Autonomous road traffic system safety situation dynamic evaluation method and system

Abstract

The invention provides an autonomous road traffic system safety situation dynamic evaluation method, provides an evaluation index system and an evaluation method for classified and graded safety situation evaluation, overcomes the defects of an autonomous road traffic system at the present stage, solves the problem that emerging traffic elements are difficult to monitor and evaluate, and improves the safety of the road traffic system. And technical support is provided for supervision of intelligent network connection automobiles and intelligent roads. According to the method, elements contained in a traffic system are sliced and classified on the basis of the influence on traffic safety, all levels and element indexes which need to be considered efficiently during evaluation are obtained, and a hierarchical element library table is constructed; all road traffic conditions included in a traffic system are classified according to traffic characteristics to obtain environment categories, a classification evaluation index system is constructed based on an element library table, and safety situation evaluation of traffic scenes of all the environment categories is covered. Meanwhile, the invention also provides an autonomous road traffic system safety situation dynamic evaluation system.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation management, and specifically to a method and system for dynamically evaluating the safety situation of an autonomous road traffic system. Background Art

[0002] Driven by intelligent enabling technologies, the roles and functions of humans as perceivers, decision-makers, and executors in the transportation system are gradually degenerating and then disappearing. At the same time, the transportation system is gradually evolving towards a direction with less human participation (intelligent) or no human participation (autonomous). The Intelligent Transportation System (ITS) is also transforming towards a cross-border integration development model for different transportation main elements. The Autonomous Road Transportation System (ARTS) has become a new trend in the development of future road traffic systems. The autonomous road traffic system is a road traffic system that uses technologies such as information and communication, artificial intelligence, etc. for interoperability of elements such as vehicles, roads, and management and control, and has the capabilities of self-perception, self-adaptation, self-learning, self-decision-making, self-repair, and self-evolution to achieve autonomous operation. From the perspective of the autonomous level, the structure of the autonomous road traffic system can be divided into several stages, including the existing system, the partially autonomous system, the highly autonomous system, and the ideal system. In the existing road traffic system, people (drivers, managers, operators) constitute the decision-making body of the existing traffic system. In the partially autonomous traffic system, some tasks are automatically completed by machines instead of humans. In the highly autonomous traffic system, the operation is completed by machines, and people have changed from operators to supervisors. In the ideal traffic system, the traffic decision-making system is distributed in the carriers and infrastructure and is integrated into an organic whole through communication.

[0003] In the current autonomous road traffic management system, with the continuous emergence and integration of emerging traffic elements such as intelligent roadside facilities, autonomous driving vehicles, high-precision electronic maps, and intelligent supervision platforms, while bringing convenience to travel, many problems have also emerged.

[0004] First, the emerging traffic elements have a great impact on traditional traffic, and the potential safety risks of road traffic are constantly increasing. The traffic safety situation is a concept that comprehensively describes the road traffic safety situation, which involves many aspects, including the accident incidence rate, traffic violations, road traffic conditions, and the impact of factors such as weather and holidays on traffic safety. The perception accuracy of the traffic safety situation of the current autonomous road traffic system is insufficient, the types of dangerous and emergency states perceived are not many, and the safety control efficiency for the complex giant system of road traffic is low, which is extremely likely to cause the continuous emergence of traffic safety hazards, the deterioration and out-of-control of traffic risks, and the spread to the whole area.

[0005] Second, a large amount of traffic data is not fully utilized, and the depth of "vehicle-road-cloud" data fusion is insufficient. Intelligent vehicles can collect and upload accurate perception data, status data, operation data, location routes, and other information of vehicles. Intelligent roadside units can achieve functions such as vehicle blind spot perception, traffic flow data statistics, and scenario edge computing. The cloud platform integrates map navigation data, vehicle management data, traffic flow control, and other contents. However, there is still the phenomenon of "information silos" in the massive data, and the potential of the data has not been fully explored and utilized.

[0006] Third, there is a lack of an evaluation system for traffic safety indicators, making it difficult to conduct quantitative situation evaluation. There is little research on the real-time monitoring and evaluation of the safety situation of autonomous road traffic systems, and relevant theoretical methods and technical standards have not been formed, making it difficult to scientifically evaluate the safety of the large-scale operation of autonomous road traffic systems with a large number of emerging traffic elements. Summary of the Invention

[0007] In order to solve the problems of insufficient traffic data mining, prominent safety hazards, and lack of an evaluation system for the safety situation in the current autonomous road traffic system, the present invention provides a method for dynamically evaluating the safety situation of an autonomous road traffic system, which proposes an evaluation index system and an evaluation method for classified and hierarchical safety situation assessment, fills the deficiencies of the current autonomous road traffic system, solves the problem of difficult monitoring and evaluation of emerging traffic elements, and provides technical support for the supervision of intelligent networked vehicles and intelligent roads. At the same time, the present application also provides a system for dynamically evaluating the safety situation of an autonomous road traffic system.

[0008] The technical solution of the present invention is as follows: A method for dynamically evaluating the safety situation of an autonomous road traffic system, characterized in that it includes the following steps:

[0009] S1: Construct a hierarchical element library table;

[0010] Based on the impact on the traffic safety of the autonomous road traffic system, slice and classify various elements included in the traffic system to construct a hierarchical element library table;

[0011] S2: Set element indicators for each element in each level of the element library table, and at the same time set corresponding index attributes for each element indicator;

[0012] The jth element indicator in the ith level of the element library table is represented as f ij ;

[0013] where j is the label of the element indicator corresponding to each level, j≥1; i is the level label of the element library table, 1≤i≤N 0 , N 0 is the number of levels of the element library;

[0014] Then, f ij The corresponding index attributes include:

[0015] Element effective enabling attribute en ij , traffic characteristic index attribute va ij , safety situation weight wt ij and safety situation normalized weight

[0016] S3: Establish an index system;

[0017] Classify all road traffic conditions according to traffic characteristics to obtain environmental categories; The environmental categories include: typical intersections, complex sections, and regional road networks;

[0018] For the traffic characteristics under different environmental categories, extract specific scenarios, select the corresponding element indicators and the index attributes corresponding to each element indicator, and construct a classification evaluation index system corresponding to each environmental category; The specific scenarios include: intersections, sections, and road networks;

[0019] The evaluation types for each specific scenario include: real-time evaluation and periodic evaluation; The real-time evaluation is an immediate response evaluation within 5 seconds, for traffic safety events with high danger, suddenness, and persistence; The periodic evaluation is a periodic evaluation of 5 - 20 minutes, for ordinary road traffic conditions;

[0020] The classification evaluation index system is implemented based on an evaluation index table; The content recorded in the evaluation index table includes: for each specific scenario in each evaluation type, specify the corresponding element indicators to be evaluated;

[0021] S4: According to the index system, define the value-taking method of the element effective enabling attribute en ij for each category of traffic scenarios in different evaluation types;

[0022] S5: Define the quantization method of the traffic characteristic index attribute va ij ;

[0023] Define different acquisition and calculation methods to determine the quantitative value dynamic acquisition method of the traffic characteristic index attribute va ij corresponding to each element indicator f ij ;

[0024] S6: Define the quantization method of the safety situation weight wt ij of the element indicator f ij ;

[0025]

[0026] Among them, wt ij is the security situation weight of the j-th element f in the i-th layer ij , is the security situation compensation weight of the i-th layer is the security situation compensation weight of the j-th element in the i-th layer;

[0027] S7: For each element index f to be calculated ij The corresponding security situation weight wt ij is normalized to obtain the normalized security situation weight

[0028]

[0029] Among them, is the normalized security situation weight of the j-th element f in the i-th layer, wt ij is the security situation weight of the j-th element f in the i-th layer, en ij is the security situation weight of the j-th element f in the i-th layer ij , en ij is the effective enabling attribute of the element f ij ;

[0030] S8: Confirm the environmental category corresponding to the traffic scenario to be evaluated this time, and based on the evaluation index table, obtain the specific values of the element indexes that need to be evaluated corresponding to the traffic scenario to be evaluated in each evaluation type;

[0031] S9: Calculate the security situation value g under the specific scenario corresponding to the traffic scenario to be evaluated;

[0032]

[0033] Among them, g is the security situation value of the traffic scenario to be evaluated, en ij is the effective enabling attribute of the element f ij , va ij is the traffic characteristic index attribute of the element f ij ; is the normalized security situation weight of the element f ij ;

[0034] S10: Perform a hierarchical representation of the security situation of the traffic scenario to be evaluated, and complete this dynamic security situation evaluation;

[0035]

[0036] Among them, G is the security situation evaluation result level under this specific scenario.

[0037] Its further feature lies in:

[0038] In step S6, the calculation method of the compensation weight is as follows:

[0039] Construct an element scenario compensation matrix for correcting the element weight

[0040]

[0041] Where, when i = 0, is the inter-level security situation compensation weight vector, N 0 = n is the total number of element library levels, to are the inter-level security situation compensation weights from the first layer to the nth layer of elements respectively, to are the initial level weights from the first layer to the nth layer of elements respectively, to are the security situation weight compensation parameters between the first layer to the nth layer of elements respectively. If there is no compensation, then

[0042] When i > 0, is the element security situation compensation weight vector of the i-th level after compensation, N i is the total number of elements in the i-th layer, to are the security situation compensation weights of each element in the i-th layer respectively, to are the initial security situation weights of each element in the i-th layer respectively, to are the security situation weight compensation parameters of each element in the i-th layer respectively. If there is no compensation, then

[0043] The calculation method of the initial level weight and the initial security situation weight of each element in the i-th layer includes the following steps:

[0044] a1: Construct a criterion judgment matrix;

[0045] The value of the element S ij in the criterion judgment matrix represents the scale value indicating that object i is more important than object j when evaluating the security of the compared objects i and j; the criterion judgment matrix includes: an inter-level security comparison matrix and an intra-level element index security comparison matrix;

[0046] The inter-level security comparison matrix is used to record the comparison values of the importance degrees between adjacent levels of the element library table levels during security evaluation; specifically expressed as:

[0047]

[0048] Among them, S 0,m is the inter-level security comparison matrix, recording the score of the m-th or the m-th person on the inter-level security comparison matrix; represents the score of the importance degree of the n-th layer compared to the first layer in the score of the m-th or the m-th person; 1 ≤ m ≤ M, where M is the total number of scoring people or times;

[0049] The intra-level element index security comparison matrix is used to record the comparison values of the importance degrees of the element indexes included in each of the levels when evaluating security; specifically expressed as:

[0050]

[0051] Among them, S i,m is the intra-level element index security comparison matrix corresponding to the i-th layer, recording the score of the m-th or the m-th person on the intra-level element security comparison matrix of the i-th layer; represents that there are N i element indexes in the i-th layer. In the score of the m-th or the m-th person, the score of the importance degree of the first element index in the i-th layer compared to the N i -th element index;

[0052] a2: Collect the specific values of the inter-level security comparison matrix and obtain the values of the intra-level element index security comparison matrix corresponding to each of the levels;

[0053] a3: Take the average value of the inter-level security comparison matrix and the intra-level element index security comparison matrix corresponding to each of the levels respectively to obtain the inter-level judgment matrix and the intra-level judgment matrices of each layer;

[0054] The inter-level judgment matrices of each level are expressed as:

[0055]

[0056] The intra-level judgment matrix of the i-th layer is expressed as:

[0057]

[0058] a4: Calculate the weight vector between the levels to obtain the initial weights of the security situation between the first layer and the n-th layer;

[0059]

[0060] Among them, is the maximum eigenvalue of;

[0061] Obtained by solving is the maximum eigenvalue The corresponding eigenvector is also the initial weight vector of the security situation between levels; to are respectively the initial weights of the security situation between the elements of the first layer to the nth layer;

[0062] a5: Calculate the eigenvector of the first layer to the nth layer of the within-level judgment matrix of each level to the nth layer to obtain the initial weights of the security situation of each element index in the i-th layer;

[0063]

[0064] Among them, is the maximum eigenvalue;

[0065] The solved to obtain is the maximum eigenvalue i The corresponding eigenvector is also the initial weight vector of the security situation within the level, N to are respectively the initial weights of the security situation of each element index in the i-th layer;

[0066] In step a2, the method for collecting the values of the security comparison matrix between levels and the security comparison matrix of the element indexes within each corresponding level includes:

[0067] Expert scoring method or calculation and collection based on historical data;

[0068] If the expert scoring method is adopted, the number of experts M is greater than or equal to 1;

[0069] If the historical data collection method is adopted, the historical data is divided into M parts according to the specified rules, and M is greater than or equal to 1;

[0070] After step a5 is executed, the following steps need to be executed:

[0071] Perform a consistency comparison on the inter-level judgment matrix and the intra-level judgment matrix;

[0072] b1: Calculate the consistency index;

[0073]

[0074] Among them, when i = 0, CR 0 is the consistency ratio of the inter-level judgment matrix , is The maximum eigenvalue, N 0 is the number of layers of the element library, is the average random consistency index;

[0075] When i > 0, CR i is the consistency ratio of the judgment matrix at each level, and is the maximum eigenvalue, N i is the total number of elements within the i-th layer, is the average random consistency index;

[0076] b2: According to the preset judgment threshold CR, compare CR i with CR;

[0077] If CR i < CR, it is considered that the current matrix meets the consistency requirements;

[0078] Otherwise, it is judged that the current matrix does not meet the consistency requirements;

[0079] The levels of the element library table include: road environment layer, traffic facility layer, traffic control layer, lighting and weather layer, road traffic layer representing traffic events, road traffic layer representing traffic operation, and road monitoring and communication layer;

[0080] The road environment layer includes the basic elements that make up the road;

[0081] The traffic facility layer includes the basic elements that reflect traffic safety facilities;

[0082] The traffic control layer includes the basic elements that reflect traffic rules and control requirements;

[0083] The lighting and weather layer includes the basic elements related to lighting and weather in the scene environment;

[0084] The elements of the road traffic layer representing traffic events include: traffic violation events, traffic accident events, vehicle or system failure events, traffic conflict events, traffic congestion, and other traffic events;

[0085] The elements of the road traffic layer representing traffic operation include: proportion of intelligent vehicles, proportion of large vehicles, violation rate, month-on-month increase rate of violation quantity, month-on-month increase rate of accident quantity, and speed difference between sections;

[0086] The road monitoring and communication layer reflects the basic elements of traffic digital and intelligent facilities and functional service systems;

[0087] The traffic characteristic index attribute va ijQuantification method for the value of

[0088] The static index quantification: For static elements mainly including the road environment layer, traffic facility layer, and road monitoring and communication layer, the change frequency of traffic characteristic index attributes is relatively low; the on-site survey collection and calculation method is used to quantify the attribute indexes of such elements;

[0089] The gradually changing state index quantification: For gradually changing state elements mainly including the traffic control layer and the lighting and weather layer, the change frequency of traffic characteristic index attributes is moderate; the collection and calculation method of the external platform database is used to quantify the attribute indexes of such elements;

[0090] The dynamic index quantification: For dynamic elements and events mainly including the road traffic layer, the change frequency of traffic characteristic index attributes is relatively fast; the intelligent perception collection and calculation method is used to quantify the index attributes of such elements;

[0091] The value range of the scale value is [1, 9]; among them, 1, 3, 4, 5, 7, 9 respectively represent: the former is equally important as the latter, the former is slightly more important than the latter, the former is significantly more important than the latter, the former is very more important than the latter, the former is extremely more important than the latter; 2, 4, 6, 8 are between the adjacent two levels in terms of the comparison of importance.

[0092] An autonomous road traffic system safety situation dynamic evaluation system, characterized in that it includes: a data access module, a data processing module, and a service application module;

[0093] The data access module externally connects various multi-source heterogeneous data related to traffic safety and provides a data basis for all calculations and evaluations;

[0094] The data processing module includes: a data access and conversion unit, a data storage unit, and a data middle platform API;

[0095] The data access and conversion unit is communicatively connected to the data access module, cleans and formats the data, outputs the unified interface and format standard data coding required by the evaluation system, and sends the processed data to the data storage unit; the data storage unit inputs the processed data coding and classifies and stores the data according to different data types;

[0096] The data middle platform API accesses the data in the data storage unit as needed, outputs the data to each functional unit of the service application layer module, and at the same time receives the processing results of each functional unit and saves them to the data storage unit;

[0097] The service application module includes: a traffic element perception unit, a safety situation evaluation unit, and a control strategy generation unit;

[0098] The traffic element perception unit calls the required storage data through the data center API, processes the data, and outputs the traffic safety element perception and motion trajectory results and safety hazard event identification results;

[0099] The safety situation evaluation unit calls the safety factor perception and safety hazard event identification results through the data center API, conducts safety situation evaluation, and outputs specific intersection, road section, and road network scenario situation evaluation results;

[0100] The control strategy generation unit calls the safety hazard event identification results and situation assessment results through the data middle platform API, calls and generates appropriate control strategies from the control strategy library according to different situations, and outputs them to the data middle platform API.

[0101] It is further characterized by:

[0102] The business application module also includes: a business application visualization unit;

[0103] The business application visualization unit calls the generation results of each functional unit in this layer through the data middle platform API, establishes a visual business application interface according to the different business needs of third-party supervision of intelligent connected vehicles on the road, intelligent connected vehicle operation services, and relevant management departments, and displays the safety situation of the autonomous road traffic system in various forms.

[0104] The present invention provides an autonomous road traffic system safety situation dynamic evaluation method. Based on the impact on traffic safety, the elements included in the traffic system are sliced ​​and classified to obtain all levels and element indicators that need to be considered for efficiency during evaluation, and a hierarchical element library table is constructed to ensure that the evaluation results fully cover the elements in the traffic system; all road traffic conditions included in the traffic system are classified according to traffic characteristics to obtain environmental categories, and a classification evaluation index system is constructed based on the element library table to cover the safety situation evaluation of traffic scenes of all types of environmental categories, ensuring that the method can be applied to the evaluation of the safety situation of different scenes in the traffic system; the element index f is defined in the method. ij The corresponding traffic characteristic index attribute va ij The dynamic acquisition method of the quantitative value of the traffic scene to be evaluated makes the final evaluation result calculated based on the quantitative data calculated according to the actual data of the elements at each level, ensuring that the safety situation of the traffic scene to be evaluated can be accurately and targetedly realized dynamically. This method defines the safety situation weight wt for the element index at each level. ij , adjust the traffic characteristic index attribute va through the safety situation weight ijAdjust the importance in the security situation assessment to ensure the accuracy of the final calculation result. Obtain the normalized security situation weight by normalizing the security situation weight. Calculate the security situation weight wt ij When calculating, first evaluate the importance of all element indicators at different levels and within the level for the dynamic evaluation of the security situation. Each finally obtained element indicator f ij The corresponding security situation weight wt ij At the same time, it covers the importance of the level where f ij is located and the importance of f ij itself, ensuring the accuracy of the security situation weight wt ij In this application, when calculating the security situation weight wt ij it is also considered that the impact of some elements on the security situation is non-linear after superposition. By constructing an element scenario compensation matrix, the security situation weight wt ij is corrected and compensated, further ensuring the accuracy of the security situation weight wt ij BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 It is a flowchart of the autonomous road traffic system security situation grading and classification dynamic evaluation method provided by this application;

[0106] Figure 2 It is an example of the slicing classification process of the element library table;

[0107] Figure 3 It is a flowchart of calculating the element weight vector provided by the present invention;

[0108] Figure 4 It is a schematic structural diagram of the security situation evaluation system provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0109] As Figure 1 shown, the present invention includes an autonomous road traffic system security situation dynamic evaluation method, which includes the following steps.

[0110] S1: Construct a hierarchical element library table;

[0111] Based on the impact on the traffic safety of the autonomous road traffic system, slice and classify various elements included in the traffic system to construct a hierarchical element library table.

[0112] In this embodiment, the element library table has 7 layers, specifically including: road environment layer, traffic facility layer, traffic control layer, light and weather layer, road traffic layer representing traffic events, road traffic layer representing traffic operation, and road monitoring and communication layer.

[0113] ​Among them, the road environment layer includes the basic elements that make up the road;

[0114] The traffic facility layer includes the basic elements that reflect traffic safety facilities;

[0115] The traffic control layer includes the basic elements that reflect traffic rules and control requirements;

[0116] The lighting and weather layer includes the basic elements related to lighting and weather in the scene environment;

[0117] The elements of the road traffic layer that characterize traffic events include: traffic violation events, traffic accident events, vehicle or system failure events, traffic conflict events, traffic congestion, and other traffic events;

[0118] The elements of the road traffic layer that characterize traffic operation include: the proportion of intelligent vehicles, the proportion of large vehicles, the violation rate, the month-on-month increase rate of the number of violations, the month-on-month increase rate of the number of accidents, and the speed difference between sections;

[0119] The road monitoring and communication layer reflects the basic elements of traffic digital and intelligent facilities and functional service systems.

[0120] Element indicators can be set according to actual situations in each layer, and each element can include one or more element traffic characteristic index attributes. Through the 7 layers in the element library table of this application, all traffic elements required for the safety situation evaluation of the autonomous road traffic system are comprehensively covered, ensuring the accuracy of the final evaluation result.

[0121] S2: Set element indicators for the elements in each layer of the element library table respectively, and set corresponding index attributes for each element indicator at the same time;

[0122] Express the j-th element indicator in the i-th layer of the element library table as f ij ;

[0123] Among them, j is the label of the element indicator corresponding to each layer, j≥1; i is the layer label of the element library table, 1≤i≤N 0 , N 0 is the number of layers of the element library. In this embodiment, the element library table includes 7 layers, that is, the value of N 0 is 7;

[0124] Then, the index attributes corresponding to f ij include:

[0125] Element effective enabling attribute en ij 、Traffic characteristic index attribute va ij 、Safety situation weight wt ij and safety situation normalized weight

[0126] Element effective enabling attribute en ij : Indicates whether the security element belongs to the evaluation index of the current security situation evaluation method. For example, when conducting real-time evaluation of the intersection scenario, the effective enabling attribute en 11 of the road surface flatness element f 11 has a value of 0; when conducting periodic evaluation of the intersection scenario, the effective enabling attribute en 11 of the road surface flatness element f 11 has a value of 1.

[0127] Element traffic characteristic index attribute va ij : Represents the quantitative assignment of the traffic characteristic index attribute characterized by the security element. The specific method for collecting the quantitative value is dynamically collected according to the type of the element index f ij , using different methods, ensuring that the final security situation evaluation result obtained by this method is a dynamic evaluation result.

[0128] Element security situation weight wt ij : Represents the assigned value after quantifying the importance of the security element in the security situation evaluation. Element security situation normalized weight Represents the element security situation weight wt ij corresponding to the security element, which is the quantified assignment of the importance after normalization processing in the security situation evaluation.

[0129] When evaluating the security situation of an autonomous road traffic system, three key conditions are required: which security elements need to be considered, the attribute status of each security element itself, and the importance of each security element. Among them, the element effective enabling attribute en ij represents the security elements to be included in the evaluation index, the element traffic characteristic index attribute va ij represents the attribute status of each element index f ij itself, and the element security situation weight wt ij and represent the importance of each security element. In this application, through the security situation weight wt ij based on the importance of each element index f ij , the degree of participation of the specific numerical value of the traffic characteristic index attribute va ij in the security situation evaluation is effectively adjusted, ensuring that the final calculated dynamic security situation evaluation result is more accurate.

[0130] The extraction of the element indicators related to the road traffic safety situation evaluation included in each level is analyzed and extracted based on the historical data such as past traffic accidents and traffic incidents. The historical data can be extracted from existing big data platforms to ensure that the finally extracted element indicators meet the requirements of relevant regulations and also conform to the actual traffic safety situation. In this embodiment, the element indicators in the 7 levels are exemplified as shown in Tables 1 to 7 below.

[0131] Table 1 Examples of Element Indicators in the Road Environment Layer

[0132]

[0133] The traffic facility layer reflects the basic elements such as traffic safety facilities. In some embodiments, elements such as road isolation types, construction facilities, and bus stops are set in the traffic facility layer. According to different actual scenario situations and situation evaluation requirements, the element indicators can be further expanded and refined.

[0134] Table 2 Examples of Element Indicators in the Traffic Facility Layer

[0135]

[0136] The traffic control layer reflects the basic elements of traffic rules and control requirements. In some embodiments, elements such as traffic sign settings, traffic marking settings, and traffic signal settings are set in the traffic control layer. According to different actual scenario situations and situation evaluation requirements, the element indicators can be further expanded and refined.

[0137] Table 3 Examples of Element Indicators in the Traffic Control Layer

[0138]

[0139] The lighting and weather layer is the basic element related to lighting and weather in the test scenario environment. In some embodiments, elements such as rainfall, snowfall, fog level, wind speed level, and the impact of meteorology on road skid resistance are set in the lighting and weather layer. According to different actual scenario situations and situation evaluation requirements, the element indicators can be further expanded and refined.

[0140] Table 4 Examples of Element Indicators in the Lighting and Weather Layer

[0141]

[0142] The road traffic layer is the basic element related to lighting and weather in the test scenario environment. In some embodiments, two types of elements, namely traffic events and traffic operations, are set in the road traffic layer. According to the actual scenario situation and the requirements of situation evaluation, the traffic event type elements can be further expanded and refined into traffic violation events, traffic accident events, vehicle or system failure events, traffic conflict events, traffic congestion, other traffic events, etc., and the traffic operation type elements can be further expanded and refined into the proportion of intelligent vehicles, the proportion of large vehicles, the violation rate, the month-on-month increase rate of the number of violations, the month-on-month increase rate of the number of accidents, the speed difference between sections, etc.

[0143] Table 5 Examples of Element Indicators of the Road Traffic Layer (Traffic Events)

[0144]

[0145] Table 6 Examples of Element Indicators of the Road Traffic Layer (Traffic Operations)

[0146]

[0147] The road monitoring and communication layer reflects the basic elements of traffic digital and intelligent facilities and functional service systems. In some embodiments, elements such as vehicle networking communication facilities, high-precision map settings, and video monitoring settings are set in the road monitoring and communication layer. According to different actual scenario situations and the requirements of situation evaluation, the element indicators can be further expanded and refined.

[0148] Table 7 Examples of Element Indicators of the Road Monitoring and Communication Layer

[0149]

[0150] S3: Establish an index system;

[0151] Classify all road traffic conditions according to traffic characteristics to obtain environmental categories. For the traffic characteristics under different environmental categories, extract specific scenarios, select the corresponding element indicators and the indicator attributes corresponding to each element indicator, and construct a classification evaluation index system corresponding to each environmental category. In this embodiment, the environmental categories of the traffic scenarios to be evaluated are set as: typical intersections, complex sections, and regional road networks; corresponding to specific intersection, section, and road network scenarios respectively.

[0152] The evaluation types for each specific scenario include: real-time evaluation and periodic evaluation; real-time evaluation is an immediate response evaluation within 5 seconds, for traffic safety events with high danger, suddenness, and persistence; periodic evaluation is a periodic evaluation within 5 - 20 minutes, for ordinary road traffic conditions. In this application, through real-time evaluation and periodic evaluation of the two types of evaluations, different types of element indicators f ij of the element traffic characteristic index attributes va ijThe quantization value ensures that the safety situation of the traffic scene to be evaluated can be comprehensively and accurately dynamically evaluated.

[0153] The classification evaluation index system is implemented based on the evaluation index table; the content recorded in the evaluation index table includes: for each specific scene in each evaluation type, the corresponding element index to be evaluated is specified.

[0154] The evaluation index table in this embodiment refers to Table 8 below.

[0155] Table 8 Autonomous Road Traffic System Safety Situation Evaluation Index Table

[0156]

[0157] S4: According to the index system, for each category of traffic scenes in different evaluation types, define the corresponding element effective enabling attribute en ij The value-taking method.

[0158] In Table 8, the first column is the 7 levels in this embodiment, and the second column is the corresponding element index for each level; following are three specific scenes under two evaluation types of "real-time evaluation" and "periodic evaluation". At the intersection of the row and column formed by the three specific scenes and the element index, it indicates whether the element index needs to be evaluated under the current "evaluation type".

[0159] In Table 8, "-" indicates not participating in the evaluation, and "●" indicates participating in the evaluation. During actual calculation, determine the specific value of the element effective enabling attribute en ij According to the information in Table 8 through the following formula to assign values to the element effective enabling attribute in the specific scene evaluation:

[0160]

[0161] Specifically, assume that the traffic scene to be evaluated this time is an intersection. During real-time evaluation, refer to the "intersection" column of "real-time evaluation":

[0162] The effective enabling attribute en 11 of the road surface flatness element f 11 is 0; the en 22 value corresponding to the element index "construction facility f 22 " in the traffic facility layer is 1.

[0163] If the traffic scene to be evaluated is an intersection and it is a periodic evaluation, then refer to the "intersection" column of "periodic evaluation", and the en 11 value is 1, and the en 22 value corresponding to the element index "construction facility f 22 " in the traffic facility layer is 1.

[0164] S5: Define the traffic feature index attribute va ij quantification method.

[0165] In this application, different acquisition and calculation methods are defined to determine each element index f ij corresponding dynamic acquisition method of quantitative value for the traffic feature index attribute va ij

[0166] The quantification method of the value of the traffic feature index attribute va ij specifically includes: static index quantification, gradually changing state index quantification, and dynamic index quantification;

[0167] Static index quantification: For static elements mainly including the road environment layer, traffic facility layer, and road monitoring and communication layer, the change frequency of the traffic feature index attribute is relatively low; the acquisition and calculation method of on-site survey is used to quantify the attribute indexes of such elements.

[0168] The static index specifically refers to element indexes that do not change, such as the length of a road section, the position of a bus stop on a road section, the position and number of traffic lights set at an intersection, etc. Specifically in implementation, for the va of the static index of an existing intersection ij quantification data acquisition can be directly connected to databases such as high-precision maps and management platforms to be acquired. If it is a newly built intersection and there is no data in the database, the on-site survey method can be used for acquisition.

[0169] Gradually changing state index quantification: For gradually changing state elements mainly including the traffic control layer and the lighting and weather layer, the change frequency of the traffic feature index attribute is moderate; the acquisition and calculation method of an external platform database is used to quantify the attribute indexes of such elements. For example, for the element indexes of the traffic control layer, the signal data related to traffic lights of the current intersection, road section, and road network to be evaluated can be obtained by connecting to the management platform.

[0170] Dynamic index quantification: For dynamic elements and events mainly including the road traffic layer, the change frequency of the traffic feature index attribute is relatively fast; the acquisition and calculation method of intelligent perception is used to quantify the attribute indexes of such elements.

[0171] The intelligent perception devices include traffic monitoring devices installed on the roadside, operation data uploaded by intelligent vehicles, and real-time data in the vehicle networking. In this application, by connecting to these intelligent perception devices in real time, it is ensured that the va of the element indexes corresponding to the dynamically changing ones can be accurately known ij quantification data, thereby ensuring the realization of dynamic evaluation of the safety situation.

[0172] According to the already quantified indexes, according to the descriptions in Tables 1 to 7 for the traffic feature index attribute va of the element f ij ​​ij Perform assignment, and the formula is as follows:

[0173]

[0174] For example, through the external platform database, the foggy weather level of the hemimetabolous index is collected as visibility of 300m. According to the description in Table 4, the element traffic characteristic index attribute va 43 of the element f 43 is 0.5. The quantified data of each index attribute are saved to the autonomous road traffic system safety situation grading and classification dynamic evaluation system.

[0175] S6: Define the quantization method of the safety situation weight wt ij of the element index f ij ;

[0176]

[0177] Among them, wt ij is the safety situation weight of the jth element f ij in the ith layer, is the safety situation compensation weight of the ith layer, is the safety situation compensation weight of the jth element in the ith

[0178] layer.

[0179] As Figure 3 shown, to calculate the safety situation weight wt ij , it is necessary to calculate and And to calculate and it is also necessary to first calculate the initial safety situation weights between the layers from the 1st layer to the nth layer, and the initial safety situation weights of each element index in the ith layer. The specific calculation methods of these two initial weights include the following steps.

[0180] a1: Construct a criterion judgment matrix;

[0181] The value of the element S ij in the criterion judgment matrix represents the scale value indicating that object i is more important than object j during the safety evaluation of the compared objects i and j; the criterion judgment matrix includes: the inter-layer safety comparison matrix and the intra-layer element index safety comparison matrix;

[0182] The inter-layer safety comparison matrix is used to record the comparison values of the importance degrees between adjacent layers of the element library table layer during the safety evaluation; specifically expressed as:

[0183]

[0184] Among them, S 0,m is the inter - layer security comparison matrix, recording the score of the m - th time or the m - th person for the inter - layer security comparison matrix; represents the score of the importance degree of the n - th layer compared to the first layer in the score of the m - th time or the m - th person; 1 ≤ m ≤ M, where M is the total number of scoring people or times.

[0185] The intra - layer element index security comparison matrix is used to record the comparison values of the importance degrees of the element indexes included within each layer when evaluating security; specifically expressed as:

[0186]

[0187] Among them, S i,m is the intra - layer element index security comparison matrix corresponding to the i - th layer, recording the score of the m - th time or the m - th person for the intra - layer element security comparison matrix of the i - th layer; represents that there are N i element indexes within the i - th layer. In the score of the m - th time or the m - th person, it is the score of the importance degree of the first element index and the N i - th element index within the i - th layer for security.

[0188] a2: Collect the specific values of the inter - layer security comparison matrix, and obtain the values of the intra - layer element index security comparison matrix corresponding to each layer.

[0189] In this application, the value of each element of the criterion judgment matrix refers to the importance scale table, as shown in Table 9 below.

[0190] Table 9 Example of importance scale representation

[0191]

[0192] In this application, the methods for collecting the values of the inter - layer security comparison matrix and the intra - layer element index security comparison matrix corresponding to each layer include: the expert scoring method or calculation and collection based on historical data.

[0193] If the traffic scenario to be evaluated is an existing intersection, road section or road network, with sufficient historical data, or if a similar existing traffic scenario that can be referred to can be found in the same city, historical data can be directly collected from databases such as big data platforms. After annotation, a training data set is formed. Based on the training data set, a historical data matrix is constructed to obtain the models of the inter - layer security comparison matrix and the intra - layer element index security comparison matrix.

[0194] However, if the traffic scenario to be evaluated is a newly built intersection, road section or road network, and there is no available or reference historical data for the traffic scenario to be evaluated, a corresponding safety comparison matrix between levels and a safety comparison matrix of element indicators within levels can be constructed based on the expert scoring method.

[0195] It should be noted that to ensure the accuracy of the values, if the expert scoring method is adopted, the number of experts M is greater than or equal to 1; if the historical data collection method is adopted, M reference models need to be found, or the historical data is divided into M parts according to the specified rules, and M takes a value greater than or equal to 1.

[0196] If the expert scoring method is adopted, taking as an example, there are N i elements in the i-th layer, represents the score given by the m-th expert according to the scale in Table 9 for the safety importance of the first element and the N i -th element in the i-th layer. Taking as an example, it represents the score given by the m-th expert according to the scale in Table 9 for the safety importance of the first layer compared to the seventh layer.

[0197] a3: For the safety comparison matrix between levels and the safety comparison matrix of element indicators within each corresponding level, take the average value respectively to obtain the judgment matrix between levels and the judgment matrix within each layer; the difference between multiple scores is balanced by taking the average value to reduce the impact of extreme scoring values or extreme data on the final result.

[0198] The judgment matrix between each level is expressed as:

[0199]

[0200] The judgment matrix within the i-th layer is expressed as:

[0201]

[0202] a4: Calculate the weight vector between levels to obtain the initial weight of the safety situation between the first layer and the n-th layer;

[0203]

[0204] where, is 's largest eigenvalue;

[0205] The solved for the largest eigenvalue corresponding eigenvector, which is also the initial weight vector of the safety situation between levels; to They are the initial weights of the security postures between the first layer and the nth layer of elements respectively.

[0206] a5: Calculate the eigenvector of the within-layer judgment matrix for each layer from the first layer to the nth layer to obtain the initial weights of the security postures of each element index in the ith layer;

[0207]

[0208] Among them, is the maximum eigenvalue;

[0209] The obtained by solving is the maximum eigenvalue corresponding eigenvector, that is, the initial weight vector of the security posture within the layer, N i is the total number of element indexes within the ith layer, to are the initial weights of the security postures of each element index in the ith layer respectively.

[0210] The specific solution process is implemented based on the existing methods for solving the maximum eigenvalue. Thus, the weight vector between layers and the initial weights of the security postures corresponding to each element index of each layer are obtained.

[0211] However, since the judgment of the initial weights of the security postures in this application is based on the comparison of importance between layers and the comparison of importance between paired element indexes within layers to obtain the judgment result; in actual applications, when the data volume is large enough, there will be logical errors caused by abnormal data, or logical errors will also occur due to human scoring errors when scoring based on expert opinions, which will affect the accuracy of the weight results. Therefore, in order to avoid abnormal phenomena such as "1 is more important than 2, 2 is more important than 3, and 3 is more important than 1" in the judgment matrix, a consistency test with logically consistent front and back is carried out. Calculate the consistency index according to the following formula.

[0212] Compare the consistency between the inter-layer judgment matrix and the intra-layer judgment matrix;

[0213] b1: Calculate the consistency index;

[0214]

[0215] Among them, in this embodiment, there are 7 layers. Then when i = 0, CR 0 is the consistency ratio of the inter-layer judgment matrix , is the maximum eigenvalue, N 0 value is 7, corresponding to 7 layers of elements, is the average random consistency index;

[0216] When i = 1…7, CR i is the judgment matrix of each level to is the consistency ratio of is to is the maximum eigenvalue of i N is the total number of elements in the i-th layer, is the average random consistency index.

[0217] Among them, the value of is obtained based on the consistency test method in the prior art, as shown in Table 10 specifically.

[0218] Table 10 Values of average random consistency index

[0219]

[0220] b2: According to the preset judgment threshold CR, compare CR i with CR;

[0221] If CR i < CR, it is considered that the current matrix meets the consistency requirement;

[0222] Otherwise, it is judged that the current matrix does not meet the consistency requirement.

[0223] Specifically, the value of CR is set according to the actual calculation accuracy requirements. In this embodiment, the value of CR is set to 0.2.

[0224] When CR 0 < 0.2, it is considered that the consistency of the inter-level judgment matrix meets the requirement, otherwise it is necessary to re-screen the expert scoring opinions and construct the inter-level criterion judgment matrix until the consistency test passes; when CR i < 0.15 (i = 1...7), it is considered that the consistency of the judgment matrix within each level meets the requirement, otherwise it is necessary to re-screen the expert scoring opinions and construct the criterion judgment matrix within the i-th level until the consistency test passes.

[0225] After calculating the initial weights of the security postures between the first layer and the nth layer, as well as the initial weights of the security postures of each element index in the ith layer, in practical applications, it is necessary to consider that the impact of some elements on the security posture after superposition is non-linear. For example, when rainfall and a large speed difference occur simultaneously in a road section scenario, the danger level of road traffic security posture will increase exponentially. Therefore, specific element index pairs and compensation parameter values that will generate additional security risks are listed to construct an element scenario compensation matrix, which is used to correct the element weights to ensure that the weights of the element indicators between layers and within layers are more accurate.

[0226] Specifically construct the element scenario compensation matrix:

[0227]

[0228] Among them, when i = 0, is the security posture compensation weight vector between layers, N 0 = 7 is the total number of layers in the element library, to are the security posture compensation weights between the element layers of the first layer to the seventh layer respectively, to are the initial layer weights of the elements of the first layer to the seventh layer respectively, to are the security posture weight compensation parameters between the element layers of the first layer to the seventh layer respectively. If there is no compensation, then

[0229] Regarding the specific values of the security posture weight compensation parameters to are obtained based on historical data. In specific implementation, based on historical data, the number of traffic accidents related to a single-layer index in the system is counted according to the main factors causing accidents to form a data set, and the number of traffic accidents corresponding to the combination of parameter indicators of any two layers is also formed into a data set respectively. Using the layer as the keyword, comparing the two data sets, all two layers with security posture relevance can be collected; according to the difference in the probability of traffic accidents caused by the impact on traffic accidents, after normalization processing, the corresponding compensation parameter values are calculated; finally, a security posture weight compensation parameter table is constructed, with the first layer to the seventh layer element layers set for both the abscissa and the ordinate, and the values corresponding to the abscissa and ordinate are the compensation parameters for the two layers. When there is no historical data available for calculation, it can also be set manually based on experience.

[0230] In this embodiment, the calculated compensation parameter is between [0.1, 0.2]. If there is no compensation, then For example, when the proportion of intelligent vehicles is relatively high and the road monitoring and communication layer is set up completely, intelligent vehicles can make full use of the perception information of roadside intelligent facilities to improve operation safety. Set

[0231] When i = 1...7, is the safety situation compensation weight vector of the i-th level after compensation, N i is the total number of elements in the i-th layer, to are the safety situation compensation weights of each element in the i-th layer respectively, to are the initial safety situation weights of each element in the i-th layer respectively, to are the safety situation weight compensation parameters of each element in the i-th layer respectively. If there is no compensation, then

[0232] Similarly, regarding to The specific values of are also obtained by calculating historical data. According to the main factors causing accidents, the number of traffic accidents corresponding to the single-element index is counted to form a data set, and the number of traffic accidents corresponding to the combination of any two element indexes within any unified layer is counted to form a data set respectively. Using the element index as the keyword and comparing the two data sets, all two element indexes with safety situation correlation within the same layer can be collected; according to the difference in the probability of traffic accidents caused by the impact on traffic accidents, after normalization processing, the corresponding compensation parameter values are calculated; finally, a safety situation weight compensation parameter table is constructed, where the abscissa and ordinate are all the element indexes in the same layer, and the values corresponding to the two element indexes are the compensation parameters corresponding to the two element indexes. When there is no historical data available for calculation, it can also be set manually based on experience.

[0233] In this embodiment, the calculated compensation parameter is between [0.1, 0.2]. If there is no compensation, then For example, when it is rainy, the danger level of traffic safety situation on curves and slopes will increase exponentially, that is, the road section line type va in the road environment layer 12 = 0.75 or 1, and the rainfall va in the lighting and weather layer 41 = 0.75 or 1, set

[0234] S7: Normalize the safety situation weight wt ij corresponding to each element index f to be calculated to obtain the safety situation normalized weight ij Normalize all wt to obtain the safety situation normalized weight ijNormalize to the same numerical range for subsequent calculations;

[0235]

[0236] Among them, is the normalized weight of the security posture of the j-th element f in the i-th layer, wt ij ij is the security posture weight of the j-th element f in the i-th layer, en ij ij is the effective enabling attribute of the element f ij

[0237] S8: Confirm the environmental category corresponding to the traffic scene to be evaluated this time. Based on the evaluation index table, obtain the specific values of the element indicators that need to be participated in the evaluation corresponding to the traffic scene to be evaluated in each evaluation type.

[0238] S9: Calculate the security posture value g under the specific scenario corresponding to the traffic scene to be evaluated;

[0239]

[0240] Among them, g is the security posture value of the traffic scene to be evaluated, en ij is the effective enabling attribute of the element f ij , va ij is the traffic characteristic index attribute of the element f ij is the normalized security posture weight of the element f ij

[0241] S10: Perform hierarchical characterization of the security posture of the traffic scene to be evaluated, and complete the dynamic evaluation of the security posture this time;

[0242]

[0243] Among them, G is the level of the security posture evaluation result under this specific scenario.

[0244] In this embodiment, when the evaluation result is displayed based on the visualization method, the evaluation results are divided into: Level I, the situation is safe, represented by dark green; Level II, the situation is basically safe, represented by light green; Level III, the situation is slightly dangerous, represented by yellow; Level IV, the situation is seriously dangerous, represented by red, etc. 4 levels.

[0245] This application is based on a Figure 4 shown autonomous road traffic system security posture dynamic evaluation system to implement the above security posture dynamic rating method. The autonomous road traffic system security posture dynamic evaluation system in this application includes: a data access module, a data processing module, and a service application module. ​​​​​

[0246] The data access module externally connects to various multi-source heterogeneous data related to traffic safety, providing a data foundation for all calculations and evaluations. The databases externally connected by the data access module include: various databases of the management platform, various sensing devices installed in the traffic system, various databases such as network maps, Internet databases, vehicle networks, and traffic flows.

[0247] The data processing module includes: a data access and conversion unit, a data storage unit, and a data middle platform API.

[0248] The data access and conversion unit is communicatively connected to the data access module, cleans and formats the data, outputs the unified interface and format standard data encoding required by the evaluation system, and classifies and saves it to the data storage unit according to different data types such as unstructured data, structured data, relational data, and electronic map data; the data storage unit inputs the processed data encoding and classifies and saves the data according to different data types.

[0249] The data middle platform API accesses the data in the data storage unit as needed, outputs the data to each functional unit of the business application layer module, and at the same time receives the processing results of each functional unit and saves them to the data storage unit.

[0250] The business application module includes: a traffic element perception unit, a safety situation evaluation unit, a control strategy generation unit, and a business application visualization unit;

[0251] The traffic element perception unit calls the required stored data through the data middle platform API, processes the data, and outputs the traffic safety element perception and movement trajectory results and the safety hazard event identification results;

[0252] The safety situation evaluation unit calls the safety element perception and safety hazard event identification results through the data middle platform API, conducts a safety situation evaluation, and outputs the specific intersection, road section, and road network scenario situation evaluation results;

[0253] The control strategy generation unit calls the safety hazard event identification results and situation evaluation results through the data middle platform API, calls and generates appropriate control strategies from the control strategy library according to different situations, and outputs them to the data middle platform API.

[0254] The business application visualization unit calls the generation results of each functional unit of this layer through the data middle platform API, establishes a visual business application interface for the different business needs of the third-party supervision of intelligent connected vehicles on the road, the operation services of intelligent connected vehicles, and relevant management departments, and displays the safety situation of the autonomous road traffic system in the form of intuitive charts, reports, and dashboards.

[0255] After adopting the technical solution of the present invention, the gap in the safety situation evaluation of autonomous road traffic systems is filled, providing a theoretical basis for promoting the development of autonomous road traffic systems. At the technical level, the safety situation evaluation method and control strategy proposed by the present invention can effectively cope with road traffic risks, ensure traffic safety and efficient operation, prevent accidents from occurring, and reduce losses of life and property. At the application level, the hierarchical and classified dynamic evaluation system proposed by the present invention can be seamlessly connected to the national and provincial vehicle networking pilot area platforms, provide visual application services for safety situation evaluation and supervision services, has engineering practicability, and meets the actual needs of the construction of intelligent connected vehicle operation safety supervision platforms and vehicle-road-cloud platforms in various places.

Claims

1. A method for dynamically evaluating the safety situation of an autonomous road traffic system, characterized in that: It includes the following steps: S1: Construct hierarchical feature library tables; Based on the impact on traffic safety of the autonomous road traffic system, the various elements contained in the traffic system are sliced ​​and classified to construct a hierarchical element library table; S2: setting element indicators for the elements in each level of the element library table respectively, and setting corresponding indicator attributes for each element indicator; The jth element index in the i-th level of the element library table is represented as f ij ; Wherein, j is the label of the element index corresponding to each level, j≥1; i is the level label of the element library table, 1≤i≤N0, N0 is the number of levels of the element library; Then, f ij The corresponding indicator attributes include: Element effective enabling attribute en ij , Traffic characteristic index attribute va ij , security situation weight wt ij and security posture normalized weight S3: Establish an indicator system; All road traffic conditions are classified into environmental categories according to traffic characteristics; the environmental categories include: typical intersections, complex road sections and regional road networks; According to the traffic characteristics under different environmental categories, specific scenes are extracted, the corresponding element indicators and the indicator attributes corresponding to each element indicator are selected, and a classification evaluation indicator system corresponding to each environmental category is constructed; the specific scenes include: intersections, road sections and road networks; The assessment types for each specific scenario include: real-time assessment and periodic assessment; the real-time assessment is an immediate response assessment within 5 seconds, targeting dangerous, sudden and persistent traffic safety incidents; the periodic assessment is a periodic assessment of 5-20 minutes, targeting ordinary road traffic conditions; The classification evaluation index system is implemented based on an evaluation index table; the contents recorded in the evaluation index table include: specifying the corresponding element index to be evaluated for each specific scenario in each evaluation type; S4: Define the traffic scene of each category according to the indicator system and define the corresponding element effective enabling attributes in different evaluation types. ij The value method of ; S5: Define the traffic characteristic index attribute va ij Quantitative methods; Define different collection and calculation methods to determine the index f of each element ij The corresponding traffic characteristic index attribute va ij A quantitative numerical dynamic acquisition method; S6: Define the element index f ij The security situation weight wt ij Quantitative methods; Among them, wt ij is the jth element f in the i-th level ij The security posture weight of is the security situation compensation weight of the i-th layer, is the security situation compensation weight of the jth element in the i-th layer; S7: For each factor index to be calculated f ij The corresponding security situation weight wt ij Normalization is performed to obtain the normalized weight of the security situation in, is the jth element f in the i-th level ij The security situation normalized weight, wt ij is the jth element f in the i-th level ij Security posture weight, en ij For the factor f ij Effective enabling attributes; S8: confirming the environment category corresponding to the traffic scene to be evaluated this time, and obtaining specific values ​​of the element indicators that need to participate in the evaluation corresponding to the traffic scene to be evaluated in each evaluation type based on the evaluation index table; S9: Calculate the safety situation value g under the specific scenario corresponding to the traffic scenario to be evaluated; Among them, g is the safety situation value of the traffic scene to be evaluated, en ij For the factor f ij Effective enable attribute, va ij For the factor f ij Traffic characteristic index attributes, For the factor f ij Normalized security posture weight of S10: Perform hierarchical characterization of the safety situation of the traffic scenario to be evaluated and complete the dynamic evaluation of the safety situation; Among them, G is the level of security situation evaluation results in this specific scenario.

2. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 1, characterized in that: In step S6, the compensation weight is calculated as follows: Constructing the element scene compensation matrix to correct the element weights Among them, when i = 0, is the inter-level security situation compensation weight vector, N0=n is the total number of factor library layers, to are the security situation compensation weights between the 1st to the nth layer elements, to are the initial level weights of the elements from the 1st to the nth level, to They are the security situation weight compensation parameters between the 1st to the nth layer elements. If there is no compensation, When i>0, is the compensation weight vector of the security situation of the i-th level elements after compensation, N i is the total number of elements in the i-th layer, to are the security situation compensation weights of each element in the i-th layer, to are the initial weights of the security situation of each element in the i-th layer, to are the security situation weight compensation parameters of each element in the i-th layer. If there is no compensation, 3. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 2, characterized in that: The method for calculating the initial level weight and the initial weight of the security situation of each element in the i-th layer includes the following steps: a1: Construct the criterion judgment matrix; The element S in the criterion judgment matrix ij The value of represents the scale value of object i being more important than object j in the safety evaluation of the objects i and j involved in the comparison; the criterion judgment matrix includes: an inter-level safety comparison matrix and an intra-level factor index safety comparison matrix; The inter-level security comparison matrix is ​​used to record the comparison values ​​of the importance between adjacent levels of the element library table level during security evaluation; it is specifically expressed as: Among them, S 0,m It is the security comparison matrix between levels, recording the score of the level security comparison matrix given by the mth time or the mth person; Indicates the security importance of the nth layer compared to the first layer in the mth time or the mth person's scoring; 1≤m≤M, M is the total number of scorers or times; The safety comparison matrix of the factor indicators within the said level is used to record the comparison values ​​of the importance of the factor indicators included in each level when evaluating safety; it is specifically expressed as: Among them, S i,m is the security comparison matrix of the factors within the level corresponding to the i-th level, recording the score of the security comparison matrix of the factors within the i-th level given by the m-th time or the m-th person; Indicates that there are N i In the mth time or the mth person's scoring, the first factor index in the i-th layer and the N-th factor index are compared. i Scoring of the safety importance of each factor indicator; a2: collecting specific values ​​of the inter-level security comparison matrix, and obtaining the values ​​of the intra-level factor index security comparison matrix corresponding to each level; a3: taking the average values ​​of the inter-level security comparison matrix and the intra-level factor index security comparison matrix corresponding to each level, respectively, to obtain the inter-level judgment matrix and the intra-level judgment matrix; The judgment matrix between the levels It is expressed as: The intra-level judgment matrix of the i-th layer It is expressed as: a4: Calculate the weight vectors between the layers to obtain the initial weights of the security situation between the first layer to the nth layer; in, for The largest characteristic root of The solution obtained The largest characteristic root The corresponding eigenvector is also the initial weight vector of the security situation between the layers; to They are the initial weights of security posture between the 1st to the nth layer of elements; a5: Calculate the first level of the judgment matrix within each level To level n The characteristic vector of is obtained to obtain the initial weight of the security situation of each factor indicator in the i-th layer; in, for The largest characteristic root of The solution obtained The largest characteristic root The corresponding feature vector, that is, the initial weight vector of the security situation within the level, N i is the total number of factor indicators in the i-th layer, to are the initial weights of the security situation of each factor indicator in the i-th layer.

4. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 3, characterized in that: In step a2, the method for collecting the values ​​of the inter-level security comparison matrix and the intra-level factor index security comparison matrix corresponding to each level includes: Expert scoring method or calculation and collection based on historical data; If the expert scoring method is used, the number of experts M is greater than or equal to 1; If the historical data collection method is adopted, the historical data will be divided into M parts according to the specified rules, and the value of M is greater than or equal to 1.

5. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 3, characterized in that: After step a5 is executed, the following steps need to be performed: Performing consistency comparison on the inter-level judgment matrix and the intra-level judgment matrix; b1: Calculate consistency index; Among them, when i = 0, CR 0 is the inter-level judgment matrix The consistency ratio, for The largest characteristic root, N 0 is the number of feature library layers, is the average random consistency index; When i>0, CR i For each level of judgment matrix The consistency ratio, for The largest characteristic root, N i is the total number of elements in the i-th layer, is the average random consistency index; b2: According to the preset judgment threshold CR, CR i Compare with CR; If CR i <CR, the current matrix is considered to meet the consistency requirements; Otherwise, it is determined that the current matrix does not meet the consistency requirements.

6. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 1, characterized in that: The levels of the element library table include: road environment layer, traffic facility layer, traffic control layer, lighting and weather layer, road traffic layer representing traffic events, road traffic layer representing traffic operation, and road monitoring and communication layer; The road environment layer includes basic elements constituting the road; The traffic facilities layer includes basic elements that reflect traffic safety facilities; The traffic control layer includes basic elements that reflect traffic rules and control requirements; The lighting and weather layer includes basic elements related to lighting and weather in the scene environment; The elements of the road traffic layer that characterize traffic events include: traffic violation events, traffic accident events, vehicle or system failure events, traffic conflict events, traffic congestion and other traffic events; The elements of the road traffic layer that characterize traffic operation include: the proportion of smart cars, the proportion of large vehicles, the violation rate, the month-on-month increase rate of the number of violations, the month-on-month increase rate of the number of accidents, and the speed difference of road sections; The road monitoring and communication layer embodies the basic elements of digital and intelligent transportation facilities and equipment and functional service systems.

7. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 1, characterized in that: The traffic characteristic index attribute va ij The quantification method of the value of includes: static index quantification, gradual index quantification and dynamic index quantification; The static index quantification: for the static elements mainly in the road environment layer, traffic facilities layer, road monitoring and communication layer, the traffic characteristic index attribute change frequency is low; the collection and calculation method of on-site survey is used to quantify the attribute index of such elements; The quantification of the gradual change index: for the gradual change elements mainly in the traffic control layer, light and weather layer, the frequency of attribute change of traffic characteristic index is moderate; the collection and calculation method of the external platform database is used to quantify the attribute index of such elements; The dynamic index quantification: for dynamic elements and events mainly in the road traffic layer, the traffic characteristic index attributes change frequency is relatively fast; the intelligent perception collection and calculation method is used to quantify the index attributes of such elements.

8. The method for dynamically evaluating the safety situation of an autonomous road traffic system according to claim 3, characterized in that: The scale value range is [1,9]; 1, 3, 4, 5, 7, 9 respectively represent: the former is equally important as the latter, the former is slightly more important than the latter, the former is obviously more important than the latter, the former is very important than the latter, and the former is extremely important than the latter; 2, 4, 6, 8 are between the above two adjacent levels of importance.

9. An autonomous road traffic system safety situation dynamic evaluation system, characterized in that: It includes: Data access module, data processing module and business application module; The data access module is connected to various types of multi-source heterogeneous data related to traffic safety, providing a data basis for all calculations and evaluations; The data processing module includes: a data access and conversion unit, a data storage unit and a data middle platform API; The data access and conversion unit is communicatively connected to the data access module, cleans and converts the data format, outputs the unified interface and format standard data code required by the evaluation system, and sends the processed data to the data storage unit; the data storage unit inputs the processed data code and classifies and saves the data according to different data types; The data middle platform API accesses the data in the data storage unit as needed, and outputs the data to each functional unit of the business application layer module, and at the same time receives the processing results of each functional unit and saves them to the data storage unit; The business application module includes: a traffic element perception unit, a safety situation evaluation unit and a control strategy generation unit; The traffic element perception unit calls the required storage data through the data center API, processes the data, and outputs the traffic safety element perception and motion trajectory results and safety hazard event identification results; The safety situation evaluation unit calls the safety factor perception and safety hazard event identification results through the data center API, conducts safety situation evaluation, and outputs specific intersection, road section, and road network scenario situation evaluation results; The control strategy generation unit calls the safety hazard event identification results and situation assessment results through the data middle platform API, calls and generates appropriate control strategies from the control strategy library according to different situations, and outputs them to the data middle platform API.

10. The autonomous road traffic system safety situation dynamic evaluation system according to claim 9, characterized in that: The business application module also includes: a business application visualization unit; The business application visualization unit calls the generation results of each functional unit in this layer through the data middle platform API, establishes a visual business application interface according to the different business needs of third-party supervision of intelligent connected vehicles on the road, intelligent connected vehicle operation services, and relevant management departments, and displays the safety situation of the autonomous road traffic system in various forms.

Images