A method and system for identifying a highway accident-prone point

Abstract

The application relates to a highway accident-prone point identification method and system, which comprises the following steps: acquiring road section characteristic data, vehicle braking performance parameters and historical accident data of a target highway; performing simulation based on the road section characteristic data and the vehicle braking performance parameters, calculating braking distances and acquiring braking efficiency indexes under different road section characteristics; classifying the historical accident data based on the braking efficiency indexes, extracting braking accidents, analyzing the clustering characteristics of the braking accidents under different road section characteristics, and acquiring risk road sections; extracting the characteristic data of the risk road sections, grouping the risk road sections with similar road section characteristics and braking efficiency decline characteristics by using a clustering analysis method, and determining road section categories with common risks; acquiring key influence factors of braking distances and high-risk road sections, i.e. accident-prone points of the target highway. The application can identify dangerous road sections with braking efficiency decline caused by road characteristics.

Description

Technical Field

[0001] This invention relates to the field of traffic safety technology, and in particular to a method and system for identifying accident hotspots on highways. Background Technology

[0002] Highway safety is a crucial research area in the transportation sector, directly impacting life and property safety as well as traffic efficiency. Accurate identification of accident-prone areas can effectively reduce accident rates and provide a scientific basis for road design and safety management. However, existing methods largely rely on accident statistics, neglecting the underlying causes of accidents, particularly road segment characteristics related to vehicle braking performance. This makes it difficult to accurately identify potential hazards and provides specific guidance for targeted improvements. Consequently, road safety management has limited effectiveness in complex environments.

[0003] Currently, the core challenge in identifying accident-prone areas on highways lies in effectively combining the relationship between braking performance and road segment characteristics. Braking distance is significantly affected by factors such as gradient, curve radius, and road surface friction coefficient, but the dynamic interaction of these factors has not been fully analyzed. For example, steep slopes may significantly increase braking distance, while at sharp curves, braking effectiveness is further limited due to decreased vehicle stability. This complexity makes single-factor analysis insufficient to capture the true cause of accidents. Furthermore, existing research lacks in-depth exploration of the spatial distribution patterns of braking-related accidents, failing to reveal the clustering characteristics of dangerous road segments through data analysis. This not only limits the accurate identification of high-risk road segments but also makes the optimized placement of safety facilities lack a targeted basis. Therefore, this invention proposes a method and system for identifying accident-prone areas on highways. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for identifying accident-prone points on highways. By establishing the correlation between braking distance and road segment characteristics, and combining big data analysis, the spatial clustering of braking-related accidents is revealed, and dangerous road segments with reduced braking performance due to insufficient slope, curve radius, or road surface friction coefficient are identified.

[0005] To achieve the above objectives, on the one hand, the present invention provides a method for identifying accident-prone areas on highways, comprising:

[0006] Acquire road segment characteristic data, vehicle braking performance parameters, and historical accident data of the target highway;

[0007] Simulations were performed based on the road segment characteristic data and vehicle braking performance parameters to calculate braking distance and obtain braking performance indicators under different road segment characteristics.

[0008] Based on the braking performance index, the historical accident data is classified and processed to extract braking accidents, and the clustering characteristics of the braking accidents under different road segment characteristics are analyzed to identify risky road segments.

[0009] Extract the characteristic data of the risky road sections, and use cluster analysis to group the risky road sections with similar road section characteristics and reduced braking performance to determine the road section categories with common risks;

[0010] Identify the key factors affecting braking distance in each category and determine the high-risk road sections in each category, i.e., the accident-prone points of the target highway.

[0011] Optionally, simulations are performed based on the road segment characteristic data and vehicle braking performance parameters to calculate the braking distance and obtain braking performance indicators under different road segment characteristics, including:

[0012] A correlation analysis was performed on the road segment characteristic data and vehicle braking performance parameters to determine the correlation coefficient and influencing factors between the characteristics of each road segment and braking performance.

[0013] Based on the influencing factors, the road segment characteristic data are grouped to obtain road segment characteristic distribution information;

[0014] The road segment characteristic distribution information is input into a pre-built braking distance calculation model to simulate the braking distance changes under different scenarios, and the braking efficiency of each road segment under different scenarios is judged based on the braking distance changes.

[0015] Optionally, determining the braking performance of each road segment under different scenarios based on the braking distance change includes:

[0016] Based on the aforementioned braking distance changes, regression analysis was performed using corresponding road segment characteristic data to determine the main influencing factors.

[0017] Based on the main influencing factors and combined with simulation data from various scenarios, the mapping relationship between braking performance and performance indicators is analyzed to determine the braking performance level of each road segment under different scenarios and obtain the braking performance evaluation results of the road segment.

[0018] Optionally, the historical accident data is classified based on the braking performance index to extract braking accidents, and the clustering characteristics of the braking accidents under different road segment characteristics are analyzed to obtain risky road segments, including:

[0019] Set the braking performance index threshold according to the braking performance index;

[0020] Based on the aforementioned braking performance index threshold, historical accident data is classified and processed to extract records containing braking-related information, thereby obtaining a braking accident dataset.

[0021] Spatial distribution information is obtained from the braking accident dataset, and the kernel density estimation algorithm is used to analyze the clustering characteristics of braking accidents on different road sections to obtain the braking accident density distribution.

[0022] The braking accident density distribution is correlated with road segment characteristic data to obtain a set of risky road segment characteristics, and a list of risky road segments is marked based on the set of risky road segment characteristics.

[0023] Based on the list of risky road segments, the K-means clustering algorithm is used to perform spatial clustering to determine the spatial distribution pattern of risky road segments and the areas of concentrated risk.

[0024] Optionally, characteristic data of the risky road segments are extracted, and cluster analysis is used to group risky road segments with similar characteristics and reduced braking performance to determine road segment categories with common risks, including:

[0025] Extract the characteristic data corresponding to the risky road sections and construct a characteristic combination dataset;

[0026] Based on the aforementioned characteristic combination dataset, cluster analysis is used to group road segment characteristics and braking performance degradation features to determine preliminary similar road segments.

[0027] The braking performance degradation characteristic distribution of the preliminary similar road sections is judged to obtain a set of risk-marked road sections, and the correlation pattern between common risks and road section characteristics is analyzed to determine the common risk feature set;

[0028] By combining the common risk feature set with road segment category information, the risk-marked road segment set is classified a second time to obtain road segment category groups.

[0029] Optionally, key factors affecting braking distance in each category are identified, and high-risk road sections in each category are determined, including:

[0030] Extract the road segment characteristic data and braking distance for each road segment category, and use statistical analysis methods to calculate the correlation strength between the characteristic parameters of each road segment and the braking distance to identify key influencing factors;

[0031] Based on the aforementioned key influencing factors, a preliminary high-risk road segment is identified in each road segment category, and a preliminary set of high-risk road segments is obtained.

[0032] Spatial proximity analysis is performed on the preliminary set of high-risk road segments to obtain the correlation of distribution characteristics between road segments, and priority ranking is performed based on the correlation to construct the final set of high-risk road segments.

[0033] Optionally, after obtaining the final set of high-risk road sections, the method further includes: generating a corresponding risk level assessment report for each high-risk road section based on historical accident clustering characteristics, and visually presenting the spatial distribution and key influencing factors of the high-risk road sections based on the risk level assessment report.

[0034] On the other hand, the present invention also provides a highway accident multi-occurrence point identification system, comprising:

[0035] The road segment data acquisition module is used to acquire road segment characteristic data, historical accident data, and vehicle braking performance parameters of the target highway.

[0036] The braking performance analysis module is used to perform simulation based on the road segment characteristic data and vehicle braking performance parameters, calculate braking distance, and obtain braking performance indicators under different road segment characteristics.

[0037] The risk road segment extraction module is used to classify the historical accident data based on the braking performance index, extract braking accidents, and analyze the clustering characteristics of the braking accidents under different road segment characteristics to obtain risk road segments.

[0038] The risk road segment classification module is used to extract the characteristic data of the risk road segments and use cluster analysis to group risk road segments with similar road segment characteristics and braking performance degradation characteristics to determine the road segment categories with common risks.

[0039] The high-risk assessment module is used to obtain the key influencing factors on braking distance in each category and to identify high-risk road sections in each category, i.e., accident-prone points on the target highway.

[0040] The beneficial effects of this invention are as follows:

[0041] This invention uses simulation analysis of road segment characteristic data and vehicle braking performance parameters, combined with historical accident data, to identify the spatial distribution patterns of risky road segments. For risky road segments, it extracts characteristic combination datasets and employs cluster analysis to group road segments with similar characteristics, determining common risk categories and further identifying high-risk road segments within each category. This invention achieves accurate identification and classification of risky road segments on highways, providing detailed data for subsequent safety facility optimization, contributing to improved road safety and reduced traffic accidents. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a method for identifying accident-prone areas on highways according to an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of a highway accident multi-occurrence point identification system according to an embodiment of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] On the one hand, this embodiment provides a method for identifying accident-prone areas on highways, such as... Figure 1 As shown, it includes:

[0048] Acquire road segment characteristic data, vehicle braking performance parameters, and historical accident data of the target highway;

[0049] Simulations were performed based on the road segment characteristic data and vehicle braking performance parameters to calculate braking distance and obtain braking performance indicators under different road segment characteristics.

[0050] Based on the braking performance index, the historical accident data is classified and processed to extract braking accidents, and the clustering characteristics of the braking accidents under different road segment characteristics are analyzed to identify risky road segments.

[0051] Extract the characteristic data of the risky road sections, and use cluster analysis to group the risky road sections with similar road section characteristics and reduced braking performance to determine the road section categories with common risks;

[0052] Identify the key factors affecting braking distance in each category and determine the high-risk road sections in each category, i.e., the accident-prone points of the target highway.

[0053] Specifically, this embodiment uses simulation analysis to acquire road segment characteristic data and vehicle braking performance parameters, combined with historical accident data, to identify the spatial distribution patterns of risky road segments. For risky road segments, a characteristic combination dataset is extracted, and clustering analysis is used to group road segments with similar characteristics, determining common risk categories, and further identifying high-risk road segments within each category. This invention achieves accurate identification and classification of risky road segments on highways, providing detailed basis for subsequent safety facility optimization, helping to improve road safety and reduce traffic accidents.

[0054] This embodiment acquires road segment characteristic data, including gradient, curve radius, and road surface friction coefficient, from highway monitoring systems and vehicle sensors, and combines this data with braking performance parameters from vehicle driving records to form an initial comprehensive dataset. First, road segment characteristic data is acquired from the highway monitoring system. For example, a section of highway has a gradient of 5.2%, a curve radius of 300 meters, and a road surface friction coefficient of 0.75. This data is collected in real-time by monitoring cameras combined with image recognition algorithms and sensors, and automatically stored in a cloud database. The data acquisition frequency is once per minute to ensure data timeliness. Next, braking performance parameters are extracted from the vehicle sensors' driving records. For example, a vehicle's braking force on this road segment is 3200 Newtons, and its braking distance is 45.6 meters. This data is uploaded to the data center every second via the vehicle's OBD interface, forming timestamp-matched records. Subsequently, the road segment characteristic data and vehicle braking parameters are integrated to form a comprehensive dataset. Using a data fusion algorithm such as the weighted average method, the gradient, curve radius, and friction coefficient are assigned weights of 0.3, 0.3, and 0.4, respectively, to calculate the comprehensive road segment difficulty index. For example, the difficulty index = 0.3*5.2 + 0.3*300 / 1000 + 0.4*0.75 = 2.16, which is used for subsequent analysis.

[0055] Furthermore, based on the road segment characteristic data and vehicle braking performance parameters, simulations are performed to calculate braking distances and obtain braking performance indicators under different road segment characteristics, including:

[0056] A correlation analysis was performed on the road segment characteristic data and vehicle braking performance parameters to determine the correlation coefficient and influencing factors between the characteristics of each road segment and braking performance.

[0057] Based on the influencing factors, the road segment characteristic data are grouped to obtain road segment characteristic distribution information;

[0058] The road segment characteristic distribution information is input into a pre-built braking distance calculation model to simulate the braking distance changes under different scenarios, and the braking efficiency of each road segment under different scenarios is judged based on the braking distance changes.

[0059] Among them, judging the braking performance of each road segment under different scenarios based on the braking distance change includes:

[0060] Based on the aforementioned braking distance changes, regression analysis was performed using corresponding road segment characteristic data to determine the main influencing factors.

[0061] Based on the main influencing factors and combined with simulation data from various scenarios, the mapping relationship between braking performance and performance indicators is analyzed to determine the braking performance level of each road segment under different scenarios and obtain the braking performance evaluation results of the road segment.

[0062] Specifically, this embodiment calculates the braking distance using a pre-established braking distance calculation model based on specific parameters of various road sections, such as slope, curve radius, and road surface friction coefficient. Assuming a road section has a slope of 5%, a curve radius of 50 meters, and a road surface friction coefficient of 0.7, the braking distance is calculated using the formula D = v 2 The formula is: / (2μg)+vt, where D is the braking distance, v is the initial velocity set to 20 m / s, μ is the coefficient of friction 0.7, g is the gravitational acceleration 9.8 m / s², and t is the reaction time set to 1 second. The calculated braking distance is approximately 32.65 meters.

[0063] Computer simulation software was used to simulate braking distance changes under different slopes (e.g., 3%, 5%, 8%), curve radii (e.g., 30 meters, 50 meters, 80 meters), and road surface friction coefficients (e.g., 0.5, 0.7, 0.9), generating multiple sets of data. For example, the braking distance increased to 35.12 meters at an 8% slope and to 40.23 meters at a friction coefficient of 0.5, forming comparative analysis charts. Further regression analysis was performed on the simulation data using algorithms to determine the influence weight of each parameter on braking distance; for example, the influence factor for slope was 0.3, the influence factor for friction coefficient was 0.5, and the influence factor for curve radius was 0.2. Finally, based on the simulation results and weight analysis, the mapping relationship between braking performance and performance indicators was analyzed to determine the braking performance level of each road segment under different scenarios, obtaining the road segment braking performance evaluation results.

[0064] Furthermore, based on the braking performance index, the historical accident data is classified and processed to extract braking accidents, and the clustering characteristics of the braking accidents under different road segment characteristics are analyzed to obtain risky road segments, including:

[0065] Set the braking performance index threshold according to the braking performance index;

[0066] Based on the aforementioned braking performance index threshold, historical accident data is classified and processed to extract records containing braking-related information, thereby obtaining a braking accident dataset.

[0067] Spatial distribution information is obtained from the braking accident dataset, and the kernel density estimation algorithm is used to analyze the clustering characteristics of braking accidents on different road sections to obtain the braking accident density distribution.

[0068] The braking accident density distribution is correlated with road segment characteristic data to obtain a set of risky road segment characteristics, and a list of risky road segments is marked based on the set of risky road segment characteristics.

[0069] Based on the list of risky road segments, the K-means clustering algorithm is used to perform spatial clustering to determine the spatial distribution pattern of risky road segments and the areas of concentrated risk.

[0070] Specifically, this embodiment first uses data mining technology to extract braking-related records from a historical accident database. Assuming the database contains 100,000 accident records, about 10,000 relevant records are filtered out by setting braking performance index thresholds (such as braking distance exceeding 5.5 meters or braking time exceeding 3 seconds). Decision tree algorithm is used to classify these data according to categories such as brake failure and braking delay. The classification accuracy rate reaches more than 85%, ensuring data accuracy.

[0071] Next, by combining spatial distribution information, these records are associated with a Geographic Information System (GIS). Assuming that each record contains latitude and longitude coordinates, an accident density heat map is generated using a spatial interpolation algorithm (such as Kriging interpolation). It is found that the accident density of section A (5 km long) of a certain city's main road is 30 accidents per square kilometer, which is much higher than the average of 10 accidents per square kilometer of other road sections, thus initially locating the high-risk road sections.

[0072] Subsequently, the clustering characteristics of accidents under different road segment characteristics were analyzed, and road segment characteristic data (such as slope, curvature, and traffic flow) were extracted. Assuming that the slope of segment A is 5%, the curvature is 0.2, and the daily traffic flow is 20,000 vehicles, spatial autocorrelation analysis (calculated by Moran's I index, the result is 0.65, indicating significant clustering) was used to confirm that accidents in segment A are highly concentrated and positively correlated with steep slope and high traffic flow.

[0073] Finally, the spatial distribution pattern of risky road sections was determined. Based on the above analysis, the kernel density estimation (KDE) algorithm was used to further refine the heat map, which showed that the 2-kilometer range within section A was the core high-incidence area of ​​accidents, with a risk value of 0.8 (out of 1.0). Combined with historical data, it was inferred that the accident rate in this area increased by 30% on rainy days, thus forming a spatial distribution pattern: high-risk road sections are mostly concentrated on urban main roads with steep slopes, high curvature, and high traffic volume.

[0074] Furthermore, characteristic data of the risky road segments are extracted, and cluster analysis is used to group risky road segments with similar characteristics and reduced braking performance, determining road segment categories with common risks, including:

[0075] Extract the characteristic data corresponding to the risky road sections and construct a characteristic combination dataset;

[0076] Based on the aforementioned characteristic combination dataset, cluster analysis is used to group road segment characteristics and braking performance degradation features to determine preliminary similar road segments.

[0077] The braking performance degradation characteristic distribution of the preliminary similar road sections is judged to obtain a set of risk-marked road sections, and the correlation pattern between common risks and road section characteristics is analyzed to determine the common risk feature set;

[0078] By combining the common risk feature set with road segment category information, the risk-marked road segment set is classified a second time to obtain road segment category groups.

[0079] Specifically, this embodiment integrates the characteristic data corresponding to risky road sections to form a characteristic combination dataset. Through in-depth mining of this dataset, the dataset containing road section characteristics (such as slope, curvature, and road surface friction coefficient) and braking performance degradation features (such as braking distance and braking time) is preprocessed. Assuming the dataset contains 1000 road section records, each record has 5 features: slope (0-15 degrees), curvature (0-0.5), friction coefficient (0.3-0.8), braking distance (10-50 meters), and braking time (1-5 seconds). Principal component analysis (PCA) is used for dimensionality reduction, compressing the 5-dimensional features to 2 dimensions while retaining 90% of the variance, resulting in principal components PC1 and PC2.

[0080] Next, the K-means clustering algorithm was used to group the dimensionality-reduced data. The initial number of clusters K=3 was set, and the optimal K value was determined by the elbow rule: the sum of squares within the clusters (WSS) was calculated from K 1 to 10. When K=3, the decrease of WSS tended to be gradual, confirming that K=3.

[0081] The clustering results divided the road segments into three groups:

[0082] The first group (400 items) is characterized by high slope (average 12 degrees), low friction coefficient (average 0.4), and long braking distance (average 45 meters).

[0083] The second group (350 items) features high curvature (mean 0.4), medium coefficient of friction (mean 0.6), and long braking time (mean 4 seconds);

[0084] The third group (250 items) is characterized by low slope (average 3 degrees), high friction coefficient (average 0.75), and good braking performance (average braking distance 15 meters).

[0085] Furthermore, the key influencing factors on braking distance in each category are identified, and high-risk road sections in each category are determined, including:

[0086] Extract the road segment characteristic data and braking distance for each road segment category, and use statistical analysis methods to calculate the correlation strength between the characteristic parameters of each road segment and the braking distance to identify key influencing factors;

[0087] Based on the aforementioned key influencing factors, a preliminary high-risk road segment is identified in each road segment category, and a preliminary set of high-risk road segments is obtained.

[0088] Spatial proximity analysis is performed on the preliminary set of high-risk road segments to obtain the correlation of distribution characteristics between road segments, and priority ranking is performed based on the correlation to construct the final set of high-risk road segments.

[0089] Specifically, this embodiment identifies key influencing factors that significantly extend braking distance based on road segment grouping results, and uses a random forest regression model to analyze the key influencing factors for each type of road segment. Combining spatial distribution patterns, kernel density estimation (KDE, bandwidth 0.5 km) is used to analyze accident hotspots. For example, in urban expressways, segments with extended braking distances are concentrated near intersections (density value > 0.8); in suburban roads, they are distributed on straight sections with a friction coefficient below 0.6; in mountainous areas, steep slopes are concentrated on downhill sections with a gradient greater than 8%; and in rural areas, they are at intersections without traffic lights.

[0090] High-risk road sections are prioritized and sorted by a comprehensive score (braking distance * 0.6 + accident density * 0.4). The top 10% of road sections in each category are selected, such as 100 road sections in urban expressways with a score of over 0.9 (e.g., road section ID001, braking distance 80 meters, density 0.85), to generate the final set of high-risk road sections.

[0091] Furthermore, after obtaining the final set of high-risk road sections, the method further includes: generating a corresponding risk level assessment report for each high-risk road section based on historical accident clustering characteristics, and visually presenting the spatial distribution and key influencing factors of the high-risk road sections based on the risk level assessment report.

[0092] Specifically, this embodiment obtains road segment location information and historical accident data from the list of dangerous road segments, and determines the unique identifier and accident record of each dangerous road segment; it uses cluster analysis of historical accident data and the K-means algorithm to extract accident clustering features, obtaining the accident frequency and severity; based on the accident clustering features, combined with road segment traffic flow and road segment environmental factors, it calculates the risk level of each dangerous road segment and generates a risk level assessment; if the risk level is higher than a preset threshold, it determines the safety facility optimization requirements based on the road segment location information and accident severity; by analyzing the safety facility optimization requirements and combining them with road segment environmental factors, it generates optimization basis and obtains a safety facility deployment plan; based on the safety facility deployment plan, it generates an assessment report for each dangerous road segment, obtaining a detailed description of the optimization basis.

[0093] Using data visualization technology, based on comprehensive risk distribution data, a distribution map of high-risk road sections is drawn, presenting spatial distribution and key risk characteristics, and generating the final map. Through an interactive visualization interface, the final map is loaded, and high-risk areas and related influencing factors are dynamically displayed to obtain the visualized distribution results.

[0094] On the other hand, this embodiment also provides a highway accident multi-occurrence point identification system, such as Figure 2 As shown, it includes:

[0095] The road segment data acquisition module is used to acquire road segment characteristic data, historical accident data, and vehicle braking performance parameters of the target highway.

[0096] The braking performance analysis module is used to perform simulation based on the road segment characteristic data and vehicle braking performance parameters, calculate braking distance, and obtain braking performance indicators under different road segment characteristics.

[0097] The risk road segment extraction module is used to classify the historical accident data based on the braking performance index, extract braking accidents, and analyze the clustering characteristics of the braking accidents under different road segment characteristics to obtain risk road segments.

[0098] The risk road segment classification module is used to extract the characteristic data of the risk road segments and use cluster analysis to group risk road segments with similar road segment characteristics and braking performance degradation characteristics to determine the road segment categories with common risks.

[0099] The high-risk assessment module is used to obtain the key influencing factors on braking distance in each category and to identify high-risk road sections in each category, i.e., accident-prone points on the target highway.

[0100] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for identifying accident-prone areas on highways, characterized in that, include: Acquire road segment characteristic data, vehicle braking performance parameters, and historical accident data of the target highway. The road segment characteristic data includes gradient, curve radius, and road surface friction coefficient, while the vehicle braking performance parameters include braking force and braking distance. Simulations are performed based on the road segment characteristic data and vehicle braking performance parameters to calculate braking distance and obtain braking performance indicators under different road segment characteristics, including: A correlation analysis was performed on the road segment characteristic data and vehicle braking performance parameters to determine the correlation coefficient and influencing factors between the characteristics of each road segment and braking performance. Based on the influencing factors, the road segment characteristic data are grouped to obtain road segment characteristic distribution information; The road segment characteristic distribution information is input into a pre-built braking distance calculation model to simulate the braking distance changes under different scenarios, and the braking efficiency of each road segment under different scenarios is judged based on the braking distance changes. Based on the braking performance index, the historical accident data is classified and processed to extract braking accidents, and the clustering characteristics of the braking accidents under different road segment characteristics are analyzed to identify risky road segments. Extract the characteristic data of the risky road sections, and use cluster analysis to group the risky road sections with similar road section characteristics and reduced braking performance to determine the road section categories with common risks; Identify the key factors affecting braking distance in each category and determine the high-risk road sections in each category, i.e., the accident-prone points of the target highway.

2. The method for identifying accident-prone areas on highways according to claim 1, characterized in that, Determining the braking performance of each road segment under different scenarios based on the aforementioned braking distance changes includes: Based on the aforementioned braking distance changes, regression analysis was performed using corresponding road segment characteristic data to determine the main influencing factors. Based on the main influencing factors and combined with simulation data from various scenarios, the mapping relationship between braking performance and performance indicators is analyzed to determine the braking performance level of each road segment under different scenarios and obtain the braking performance evaluation results of the road segment.

3. The method for identifying accident-prone areas on highways according to claim 1, characterized in that, Based on the braking performance index, the historical accident data is classified and processed to extract braking accidents. The clustering characteristics of these braking accidents under different road segment characteristics are analyzed to identify high-risk road segments, including: Set the braking performance index threshold according to the braking performance index; Based on the aforementioned braking performance index threshold, historical accident data is classified and processed to extract records containing braking-related information, thereby obtaining a braking accident dataset. Spatial distribution information is obtained from the braking accident dataset, and the kernel density estimation algorithm is used to analyze the clustering characteristics of braking accidents on different road sections to obtain the braking accident density distribution. The braking accident density distribution is correlated with road segment characteristic data to obtain a set of risky road segment characteristics, and a list of risky road segments is marked based on the set of risky road segment characteristics. Based on the list of risky road segments, the K-means clustering algorithm is used to perform spatial clustering to determine the spatial distribution pattern of risky road segments and the areas of concentrated risk.

4. The method for identifying accident-prone areas on highways according to claim 1, characterized in that, The characteristic data of the risky road segments are extracted, and cluster analysis is used to group risky road segments with similar characteristics and reduced braking performance to determine the categories of road segments with common risks, including: Extract the characteristic data corresponding to the risky road sections and construct a characteristic combination dataset; Based on the aforementioned characteristic combination dataset, cluster analysis is used to group road segment characteristics and braking performance degradation features to determine preliminary similar road segments. The braking performance degradation characteristic distribution of the preliminary similar road sections is judged to obtain a set of risk-marked road sections, and the correlation pattern between common risks and road section characteristics is analyzed to determine the common risk feature set; By combining the common risk feature set with road segment category information, the risk-marked road segment set is classified a second time to obtain road segment category groups.

5. The method for identifying accident-prone areas on highways according to claim 1, characterized in that, Identify the key factors affecting braking distance in each category and determine the high-risk road sections within each category, including: Extract the road segment characteristic data and braking distance for each road segment category, and use statistical analysis methods to calculate the correlation strength between the characteristic parameters of each road segment and the braking distance to identify key influencing factors; Based on the aforementioned key influencing factors, a preliminary high-risk road segment is identified in each road segment category, and a preliminary set of high-risk road segments is obtained. Spatial proximity analysis is performed on the preliminary set of high-risk road segments to obtain the correlation of distribution characteristics between road segments, and priority ranking is performed based on the correlation to construct the final set of high-risk road segments.

6. The method for identifying accident-prone areas on highways according to claim 5, characterized in that, After obtaining the final set of high-risk road sections, the method further includes: generating a corresponding risk level assessment report for each high-risk road section based on the characteristics of historical accident clusters, and visually presenting the spatial distribution and key influencing factors of the high-risk road sections based on the risk level assessment report.

7. A system for identifying accident-prone locations on highways, characterized in that, include: The road segment data acquisition module is used to acquire road segment characteristic data, vehicle braking performance parameters, and historical accident data of the target highway. The road segment characteristic data includes slope, curve radius, and road surface friction coefficient, while the vehicle braking performance parameters include braking force and braking distance. The braking performance analysis module is used to perform simulations based on the road segment characteristic data and vehicle braking performance parameters, calculate braking distance, and obtain braking performance indicators under different road segment characteristics, including: A correlation analysis was performed on the road segment characteristic data and vehicle braking performance parameters to determine the correlation coefficient and influencing factors between the characteristics of each road segment and braking performance. Based on the influencing factors, the road segment characteristic data are grouped to obtain road segment characteristic distribution information; The road segment characteristic distribution information is input into a pre-built braking distance calculation model to simulate the braking distance changes under different scenarios, and the braking efficiency of each road segment under different scenarios is judged based on the braking distance changes. The risk road segment extraction module is used to classify the historical accident data based on the braking performance index, extract braking accidents, and analyze the clustering characteristics of the braking accidents under different road segment characteristics to obtain risk road segments. The risk road segment classification module is used to extract the characteristic data of the risk road segments and use cluster analysis to group risk road segments with similar road segment characteristics and braking performance degradation characteristics to determine the road segment categories with common risks. The high-risk assessment module is used to obtain the key influencing factors on braking distance in each category and to identify high-risk road sections in each category, i.e., accident-prone points on the target highway.

Images