Methods and Systems for Safety Assessment of Merged Traffic at Interchange Service Areas Based on Traffic Flow Patterns

Abstract

This invention discloses a method and system for assessing the safety of merging traffic in grade-separated interchange service areas based on traffic flow patterns. The method includes: collecting multi-dimensional parameters to construct an initial dataset using a traffic conflict data intelligent analysis platform; generating a dynamic traffic flow evolution sequence using a spatiotemporal coupling inference algorithm based on traffic flow patterns; determining safe operating boundary conditions under different traffic flow densities using a service area merging safety threshold model; quantifying risk levels and identifying potential hazards using a sudden traffic flow risk intelligent discrimination model; constructing an assessment index system and weights through multi-source data fusion analysis; and outputting the assessment results through comprehensive calculation. The method optimizes each algorithm model specifically, and the system achieves efficient data processing through the orderly connection of multiple units. It utilizes a multi-source sensor array and heterogeneous computing architecture to ensure data acquisition and processing efficiency, improving assessment accuracy and adaptability, and providing comprehensive intelligent technical support for the safety of merging traffic in grade-separated interchange service areas.

Description

Technical Field

[0001] This invention relates to the field of safety assessment technology for grade-separated service areas, and in particular to a method and system for assessing the safety of combined traffic flow in grade-separated service areas based on traffic flow patterns. Background Technology

[0002] With the acceleration of urbanization, highway interchange service areas, as key nodes where traffic flows converge, are experiencing a continuous increase in traffic volume, especially during holidays and other periods when sudden traffic surges occur frequently. This significantly increases the risk of traffic conflicts in merging areas, placing higher demands on the accuracy and timeliness of traffic safety assessments. Merging traffic at interchange service areas involves complex interactions between mainline and ramp traffic flows, and the spatiotemporal dynamics of traffic flow directly affect merging safety. Therefore, there is an urgent need to construct an assessment system that integrates multi-dimensional parameters and intelligent algorithms to address the safety requirements in complex traffic environments and provide scientific support for traffic management decisions.

[0003] Currently, safety assessments are primarily achieved by collecting basic data such as vehicle speed and density using sensors and cameras, combined with traditional traffic flow analysis methods. Some technologies introduce simple threshold judgments or single prediction models, performing preliminary statistical analysis on the collected data before outputting safety assessment results. These technologies largely rely on fixed parameter settings, establishing assessment standards through historical data fitting. Some employ a segmented analysis approach, processing single-dimensional information such as vehicle flow and merging conflicts separately, and then performing simple integration, attempting to meet safety assessment needs through superficial analysis of basic data.

[0004] Existing technologies have two main drawbacks: First, they lack in-depth analysis of the spatiotemporal coupling characteristics of traffic flow, failing to fully consider the dynamic correlation of traffic flow operation status at different times and in different areas, making it difficult for the assessment results to reflect the true safety situation in complex traffic environments. Second, they lack the ability to identify risks from sudden traffic flows, have not established a targeted intelligent discrimination mechanism, cannot accurately quantify the safety hazards and impact range caused by sudden traffic flows, and the safety thresholds are mostly fixed values, making it difficult to dynamically adjust according to real-time traffic flow conditions, thus limiting the adaptability and accuracy of the assessment. Summary of the Invention

[0005] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides a method and system for assessing the safety of merging traffic in grade-separated service areas based on traffic flow patterns.

[0006] The technical solution adopted in this invention is a method for assessing the safety of merging traffic in grade-separated service areas based on traffic flow patterns, comprising the following steps: S1, collecting traffic flow operation parameters, road geometric parameters, and traffic environment parameters of merging traffic sections in grade-separated service areas through a traffic conflict data intelligent analysis platform to construct a multi-dimensional raw data set; S2, using a traffic flow pattern spatiotemporal coupling inference algorithm to perform dynamic correlation analysis on the spatiotemporal related data in the raw data set, mining the coupling correlation characteristics of traffic flow operation status in different time periods and different areas, and generating a dynamic traffic flow pattern evolution sequence; S3, based on the traffic flow pattern evolution sequence, calculating the traffic flow operation in the merging area using a service area merging safety threshold model. S4. Determine the critical control parameters to establish safe operating boundary conditions under different traffic flow densities; S5. Use a sudden traffic flow risk intelligent discrimination model to extract abnormal features and quantify risk levels of the traffic flow situation evolution sequence, and identify the types and impact range of safety hazards caused by sudden traffic flows; S6. Use a traffic conflict data intelligent analysis platform to perform multi-source data fusion analysis on coupled correlation features, critical control parameters, and risk level quantification results, construct a safety assessment index system, and calculate the weight of each index; S7. Combine the safety assessment index system and index weights, perform comprehensive calculations using the grade-separated interchange service area merged traffic safety assessment parameters based on traffic flow situation, and output the merged traffic safety assessment results.

[0007] Furthermore, the expression for the spatiotemporal coupling inference algorithm of the traffic flow situation is:

[0008]

[0009]

[0010] in, It is a spatial position vector. For time variables, Let be the instantaneous velocity vector of the i-th type of traffic flow. Let be the density distribution function of the i-th type of traffic flow. Let be the spatiotemporal coupling coefficient of the i-th type of traffic flow. To deduce the correction coefficient, For the time integration interval, The function representing the influence of historical traffic flow patterns. This represents the traffic flow spatial diffusion coefficient.

[0011] Furthermore, the expression for the service area merging safety threshold model is:

[0012]

[0013]

[0014] in, Based on the basic merging safety threshold, For model calibration parameters, The weighting coefficient for the k-th merging conflict point. Let k be the distance parameter of the kth merging conflict point. Let be the traffic convergence angle parameter at the k-th merging conflict point. To maximize traffic capacity, For merging lane width parameters, For the green light signal cycle parameters, The dynamically adjusted merging safety threshold. For real-time traffic density, The baseline traffic density is used.

[0015] Furthermore, the expression for the intelligent judgment model for sudden traffic flow risks is:

[0016]

[0017] in, To quantify the risk of sudden traffic congestion, , For risk discrimination coefficient, Let p be the traffic flow mutation value of the p-th type of traffic flow. For the time span of the sudden change in traffic flow of type p, This represents the real-time traffic flow of the p-th type of traffic. Let p be the maximum throughput for traffic flow type p. This is the spatial risk diffusion function.

[0018] Furthermore, the intelligent traffic conflict data analysis platform includes a multi-source sensor array, edge computing nodes, and a cloud data processing module. The multi-source sensor array consists of a millimeter-wave radar, a high-definition video acquisition unit, and a microwave detector. The millimeter-wave radar has a sampling frequency of 50Hz, the high-definition video acquisition unit has a resolution of 4K and a frame rate of 30fps, and the microwave detector has a detection range of 5-200m. The edge computing node adopts a heterogeneous computing architecture, integrating FPGA and ARM processors, and supports real-time data preprocessing and feature extraction. The cloud data processing module adopts a distributed storage architecture, and the data transmission protocol adopts a custom transmission protocol based on UDP, with transmission latency controlled within 10ms. The weights of each indicator in the safety assessment index system are determined by combining the analytic hierarchy process (AHP) with the entropy weight method, with a weight allocation range of 0.05-0.3.

[0019] Further, S2 includes the following sub-steps: S21, selecting traffic flow speed, density, and flow data corresponding to the spatiotemporal coordinates in the original dataset, establishing a spatiotemporal data matrix, and determining the spatiotemporal correlation range of different data points through a neighborhood search algorithm; S22, inputting the data within the spatiotemporal correlation range into a traffic flow situation spatiotemporal coupling inference algorithm, and obtaining the spatiotemporal evolution trend of the traffic flow situation through time series decomposition and spatial interpolation operations; S23, based on the spatiotemporal evolution trend, analyzing the dynamic law of traffic flow operation status changing with time and the coupling characteristics of spatial distribution, and identifying the precursor information of sudden changes in traffic flow situation; S24, integrating the dynamic law, coupling characteristics, and precursor information to generate a dynamic traffic flow situation evolution sequence including speed change rate, density gradient, and flow fluctuation amplitude.

[0020] Further, S3 includes the following sub-steps: S31, extracting traffic density, merging speed, and lane occupancy rate data of the merging area from the dynamic traffic flow evolution sequence, and constructing a threshold calculation dataset; S32, inputting the dataset into the service area merging safety threshold model, determining the model parameters through variable correlation analysis, and determining the degree of influence of each parameter on merging safety; S33, based on the parameters and the degree of influence, calculating the merging critical speed, safe distance, and lane switching frequency corresponding to different traffic densities through boundary condition constraint calculation; S34, integrating the critical speed, safe distance, and lane switching frequency to form complete boundary conditions for safe operation of the merging area.

[0021] Further, S4 includes the following sub-steps: S41, performing sliding window processing on the dynamic traffic flow situation evolution sequence, extracting traffic flow, speed fluctuation, and density change data within each window to construct a risk identification dataset; S42, inputting the dataset into the emergency traffic flow risk intelligent discrimination model, identifying abnormal features in traffic flow operation through feature normalization operations and outlier detection; S43, based on the abnormal features, quantifying the risk level corresponding to the abnormal features through risk level classification rules, and determining the risk level; S44, combining the risk level and the spatial distribution of abnormal features, analyzing the types of safety hazards caused by the emergency traffic flow, and defining the spatial range and time span of the hazard's impact.

[0022] Further, S5 includes the following sub-steps: S51, collecting coupling correlation features, critical control parameters, and risk level quantification results, and performing format standardization processing on multi-source data to ensure data consistency; S52, using the multi-source data fusion module of the traffic conflict data intelligent analysis platform, employing a weighted fusion algorithm to perform fusion operations on the standardized data to generate a fused data set; S53, based on the fused data set, selecting indicators such as traffic flow conflict frequency, merging efficiency, and risk duration to construct a multi-level safety assessment indicator system; S54, using the indicator weight calculation method, combined with the statistical characteristics of the fused data, determining the weight of each assessment indicator to form a complete indicator weight set.

[0023] A traffic flow-based safety assessment system for merging traffic in grade-separated service areas is implemented. This system comprises: a multi-dimensional traffic flow parameter acquisition unit, connected to a traffic conflict data intelligent analysis platform, used to collect traffic flow operation parameters, road geometric parameters, and traffic environment parameters of the merging traffic sections in the grade-separated service area and transmit them to a data preprocessing unit; a data preprocessing unit, connected to both the multi-dimensional traffic flow parameter acquisition unit and the traffic flow situation spatiotemporal coupling inference unit, used to perform format conversion and redundant data removal on the collected multi-dimensional parameters before transmitting them to the traffic flow situation spatiotemporal coupling inference unit; and a traffic flow situation spatiotemporal coupling inference unit, connected to the data preprocessing unit and the service area merging safety threshold calculation unit. The system processes data and generates a dynamic traffic flow evolution sequence through a traffic flow situation spatiotemporal coupling inference algorithm. A service area merging safety threshold calculation unit, connected to both the traffic flow situation spatiotemporal coupling inference unit and the sudden traffic flow risk discrimination unit, calculates the safe operation boundary conditions of the merging area based on the evolution sequence. A sudden traffic flow risk discrimination unit, connected to both the service area merging safety threshold calculation unit and the multi-source data fusion evaluation unit, identifies safety hazards and their impact range through a sudden traffic flow risk intelligent discrimination model. A multi-source data fusion evaluation unit, connected to both the sudden traffic flow risk discrimination unit, the traffic flow situation spatiotemporal coupling inference unit, and the service area merging safety threshold calculation unit, integrates multi-source data to construct an evaluation index system and calculates weights, outputting the merging traffic safety evaluation results.

[0024] The present invention has the following beneficial effects:

[0025] Through a multi-source sensor array and heterogeneous computing architecture, the intelligent traffic conflict data analysis platform accurately collects and efficiently processes multi-dimensional parameters of traffic flow, road geometry, and traffic environment. Combined with a traffic flow situation spatiotemporal coupling inference algorithm, it deeply mines the spatiotemporal coupling correlation characteristics of traffic flow in different time periods and areas, generating dynamic evolution sequences. This addresses the shortcomings of existing technologies in providing in-depth analysis of the spatiotemporal dynamic correlation of traffic flow. The system calculates dynamic safe operation boundary conditions under different traffic flow densities using a service area merging safety threshold model, replacing traditional fixed threshold settings. Simultaneously, it utilizes an intelligent judgment model for sudden traffic flow risks to accurately extract abnormal features and quantify risk levels, determining the type of hazard and its impact range, thus solving the problem of insufficient ability of existing technologies to identify sudden traffic flow risks. The system's various units achieve efficient data flow and processing through orderly connections. The phased and detailed steps ensure the technology's practicality. A scientific evaluation index system and weight allocation constructed through multi-source data fusion analysis, combined with comprehensive calculations using dedicated evaluation parameters, improves the accuracy and adaptability of the evaluation results, providing more comprehensive and intelligent technical support for the safety of merging traffic in interchange service areas. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention.

[0027] Figure 2 This is a flowchart of method step S2 of the present invention;

[0028] Figure 3 This is a flowchart of method step S3 of the present invention;

[0029] Figure 4 This is a flowchart of method step S4 of the present invention;

[0030] Figure 5 This is a flowchart of step S5 of the method of the present invention;

[0031] Figure 6 This is a diagram showing the system unit composition of the present invention. Detailed Implementation

[0032] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] like Figure 1 As shown, the safety assessment method for merging traffic in grade-separated service areas based on traffic flow patterns includes the following steps: S1, collecting traffic flow operation parameters, road geometric parameters, and traffic environment parameters of merging traffic sections in grade-separated service areas through a traffic conflict data intelligent analysis platform to construct a multi-dimensional raw data set; S2, using a traffic flow pattern spatiotemporal coupling inference algorithm to perform dynamic correlation analysis on the spatiotemporal related data in the raw data set, mining the coupling correlation characteristics of traffic flow operation status in different time periods and different areas, and generating a dynamic traffic flow pattern evolution sequence; S3, based on the traffic flow pattern evolution sequence, calculating the critical control of traffic flow operation in the merging area through a service area merging safety threshold model. S4. Parameters are used to determine the safe operation boundary conditions under different traffic flow densities; S5. An intelligent judgment model for sudden traffic flow risks is used to extract abnormal features and quantify the risk level of the traffic flow situation evolution sequence, and to identify the types and impact range of safety hazards caused by sudden traffic flow; S6. Through the intelligent analysis platform for traffic conflict data, multi-source data fusion analysis is performed on the coupled correlation features, critical control parameters and risk level quantification results to construct a safety assessment index system and calculate the weight of each index; S7. Combining the safety assessment index system and index weights, comprehensive calculations are performed using the grade-separated interchange service area combined traffic safety assessment parameters based on traffic flow situation, and the combined traffic safety assessment results are output.

[0034] Step S1 involves data collection using a multi-source sensor array mounted on the intelligent traffic conflict data analysis platform. This array consists of a millimeter-wave radar, a high-definition video acquisition unit, and a microwave detector. The millimeter-wave radar continuously captures real-time traffic flow data at a sampling frequency of 50Hz. The high-definition video acquisition unit records road scenes and traffic flow dynamics at 4K resolution and 30fps. The microwave detector collects relevant parameters within a detection range of 5 to 200 meters. The collected traffic flow parameters include core data such as traffic speed, density, flow rate, and lane occupancy. Road geometric parameters include key information such as merging lane width, merging conflict point distance, and traffic convergence angle. Traffic environment parameters involve environmental influencing factors such as weather conditions and lighting conditions. During the data collection process, the principle of data integrity was strictly followed to ensure that no parameters were missed or biased. Through the collaborative work of multiple source sensors, comprehensive and multi-dimensional data coverage was achieved. Subsequently, the collected raw data were classified and organized, obviously invalid data was removed, and the data was integrated according to the preset data format specifications. Finally, a multi-source raw data set including three dimensions of traffic flow, roads, and environment was constructed, which provides comprehensive and accurate data support for subsequent traffic flow situation analysis and safety assessment. The quality of the construction of this data set directly affects the reliability and accuracy of the calculation results of subsequent algorithm models.

[0035] Step S2 utilizes a traffic flow situation spatiotemporal coupling inference algorithm to perform deep dynamic correlation analysis on data with spatiotemporal correlation attributes in the original dataset. First, core data such as traffic speed, density, and flow rate corresponding to different spatiotemporal coordinates are selected from the original dataset to establish a structured spatiotemporal data matrix. A neighborhood search algorithm is used to define the spatiotemporal correlation range of different data points, ensuring the relevance and effectiveness of the correlation analysis. The selected spatiotemporal correlation data is input into the traffic flow situation spatiotemporal coupling inference algorithm. Time series decomposition operations are used to analyze the changes in data over time, and spatial interpolation operations are used to supplement data details in the spatial dimension, comprehensively obtaining the evolution trend of traffic flow situation in the spatiotemporal dimension. Based on the obtained spatiotemporal evolution trend, the system analyzes the dynamic changes in traffic flow status over time, deeply explores the coupling correlation characteristics between traffic flow statuses in different regions, and accurately identifies various precursory information before sudden changes in traffic flow status, including key signals such as abnormal speed change rate, sudden density gradient changes, and excessive flow rate fluctuations. Finally, the dynamic change patterns, spatial coupling characteristics, and various precursor information are systematically integrated to generate a dynamic traffic flow situation evolution sequence including core indicators such as speed change rate, density gradient, and traffic fluctuation amplitude. This sequence fully presents the trajectory and correlation characteristics of traffic flow changes in the spatiotemporal dimension, providing accurate situation data support for subsequent safety threshold calculation and risk assessment.

[0036] Step S3 uses the dynamic traffic flow evolution sequence generated in Step S2 as the core data foundation to calculate the critical control parameters for traffic flow operation in the merging area through the service area merging safety threshold model. First, key data for the merging area is accurately extracted from the dynamic traffic flow evolution sequence, including core parameters such as traffic density, merging speed, and lane occupancy rate at different times and in different areas. This data is categorized and organized according to data type and assessment requirements to construct a dataset specifically for threshold calculation, ensuring a high degree of matching between the data and the model's calculation needs. The constructed dataset is input into the service area merging safety threshold model. A variable correlation analysis method is used to systematically analyze the correlation between each input parameter and merging safety, identifying the key parameters in the model and determining the weight of each key parameter's impact on merging safety. Based on the determined key parameters and their corresponding weights, boundary condition constraint calculations are performed, combined with the road geometry characteristics of the merging area, to calculate the corresponding critical merging speed, safe following distance, and lane switching frequency at different traffic density levels, ensuring that each traffic density level has a corresponding precise control standard. Finally, the calculated parameters such as critical speed, safe following distance, and lane switching frequency are integrated and optimized to form a complete and systematic boundary condition for safe operation of merging areas. This boundary condition provides a definite and quantitative reference standard for judging whether the traffic flow in the merging area is safe, and achieves accurate definition of the merging safety status.

[0037] Step S4 employs a sudden traffic flow risk intelligent discrimination model to comprehensively extract abnormal features and quantify risk levels in the dynamic traffic flow situation evolution sequence generated in Step S2. First, a sliding window process is applied to the dynamic traffic flow situation evolution sequence, setting a reasonable window size and sliding step to ensure comprehensive coverage of all data segments in the sequence. Key data such as traffic flow volume, speed fluctuations, and density abrupt changes are extracted from each sliding window. The extracted data undergoes preliminary screening and organization, eliminating meaningless redundant data to construct a clearly structured and effective risk identification dataset. This risk identification dataset is then input into the sudden traffic flow risk intelligent discrimination model. Feature normalization is used to standardize the data, eliminating dimensional differences between different parameters. An outlier detection algorithm then accurately identifies various abnormal features in traffic flow operation, including sudden changes in flow volume, sudden drops in speed, and surges in density—signature signals that deviate from normal operating patterns. Based on the identified abnormal features, and according to preset risk level classification rules, combined with factors such as the severity, duration, and scope of impact of the abnormal features, the risk level corresponding to the abnormal features is quantitatively assessed to determine the specific risk level. The risk level classification follows the principles of precise quantification and clear hierarchy to ensure the distinguishability between different levels. Finally, based on the determined risk level and the spatial distribution of abnormal characteristics, we conducted an in-depth analysis of the types of safety hazards caused by sudden traffic flow, determined the specific manifestations of the hazards, and accurately defined the spatial scope and time span of the hazard's impact, providing definite goals and basis for subsequent safety assessments and risk management.

[0038] Step S5 utilizes the multi-source data fusion module of the traffic conflict data intelligent analysis platform to conduct a comprehensive multi-source data fusion analysis on the coupling correlation features obtained in Step S2, the critical control parameters calculated in Step S3, and the risk level results quantified in Step S4. First, the three types of core data mentioned above are collected, and the formats of each type of data are standardized to unify the data storage format, representation method, and accuracy requirements, ensuring consistency and compatibility between data from different sources and of different types, and avoiding the impact of data format differences on the fusion effect. The standardized multi-source data is then fused using the weighted fusion algorithm on the traffic conflict data intelligent analysis platform. This platform adopts a heterogeneous computing architecture integrating FPGA and ARM processors, supporting real-time data processing. Data transmission uses a custom transmission protocol based on UDP, with transmission latency controlled within 10ms, ensuring the efficiency and real-time nature of the fusion operation, ultimately generating a comprehensive and unified fused data set. Based on a fused dataset and considering the core requirements of safety assessment for merging traffic in interchange service areas, a multi-level safety assessment indicator system with distinct layers and comprehensive coverage is constructed, selecting key indicators such as traffic conflict frequency, merging efficiency, risk duration, and critical parameter compliance rate. This system includes multiple dimensions such as traffic flow operation, merging process, and risk status. The weights of each assessment indicator are determined using the analytic hierarchy process (AHP) combined with the entropy weight method, with the weight allocation ranged between 0.05 and 0.3 to ensure the scientific and reasonable nature of the weight allocation. This results in a complete and accurate set of indicator weights, providing reliable indicator support for the final comprehensive safety assessment calculation.

[0039] Step S6, based on the safety assessment indicator system and determined indicator weights established in Step S5, conducts comprehensive calculations using dedicated parameters for the safety assessment of merged traffic in interchange service areas based on traffic flow patterns. The final output is an accurate and reliable safety assessment result for merged traffic. During the comprehensive calculation process, the preset assessment logic and calculation rules are strictly followed, associating and matching each indicator in the safety assessment indicator system with its corresponding weight to ensure that each indicator reflects its corresponding impact in the assessment result according to its importance. The calculation process fully utilizes the cloud data processing module of the intelligent traffic conflict data analysis platform. This module adopts a distributed storage architecture, enabling efficient processing of large-scale fused data. Combined with the characteristics of dedicated assessment parameters, targeted calculation strategies are employed to ensure the stability and efficiency of the calculation process. During the comprehensive calculation process, multiple factors are fully considered, including the spatiotemporal coupling characteristics of traffic flow patterns, merging safety threshold boundaries, and the risk level of sudden traffic flows, to achieve a comprehensive and multi-dimensional assessment of the safety status of merged traffic, avoiding result deviations caused by single-factor assessments. After the calculation is completed, the results are systematically organized and analyzed. Following pre-defined evaluation result presentation standards, the results are categorized into different safety levels. Key influencing factors and risk points corresponding to each safety level are identified and marked. Targeted safety optimization suggestions are also provided to ensure the practicality and guidance of the evaluation results. Finally, the merged traffic safety evaluation results are output in a clear and intuitive format, providing traffic management departments with a scientific and accurate decision-making basis for formulating safety control strategies and optimizing the merging design of interchange service areas. This effectively supports the intelligent and refined management of merged traffic safety at interchange service areas.

[0040] Preferably, the expression for the spatiotemporal coupling inference algorithm of traffic flow situation is:

[0041]

[0042]

[0043] in, It is a spatial position vector. For time variables, Let be the instantaneous velocity vector of the i-th type of traffic flow. Let be the density distribution function of the i-th type of traffic flow. Let be the spatiotemporal coupling coefficient of the i-th type of traffic flow. To deduce the correction coefficient, For the time integration interval, The function representing the influence of historical traffic flow patterns. This represents the traffic flow spatial diffusion coefficient.

[0044] Specifically, the spatiotemporal coupling inference algorithm for traffic flow is built upon the dynamic correlation characteristics of traffic flow in the spatiotemporal dimension. The calculation process first revolves around the temporal rate of change of instantaneous speed and the spatial second derivative of density distribution of traffic flow. These two factors reflect the dynamic evolution of traffic flow in the temporal dimension and the distributional differences in the spatial dimension, respectively. By introducing spatiotemporal coupling coefficients specific to various types of traffic flow, the differentiated consideration of the characteristics of different types of traffic flow is achieved. Then, the spatiotemporal evolution contribution values ​​of all traffic flow types are integrated using a product operation to form the basic inference function. On this basis, the impact of historical traffic flow patterns on the current state is combined, and the cumulative effect of historical data is quantified through time integration. At the same time, the gradient term of the traffic flow spatial diffusion coefficient is introduced to characterize the diffusion trend of spatial distribution. Combined with the inference correction coefficient, the model accuracy is optimized, and finally, a complete inference algorithm is formed. The spatiotemporal coupling coefficient is calibrated to a differentiated value between 0.1 and 0.8 based on a large amount of measured data for different traffic flow types (such as small cars and large cars). The inference correction coefficient is set to 0.3 to 1.2 based on road level and merging scenario type. The time integration interval is selected from historical data periods of the past 15 to 60 minutes. This formula considers the coupled correlation of traffic flow spatiotemporal evolution, reflecting both the dynamic changes in the time dimension and the distribution characteristics in the spatial dimension. During implementation, multi-source data is acquired in real time through the edge computing nodes of the traffic conflict data intelligent analysis platform, and the data is substituted into the formula to complete the calculation. The generated results provide core algorithm support for the construction of subsequent dynamic traffic flow situation evolution sequences, ensuring that the evolution sequence can accurately reflect the spatiotemporal dynamic changes of traffic flow.

[0045] Preferably, the expression for the service area merging safety threshold model is:

[0046]

[0047]

[0048] in, Based on the basic merging safety threshold, For model calibration parameters, The weighting coefficient for the k-th merging conflict point. Let k be the distance parameter of the kth merging conflict point. Let be the traffic convergence angle parameter at the k-th merging conflict point. To maximize traffic capacity, For merging lane width parameters, For the green light signal cycle parameters, The dynamically adjusted merging safety threshold. For real-time traffic density, The baseline traffic density is used.

[0049] Specifically, the service area merging safety threshold model focuses on the conflict risk influencing factors in merging areas. First, for each merging conflict point, distance and intersection angle parameters are selected as key influencing factors. Differentiated weighting coefficients are assigned to reflect the varying risk contributions of different conflict points. The combined impact of all conflict points is accumulated using a summation operation, and the data distribution characteristics are optimized through square and square root operations. This, combined with model calibration parameters, yields the basic merging safety threshold. Further consideration is given to the constraints of maximum capacity, merging lane width, and green light cycle on traffic efficiency. The impact of traffic efficiency on the safety threshold is quantified through the correlation calculation of these three factors and incorporated into the basic threshold formula. Simultaneously, the difference between real-time traffic density and baseline traffic density is introduced, and an exponential function is used to construct a dynamic adjustment mechanism. Combined with dedicated calibration parameters, this achieves dynamic adaptation of the safety threshold, forming a complete model. The conflict point weighting coefficients are set from 0.2 to 0.9 based on the conflict risk level determined by on-site surveys. The model calibration parameters are fitted to 0.4 to 1.5 using multi-scenario measured data. The baseline traffic density is determined to be a fixed value between 12 and 35 based on road design standards and historical traffic data. This formula combines the core influencing factors of merging safety with a combination of static basic thresholds and dynamic adjustment mechanisms. During implementation, relevant parameters are extracted from the dynamic traffic flow evolution sequence and substituted into the model to calculate the dynamic safety thresholds under different traffic densities. This provides a quantitative basis for determining the boundary conditions for safe operation in merging areas, ensuring the scientific nature and adaptability of the boundary conditions.

[0050] Preferably, the expression of the intelligent judgment model for sudden traffic flow risks is:

[0051]

[0052] in, To quantify the risk of sudden traffic congestion, , For risk discrimination coefficient, Let p be the traffic flow mutation value of the p-th type of traffic flow. For the time span of the sudden change in traffic flow of type p, This represents the real-time traffic flow of the p-th type of traffic. Let p be the maximum throughput for traffic flow type p. This is the spatial risk diffusion function.

[0053] Specifically, the intelligent model for identifying sudden traffic flow risks correlates traffic flow mutation characteristics with risk quantification. It characterizes the mutation intensity by the ratio of the mutation value to the time span of each type of traffic flow, and reflects the traffic saturation level by combining the ratio of real-time traffic flow to the maximum allowable traffic flow. The influence weights of these two indicators are adjusted by assigning risk discrimination coefficients. An integration operation is used to consolidate the mutation risk contribution of all traffic flow types. The second derivative of the spatial risk diffusion function is then introduced to quantify the spatial propagation effect of risk. Combined with a global risk discrimination coefficient, the overall accuracy of the model is optimized, ultimately forming a risk quantification formula. Regarding parameter values, the risk discrimination coefficient is set from 0.3 to 1.1 based on traffic flow type and road capacity, the global risk discrimination coefficient is calibrated to 0.5 to 1.3 using historical sudden traffic flow event data, and the maximum allowable traffic flow is determined to be a fixed value between 20 and 45 based on road design standards, number of lanes, and merging scenarios. This formula combines the abrupt changes and spatial diffusion patterns of sudden traffic flows to achieve precise risk quantification. During implementation, relevant parameters such as traffic flow mutations and real-time traffic flow are extracted from the dynamic traffic flow evolution sequence and substituted into the model to calculate the risk quantification value. Based on the quantification results, the types and scope of safety hazards caused by sudden traffic flows are identified, providing accurate risk data support for safety assessment and ensuring the timeliness and accuracy of risk identification.

[0054] Preferably, the intelligent traffic conflict data analysis platform includes a multi-source sensor array, edge computing nodes, and a cloud data processing module. The multi-source sensor array consists of a millimeter-wave radar, a high-definition video acquisition unit, and a microwave detector. The millimeter-wave radar has a sampling frequency of 50Hz, the high-definition video acquisition unit has a resolution of 4K and a frame rate of 30fps, and the microwave detector has a detection range of 5-200m. The edge computing node adopts a heterogeneous computing architecture, integrating FPGA and ARM processors, and supports real-time data preprocessing and feature extraction. The cloud data processing module adopts a distributed storage architecture, and the data transmission protocol adopts a custom transmission protocol based on UDP, with transmission latency controlled within 10ms. The weights of each indicator in the safety assessment index system are determined by combining the analytic hierarchy process (AHP) with the entropy weight method, and the weight allocation range is 0.05-0.3.

[0055] Preferred, such as Figure 2As shown, step S2 includes the following sub-steps: S21, selecting traffic flow speed, density, and flow data corresponding to the spatiotemporal coordinates in the original dataset, establishing a spatiotemporal data matrix, and determining the spatiotemporal correlation range of different data points through a neighborhood search algorithm; S22, inputting the data within the spatiotemporal correlation range into a traffic flow situation spatiotemporal coupling inference algorithm, and obtaining the spatiotemporal evolution trend of the traffic flow situation through time series decomposition and spatial interpolation operations; S23, based on the spatiotemporal evolution trend, analyzing the dynamic law of traffic flow operation status changing with time and the coupling characteristics of spatial distribution, and identifying the precursor information of sudden changes in traffic flow situation; S24, integrating the dynamic law, coupling characteristics, and precursor information to generate a dynamic traffic flow situation evolution sequence including speed change rate, density gradient, and flow fluctuation amplitude.

[0056] Specifically, step S involves generating a dynamic traffic flow situation evolution sequence: S21 accurately selects core data on traffic speed, density, and flow corresponding to spatiotemporal coordinates from a multi-dimensional raw data set, constructs a structured spatiotemporal data matrix according to time sequence and spatial location, and uses a neighborhood search algorithm to set a spatial search radius of 30 meters and a time search window of 5 seconds to determine the spatiotemporal correlation range of different data points, ensuring the timeliness and spatial relevance of the correlated data; S22 inputs the selected data within the spatiotemporal correlation range into the traffic flow situation spatiotemporal coupling inference algorithm, decomposes the traffic flow data change pattern in a 5-minute cycle through time series decomposition operations, and uses linear interpolation operations to supplement the 10-meter interval data gaps in the spatial dimension, comprehensively acquiring the traffic flow situation in the spatiotemporal dimension. Evolutionary Trends; S23 Based on the acquired spatiotemporal evolutionary trends, the dynamic patterns of traffic flow changes over time are extracted through statistical analysis. Correlation analysis is used to uncover the coupling characteristics of traffic flow in different spatial areas. Abrupt change thresholds are set for speed change rate exceeding 15% and density gradient exceeding 0.8 to accurately identify precursory information of sudden changes in traffic flow. S24 The dynamic patterns, coupling characteristics, and precursory information are systematically integrated to extract three core indicators: speed change rate, density gradient, and flow fluctuation amplitude. The speed change rate is calculated at a 2-second interval, the density gradient is statistically analyzed in 50-meter segments, and the flow fluctuation amplitude is calculated in 1-minute cycles. Finally, an evolutionary sequence that comprehensively reflects the spatiotemporal dynamic changes of traffic flow is generated, providing core data support for subsequent safety assessments.

[0057] Preferred, such as Figure 3As shown, step S3 includes the following sub-steps: S31, extracting traffic density, merging speed, and lane occupancy rate data of the merging area from the dynamic traffic flow evolution sequence, and constructing a threshold calculation dataset; S32, inputting the dataset into the service area merging safety threshold model, determining the model parameters through variable correlation analysis, and determining the degree of influence of each parameter on merging safety; S33, based on the parameters and the degree of influence, calculating the merging critical speed, safe distance, and lane switching frequency corresponding to different traffic densities through boundary condition constraint calculation; S34, integrating the critical speed, safe distance, and lane switching frequency to form complete boundary conditions for safe operation of the merging area.

[0058] Specifically, step S3 determines the boundary conditions for safe operation of the merging area: S31 extracts key data of the merging area from the dynamic traffic flow evolution sequence, including traffic density, merging speed, and lane occupancy rate with a statistical period of 10 seconds. The extracted data is deduplicated and sorted to construct a standardized threshold calculation dataset, ensuring a high degree of data compatibility with the model's calculation requirements. S32 inputs the constructed dataset into the service area merging safety threshold model, uses Pearson correlation coefficient analysis to calculate the correlation between each parameter and merging safety, and sets an absolute correlation coefficient greater than 0.6 as the key parameter selection criterion to determine the impact of each key parameter on merging safety, thus providing a basis for subsequent... The subsequent calculations provide a priority basis; S33, based on the selected key parameters and their corresponding impact levels, calculates the corresponding merging critical speed, safe following distance, and lane switching frequency through boundary condition constraint calculations, with each 5-unit increase in traffic density as a gradient. The critical speed calculation is accurate to one decimal place, the safe following distance is taken as an integer value, and the lane switching frequency is statistically calculated in minutes; S34 integrates and optimizes the critical speed, safe following distance, and lane switching frequency under different traffic density gradients, eliminates logically contradictory data, and forms complete boundary conditions for safe operation of merging areas covering the traffic density range of 0 to 40, providing a definite quantitative standard for judging the safety of traffic operation in merging areas.

[0059] Preferred, such as Figure 4 As shown, S4 includes the following sub-steps: S41, performing sliding window processing on the dynamic traffic flow situation evolution sequence, extracting traffic flow, speed fluctuation, and density change data within each window to construct a risk identification dataset; S42, inputting the dataset into the emergency traffic flow risk intelligent discrimination model, identifying abnormal features in traffic flow operation through feature normalization operations and outlier detection; S43, based on the abnormal features, quantifying the risk level corresponding to the abnormal features through risk level classification rules, and determining the risk level; S44, combining the risk level and the spatial distribution of abnormal features, analyzing the types of safety hazards caused by the emergency traffic flow, and defining the spatial range and time span of the hazard's impact.

[0060] Specifically, step S4 involves the identification and quantification of sudden traffic flow risks: S41 applies a sliding window to the dynamic traffic flow evolution sequence, setting the window size to 30 seconds and the sliding step to 10 seconds to ensure comprehensive coverage of the sequence data. From each window, three core data types—traffic volume, speed fluctuations, and density mutations—are extracted. Anomalies are preliminarily screened, and data exceeding three times the normal range standard deviation are removed, constructing a risk identification dataset that meets data quality standards. S42 inputs the dataset into the sudden traffic flow risk intelligent discrimination model, using the min-max normalization method to standardize the data, mapping it to the 0-1 interval to eliminate the influence of dimensional differences. Then, anomaly detection is performed using a local anomaly factor algorithm, setting the anomaly factor threshold to 2.5 for further refinement. S43 accurately identifies abnormal characteristics in traffic flow; based on the identified abnormal characteristics, according to the preset risk level classification rules, it quantifies and scores the abnormality according to three dimensions: duration, scope of impact, and severity. Each dimension has a score of 0 to 10, and the total score is 0 to 30, corresponding to three risk levels: low, medium, and high. Among them, 0 to 10 is low risk, 11 to 20 is medium risk, and 21 to 30 is high risk; S44 combines the determined risk level with the spatial distribution of abnormal characteristics, and analyzes the clustering area of ​​abnormal characteristics through spatial clustering algorithm to determine the type of safety hazard caused by sudden traffic flow. The spatial range of the hazard's impact is defined in 50-meter units, and the time span of the hazard's impact is calculated in minutes, providing a definite target for subsequent safety management.

[0061] Preferred, such as Figure 5 As shown, step S5 includes the following sub-steps: S51, collecting coupling correlation features, critical control parameters, and risk level quantification results, and standardizing the format of multi-source data to ensure data consistency; S52, using the multi-source data fusion module of the traffic conflict data intelligent analysis platform, employing a weighted fusion algorithm to perform fusion operations on the standardized data to generate a fused data set; S53, based on the fused data set, selecting indicators such as traffic flow conflict frequency, merging efficiency, and risk duration to construct a multi-level safety assessment indicator system; S54, using the indicator weight calculation method and combining the statistical characteristics of the fused data, determining the weight of each assessment indicator to form a complete indicator weight set.

[0062] Specifically, step S5 involves multi-source data fusion and the construction of an evaluation index system: S51 comprehensively collects the coupling correlation features obtained in step S2, the critical control parameters calculated in step S3, and the risk level results quantified in step S4. The three types of data are then standardized in format, with data storage format unified to CSV, data precision retained to two decimal places, and a unified coding rule used in the representation to ensure consistency and compatibility between data from different sources and of different types. S52 utilizes the multi-source data fusion module of the traffic conflict data intelligent analysis platform, employing a weighted fusion algorithm to perform fusion operations on the standardized data. Weights are assigned based on data reliability, with sensor-collected data weighted at 0.6 and model-calculated data weighted at 0.4. The fusion operation is completed at the edge computing node, ensuring a 10m... Within seconds, the output results are generated into a comprehensive and unified fusion data set. Based on the fusion data set and combined with the safety assessment requirements for merging traffic in interchange service areas, S53 selects eight key indicators, including traffic flow conflict frequency, merging efficiency, risk duration, and critical parameter compliance rate. A multi-level safety assessment indicator system is constructed in three levels: target layer, criterion layer, and indicator layer, covering core dimensions such as traffic flow operation, merging process, and risk status. S54 uses the analytic hierarchy process (AHP) combined with the entropy weight method to determine the weights of each assessment indicator. In the AHP, the consistency ratio of the judgment matrix is ​​controlled within 0.1, and in the entropy weight method, the data entropy value calculation precision is retained to three decimal places. The final determined weights of each indicator are allocated in the range of 0.05 to 0.3, forming a scientific and reasonable set of indicator weights, which provides reliable support for the final comprehensive safety assessment calculation.

[0063] like Figure 6As shown, a safety assessment system for merging traffic in grade-separated service areas based on traffic flow patterns is applied to this method. The system includes: a multi-dimensional traffic flow parameter acquisition unit connected to a traffic conflict data intelligent analysis platform, used to collect traffic flow operation parameters, road geometric parameters, and traffic environment parameters of the merging traffic sections in the grade-separated service area and transmit them to a data preprocessing unit; a data preprocessing unit connected to both the multi-dimensional traffic flow parameter acquisition unit and the traffic flow pattern spatiotemporal coupling inference unit, used to perform format conversion and redundant data removal on the collected multi-dimensional parameters before transmitting them to the traffic flow pattern spatiotemporal coupling inference unit; and a traffic flow pattern spatiotemporal coupling inference unit connected to both the data preprocessing unit and the service area merging safety threshold calculation unit. Next, the system processes data and generates a dynamic traffic flow evolution sequence through a traffic flow situation spatiotemporal coupling inference algorithm. The service area merging safety threshold calculation unit is connected to both the traffic flow situation spatiotemporal coupling inference unit and the sudden traffic flow risk discrimination unit, calculating the safe operation boundary conditions of the merging area based on the evolution sequence. The sudden traffic flow risk discrimination unit is connected to both the service area merging safety threshold calculation unit and the multi-source data fusion evaluation unit, identifying safety hazards and their impact range through a sudden traffic flow risk intelligent discrimination model. The multi-source data fusion evaluation unit is connected to both the sudden traffic flow risk discrimination unit, the traffic flow situation spatiotemporal coupling inference unit, and the service area merging safety threshold calculation unit, fusing multi-source data to construct an evaluation index system and calculate weights, outputting the merging traffic safety evaluation results.

[0064] This paper presents a method and system for assessing the safety of merging traffic in grade-separated interchange service areas based on traffic flow patterns. Utilizing a multi-source sensor array and heterogeneous computing architecture of a traffic conflict data intelligent analysis platform, it achieves high-precision acquisition and low-latency processing of traffic flow, road geometry, and traffic environment parameters. Combined with a spatiotemporal coupling inference algorithm for traffic flow patterns, it deeply mines spatiotemporally correlated data, analyzing the coupled correlation characteristics of traffic flow operation states in different time periods and areas, generating dynamic evolution sequences. This addresses the lack of in-depth analysis of the spatiotemporal dynamic correlation of traffic flow patterns in existing technologies, making the assessment more closely aligned with real traffic environments. Simultaneously, it dynamically calculates safe operating boundary conditions under different traffic flow densities using a service area merging safety threshold model, replacing traditional fixed threshold settings. Furthermore, it utilizes an intelligent judgment model for sudden traffic flow risks to accurately extract abnormal features and quantify risk levels, determining the type of hazard and its impact range, thus overcoming the shortcomings of existing technologies in identifying sudden traffic flow risks.

[0065] This invention constructs a multi-level safety assessment index system through multi-source data fusion analysis, combined with a scientific weight allocation method, to achieve comprehensive coverage of assessment dimensions. The heterogeneous computing architecture and distributed storage design at the hardware level, and the dedicated data transmission protocol and weight calculation method at the software level, jointly improve the system's computing efficiency and data processing capabilities, making the assessment results both accurate and dynamically adaptable. This provides comprehensive and intelligent technical support for the safety of merged traffic in interchange service areas, which is superior to the single-dimensional analysis and fixed parameter assessment mode of existing technologies.

[0066] In the description of this invention, it should be noted that, unless otherwise specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for assessing the safety of combined traffic flow in grade-separated interchange service areas, characterized in that: Includes the following steps: S1 collects traffic flow parameters, road geometry parameters, and traffic environment parameters of the merged traffic sections in the interchange service area through the traffic conflict data intelligent analysis platform, and constructs a multi-dimensional raw data set; S2 utilizes a traffic flow situation spatiotemporal coupling inference algorithm to perform dynamic correlation analysis on spatiotemporal related data in the original dataset, mine the coupling correlation characteristics of traffic flow operation status in different time periods and different regions, and generate a dynamic traffic flow situation evolution sequence. S3, based on the traffic flow situation evolution sequence, calculates the critical control parameters of traffic flow operation in the merging area through the service area merging safety threshold model, and determines the safe operation boundary conditions under different traffic flow densities; S4 employs an intelligent judgment model for sudden traffic flow risks to extract abnormal features and quantify risk levels in the traffic flow situation evolution sequence, identifying the types and scope of safety hazards caused by sudden traffic flow. S5 uses a traffic conflict data intelligent analysis platform to perform multi-source data fusion analysis on coupled correlation characteristics, critical control parameters and risk level quantification results, construct a safety assessment index system and calculate the weight of each index. S6 combines the safety assessment indicator system and indicator weights, and performs comprehensive calculations based on the safety assessment parameters for combined traffic flow in interchange service areas based on traffic flow patterns, to output the combined traffic safety assessment results.

2. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, The expression for the spatiotemporal coupling inference algorithm of traffic flow situation is: ； ； in, It is a spatial position vector. For time variables, Let be the instantaneous velocity vector of the i-th type of traffic flow. Let be the density distribution function of the i-th type of traffic flow. Let be the spatiotemporal coupling coefficient of the i-th type of traffic flow. To deduce the correction coefficient, For the time integration interval, The function representing the influence of historical traffic flow patterns. This represents the traffic flow spatial diffusion coefficient.

3. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, The expression for the service area merging safety threshold model is: ； ； in, Based on the basic merging safety threshold, For model calibration parameters, The weighting coefficient for the k-th merging conflict point. Let k be the distance parameter of the kth merging conflict point. Let be the traffic convergence angle parameter at the k-th merging conflict point. To maximize traffic capacity, For merging lane width parameters, For the green light signal cycle parameters, The dynamically adjusted merging safety threshold. For real-time traffic density, The baseline traffic density is used.

4. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, The expression for the intelligent judgment model of sudden traffic flow risk is: ； in, To quantify the risk of sudden traffic congestion, , For risk discrimination coefficient, Let p be the traffic flow mutation value of the p-th type of traffic flow. For the time span of the sudden change in traffic flow of type p, This represents the real-time traffic flow of the p-th type of traffic. Let p be the maximum throughput for traffic flow type p. This is the spatial risk diffusion function.

5. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, The intelligent traffic conflict data analysis platform includes a multi-source sensor array, edge computing nodes, and a cloud data processing module. The multi-source sensor array consists of a millimeter-wave radar, a high-definition video acquisition unit, and a microwave detector. The millimeter-wave radar has a sampling frequency of 50Hz, the high-definition video acquisition unit has a resolution of 4K and a frame rate of 30fps, and the microwave detector has a detection range of 5-200m. The edge computing nodes adopt a heterogeneous computing architecture, integrating FPGA and ARM processors, and support real-time data preprocessing and feature extraction. The cloud data processing module adopts a distributed storage architecture, and the data transmission protocol adopts a custom transmission protocol based on UDP, with transmission latency controlled within 10ms. The weights of each indicator in the safety assessment index system are determined by combining the analytic hierarchy process (AHP) with the entropy weight method, with a weight allocation range of 0.05-0.

3.

6. The method for assessing the safety of merged traffic in grade-separated service areas based on traffic flow patterns according to claim 1, characterized in that, S2 includes the following sub-steps: S21, selecting traffic flow speed, density, and flow data corresponding to spatiotemporal coordinates in the original dataset, establishing a spatiotemporal data matrix, and determining the spatiotemporal correlation range of different data points through a neighborhood search algorithm; S22, inputting the data within the spatiotemporal correlation range into a traffic flow situation spatiotemporal coupling inference algorithm, and obtaining the spatiotemporal evolution trend of the traffic flow situation through time series decomposition and spatial interpolation operations; S23, based on the spatiotemporal evolution trend, analyzing the dynamic law of traffic flow operation status changing with time and the coupling characteristics of spatial distribution, and identifying the precursor information of sudden changes in traffic flow situation; S24, integrating the dynamic law, coupling characteristics, and precursor information to generate a dynamic traffic flow situation evolution sequence including speed change rate, density gradient, and flow fluctuation amplitude.

7. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, S3 includes the following steps: S31, extracting traffic density, merging speed, and lane occupancy rate data of the merging area from the dynamic traffic flow evolution sequence, and constructing a threshold calculation dataset; S32, inputting the dataset into the service area merging safety threshold model, determining the model parameters through variable correlation analysis, and determining the degree of influence of each parameter on merging safety; S33, based on the parameters and the degree of influence, calculating the merging critical speed, safe distance, and lane switching frequency corresponding to different traffic densities through boundary condition constraint calculation; S34, integrating the critical speed, safe distance, and lane switching frequency to form complete boundary conditions for safe operation of the merging area.

8. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, S4 includes the following steps: S41, performing sliding window processing on the dynamic traffic flow situation evolution sequence, extracting traffic flow, speed fluctuation, and density change data in each window, and constructing a risk identification dataset; S42, inputting the dataset into the emergency traffic flow risk intelligent discrimination model, and identifying abnormal features in traffic flow operation through feature normalization operation and outlier detection. S43, Based on abnormal characteristics, the risk level is determined by quantifying the risk level corresponding to the abnormal characteristics through risk level classification rules; S44, combining the spatial distribution of risk levels and abnormal characteristics, analyzes the types of safety hazards caused by sudden traffic flows, and defines the spatial range and time span of the hazard's impact.

9. The method for assessing the safety of combined traffic flow in grade-separated interchange service areas based on traffic flow patterns according to claim 1, characterized in that, S5 includes the following steps: S51, collecting coupling correlation features, critical control parameters, and risk level quantification results, and standardizing the format of multi-source data to ensure data consistency; S52, using the multi-source data fusion module of the traffic conflict data intelligent analysis platform, performing fusion calculations on the standardized data using a weighted fusion algorithm to generate a fused data set; S53, based on the fused data set, selecting indicators such as traffic flow conflict frequency, merging efficiency, and risk duration to construct a multi-level safety assessment indicator system. S54 uses the indicator weight calculation method and combines the statistical characteristics of the fused data to determine the weight of each evaluation indicator, forming a complete set of indicator weights.

10. A grade-separated interchange service area merging traffic safety assessment system based on traffic flow patterns, characterized in that: The system is applied to the grade-separated interchange service area merging traffic safety assessment method based on traffic flow patterns as described in claim 1, including: The traffic flow multi-dimensional parameter acquisition unit is connected to the traffic conflict data intelligent analysis platform. It is used to collect traffic flow operation parameters, road geometric parameters and traffic environment parameters of the merged traffic sections in the interchange service area and transmit them to the data preprocessing unit. The data preprocessing unit is connected to the multi-dimensional parameter acquisition unit and the spatiotemporal coupling inference unit of traffic flow, respectively. It is used to convert the format of the acquired multi-dimensional parameters and remove redundant data before transmitting them to the spatiotemporal coupling inference unit of traffic flow. The traffic flow situation spatiotemporal coupling inference unit is connected to the data preprocessing unit and the service area merging safety threshold calculation unit. It processes data and generates a dynamic traffic flow situation evolution sequence through the traffic flow situation spatiotemporal coupling inference algorithm. The service area merging safety threshold calculation unit is connected to the traffic flow situation spatiotemporal coupling simulation unit and the sudden traffic flow risk discrimination unit, respectively, and calculates the safe operation boundary conditions of the merging area based on the evolution sequence; The sudden traffic flow risk judgment unit is connected to the service area merging safety threshold calculation unit and the multi-source data fusion evaluation unit. It identifies safety hazards and their impact range through the sudden traffic flow risk intelligent judgment model. The multi-source data fusion assessment unit is connected to the sudden traffic flow risk judgment unit, the traffic flow situation spatiotemporal coupling inference unit, and the service area merging safety threshold calculation unit, respectively. It integrates multi-source data to construct an assessment index system and calculates weights, and outputs the merged traffic safety assessment results.

Images