A safety situation judgment and early warning system and method for urban road traffic

Abstract

The application is a safety situation judgment and early warning system and method for urban road traffic, belonging to the technical field of intelligent transportation system, comprising the following steps: S1, collecting real-time sensor data in the urban road traffic network, and abstracting the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data; S2, determining the topological structure and real-time traffic saturation of the real-time dynamic weighted graph, combining the topological structure and real-time traffic saturation, calculating the structural vulnerability index of each node in the network, and screening out a high-risk node set according to the structural vulnerability index. The application can accurately capture the urgent trend of the evolution of traffic flow from a stable state to a congestion state by detecting the physical precursors of congestion formation such as the normalized density growth rate and speed attenuation degree in the neighborhood.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation systems, specifically to a safety situation assessment and early warning system and method for urban road traffic. Background Technology

[0002] In the field of urban traffic congestion early warning, traditional traffic situation analysis methods mainly rely on static network topology analysis or threshold judgment based on real-time traffic data. These methods often suffer from insufficient network risk assessment and delayed early warning information dissemination, making it difficult to effectively address systemic traffic collapses caused by sudden failures of critical nodes.

[0003] This situation makes it difficult for traffic managers to obtain effective early warnings before congestion occurs, affecting the timeliness and effectiveness of proactive intervention measures. In addition, traditional methods lack the ability to comprehensively assess structural and dynamic risks, and cannot provide scientific decision support for preventing and mitigating the risk of systemic paralysis.

[0004] The aforementioned situation and shortcomings are mainly due to limitations in technical means. Static topology analysis methods ignore the impact of real-time traffic pressure and cannot dynamically identify key nodes that are truly in a high-risk state; while judgment methods based on traffic flow data usually only trigger alarms after congestion has already occurred, resulting in insufficient advance warning. As a result, when local traffic conditions are about to deteriorate and may trigger widespread congestion, managers cannot quickly obtain forward-looking and accurate risk prediction information, thus missing the best opportunity to take preventive measures.

[0005] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a safety situation assessment and early warning system and method for urban road traffic, so as to solve the problems mentioned in the background art.

[0007] The technical solution of the present invention includes the following steps:

[0008] S1. Collect real-time sensor data in the urban road traffic network, and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data;

[0009] S2. Determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combine the topology and real-time traffic saturation to calculate the structural vulnerability index of each node in the network, and select the set of high-risk nodes based on the structural vulnerability index.

[0010] S3. For each node in the high-risk node cluster, determine the normalized density growth rate and velocity decay rate in the neighborhood, and calculate and generate the phase transition proximity index based on the normalized density growth rate and velocity decay rate.

[0011] S4. Determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node.

[0012] S5. Determine the overall risk score against the preset risk threshold:

[0013] If the overall risk score of any high-risk node exceeds the risk threshold, an early warning mechanism will be triggered to generate an early warning message for that high-risk node.

[0014] If the overall risk score of all high-risk nodes does not exceed the risk threshold, the current monitoring status will be maintained.

[0015] Preferably, S1 specifically includes:

[0016] S11. Abstract the intersections in the urban road traffic network into a set of nodes, and the road segments into a set of edges;

[0017] S12. Collect sensor data deployed in the road network to obtain the real-time traffic capacity or travel time of the road segment;

[0018] S13. Based on real-time traffic capacity or travel time, construct a weight matrix that changes over time, and combine the node set and edge set to generate a real-time dynamic weighted graph.

[0019] Preferably, S2 specifically includes:

[0020] S21. Calculate the normalized network centrality of each node in the network to characterize its topological importance.

[0021] S22. Collect the real-time traffic saturation of each node;

[0022] S23. Combine normalized network centrality with real-time traffic saturation to calculate and generate a structural vulnerability index;

[0023] S24. Sort the structural vulnerability indices of all nodes and identify the nodes with the highest indexes as the high-risk node set.

[0024] Preferably, S3 specifically includes:

[0025] S31. Obtain the real-time average traffic density and real-time average speed in the neighborhood of each high-risk node.

[0026] S32. Calculate the normalized density growth rate based on the real-time average traffic density, the preset critical traffic density, and the preset characteristic time scale.

[0027] S33. Calculate the velocity decay rate based on the real-time average velocity and the preset free-flow velocity;

[0028] S34. Multiply the normalized density growth rate by the velocity decay rate to generate a phase transition proximity index.

[0029] Preferably, the dynamic coupling coefficient is determined as follows:

[0030] If the average saturation of the entire road network is lower than the preset saturation threshold, the value of the dynamic coupling coefficient will be reduced to increase the weight of the structural vulnerability index in the dynamic weighted fusion.

[0031] If the average saturation of the entire road network is not lower than the preset saturation threshold, the value of the dynamic coupling coefficient is increased to improve the weight of the phase transition convergence index in the dynamic weighted fusion.

[0032] Preferably, S5 also includes:

[0033] Based on the early warning information, proactive intervention strategy suggestions are output, which include at least one of the following: signal timing adjustment strategy, traffic guidance strategy, and emergency resource dispatch strategy.

[0034] Preferably, the preset risk threshold is dynamically optimized by statistically learning from successful and failed warning cases in the system's operating history.

[0035] A safety situation assessment and early warning system for urban road traffic includes:

[0036] The real-time network modeling module is used to collect real-time sensor data in the urban road traffic network and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data.

[0037] The structural risk calculation module is used to determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combining the topology and real-time traffic saturation, it calculates the structural vulnerability index of each node in the network and selects a set of high-risk nodes based on the structural vulnerability index.

[0038] The phase transition precursor detection module is used to determine the normalized density growth rate and velocity decay rate in the neighborhood of each node in the high-risk node cluster, and to calculate and generate a phase transition proximity index based on the normalized density growth rate and velocity decay rate.

[0039] The comprehensive risk assessment module is used to determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and to dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node.

[0040] The early warning and decision generation module is used to compare the comprehensive risk score with the preset risk threshold. If the comprehensive risk score of any high-risk node exceeds the risk threshold, the early warning mechanism is triggered and an early warning message is generated. If the comprehensive risk scores of all high-risk nodes do not exceed the risk threshold, the control system maintains the current monitoring status.

[0041] This invention provides an improved safety situation assessment and early warning system and method for urban road traffic, which has the following improvements and advantages compared with the prior art:

[0042] 1. By deeply coupling network topology analysis with a traffic flow phase transition physical model, this approach overcomes the limitations of traditional technologies that rely on static topology or delayed traffic data for judgment, achieving a fundamental shift from post-event response to traffic congestion to pre-event prediction of systemic collapse risks. It abstracts complex urban traffic networks into real-time dynamic weighted graphs, laying a model foundation for accurately reflecting the real-time operating status of the road network. The core advancement lies in constructing a dual-dimensional risk identification framework: on the one hand, by calculating a structural vulnerability index that integrates topological importance and real-time traffic saturation, it can proactively identify potential high-risk nodes that are crucial to the stability of the entire network under the dual pressures of static structure and dynamic load; on the other hand, for these identified high-risk nodes, a phase transition proximity index is introduced. By detecting physical precursors to congestion such as the normalized density growth rate and speed decay rate within their neighborhood, it accurately captures the urgent trend of traffic flow evolving from a stable state to a congested state.

[0043] 2. An adaptive risk assessment mechanism was established. By determining a dynamic coupling coefficient associated with the average saturation of the entire road network, the aforementioned structural vulnerability index and phase transition proximity index were dynamically weighted and fused. This mechanism enables the early warning system to focus more on inherent structural risks determined by the network topology during off-peak traffic periods, while automatically shifting its focus to immediate phase transition risks triggered by local dynamic disturbances during peak traffic periods. This intelligent shift in risk assessment focus greatly improves the robustness and accuracy of the early warning system under different traffic conditions.

[0044] 3. This invention goes beyond early warning; it can also output proactive intervention strategy suggestions, including signal timing adjustment, traffic guidance, and emergency resource dispatch, after determining that the comprehensive risk score exceeds the risk threshold. This transforms risk prediction into actionable management actions, giving traffic managers valuable time to respond. Simultaneously, by dynamically optimizing the risk threshold through statistical learning from successful and failed early warning cases, the system is endowed with self-learning and continuous evolution capabilities, ensuring its long-term reliability and optimal performance. This provides urban traffic management with a powerful, reliable, and intelligent systemic risk control platform. Attached Figure Description

[0045] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0046] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0048] Example 1

[0049] Please see Figure 1 This invention provides a method for assessing and issuing early warnings of safety conditions in urban road traffic, comprising the following steps:

[0050] S1. Collect real-time sensor data in the urban road traffic network, and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data;

[0051] S2. Determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combine the topology and real-time traffic saturation to calculate the structural vulnerability index of each node in the network, and select the set of high-risk nodes based on the structural vulnerability index.

[0052] S3. For each node in the high-risk node cluster, determine the normalized density growth rate and velocity decay rate in the neighborhood, and calculate and generate the phase transition proximity index based on the normalized density growth rate and velocity decay rate.

[0053] S4. Determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node.

[0054] S5. Determine the overall risk score against the preset risk threshold:

[0055] If the overall risk score of any high-risk node exceeds the risk threshold, an early warning mechanism will be triggered to generate an early warning message for that high-risk node.

[0056] If the overall risk score of all high-risk nodes does not exceed the risk threshold, the current monitoring status will be maintained.

[0057] This embodiment discloses a method for assessing and warning the safety situation of urban road traffic. The method aims to break through the assumption of static stability of traffic network topology and solve the problem of delayed warning when dealing with sudden failures of key nodes by deeply coupling network topology analysis and traffic flow phase transition physical model, so as to achieve early prediction of systemic collapse risk.

[0058] The implementation steps of this method are as follows:

[0059] The implementation of this method begins with step S1, which involves collecting real-time sensor data from the urban road traffic network and abstracting the complex urban road traffic network into a real-time dynamic weighted graph based on this data. The purpose of this step is to construct a mathematical model that can accurately reflect the real-time operating status of the road network, providing a foundation for subsequent risk analysis.

[0060] Based on the constructed real-time dynamic weighted graph, step S2 is executed to determine its topology and the real-time traffic saturation of the nodes. The topology refers to the physical connection between intersections and road segments, while the real-time traffic saturation refers to the ratio of traffic flow through a node to its design capacity within a specific time period. By combining these two aspects of information, the structural vulnerability index of each node in the network is calculated. This index is used to identify nodes that are critical to the stability of the entire network before failure occurs. Based on the calculated structural vulnerability index values, all nodes are sorted, and a subset of nodes with the highest index rankings are selected to form a high-risk node set for focused monitoring.

[0061] For each node in the high-risk node set identified in the previous step, step S3 is then executed to continuously monitor the traffic flow dynamics of its neighboring area. This step determines the normalized density growth rate and speed decay rate within its neighborhood. The normalized density growth rate characterizes the relative severity of traffic congestion, while the speed decay rate reflects the decline in traffic efficiency. Based on these two indicators, a phase transition proximity index is calculated. The purpose of this index is to quantify the urgency of the local traffic flow evolving from a steady state to a congested state, and to capture dynamic precursors of traffic collapse.

[0062] To achieve a comprehensive risk assessment, step S4 is executed to determine the dynamic coupling coefficient associated with the average saturation of the entire road network. The average saturation of the entire road network is an indicator that measures the macroscopic load level of the entire transportation system. The role of this dynamic coupling coefficient is to dynamically adjust the proportion of static structural risk and dynamic evolution risk in the final score under different traffic conditions, such as off-peak and peak periods. Based on this dynamic coupling coefficient, the structural vulnerability index and phase transition convergence index calculated in the previous steps are dynamically weighted and fused to generate a comprehensive risk score for each high-risk node.

[0063] The final step of the method is step S5, which compares the comprehensive risk score of each high-risk node with a preset risk threshold in real time. If the comprehensive risk score of any high-risk node exceeds the risk threshold, it indicates that the risk of the node causing a systemic collapse has reached a critical level. At this time, the early warning mechanism is triggered, and the system generates an early warning message for the high-risk node. If the comprehensive risk scores of all high-risk nodes do not exceed the risk threshold, the system maintains the current continuous monitoring state and repeats the above steps.

[0064] This embodiment constructs a complete technical closed loop from static structural vulnerability identification to dynamic phase transition precursor detection, and then to comprehensive risk dynamic assessment. It can provide early and accurate warnings of key node risks that may lead to the systemic collapse of urban road traffic. It overcomes the shortcomings of traditional methods that rely solely on static topology or traffic data that lags behind events for judgment, and realizes the transformation from post-event response to pre-event prediction, providing valuable time windows for urban traffic managers to take proactive intervention measures.

[0065] S1 specifically includes:

[0066] S11. Abstract the intersections in the urban road traffic network into a set of nodes, and the road segments into a set of edges;

[0067] S12. Collect sensor data deployed in the road network to obtain the real-time traffic capacity or travel time of the road segment;

[0068] S13. Based on real-time traffic capacity or travel time, construct a weight matrix that changes over time, and combine the node set and edge set to generate a real-time dynamic weighted graph.

[0069] In this embodiment, step S1, which is the process of abstracting the urban road traffic network into a real-time dynamic weighted graph, can be further refined;

[0070] Step S11: Abstracting network elements; The purpose of constructing the node set V is to digitize the intersections in the physical world. In this embodiment, each intersection in the urban road traffic network is abstracted as a node in graph theory; The purpose of constructing the edge set E is to represent the travel path between intersections. In this embodiment, the physical road segment connecting two intersections is abstracted as an edge in graph theory.

[0071] Step S12: Real-time status data is collected. The purpose of collecting sensor data is to obtain the dynamic operating parameters of the road network. In this embodiment, data from various sensors deployed in the road network, such as geomagnetic coils, microwave detectors, or floating car data, are collected to obtain the traffic capacity or travel time of each road segment in real time. Traffic capacity refers to the maximum traffic flow that a road segment can pass through per unit time, while travel time refers to the average time required for a vehicle to pass through the road segment.

[0072] Step S13 involves generating a dynamic graph and constructing a weight matrix Wt to ​​quantify the real-time traffic status of road segments. In this embodiment, based on the real-time traffic capacity or travel time of the road segments obtained in step S12, a time-varying dynamic graph is constructed. Dynamically changing weight matrix ; Elements of the matrix Characterizes the connection nodes and nodes The section of road in The current traffic status; finally, combined with the node set generated in step S11. With edge set Together, they generate a real-time dynamic weighted graph. ;

[0073] Through the above-described specific modeling steps, it is ensured that the constructed real-time dynamic weighted graph can reflect the physical topology and instantaneous operating status of the urban transportation network with high fidelity and precision; this refined model lays a solid data foundation for the accuracy and reliability of subsequent structural vulnerability analysis and risk assessment.

[0074] S2 specifically includes:

[0075] S21. Calculate the normalized network centrality of each node in the network to characterize its topological importance.

[0076] S22. Collect the real-time traffic saturation of each node;

[0077] S23. Combine normalized network centrality with real-time traffic saturation to calculate and generate a structural vulnerability index;

[0078] S24. Sort the structural vulnerability indices of all nodes and identify the nodes with the highest index ranking as the high-risk node set.

[0079] In this embodiment, step S2, namely the process of calculating the structural vulnerability index and screening the high-risk node set, can be further refined;

[0080] Step S21: Calculate the topological importance of nodes; normalize network centrality. The purpose of this calculation is to quantify the importance of nodes as transportation hubs from the perspective of network topology. In this embodiment, the betweenness centrality algorithm, which is well-known in the art, can be used to calculate the shortest path between all pairs of nodes. The frequency of is determined, reflecting its centrality; the calculation result is normalized to ensure its value falls within a certain range. interval;

[0081] Step S22: Collect the real-time load status of the nodes; real-time traffic saturation. The purpose of this data collection is to quantify the current traffic pressure at the node; in this embodiment, Defined as through nodes The ratio of real-time traffic density to design critical density for all connecting road segments is limited to a range within which... ;

[0082] Step S23 introduces a structural vulnerability index to more comprehensively assess the potential risks of nodes by combining their static topological importance with their real-time dynamic load. The computational model is defined as follows:

[0083]

[0084] in, For nodes The dimensionless structural fragility index; This is a dimensionless system calibration coefficient, whose value is determined and adjusted through retrospective analysis of historical network outage events. To maximize the occurrence of the incident before it happens The degree to which the score distinguishes itself from other nodes; Normalized network centrality of node i quantifies the importance of a node as a transportation hub from the perspective of network topology. The natural logarithm function acts as a function of saturation. As it approaches 1, this term tends to infinity, thus non-linearly amplifying the vulnerability score of high-load critical nodes; The real-time traffic saturation of node i is defined as the ratio of the real-time traffic density of all connecting road segments passing through that node to the design critical density.

[0085] To achieve this calibration, the following optimization objective function can be constructed, assuming that historical data analysis has identified... Each network outage event corresponds to a set of offending nodes. and a set of non-incident nodes For the first For each event, within a time window preceding the event, the average vulnerability index of the offending node was [value missing]. The average vulnerability index of non-incident nodes is The goal of optimization is to find a The objective function aims to maximize the sum of the score differences between the offending and non-offending nodes across all historical events; it can be defined as follows:

[0086]

[0087] in, The optimal calibration coefficient with clear statistical significance; A mathematical function whose goal is to find a... The value that maximizes the subsequent objective function value; Index of historical network outage events; The total number of historical network outages identified; The average vulnerability index of the node responsible for the incident in the j-th event within a time window preceding the incident. The average vulnerability index of non-initiating nodes in the j-th event within a time window preceding the incident. The system calibration coefficients are the variables that this optimization function needs to solve for.

[0088] This optimization problem can be solved using the gradient ascent method or other numerical optimization algorithms to obtain an optimal calibration coefficient with clear statistical significance. ;

[0089] The node obtained through step S21 Normalized network centrality; The nodes obtained through step S22 Real-time traffic saturation; for several items The introduction of this is based on traffic flow theory, and its function is to adjust the saturation level. As it approaches its limit of 1, the term tends to infinity, thus non-linearly amplifying the vulnerability score of high-load critical nodes, making the index particularly sensitive to high-risk nodes during peak periods.

[0090] Step S24: Screen high-risk nodes; assess the structural vulnerability index of all nodes in the network. The nodes are sorted in descending order, and those with the highest index rankings, such as the top 5%, are identified as the high-risk node set. This aims to concentrate limited monitoring and computing resources on the critical nodes most likely to cause systemic problems.

[0091] This specific implementation introduces a structural vulnerability index that integrates static topology and dynamic load, enabling accurate and proactive identification of critical network nodes. Compared to traditional methods that rely solely on static topological centrality, this method is more sensitive in capturing critical nodes that become vulnerable due to surges in real-time traffic pressure, thus improving the dynamism and accuracy of risk identification.

[0092] S3 specifically includes:

[0093] S31. Obtain the real-time average traffic density and real-time average speed in the neighborhood of each high-risk node.

[0094] S32. Calculate the normalized density growth rate based on the real-time average traffic density, the preset critical traffic density, and the preset characteristic time scale.

[0095] S33. Calculate the velocity decay rate based on the real-time average velocity and the preset free-flow velocity;

[0096] S34. Multiply the normalized density growth rate by the velocity decay rate to generate a phase transition proximity index.

[0097] In this embodiment, step S3, namely the process of calculating the phase transition convergence index for high-risk nodes, can be further refined;

[0098] Step S31: Obtain neighborhood traffic flow parameters; for each node in the high-risk node set. Obtain the real-time average traffic density in its neighborhood. With real-time average speed These two parameters are obtained by real-time collection of data from sensors deployed on road sections near the node and by calculating their average value.

[0099] Step S32: Calculate the normalized density growth rate. The purpose of calculating the normalized density growth rate is to quantify the relative severity of traffic congestion per unit time. It is calculated as follows: First, obtain the time change rate of the real-time average traffic density. Then, it is normalized by combining two preset parameters; the preset critical traffic density It is an inherent physical property of the target road network, representing the critical density value at which traffic flow transitions from free flow to congested flow; the preset characteristic time scale This represents the typical timeframe from the formation to the worsening of traffic congestion. For example, it can be the average time taken for traffic density to increase from the upper limit of normal fluctuation to the critical density when congestion occurs in the area historically. Both of the above preset parameters can be obtained through statistical analysis of historical traffic data. The normalized density growth rate is ultimately calculated as... The result is a dimensionless value;

[0100] Step S33, calculate the velocity decay rate; the purpose of calculating the velocity decay rate is to quantify the degree of decrease in real-time velocity relative to the ideal state; through The calculation yielded the following result: where the preset free-flow velocity... It is another inherent physical property of the road network, representing the maximum design speed under undisturbed conditions, and can be obtained through statistical analysis of historical data; the result is also a dimensionless value.

[0101] Step S34: Generate a phase transition proximity index; multiply the normalized density growth rate by the velocity decay rate and introduce an adjustable weighting coefficient to generate the phase transition proximity index. This indicator is used to capture dynamic signs preceding traffic collapses, and the calculation model has been revised as follows:

[0102]

[0103] in, For nodes A dimensionless phase transition proximity index for the neighborhood; It is an adjustable dimensionless weighting coefficient, the value of which can be optimized and adjusted in a traffic simulation environment for different types of failure scenarios in order to seek the most sensitive precursor detection effect; The preset characteristic time scale represents the typical time from the formation to the deterioration of traffic congestion; The preset critical traffic density represents the critical density value at which traffic flow transitions from free flow to congested flow. The rate of change of real-time average traffic density over time is used to quantify the relative severity of traffic congestion per unit time. : The real-time average traffic density within the neighborhood of node i; : The real-time average velocity within the neighborhood of node i; The preset free-flow velocity represents the maximum design velocity under undisturbed conditions.

[0104] When calculating the phase transition proximity index, only the case of density growth is considered, that is, the time change rate of the real-time average traffic density. When the time is right, the value is 0;

[0105] Optimization and adjustment steps may include:

[0106] Build simulation models: Use industry-recognized traffic simulation software, such as SUMO and VISSIM, to build simulation models that match the target road network topology, traffic signal control, and traffic demand.

[0107] Define failure scenarios: Design a set of typical traffic failure scenarios, for example, simulate a single lane closure event lasting 15 minutes on the upstream section of a high-risk node, resulting in a 50% decrease in traffic capacity;

[0108] Perform iterative simulation: Set a A reasonable search range for values, for example Take a series of discrete values ​​within this interval, for each The value was used to run failure scenario simulations multiple times, and the phase transition convergence index was recorded. The temporal changes;

[0109] Establish evaluation indicators: Define the criteria for successful and unsuccessful early warnings; for example, if If a preset test threshold is exceeded before a simulated traffic congestion occurs, for example, when the average speed on a road segment falls below 30% of the free-flow speed, it is recorded as a true positive. If the threshold is exceeded in a normal simulation where no congestion occurs, it is recorded as a false positive.

[0110] Determining the optimal value: by plotting different Calculate the area under the receiver operating characteristic curve corresponding to the value; select the curve that maximizes the AUC value. As the optimal weighting coefficient, this value represents the best balance between sensitivity and specificity;

[0111] , and These are preset parameters; and These are parameters acquired in real time; the model's construction enables... Its value only increases significantly when rapid density increase and significant velocity decrease occur simultaneously.

[0112] This specific implementation method is based on the phase transition theory in traffic flow physics, and constructs an index that can accurately quantify the precursors of congestion. By ensuring the dimensionlessness of each component of the model and integrating the changes in both density and speed, it can effectively filter out normal traffic fluctuation noise, accurately capture the key precursors of the phase transition from steady flow to congested flow, and greatly improve the timeliness and reliability of the warning.

[0113] Example 2

[0114] The dynamic coupling coefficient is determined as follows:

[0115] If the average saturation of the entire road network is lower than the preset saturation threshold, the value of the dynamic coupling coefficient will be reduced to increase the weight of the structural vulnerability index in the dynamic weighted fusion.

[0116] If the average saturation of the entire road network is not lower than the preset saturation threshold, the value of the dynamic coupling coefficient will be increased to improve the weight of the phase transition proximity index in the dynamic weighted fusion.

[0117] In this embodiment, the dynamic coupling coefficient in step S4 The determination method is specified in order to enable the comprehensive risk assessment to adapt to the macro traffic conditions of the entire road network.

[0118] The value was designed to be similar to the average saturation of the entire road network. Relatedly, during off-peak traffic hours, the road network operates smoothly overall. The systemic risk is relatively low, and is largely determined by static factors such as network topology; at this point, if the average saturation of the entire network is low... Below the preset saturation threshold This reduces the dynamic coupling coefficient. The value of this makes the structural vulnerability index, representing potential structural risk, more accurate during dynamic weighted fusion. Occupy a higher weight;

[0119] Conversely, during peak traffic hours, the road network as a whole tends to be saturated. The higher the average saturation level, the more easily systemic risks are triggered by local dynamic disturbances; at this point, if the average saturation level of the entire road network is high... Not lower than the preset saturation threshold This increases the dynamic coupling coefficient. The value of this; makes the phase transition approximation index, which represents real-time dynamic perturbations, more suitable. It dominates in dynamic weighted fusion;

[0120] In this embodiment, to achieve a smooth transition of weights, The specific function form can be preset as an S-type function:

[0121]

[0122] in, For time The changing dynamic coupling coefficient; The average saturation level of the entire road network; This refers to the dimensionless weighted switching sensitivity parameter. It is the saturation threshold at which the network enters a critical state; parameter and All of these can be determined through long-term statistical analysis of historical urban traffic data; The base of the natural logarithm is a mathematical constant, and the formula is an sigmoid function used to achieve a smooth transition of weights.

[0123] Statistical analysis methods may include:

[0124] Determine the critical saturation threshold Collect the historical hourly average saturation of the road network over at least one year. Data. Clustering analysis algorithms, such as K-means clustering, are used to divide these data points into two or more state clusters: smooth and congested. The saturation value defined as the boundary between two main state clusters can be determined as the critical saturation threshold. For example, if the clustering results show that when If the road network condition deteriorates significantly, then a setting can be made. ;

[0125] Determine the weight switching sensitivity parameters In determining Subsequently, historical traffic events can be further analyzed. Events primarily caused by topological issues, such as critical hub failures, and events primarily caused by local dynamic disturbances, such as sudden accidents, are categorized. These categorized data are used as the dependent variable, with saturation difference as the criterion. Using these as independent variables, a logistic regression model is fitted. The correlation coefficients are extracted from the fitted model. The coefficients related to the terms can be used as parameters for weight switching sensitivity. The value of; a larger The value implies that the focus of risk assessment will change with the saturation level. Cross And a very rapid switch occurs;

[0126] Comprehensive Risk Score The calculation model is as follows:

[0127]

[0128] in, For nodes The final comprehensive risk score; and As defined above; For dynamic coupling coefficients;

[0129] By introducing a dynamic coupling coefficient linked to the overall network status, this method achieves adaptive switching of the risk assessment focus. This design enables the early warning system to pay more attention to inherent hidden dangers in the network structure during off-peak hours, while responding more sensitively to real-time traffic flow deterioration precursors during peak hours. This allows the system to provide the most relevant risk assessment under various traffic conditions, significantly improving the intelligence and robustness of the early warning system.

[0130] S5 also includes:

[0131] Based on the early warning information, proactive intervention strategy suggestions are output, which include at least one of signal timing adjustment strategy, traffic guidance strategy and emergency resource dispatch strategy.

[0132] In this embodiment, after triggering the early warning mechanism, step S5 further expands its function; in addition to generating early warning information, the system is also designed to output proactive intervention strategy suggestions; the purpose of this step is to transform the early warning information into executable management actions, forming a closed loop from risk identification to risk handling.

[0133] When the comprehensive risk score of any high-risk node Exceeding the risk threshold At the same time, while generating early warning information for the node, such as node ID, location, risk score and main risk drivers, the system will automatically generate and output one or more proactive intervention strategy suggestions based on a preset rule base or decision model.

[0134] These strategy recommendations include at least one of the following types:

[0135] Signal timing adjustment strategies: For example, it is recommended to extend the green light time at intersections downstream of high-risk nodes, or to activate green wave coordination control schemes in specific areas to accelerate the dispersal of backlogged traffic.

[0136] Traffic guidance strategies: For example, by using roadside variable message signs, mobile navigation applications, and other channels, detour guidance information can be issued to drivers who are about to enter high-risk areas to reduce traffic flow at the nodes;

[0137] Emergency resource dispatch strategies: For example, suggest to the traffic management center that tow trucks be dispatched in advance and traffic police be assigned to stand by near high-risk nodes so that a rapid response can be achieved in the event of congestion or accidents;

[0138] This implementation upgrades the system's function from a passive alarm system to a proactive decision support system. By providing specific and actionable intervention suggestions, it can help traffic managers take effective measures at the initial stage of risk occurrence, thereby resolving congestion more efficiently, avoiding systemic collapse, and significantly enhancing the initiative and effectiveness of urban traffic management.

[0139] The preset risk threshold is dynamically optimized by statistically learning from successful and failed warning cases in the system's operating history.

[0140] In this embodiment, the preset risk threshold used in step S5 The parameters are designed to be dynamically optimized rather than fixed static values; the aim is to enable the decision threshold of the early warning system to continuously evolve through self-learning in order to adapt to long-term changes in the traffic environment and maintain optimal early warning performance.

[0141] This dynamic optimization process is achieved through statistical learning of successful and failed early warning cases in the system's operating history. A successful early warning case refers to a case where the system issues an early warning and, within a subsequent predetermined time window, the node does indeed experience severe congestion or paralysis. Failed early warning cases include both missed and false alarms.

[0142] The system periodically collects this case data and uses machine learning methods, such as receiver operating characteristic curve analysis, to evaluate the true positive rate, sensitivity, false positive rate, and 1-specificity under different threshold settings. The optimization goal is to find a threshold that achieves the highest early warning accuracy within an acceptable false alarm rate for traffic managers. This optimized new threshold will replace the old one and be applied to the next stage of risk assessment.

[0143] By introducing a dynamically optimized risk threshold, this method enables the early warning system to have adaptive learning capabilities. This ensures that the system's early warning performance will not deteriorate due to seasonal changes in traffic patterns, reconstruction of urban road networks, or other factors, and can maintain its optimal operating point in the long term. This achieves the best balance between sensitivity and reliability, providing managers with continuous and reliable decision support.

[0144] Example 3

[0145] A safety situation assessment and early warning system for urban road traffic includes:

[0146] The real-time network modeling module is used to collect real-time sensor data in the urban road traffic network and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data.

[0147] The structural risk calculation module is used to determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combining the topology and real-time traffic saturation, it calculates the structural vulnerability index of each node in the network and selects a set of high-risk nodes based on the structural vulnerability index.

[0148] The phase transition precursor detection module is used to determine the normalized density growth rate and velocity decay rate in the neighborhood of each node in the high-risk node cluster, and to calculate and generate a phase transition proximity index based on the normalized density growth rate and velocity decay rate.

[0149] The comprehensive risk assessment module is used to determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and to dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node.

[0150] The early warning and decision generation module is used to compare the comprehensive risk score with the preset risk threshold. If the comprehensive risk score of any high-risk node exceeds the risk threshold, the early warning mechanism is triggered and an early warning message is generated. If the comprehensive risk scores of all high-risk nodes do not exceed the risk threshold, the control system maintains the current monitoring status.

[0151] This embodiment also discloses a safety situation assessment and early warning system for urban road traffic. The system is built based on the aforementioned method, and its internal functional modules correspond one-to-one with the steps of the method.

[0152] The system includes:

[0153] The real-time network modeling module is responsible for executing step S1 in the method. This module is used to collect real-time data from various sensors in the urban road traffic network through the interface, and abstract intersections as nodes and road segments as edges. Based on real-time road segment capacity or travel time data, it constructs and maintains a real-time dynamic weighted graph that can reflect the instantaneous state of the entire road network.

[0154] The structural risk calculation module performs step S2 in the method. This module is configured to analyze the topology of the dynamically weighted graph output by the real-time network modeling module to calculate the normalized network centrality of each node, and, in conjunction with real-time traffic saturation data, calculate it according to a specific calculation model, as described above. The model is used to calculate the structural vulnerability index of each node in the network; after the calculation is completed, the module is also responsible for sorting the indices, filtering out the set of high-risk nodes and passing the information to subsequent modules.

[0155] The phase change precursor detection module performs step S3 in the method; this module specifically works on the high-risk node set identified by the structural risk calculation module; it continuously receives real-time average traffic density and speed data in the neighborhood of these nodes, calculates the normalized density growth rate and speed decay rate, and, based on a specific calculation model, as described above. The model integrates these two dimensions to generate a phase transition proximity index that quantifies the urgency of congestion risk;

[0156] The comprehensive risk assessment module executes step S4 of the method; this module is used to perform the final risk assessment; it first determines the dynamic coupling coefficient based on the average saturation of the entire road network. Subsequently, it receives the structural vulnerability index from the structural risk calculation module. Phase transition proximity index from the phase transition precursor detection module The system then uses a dynamic coupling coefficient to dynamically weight and fuse the two factors, generating a final comprehensive risk score for each high-risk node. ;

[0157] The early warning and decision generation module performs step S5 of the method; this module receives the comprehensive risk score generated by the comprehensive risk assessment module and compares it with a dynamically optimizeable risk threshold. The module compares the scores of any node; if the score of any node exceeds the threshold, the module immediately triggers an early warning mechanism, generates and pushes detailed early warning information to the user interface or third-party systems; in addition, the module can also be configured to automatically generate and output proactive intervention strategy suggestions based on the early warning information; if the scores of all nodes do not exceed the limit, the module controls the entire system to maintain a normal monitoring state.

[0158] This system, through its modular design, solidifies complex assessment and early warning methods into a set of collaborative hardware and software entities. Each module has clearly defined responsibilities and data flows smoothly, achieving full-process automation from raw data collection and multi-dimensional risk analysis to final early warning decision output. It provides urban traffic management departments with a powerful, reliable, and intelligent systemic risk control platform.

[0159] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for assessing and issuing early warnings of safety conditions in urban road traffic, characterized in that, Includes the following steps: S1. Collect real-time sensor data in the urban road traffic network, and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data; S2. Determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combine the topology and real-time traffic saturation to calculate the structural vulnerability index of each node in the network, and select the set of high-risk nodes based on the structural vulnerability index. S3. For each node in the high-risk node cluster, determine the normalized density growth rate and velocity decay rate in the neighborhood, and calculate and generate the phase transition proximity index based on the normalized density growth rate and velocity decay rate. S4. Determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node. S5. Determine the overall risk score against the preset risk threshold: If the overall risk score of any high-risk node exceeds the risk threshold, an early warning mechanism will be triggered to generate an early warning message for that high-risk node. If the overall risk score of all high-risk nodes does not exceed the risk threshold, the current monitoring status will be maintained. S1 specifically includes: S11. Abstract the intersections in the urban road traffic network into a set of nodes, and the road segments into a set of edges; S12. Collect sensor data deployed in the road network to obtain the real-time traffic capacity or travel time of the road segment; S13. Based on real-time traffic capacity or travel time, construct a weight matrix that changes over time, and combine the node set and edge set to generate a real-time dynamic weighted graph. S2 specifically includes: S21. Calculate the normalized network centrality of each node in the network to characterize its topological importance. S22. Collect the real-time traffic saturation of each node; S23. Combine normalized network centrality with real-time traffic saturation to calculate and generate a structural vulnerability index; S24. Sort the structural vulnerability indices of all nodes and identify the nodes with the highest index ranking as the high-risk node set. S3 specifically includes: S31. Obtain the real-time average traffic density and real-time average speed in the neighborhood of each high-risk node. S32. Calculate the normalized density growth rate based on the real-time average traffic density, the preset critical traffic density, and the preset characteristic time scale. S33. Calculate the velocity decay rate based on the real-time average velocity and the preset free-flow velocity; S34. Multiply the normalized density growth rate by the velocity decay rate to generate a phase transition proximity index. The dynamic coupling coefficient is determined as follows: If the average saturation of the entire road network is lower than the preset saturation threshold, the value of the dynamic coupling coefficient will be reduced to increase the weight of the structural vulnerability index in the dynamic weighted fusion. If the average saturation of the entire road network is not lower than the preset saturation threshold, the value of the dynamic coupling coefficient will be increased to improve the weight of the phase transition proximity index in the dynamic weighted fusion. The model for calculating the structural vulnerability index is defined as follows: in, For nodes The dimensionless structural fragility index; These are dimensionless system calibration coefficients; For nodes Normalized network centrality; It is the natural logarithm function; For nodes Real-time traffic saturation; The model for calculating the phase transition approximation index is as follows: in, For nodes A dimensionless phase transition proximity index for the neighborhood; , which is an adjustable dimensionless weighting coefficient; The preset feature time scale; The preset critical traffic density; The rate of change of real-time average traffic density over time; For nodes Real-time average traffic density within the neighborhood; For nodes Real-time average velocity within the neighborhood; For the preset free-flow velocity, when calculating the phase transition convergence index, only the case of density growth is considered, which is expressed as the time change rate of the real-time average traffic density. The value is 0 at that time; Dynamic coupling coefficient The specific function form is preset to an S-shaped function: in, For time The changing dynamic coupling coefficient; The average saturation level of the entire road network; This refers to the dimensionless weighted switching sensitivity parameter. It is the saturation threshold at which the network enters a critical state; is the base of the natural logarithm; Comprehensive Risk Score The calculation model is as follows: in, For nodes The final comprehensive risk score.

2. The method for safety situation assessment and early warning of urban road traffic as described in claim 1, characterized in that, S5 also includes: Based on the early warning information, proactive intervention strategy suggestions are output, which include at least one of the following: signal timing adjustment strategy, traffic guidance strategy, and emergency resource dispatch strategy.

3. The method for assessing and issuing early warning of safety conditions for urban road traffic as described in claim 1, characterized in that, The preset risk threshold is dynamically optimized by statistically learning from successful and failed warning cases in the system's operating history.

4. A safety situation assessment and early warning system for urban road traffic, based on the safety situation assessment and early warning method for urban road traffic according to any one of claims 1-3, characterized in that, include: The real-time network modeling module is used to collect real-time sensor data in the urban road traffic network and abstract the urban road traffic network into a real-time dynamic weighted graph based on the real-time sensor data. The structural risk calculation module is used to determine the topology and real-time traffic saturation of the real-time dynamic weighted graph. Combining the topology and real-time traffic saturation, it calculates the structural vulnerability index of each node in the network and selects a set of high-risk nodes based on the structural vulnerability index. The phase transition precursor detection module is used to determine the normalized density growth rate and velocity decay rate in the neighborhood of each node in the high-risk node cluster, and to calculate and generate a phase transition proximity index based on the normalized density growth rate and velocity decay rate. The comprehensive risk assessment module is used to determine the dynamic coupling coefficient associated with the average saturation of the entire road network, and to dynamically weight and fuse the structural vulnerability index and phase transition proximity index based on the dynamic coupling coefficient to generate a comprehensive risk score for each high-risk node. The early warning and decision generation module is used to compare the comprehensive risk score with the preset risk threshold. If the comprehensive risk score of any high-risk node exceeds the risk threshold, the early warning mechanism is triggered and an early warning message is generated. If the comprehensive risk scores of all high-risk nodes do not exceed the risk threshold, the control system maintains the current monitoring status.

Images