scope of the claims and description of the invention.

Claims

1. A highway risk map construction method integrating multi-source data, characterized in that: The following steps are involved: S1: traffic conflict risk, driving behavior risk and vehicle safety risk are selected as risk dimensions, and risk indicators are selected according to the traffic conflict risk, driving behavior risk and vehicle safety risk. The traffic conflict risk selects TTC indicator and PET indicator as risk indicators; The driving behavior risk selects the driver's electrocardiogram index and electromyography index as risk indicators; The vehicle safety risk selects stability index and comfort index as risk index; and calculate the risk indicator data and the weight of the risk indicator; S2: Obtain traffic conflict risk data based on SUMO microscopic traffic flow simulation, build real road scenes based on BIM, collect driving behavior risk and vehicle safety risk data in combination with the driving simulation platform, and standardize the risk indicator data; The total risk of the road section is calculated according to the weight of the risk dimension, the weight of the risk index and the risk index data; the weight of the risk dimension and the weight of the risk index are calculated by rough set theory, and the specific steps include: Discretize the risk indicators and risk dimension data, convert the continuous values ​​in the risk indicators and risk dimension data into limited intervals or categories, use each risk dimension and risk indicator as a conditional attribute, and use risk assessment as a decision attribute; Use the positive domain in rough set theory to evaluate the importance of conditional attributes and calculate the positive domain contribution of conditional attributes relative to decision attributes; Calculate the weight of the condition attribute, the weight of the condition attribute is determined by the relative importance of the condition attribute; The formula for the total risk of the section is: in, is the total risk of the road section; For traffic conflict risks, Driving behavior risk, Risks to vehicle safety; , , are the weights of traffic conflict risk, driving behavior risk and vehicle safety risk respectively; S3: Obtain a risk map based on the total risk of the road section and the weights of the risk dimensions; S4: construct a dynamic risk map according to the risk map, wherein the dynamic risk map includes a multi-level coupling structure of a road section layer, a risk dimension layer, and a risk indicator layer, and dynamically displays the risk level and weight importance through color saturation and entity area; S5: different road sections with the same risk level in the risk dimension layer and the risk indicator layer are merged to form a unified risk node, and a coupled risk map is obtained.

2. The method for constructing a highway risk map by integrating multi-source data according to claim 1, characterized in that: The TTC index uses the TTC formula to calculate the rear-end collision of straight traffic on the main line, and the PET index uses the PET formula to calculate the lateral conflict in the merge zone.

3. The method for constructing a highway risk map by integrating multi-source data according to claim 2, characterized in that: The data acquisition of the TTC index and the PET index comprises the following steps: The road section is simulated according to the SUMO software to obtain the vehicle position, lane ID, vehicle length, speed, acceleration, headway and distance from the conflict point in the road section simulation, which are used to calculate the TTC formula and PET formula.

4. The highway risk map construction method integrating multi-source data according to claim 1 is characterized in that: The ECG index extracts the R wave in the original ECG signal and calculates the RR wave interval, and calculates the RR wave interval standard deviation according to the RR wave interval, so as to characterize the physiological fluctuation law and state stability of the driver in different driving situations; The electromyographic index uses the RMS formula to calculate the surface electromyographic signals of the main force-generating muscles of the steering operation during the simulated driving of the car.

5. The method for constructing a highway risk map by integrating multi-source data according to claim 1, characterized in that: Using the driving simulation experiment platform, we conduct BIM modeling and real vehicle driving analysis on the road, obtain the longitudinal stability and lateral stability data of the stability index based on the API interface, and calculate the stability of the vehicle under the condition that there is no lateral or longitudinal slippage; The data of the comfort index is obtained through the data interface of the physical engine, and then the comfort of the vehicle is determined by calculating the acceleration interference values ​​in the horizontal and vertical directions.

6. The highway risk map construction method integrating multi-source data according to claim 1 is characterized in that: The formula for the importance of the conditional attribute is: in, Represents conditional attributes Decision attributes The importance of represents the positive domain, is the complete set of objects; The formula for relative importance is: in, For the The weight of each attribute; is the number of risk indicators.

7. The method for constructing a highway risk map by integrating multi-source data according to claim 1, characterized in that: The calculation formula of the risk dimension in the total risk of the road section is: ， ， in, , , , , , They are the standardized risk values ​​of TTC index, PET index, ECG index, EMG index, stability index, and comfort index; , , , , , They are the weights of TTC, PET, ECG, EMG, stability and comfort in their corresponding risk dimensions.

8. A highway risk map construction device integrating multi-source data, characterized in that: It includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a highway risk map construction method integrating multi-source data as described in any one of claims 1 to 7.

Images