Safety risk assessment and early warning system for complex interwoven area in intelligent network-connected mixed driving environment

Abstract

The invention discloses a safety risk assessment and early warning system for a complex interleaving area in an intelligent network connection mixed driving environment, and the system comprises the following steps: N1, the current situation analysis of the complex interleaving area, including data collection and traffic operation characteristic analysis, N2, the construction of a traffic conflict calculation model, including the design of a conflict simulation experiment in a mixed driving environment, and the design of the traffic conflict calculation model through an orthogonal experiment design method. Constructing a simulation experiment orthogonal table; carrying out a simulation experiment on the complex interleaving area by utilizing SUMO software, exporting corresponding data, and establishing a conflict calculation model; n3, constructing a traffic safety evaluation method: establishing traffic safety evaluation indexes, constructing a traffic safety evaluation method based on a grey clustering model, introducing a sine curve type possibility degree function, and determining the weight of each evaluation index through combined weighting of a fuzzy consistency matrix and an entropy evaluation method; and N4, performing security level division and verification. According to the method, gray fixed weight clustering evaluation is improved, and the traffic safety pre-judgment capability of the complex interlaced area is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation systems, and particularly to a safety risk assessment and warning system for complex weaving areas in an intelligent connected mixed-driving environment. Background Art

[0002] With the development of intelligent connected vehicles, the coexistence of autonomous driving and manual driving will become the norm, especially in complex weaving areas with multiple lanes and short distances, bringing new traffic safety challenges. Traditional safety assessment methods mainly target conflicts of manually driven vehicles and are difficult to adapt to the new changes in the intelligent connected environment.

[0003] Currently, domestic and foreign research on the safety of mixed traffic flows mostly focuses on the impact of the penetration rate of autonomous driving, with the emphasis on urban intersections and basic sections of expressways, and less research on weaving areas. In addition, limited by the openness of autonomous driving data, existing research mainly relies on simulation modeling.

[0004] Therefore, in-depth analysis of the traffic flow and conflict characteristics in complex weaving areas, construction of a safety risk assessment and warning system for complex weaving areas in an intelligent connected mixed-driving environment, and realization of the prediction of safety levels are of great significance for ensuring the efficient and safe operation of traffic. Summary of the Invention

[0005] Aiming at the deficiencies of the above-mentioned prior art, the technical problem to be solved by this patent application is how to provide a safety risk assessment and warning system for complex weaving areas in an intelligent connected mixed-driving environment, which can quantitatively evaluate the traffic safety of complex weaving areas in the state of mixed traffic flow and provide guarantee for the efficient and safe operation of intelligent connected mixed traffic.

[0006] To solve the above technical problem, the present invention adopts the following technical solutions:

[0007] A safety risk assessment and warning system for complex weaving areas in an intelligent connected mixed-driving environment includes the following steps:

[0008] N1: Analysis of the current situation of complex weaving areas, including data collection and analysis of traffic operation characteristics:

[0009] Data collection is carried out by using an unmanned aerial vehicle to take aerial videos of the traffic flow operation in the weaving area under different traffic states at high altitude, and using Tracker software to process the traffic flow operation videos to obtain data on the road conditions, traffic operation characteristics and traffic conflict current situations in the weaving area; the investigation videos of different traffic states taken on the spot are statistically analyzed by lane and vehicle type to analyze traffic volume-related indicators;

[0010] The analysis of traffic operation characteristics includes the analysis of traffic volume characteristics, vehicle behavior characteristics and traffic conflict characteristics;

[0011] N2: Construction of Traffic Conflict Calculation Model: Design conflict simulation experiments in a mixed traffic environment. Using the orthogonal experimental design method, based on the analysis of traffic flow characteristics and vehicle behavior characteristics, select hourly traffic volume, autonomous driving penetration rate, large vehicle ratio, weaving ratio, weaving flow ratio, road longitudinal slope, and weaving length as the influencing factors of the orthogonal experiment, and construct an orthogonal table for the simulation experiment;

[0012] Use SUMO software to conduct simulation experiments on complex weaving areas and export corresponding data. Combine SSAM software to extract traffic conflict data and establish a conflict calculation model; According to the types and severity of traffic conflicts, divide the traffic conflicts in complex weaving areas into four categories: severe lane-changing conflicts, general lane-changing conflicts, severe rear-end collisions, and general rear-end collisions, and establish conflict calculation models respectively;

[0013] N3: Construction of Traffic Safety Evaluation Method: Establish traffic safety evaluation indicators, including the incidence rate of severe lane-changing conflicts, the incidence rate of general lane-changing conflicts, the incidence rate of severe rear-end collisions, and the incidence rate of general rear-end collisions; Construct a traffic safety evaluation method based on the grey clustering model, introduce a sine curve-type possibility function, and determine the weights of each evaluation indicator by combining the fuzzy consistency matrix and the entropy value method; Verify the safety evaluation method and verify the index thresholds divided by the evaluation method based on the simulation experiment data;

[0014] N4: Safety Level Classification and Verification: According to the severe rear-end collision rate, general rear-end collision rate, severe lane-changing conflict rate, and general lane-changing conflict rate, divide the traffic safety risks in complex weaving areas into four levels, Level I - the highest safety level, Level II - lower risk, Level III - medium risk, Level IV - high risk.

[0015] Preferably, in step N1, the analysis of traffic operation characteristics further includes: analyzing the vehicle behavior characteristics under different traffic states, including lane-changing behavior analysis and vehicle average speed statistics; According to the traffic conflict investigation situation, set the severe degree conflict threshold and general degree conflict threshold based on the 85% percentile cumulative frequency statistics method.

[0016] Preferably, step N2 specifically includes the following processes:

[0017] P1: Experimental Design: Use SPSS software for orthogonal experimental design, select 7-factor 5-level variables, and use the method of virtual levels to fill in the variables with less than 5 levels, and complete the experimental scheme with a 1% longitudinal slope;

[0018] P2: Simulation: Take the combinations of each factor of the experimental scheme as the input parameters of the SUMO simulation experiment, run the simulation model, and the simulation duration is 4200s, with 600s as the warm-up stage;

[0019] P3: Conflict analysis: Extract the fcd.xml file from the simulation results and convert it into a trajectory trj file. Use the SSAM software for conflict analysis, and respectively count the severe lane-changing conflicts, general lane-changing conflicts, severe rear-end collisions, and general rear-end collisions according to the conflict threshold.

[0020] P4: Correlation test: Analyze the candidate independent variables using the Pearson correlation coefficient. If the correlation coefficient is greater than 0.8, then remove one of the variables; otherwise, retain them.

[0021] P5: Construction of conflict calculation model: Conduct statistical analysis on the number of severe lane-changing conflicts. Based on the mean and variance characteristics, use the Stata software to construct a Poisson distribution model and a negative binomial distribution model, and gradually remove non-significant variables through backward regression to optimize the model.

[0022] P6: Model optimization: Evaluate the overdispersion and goodness of fit of the Poisson model and the negative binomial model according to the AIC criterion and the BIC criterion to determine the final conflict calculation model.

[0023] Preferably, the method for traffic safety evaluation in step N3 specifically includes the following process:

[0024] S1: Establishment of traffic safety evaluation indicators. Select "conflict incidence rate" as the evaluation indicator for the safety status of complex weaving areas. In the calculation of the conflict incidence rate, the traffic conflict incidence rate k i refers to the ratio of the number of traffic conflicts N i within an hour to the traffic volume Q i . According to the traffic conflict incidence rate formula, calculate the 4 types of traffic conflict incidence rates k i for 49 groups of experiments, forming a 49×4 evaluation matrix A. The calculation formula is as follows:

[0025]

[0026] Among them, k i represents the conflict incidence rate, i represents the conflict type, i = 1, 2, 3, 4, respectively representing severe lane-changing conflicts, general lane-changing conflicts, severe rear-end collisions, and general rear-end collisions.

[0027] S2: Use the cumulative percentage frequency method to determine each grey class level. Draw a cumulative percentage frequency curve graph from the data values of the objects participating in clustering, and select the 15%, 40%, 60%, and 85% percentile values of the cumulative frequency curve as the turning points of the possibility function images of each index, that is, the characteristic values k of different grey classes.

[0028] S3: Introduce a sine curve-type possibility function. Use the determination method of the sine curve-type possibility function to determine the possibility functions of each index under different grey classes. The function form is as follows:

[0029]

[0030] S4: Select the method that combines the fuzzy consistency matrix and the entropy method subjectively and objectively:

[0031] The weights of the fuzzy consistency matrix refer to the weights of the conflict indicators of different severities and types by Xu Wenkang, and the values are:

[0032]

[0033] Entropy method weight calculation: By calculating the entropy values of each indicator, the indicator weights are determined according to the differences of the indicator samples: The greater the sample difference, the smaller the information entropy of the indicator, the greater the amount of information provided by the indicator, that is, the greater the role played in the evaluation, and the higher the weight; When the probability of each state is p i (i = 1, 2,..., n), the system entropy form is as follows:

[0034]

[0035] Obtain the relative weights of each indicator according to the entropy method weight calculation method, as follows:

[0036] First, determine the original data matrix according to the indicator system Its standardized values form the decision matrix C = (C ij ) m×n ;

[0037] Then, calculate the proportion of the jth solution under the ith indicator in this indicator:

[0038]

[0039] Calculate the entropy value of the ith indicator, where the constant k = 1 / ln(n)

[0040]

[0041] Calculate the weights of different indicators:

[0042]

[0043] S5: Conduct a comprehensive weight calculation for the four types of conflict incidence rate indicators. After referring to the research results of Yang Ming, select ρ = 0.6, and the specific formula is:

[0044] w i = ρw i1 +(1 - ρ)w i2

[0045] Among them, w i is the comprehensive weight of each indicator; ρ is the correction coefficient; w i1 、wi2 For the weights of the fuzzy consistency matrix method and entropy value method for each index;

[0046] S6: According to the possibility function and index weights, a comprehensive clustering coefficient matrix can be constructed. Then, the clustering coefficient of the observed object i with respect to the k-th grey class The specific calculation formula is:

[0047]

[0048] where is the possibility function of the j-th index of the observed object i at the grey class k, and η j is the weight of the j-th index;

[0049] S7: Compare the clustering coefficients corresponding to different grey classes for each experimental scheme in turn. The grey class to which the maximum clustering coefficient belongs is the safety risk level of the experimental scheme; count the sample values of different evaluation indexes under different safety levels in the complex weaving area. For the first, second, and third levels, take the maximum evaluation index value as the threshold to determine the evaluation index thresholds for each safety level in the intelligent connected mixed traffic environment.

[0050] Preferably, for the verification of the safety evaluation method, the conventional straight-line grey variable weight clustering evaluation method and the conventional straight-line grey clustering evaluation method based on fuzzy consistency matrix weighting of previous safety evaluation methods are used to evaluate the sample data in this paper and compare the accuracy rate of the evaluation method in this paper. By comparing with the conventional straight-line grey variable weight clustering and fixed weight clustering evaluation methods, the accuracy and effectiveness of the safety evaluation method constructed in this paper are verified.

[0051] In summary, the safety risk assessment and early warning system for the complex weaving area in the intelligent connected mixed driving environment has the following beneficial effects:

[0052] The present invention uses microscopic simulation and conflict analysis to improve the accuracy of safety risk assessment for complex weaving areas; adopts an improved grey clustering model to realize the quantitative evaluation of safety levels in mixed traffic environments, providing a scientific decision-making basis for traffic safety optimization in intelligent connected environments. Description of the Drawings

[0053] Figure 1 is the flow chart of the safety risk assessment and early warning system for the complex weaving area in the intelligent connected mixed driving environment;

[0054] Figure 2 is the lane-changing distribution map of different sections. Detailed Embodiments

[0055] The present invention will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several changes and improvements can still be made. These all fall within the protection scope of the present invention.

[0056] As Figure 1 shown, a safety risk assessment and early warning system for complex weaving areas in an intelligent networked mixed driving environment includes the following steps:

[0057] N1: Analysis of the current situation of complex weaving areas, including data collection and analysis of traffic operation characteristics:

[0058] As Figure 2 Select a short-distance, multi-lane complex weaving area as the research object. This area has a large traffic flow, balanced weaving flow and non-weaving flow, and is suitable for traffic safety research in an intelligent networked mixed driving environment. Through field research and UAV aerial photography, traffic flow videos under different traffic conditions are obtained, and Tracker software is used to process the video data to extract vehicle trajectory information.

[0059] Based on SUMO software, a basic road network for complex weaving areas is constructed, and the behavior parameters of manually driven vehicles and autonomous vehicles are calibrated in combination with traffic flow characteristics.

[0060] Manually driven vehicles: Three types (Human, Human1, Human2) are set, and parameters such as the minimum headway, desired time headway, speed factor, speed deviation, strategic lane change, acceleration lane change, cooperative lane change, keep right driving, and minimum acceptable lane change distance are calibrated.

[0061] Autonomous vehicles: The CACC (Cooperative Adaptive Cruise Control) model is used as the car-following model, and the SL2013 model is selected as the lane change model. Parameters such as acceleration, deceleration, maximum deceleration in an emergency, strategic lane change, acceleration lane change, cooperative lane change, keep right driving, and minimum acceptable lane change distance are calibrated.

[0062] By comparing the measured data with the simulation results, the accuracy of parameter calibration is verified, indicating that the constructed simulation model can effectively reflect the actual traffic operation state.

[0063] N2: Construction of a traffic conflict calculation model

[0064] To study the traffic conflict change law in complex weaving areas under different traffic conditions, the orthogonal experimental design method is adopted, and the following factors are selected as experimental variables: hourly traffic volume, autonomous driving penetration rate, large vehicle ratio, weaving ratio, weaving flow ratio, weaving length, and road longitudinal slope.

[0065] Use SPSS software to generate an orthogonal experiment table, obtain data through SUMO simulation, and extract traffic conflict data in combination with SSAM software. According to the severity of conflicts, traffic conflicts are divided into four categories: severe lane-changing conflicts, general lane-changing conflicts, severe rear-end collisions, and general rear-end collisions. Based on negative binomial regression analysis, calculation models for the four types of conflicts are constructed respectively, and the model formulas are as follows:

[0066] N 1 = exp(1.9225128 + 0.0020886V - 0.5406693MR + 0.0131275L - 0.7760446lR + 2.6839670YR)

[0067] N 2 = exp(2.5629645 + 0.0009639V - 0.5953913MR + 0.0171947L + 2.0972626VR)

[0068] N 3 = exp(0.1211006 + 0.0035253V + 1.6124795VR + 0.0070063L)

[0069] N 4 = exp(3.5436384 + 0.0015160V - 0.4050758MR + 0.0120833L - 0.6183874IR)

[0070] Among them, V is the average hourly traffic volume of the lane, MR is the penetration rate of autonomous driving, L is the weaving length, IR is the weaving ratio, and VR is the weaving flow ratio.

[0071] N3: Construction of traffic safety evaluation method

[0072] Taking the traffic conflict incidence rate as the evaluation index, construct an evaluation index matrix. Introduce a sine curve-type possibility function, and combine the fuzzy consistency matrix and the entropy method for subjective and objective combined weighting to determine the comprehensive weight of each index. The specific method is as follows:

[0073] S1: Establishment of traffic safety evaluation indicators. According to the traffic conflict incidence rate formula, calculate the 4 types of traffic conflict incidence rates of 49 groups of experiments to form a 49×4 evaluation matrix A.

[0074] S2: Use the cumulative percentage frequency method to determine the risk rates of severe lane-changing conflicts, general lane-changing conflicts, severe rear-end collisions, and general rear-end collisions for the gray class levels, and select the 15%, 40%, 60%, and 85% percentile values as the turning points of the possibility functions of each index.

[0075] S3: Construction of sine curve - type possibility function. Introduce the sine curve - type possibility function and adopt the establishment method of the sine curve - type possibility function to determine the possibility functions of each index under different grey classes.

[0076] S4: Calculation of combined weights

[0077] Adopt the method combining fuzzy consistency matrix and entropy method to calculate the comprehensive weights of the indicators of severe lane - changing conflict rate, general lane - changing conflict rate, severe rear - end conflict rate, and general rear - end conflict rate.

[0078] S5: Calculate the comprehensive weights of the four types of conflict incidence rate indicators. After referring to the research results of Yang Ming, select ρ = 0.6.

[0079] S6: Construction of comprehensive clustering coefficient matrix. According to the possibility function and index weights, calculate the clustering coefficients of each experimental scheme and determine its safety level.

[0080] S7: Compare the clustering coefficients corresponding to different grey classes of each experimental scheme in turn. The grey class to which the maximum clustering coefficient belongs is the safety risk level of the experimental scheme. Statistically analyze the sample values of different evaluation indicators under different safety levels in the complex weaving area. Among them, for the first, second, and third levels, take the maximum evaluation index value as the threshold to determine the evaluation index thresholds of each safety level in the intelligent - connected mixed - traffic environment.

[0081] N4: Safety level division and verification

[0082] According to the severe rear - end conflict rate, general rear - end conflict rate, severe lane - changing conflict rate, and general lane - changing conflict rate, divide the traffic safety risks in the complex weaving area into four levels: Level Ⅰ (the highest safety level), Level Ⅱ (lower risk), Level Ⅲ (medium risk), and Level Ⅳ (high risk). Use three evaluation methods for comparison, including the conventional linear grey variable - weight clustering with objective weights, the conventional linear grey fixed - weight clustering with subjective weights, and the curve - type grey clustering method based on combined weights. Among them, the curve - type grey clustering method based on combined weights combines subjective and objective weight - assignment methods and adopts the sine curve - type possibility function. Compared with the traditional linear grey clustering method, it improves the evaluation accuracy and has stronger applicability in the safety evaluation of the mixed - traffic environment.

[0083] In summary, the traffic safety evaluation method for the complex weaving area in the intelligent - connected mixed - traffic environment proposed by the present invention can accurately and efficiently evaluate the traffic safety level of the complex weaving area and provide a scientific basis for the safety prevention strategies of intelligent - connected mixed - traffic.

[0084] Finally, it should be noted that those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to cover these modifications and variations.

Claims

1. A safety risk assessment and early warning system for complex interweaving areas in an intelligent networked mixed driving environment, characterized in that: The following steps are involved: N1: Analysis of the current situation of complex interweaving areas, including data collection and analysis of traffic operation characteristics: Data collection is done by taking aerial photos of traffic flow operation videos under different traffic conditions in the weaving area by drones, and using Tracker software to process the traffic flow operation videos to obtain data on road conditions, traffic operation characteristics and traffic conflict status in the weaving area; Traffic operation characteristics analysis includes traffic volume characteristics analysis, vehicle behavior characteristics analysis and traffic conflict characteristics analysis; N2: Construction of traffic conflict calculation model: Design a conflict simulation experiment in a mixed traffic environment. Use the orthogonal experimental design method to analyze the traffic flow characteristics and vehicle behavior characteristics. Select hourly traffic volume, autonomous driving penetration rate, large vehicle ratio, weaving ratio, weaving flow ratio, road longitudinal slope, and weaving length as influencing factors of the orthogonal experiment, and construct an orthogonal table for the simulation experiment. The SUMO software was used to simulate the complex weaving area and export the corresponding data. The SSAM software was used to extract the traffic conflict data and establish a conflict calculation model. According to the type and severity of traffic conflicts, the traffic conflicts in the complex weaving area were divided into four categories: severe lane change conflicts, general lane change conflicts, severe rear-end collision conflicts and general rear-end collision conflicts. The corresponding conflict calculation models were established. N3: Construction of traffic safety evaluation method: Establish traffic safety evaluation indicators, including the incidence rate of severe lane-changing conflicts, the incidence rate of general lane-changing conflicts, the incidence rate of severe rear-end collisions, and the incidence rate of general rear-end collisions; construct a traffic safety evaluation method based on the grey clustering model, introduce a sinusoidal possibility function, and determine the weight of each evaluation indicator by combining the fuzzy consistency matrix and the entropy method; verify the safety evaluation method, and verify the indicator thresholds divided by the evaluation method based on simulation experimental data. N4: Safety level classification and verification: Based on the severe rear-end collision rate, general rear-end collision rate, severe lane change conflict rate and general lane change conflict rate, the traffic safety risks in complex interweaving areas are divided into four levels: Level I - the highest safety level, Level II - lower risk, Level III - medium risk, and Level IV - high risk.

2. According to claim 1, a safety risk assessment and early warning system for complex interweaving areas in an intelligent networked mixed driving environment is characterized in that: In step N1, the traffic operation characteristics analysis further includes: analyzing the vehicle behavior characteristics under different traffic conditions, including lane change behavior analysis and vehicle average speed statistics; according to the traffic conflict investigation, based on the 85% cumulative frequency statistics method, setting the severity conflict threshold and the general degree conflict threshold.

3. The safety risk assessment and early warning system for complex interweaving areas in an intelligent networked mixed driving environment according to claim 2 is characterized in that: Step N2 specifically includes the following process: P1: Experimental design: Orthogonal experimental design was conducted using SPSS software, with 7 factors and 5 levels of variables selected. The variables with less than 5 levels were filled with the pseudo-level method, and the experimental scheme was completed with a 1% longitudinal slope; P2: Simulation: The various factors of the experimental scheme are combined as the input parameters of the SUMO simulation experiment, and the simulation model is run. The simulation time is 4200s, of which the first 600s is used as the warm-up stage; P3: Conflict analysis: extract the fcd.xml file in the simulation results and convert it into a trajectory trj file, use SSAM software to perform conflict analysis, and count the severe lane change conflicts, general lane change conflicts, severe rear-end collision conflicts and general rear-end collision conflicts according to the conflict threshold; P4: Correlation test: Pearson correlation coefficient is used to analyze the selected variables. If the correlation coefficient is greater than 0.8, one of the variables is eliminated, otherwise it is retained; P5: Construction of conflict calculation model: Statistical analysis was performed on the number of serious lane-changing conflicts. Based on the mean and variance characteristics, Stata software was used to construct the Poisson distribution model and the negative binomial distribution model. Non-significant variables were gradually eliminated through backward regression to optimize the model. P6: Model optimization: Based on the AIC criterion and BIC criterion, the overdispersion and goodness of fit of the Poisson model and the negative binomial model were evaluated to determine the final conflict calculation model.

4. The safety risk assessment and early warning system for complex interweaving areas in an intelligent networked mixed driving environment according to claim 3 is characterized in that: The method for traffic safety evaluation in step N3 specifically includes the following process: S1: Traffic safety evaluation index is established. The "conflict occurrence rate" is selected as the evaluation index of the safety status of complex interweaving areas. In the calculation of the conflict occurrence rate, the traffic conflict occurrence rate k i Refers to the number of traffic conflicts N per hour i With traffic volume Q i The ratio of traffic conflict occurrence rate was calculated according to the traffic conflict occurrence rate formula. i , forming a 49*4 evaluation matrix A, the calculation formula is as follows: Among them, k i represents the conflict occurrence rate, i represents the conflict type, i=1,2,3,4, representing severe lane-changing conflict, general lane-changing conflict, severe rear-end collision conflict, and general rear-end collision conflict, respectively; S2: Use the cumulative percentage frequency method to determine the level of each gray class, draw a cumulative percentage frequency curve based on the data values ​​of the objects participating in the clustering, and select the 15%, 40%, 60% and 85% values ​​of the cumulative frequency curve as the turning points of the possible degree function graph of each indicator, that is, the characteristic values ​​k of different gray classes; S3: Introduce the sinusoidal possibility function, adopt the sinusoidal possibility function establishment method, determine the possibility function of each indicator under different gray categories, and the function form is as follows: S4: Select the method combining the fuzzy consistency matrix and the entropy method with the subjective and objective method: The weight of the fuzzy consistency matrix refers to the weight of different types of conflict indicators with different severity levels given by Xu Wenkang, and the value is: Entropy method weight calculation: By calculating the entropy value of each indicator, the indicator weight is determined according to the difference of the indicator sample: the greater the sample difference, the smaller the information entropy of the indicator, the greater the amount of information provided by the indicator, that is, the greater the role it plays in the evaluation, and the higher the weight; when the probability of each state is p i (i=1,2,...,n), the system entropy is as follows: The relative weight of each indicator is obtained according to the entropy weight calculation method, as follows: First, determine the original data matrix according to the indicator system Its standardized values ​​constitute the decision matrix C = (C ij ) m×n ; Then, calculate the proportion of the jth solution under the i-th indicator: Calculate the entropy value of the i-th indicator, where the constant k = 1 / ln(n) Calculate the weights of different indicators: S5: Calculate the comprehensive weights of the four types of conflict incidence indicators, and select ρ = 0.6 based on Yang Ming’s research results. The specific formula is: w i =ρw i1 +(1-ρ)w i2 Among them, w i is the comprehensive weight of each indicator; ρ is the correction coefficient; w i1 、w i2 is the fuzzy consistency matrix method and entropy value weight of each indicator; S6: According to the possibility function and indicator weights, the comprehensive clustering coefficient matrix can be constructed, then the clustering coefficient of observation object i about the kth gray class is The specific calculation formula is: in, is the possible degree function of the jth indicator of observation object i in gray class k, η j is the weight of the j indicator; S7: Compare the clustering coefficients corresponding to different gray classes of each experimental scheme in turn, and the gray class to which the maximum clustering coefficient belongs is the safety risk level of the experimental scheme; count the sample values ​​of different evaluation indicators under different safety levels in complex interweaving areas, among which the maximum evaluation index value is used as the threshold for levels one, two and three, and determine the evaluation index thresholds for each safety level in the intelligent connected mixed traffic environment.

5. The safety risk assessment and early warning system for complex interweaving areas in an intelligent networked mixed driving environment according to claim 4 is characterized in that: Verification of the safety evaluation method The conventional linear grey variable weight clustering evaluation method and the conventional linear grey clustering evaluation method based on fuzzy consistency matrix weighting were used to evaluate the sample data in this paper and compare the accuracy of the evaluation method in this paper.

Images