Automatic matching method and system for road congestion cause study and judgment and blockage relieving strategy

Abstract

The invention belongs to the technical field of traffic management, and particularly discloses an automatic matching method and system for road congestion cause study and judgment and a congestion relieving strategy. The method disclosed by the invention comprises the following steps: firstly, acquiring and preprocessing multi-source traffic data, and establishing a congestion judgment index to carry out congestion state judgment on a road; thirdly, constructing a traffic knowledge graph, converting multi-source data into a structured knowledge network, performing cause tracing on a congested road section, and screening out congestion causes; then, establishing a multi-dimensional strategy library, automatically matching a blockage relieving strategy from the strategy library, and outputting and executing the matched blockage relieving strategy; and finally, performing closed-loop optimization on the knowledge graph, the matching model and the strategy library based on execution feedback data. According to the invention, the conversion of traffic control from static experience matching to dynamic intelligent adaptation is realized, and the accuracy and timeliness of traffic jam control are improved. The method can be widely applied to congestion treatment of traffic roads.

Description

Technical Field

[0001] This invention belongs to the field of traffic management technology, specifically relating to a method and system for automatically matching road congestion causes and congestion mitigation strategies. Background Technology

[0002] With the acceleration of urbanization and the explosive growth of motor vehicle ownership in my country, traffic congestion has become a core bottleneck restricting high-quality urban development. By the end of 2024, the number of motor vehicles in China exceeded 420 million, a 28% increase compared to 2020, while the average annual growth rate of road mileage in urban built-up areas was only 8.2%, and the road network density was only 6.8 kilometers per square kilometer, highlighting the increasingly prominent supply-demand imbalance. The average congestion index during morning and evening rush hours in 50 major cities across the country reached 1.78. In first-tier cities such as Beijing and Shanghai, the average vehicle speed during peak hours was less than 20 kilometers per hour, and commuters wasted an average of 47 minutes per day due to congestion.

[0003] Current traditional traffic management models are ill-suited to dynamically changing traffic demands, exhibiting three core pain points: first, decision-making relies on experience-based judgment; second, data fusion capabilities are insufficient; and third, strategy implementation is often "one-size-fits-all," lacking differentiation based on regional characteristics and congestion types. This passive, experience-based management model results in poor adaptability and low accuracy of existing measures when dealing with complex congestion scenarios, failing to address the timely correlation and matching of "congestion causes" with "congestion mitigation strategies." Therefore, technological upgrades are urgently needed to achieve a paradigm shift in governance. Summary of the Invention

[0004] The purpose of this invention is to provide a method for automatically matching the causes of road congestion with congestion mitigation strategies. It constructs a closed-loop framework of "data support - knowledge modeling - intelligent reasoning - application feedback", which solves the problems of reliance on experience, insufficient data integration and poor strategy adaptability in traditional governance models. It realizes the transformation from static experience matching to dynamic intelligent adaptation, and improves the accuracy and timeliness of urban traffic congestion management. The second objective of this invention is to provide a road congestion cause analysis and automatic matching system for congestion mitigation strategies, which is used to implement the aforementioned method for automatic matching of road congestion cause analysis and congestion mitigation strategies.

[0005] To achieve the above objectives, the technical solution adopted by this invention is as follows: A method for automatically matching road congestion cause analysis and congestion mitigation strategies, the method comprising the following steps performed sequentially: S1. Acquire multi-source traffic data and perform preprocessing; S2. Establish congestion assessment indicators to determine the congestion status of roads; S3. Construct a traffic knowledge graph to transform multi-source data into a structured knowledge network, trace the causes of congestion on road sections, and screen out the causes of congestion. S4. Establish a multi-dimensional strategy library, automatically match congestion mitigation strategies from the strategy library, output the matched congestion mitigation strategies and execute them; S5. Perform closed-loop optimization of the knowledge graph, matching model, and strategy library based on execution feedback data.

[0006] As a limitation, the multi-source traffic data mentioned in step S1 includes traffic status data, event data, and static basic data; Traffic status data includes traffic flow, vehicle speed, queue length, and vehicle occupancy. Event data includes traffic accidents, road construction, large events, and parking incidents; Static basic data includes road network structure, intersection type, and road segment attributes.

[0007] As a second limitation, the preprocessing described in step S1 includes data cleaning and standardization; Non-continuous data is labeled using 1s and 0s; The standardized calculation formula for dynamically changing data is: , in, Y For the standardization results, x is The input is dynamically changing data. x max and x min These represent the maximum and minimum values ​​of dynamically changing data under the same metric.

[0008] As a third limitation, the construction of the traffic knowledge graph in step S3 includes: An ontology framework for the transportation domain is constructed. Entities, attributes, and relationships are extracted from multi-source data through structured queries, regular expressions, and the BiLSTM-CRF model. Knowledge fusion and conflict resolution are then performed to form a reasonable congestion ecosystem knowledge graph.

[0009] As a fourth limitation, the cause tracing in step S3 includes: The congestion candidate cause set is obtained by using knowledge graph association query, the influence weight of each cause is calculated by using a pre-trained model, and the cause of congestion is selected by dynamic adjustment combined with reinforcement learning.

[0010] As a fifth constraint, in step S4, a weighted matching degree model is used to calculate the matching score between the strategy and the current congestion scenario, and a congestion mitigation strategy is matched based on the matching score.

[0011] As a further limitation, the calculation formula for the weighted matching degree model is as follows: , In the formula, MatchScore To match scores, W p For the first p The weight of each contributing factor S p,q For strategy q Causes p Basic compatibility, L k For congestion level coefficients, C m This is a correction factor for strategy execution costs.

[0012] As the sixth constraint, the closed-loop optimization of step S5 includes: collecting traffic status data after the strategy is implemented and evaluating the effect; if the effect does not meet expectations, backtracking and updating the entity relationships in the knowledge graph, adjusting the matching model parameters, or modifying the strategy library attributes.

[0013] A road congestion cause analysis and automatic matching system for congestion mitigation strategies is provided to implement the above-mentioned method for road congestion cause analysis and congestion mitigation strategies. The system includes a data acquisition and preprocessing module, a knowledge graph construction and management module, a cause analysis and strategy matching module, and a strategy push and execution feedback module. The data acquisition and preprocessing module is used to acquire multi-source traffic data, perform preprocessing, and then transmit the processed data to the knowledge graph construction and management module. The knowledge graph construction and management module receives processed data, constructs and maintains a traffic congestion ecosystem knowledge graph; The cause analysis and strategy matching module is used to determine the congestion status, trace the cause of congestion, and match and recommend congestion mitigation strategies from the strategy library. The strategy push and execution feedback module is used to push strategies, collect execution feedback, and drive closed-loop optimization of the system.

[0014] As a limitation, the cause analysis and strategy matching module includes a congestion determination unit, a graph query unit, a weight calculation unit, and a matching degree calculation unit.

[0015] The present invention, by adopting the above-described technical solution, achieves the following technical advancements compared to existing technologies: (1) The method of the present invention realizes the paradigm shift of traffic management from experience-driven to data-driven, from passive response to proactive prediction, and from "one-size-fits-all" to precise policy implementation by constructing a complete technical chain of "data perception - knowledge construction - intelligent diagnosis - strategy matching - closed-loop optimization". Based on the fusion and structuring of multi-source data of knowledge graph, it breaks down information silos and realizes accurate judgment of congestion status and in-depth tracing of causes. Furthermore, by constructing a multi-dimensional strategy library and automatic matching mechanism, it ensures the dynamic and accurate adaptation of congestion mitigation strategies to specific congestion scenarios, which significantly improves the pertinence and timeliness of management. Finally, the closed-loop optimization design enables the system to have the ability to continuously learn and self-evolve, thus forming an intelligent management solution that can adapt to the complex dynamic changes of urban traffic. (2) The knowledge graph constructed by the method of this invention is based on the domain ontology framework to ensure the structure and standardization of the knowledge system, and provides a unified semantic expression basis for multi-source heterogeneous data. By comprehensively using structured query, regular expression and BiLSTM-CRF model, high-precision and automated extraction of entities, attributes and relationships in multimodal data is achieved, which greatly improves the comprehensiveness and efficiency of knowledge acquisition. Through knowledge fusion and conflict resolution, a structured knowledge network with internal consistency and rich associations is formed, so that the graph can not only statically reflect the association of traffic elements, but also support dynamic semantic reasoning and in-depth analysis, thus laying an interpretable and deductive knowledge foundation for subsequent accurate determination of congestion status and tracing of complex causes. (3) In the method of the present invention, the association query based on knowledge graph can quickly lock the set of candidate causes associated with congestion in time and space, ensuring the comprehensiveness and domain rationality of the source analysis; the use of pre-trained model to calculate the influence weight introduces the complex causal pattern hidden in historical data, improving the scientificity and objectivity of the initial assessment of the importance of the cause; and the combination of reinforcement learning dynamic adjustment mechanism enables the system to continuously optimize the weight allocation of each cause based on the real-time feedback after the strategy is executed, thereby accurately capturing the evolution of the main causes in the dynamic traffic environment, and finally screening out the most likely and most critical congestion root cause, providing a solid and reliable decision basis for the accurate matching of subsequent strategies. (4) The method of the present invention uses a weighted matching degree model. By integrating multi-dimensional parameters such as cause weight, strategy fit, congestion level and execution cost, it realizes a refined and quantitative evaluation of the applicability of the strategy, ensuring that the matched strategy not only targets the root cause of congestion, but also takes into account the feasibility and timeliness of implementation, effectively avoiding the "precise mismatch" of the strategy and the waste of resources. (5) The method of the present invention constructs a self-iterative loop of "evaluation-backtracking-update", which can continuously correct the accuracy of the knowledge graph, optimize the parameters of the matching model, and improve the effectiveness of the strategy library based on the actual execution effect, so that the whole system has the ability to continuously learn and evolve in practice.

[0016] This invention belongs to the field of traffic management technology. It constructs a closed-loop framework of "data support - knowledge modeling - intelligent reasoning - application feedback", which solves the problems of reliance on experience, insufficient data integration and poor strategy adaptability in traditional governance models. It realizes the transformation from static experience matching to dynamic intelligent adaptation, and improves the accuracy and timeliness of urban traffic congestion management. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0018] In the attached diagram: Figure 1 This is a flowchart of the processing in Embodiment 1 of the present invention; Figure 2 This is a system structure block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0019] The preferred embodiments of the present invention will now be described with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the scope of the invention. Example

[0020] like Figure 1 As shown, this embodiment is a method for automatically matching the causes of road congestion with congestion mitigation strategies, which includes the following steps performed in sequence: S1. Acquire multi-source traffic data and perform preprocessing.

[0021] Specifically, multi-source traffic data includes traffic status data, event data, and static basic data; Traffic status data includes traffic flow, vehicle speed, queue length, and vehicle occupancy. Event data includes traffic accidents, road construction, large events, and parking incidents; Static basic data includes road network structure, intersection type, and road segment attributes.

[0022] After acquiring the data, preprocessing is performed, which includes data cleaning and standardization: For non-continuous data, use 1 and 0 to label it. For example, if there is a traffic accident, it can be labeled as 1 if yes and 0 if no.

[0023] The standardized calculation formula for dynamically changing data is: , in, Y For the standardization results, x is The input is dynamically changing data.x max and x min These represent the maximum and minimum values ​​of dynamically changing data under the same metric.

[0024] S2. Establish congestion assessment indicators to determine the congestion status of roads.

[0025] A comprehensive judgment rule based on multiple indicators is established. For example, in this embodiment, judgment indicators for evaluating operating speed and congestion index are established, as shown in Table 1 and Table 2, respectively.

[0026] Table 1. Average Operating Speed ​​Judgment Indicators Run level Smooth Basically unobstructed Mild congestion Moderate congestion Severe congestion expressway V＞65 50＜V≤65 35＜V≤50 20＜V≤35 V≤20 Main road V＞40 30＜V≤40 20＜V≤30 15＜V≤20 V≤15 Secondary arterial roads and branch roads V＞35 25＜V≤35 15＜V≤25 10＜V≤15 V≤10 Table 2 Congestion Index Determination Indicators Service Level Level 1 Level 2 Level 3 Level 4 Intensity Index [0，0.6) [0.6，0.8) [0.8，1) ≥1 Traffic conditions Smooth Mild congestion Moderate congestion Severe congestion S3. Construct a traffic knowledge graph to transform multi-source data into a structured knowledge network, trace the causes of congestion on road sections, and screen out the causes of congestion.

[0027] Specifically, the first step is to build an ontological framework for the transportation sector, and to refine the entity classification, such as subdividing "traffic events" into subcategories such as accidents, construction, and large-scale events; and to clarify attribute constraints, such as specifying the "range of impact" of congestion events in kilometers.

[0028] Then, entities, attributes, and relationships are extracted from multi-source data using structured queries, regular expressions, and a BiLSTM-CRF model. Entities and attributes are directly extracted from structured data using SQL queries; key information is extracted from semi-structured data using regular expressions; and entity recognition is performed on unstructured text using a BiLSTM-CRF model.

[0029] Finally, knowledge fusion and conflict resolution are performed to form a reasonable congestion ecosystem knowledge graph. Multi-source knowledge is integrated, and the final value of different attributes of the same entity is determined by weighted calculation based on the credibility of the data sources. Conflicting relationships are corrected through cross-validation of the original data and manual review. The fused knowledge is then transformed into "entity-attribute-relationship" triples, forming standardized knowledge units.

[0030] Next, we will trace the causes of congestion and identify the underlying causes. Based on knowledge graph-based association queries, we will mine all potential factors related to congested road sections, including directly related accidents and construction, as well as indirectly related factors such as weather and promotional activities of nearby large shopping malls, forming a candidate cause set containing 5-10 factors.

[0031] Candidate causes are input into a pre-trained model for training. The model learns the implicit relationships between entities in the graph and calculates the influence weight of each cause. The weights are dynamically adjusted by combining reinforcement learning, and finally the causes of congestion are selected.

[0032] The formula for training the model is:

[0033] S4. Establish a multi-dimensional strategy library, automatically match congestion mitigation strategies from the strategy library, output the matched congestion mitigation strategies, and execute them.

[0034] Traffic management strategies are organized and categorized into different dimensions to form a strategy library. For example, police deployment and detour suggestions are considered rapid response strategies, traffic light timing adjustments are considered short-term optimization strategies, and road widening is considered a long-term planning strategy. Each strategy is labeled with attributes such as applicable scenarios, execution costs, and mitigation timeliness.

[0035] A weighted matching degree model is used to calculate the matching score between the strategy and the current congestion scenario. Based on this matching score, a congestion mitigation strategy is then matched. The calculation formula is as follows: , In the formula, MatchScore To match scores, W p For the first p The weight of each contributing factor S p,q For strategy q Causes p Basic compatibility, L k For congestion level coefficients, C m This is a correction factor for strategy execution costs.

[0036] In this embodiment, C m Use 1.2 for low cost, 1.0 for medium cost, and 0.8 for high cost.

[0037] S5. Perform closed-loop optimization of the knowledge graph, matching model, and strategy library based on execution feedback data.

[0038] Real-time data collection of vehicle speed and traffic flow after strategy execution, statistical analysis of congestion relief duration, and generation of multi-dimensional effectiveness evaluation data. For cases with poor performance, backtracking is conducted to investigate the problem. If the cause is missing, entity relationships in the knowledge graph are added; if the cause is a mismatch in the matching model, the model is retrained using feedback data, and weight parameters are adjusted; if the cause is poor adaptability of the strategy itself, the strategy library attributes are updated. Feedback data is synchronized to the data support layer to supplement new training samples, update the graph relationships and attributes of the knowledge modeling layer, and optimize the rule base and model parameters of the intelligent reasoning layer, forming a complete closed loop of "data, knowledge, reasoning, application, and feedback."

[0039] In summary, this embodiment constructs a closed-loop framework of "data support - knowledge modeling - intelligent reasoning - application feedback", which solves the problems of reliance on experience, insufficient data integration, and poor strategy adaptability in traditional governance models. It realizes the transformation from static experience matching to dynamic intelligent adaptation, and improves the accuracy and timeliness of urban traffic congestion management. Example

[0040] like Figure 2 As shown, this embodiment is an automatic matching system for road congestion cause analysis and congestion mitigation strategies, used to implement Embodiment 1. It includes a data acquisition and preprocessing module, a knowledge graph construction and management module, a cause analysis and strategy matching module, and a strategy push and execution feedback module.

[0041] The data acquisition and preprocessing module is used to acquire multi-source traffic data, perform preprocessing, and then transfer the processed data to the knowledge graph construction and management module.

[0042] The knowledge graph construction and management module receives processed data and constructs and maintains a traffic congestion ecosystem knowledge graph.

[0043] The cause analysis and strategy matching module includes a congestion determination unit, a graph query unit, a weight calculation unit, and a matching degree calculation unit. It is used to determine the congestion status, trace the cause of congestion, and match and recommend congestion mitigation strategies from the strategy library.

[0044] The strategy push and execution feedback module is used to push strategies, collect execution feedback, and drive closed-loop optimization of the system.

Claims

1. A method for automatically matching road congestion cause analysis and congestion mitigation strategies, characterized in that, The method includes the following steps performed sequentially: S1. Acquire multi-source traffic data and perform preprocessing; S2. Establish congestion assessment indicators to determine the congestion status of roads; S3. Construct a traffic knowledge graph to transform multi-source data into a structured knowledge network, trace the causes of congestion on road sections, and screen out the causes of congestion. S4. Establish a multi-dimensional strategy library, automatically match congestion mitigation strategies from the strategy library, output the matched congestion mitigation strategies and execute them; S5. Perform closed-loop optimization of the knowledge graph, matching model, and strategy library based on execution feedback data.

2. The method for automatically matching road congestion causes and mitigation strategies according to claim 1, characterized in that, The multi-source traffic data mentioned in step S1 includes traffic status data, event data, and static basic data; Traffic status data includes traffic flow, vehicle speed, queue length, and vehicle occupancy. Event data includes traffic accidents, road construction, large events, and parking incidents; Static basic data includes road network structure, intersection type, and road segment attributes.

3. The method for automatically matching road congestion cause analysis and congestion mitigation strategies according to claim 1 or 2, characterized in that, The preprocessing described in step S1 includes data cleaning and standardization; Non-continuous data is labeled using 1s and 0s; The standardized calculation formula for dynamically changing data is: ， in, Y For the standardization results, x is The input is dynamically changing data. x max and x min These represent the maximum and minimum values ​​of dynamically changing data under the same metric.

4. The method for automatically matching road congestion causes and mitigation strategies according to claim 1, characterized in that, The construction of the traffic knowledge graph in step S3 includes: An ontology framework for the transportation domain is constructed. Entities, attributes, and relationships are extracted from multi-source data through structured queries, regular expressions, and the BiLSTM-CRF model. Knowledge fusion and conflict resolution are then performed to form a reasonable congestion ecosystem knowledge graph.

5. The method for automatically matching road congestion causes and mitigation strategies according to claim 1, characterized in that, The cause tracing mentioned in step S3 includes: The congestion candidate cause set is obtained by using knowledge graph association query, the influence weight of each cause is calculated by using a pre-trained model, and the cause of congestion is selected by dynamic adjustment combined with reinforcement learning.

6. The method for automatically matching road congestion causes and mitigation strategies according to claim 1, characterized in that, In step S4, a weighted matching degree model is used to calculate the matching score between the strategy and the current congestion scenario, and a congestion mitigation strategy is matched based on the matching score.

7. The method for automatically matching road congestion causes and mitigation strategies according to claim 6, characterized in that, The calculation formula for the weighted matching degree model is as follows: ， In the formula, MatchScore To match scores, W p For the first p The weight of each contributing factor S p,q For strategy q Causes p Basic compatibility, L k For congestion level coefficients, C m This is a correction factor for strategy execution costs.

8. The method for automatically matching road congestion causes and mitigation strategies according to claim 1, characterized in that, The closed-loop optimization in step S5 includes: collecting traffic status data after the strategy is implemented and evaluating the effect; if the effect does not meet expectations, backtracking and updating the entity relationships in the knowledge graph, adjusting the matching model parameters, or modifying the strategy library attributes.

9. A road congestion cause analysis and automatic matching system for congestion mitigation strategies, used to implement the road congestion cause analysis and automatic matching system for congestion mitigation strategies as described in any one of claims 1 to 8, characterized in that, It includes modules for data acquisition and preprocessing, knowledge graph construction and management, cause analysis and strategy matching, and strategy push and execution feedback. The data acquisition and preprocessing module is used to acquire multi-source traffic data, perform preprocessing, and then transmit the processed data to the knowledge graph construction and management module. The knowledge graph construction and management module receives processed data and constructs and maintains a traffic congestion ecosystem knowledge graph. The cause analysis and strategy matching module is used to determine the congestion status, trace the cause of congestion, and match and recommend congestion mitigation strategies from the strategy library. The strategy push and execution feedback module is used to push strategies, collect execution feedback, and drive closed-loop optimization of the system.

10. The automatic matching system for road congestion cause analysis and congestion mitigation strategies according to claim 9, characterized in that, The cause analysis and strategy matching module includes a congestion determination unit, a graph query unit, a weight calculation unit, and a matching degree calculation unit.

Images