Traffic signal control method and device based on data matching, equipment and medium

Abstract

The embodiment of the invention discloses a traffic signal control method and device based on data matching, equipment and a medium. A specific embodiment of the method comprises the following steps: constructing a semantic association network; generating a data supply label set based on each acquired traffic data set, and linking each data supply label in the data supply label set to a semantic association network; generating a data demand tag set based on the acquired traffic control demand information, and linking each data demand tag in the data demand tag set to a semantic association network; generating a target traffic data set according to the data supply label set and the data demand label set; generating a traffic signal control parameter set based on the target traffic data set; and controlling each traffic signal lamp based on the traffic signal control parameter set. According to the embodiment, deep mining and efficient utilization of traffic data resources are realized, and an innovative technical normal form is provided for construction of an intelligent traffic system.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to traffic signal control methods, apparatus, devices, and media based on data matching. Background Technology

[0002] With the rapid development of smart cities and intelligent transportation systems, utilizing multi-source data to achieve precise and adaptive traffic signal control has become a key means to improve road traffic efficiency. The formulation of traffic control strategies increasingly relies on the in-depth mining and intelligent application of various traffic datasets (such as intersection traffic flow, vehicle trajectories, queue lengths, etc.). Currently, in the field of intelligent transportation, finding suitable data support for specific control needs mainly relies on two methods: first, keyword-based database queries, which search by matching the description of the control requirement with the name and tags of the dataset; second, relying on the prior knowledge of domain experts, where engineers manually select and call relevant historical datasets based on experience.

[0003] However, the above methods often present the following technical problems: Keyword-based matching methods heavily rely on strict semantic consistency, making them ineffective in addressing cross-domain and divergent data discovery needs. For example, when the control objective is "optimizing traffic congestion around schools," keyword matching alone may not be able to identify potentially related but differently labeled datasets such as "parking lot vacancy rates in surrounding commercial areas" or "pedestrian crossing requests at adjacent intersections." Methods relying on expert experience are inefficient, highly subjective, and difficult to systematically reproduce and optimize, failing to meet the breadth and depth requirements of real-time, intelligent traffic control for data discovery. Therefore, there is an urgent need in this field for a method that can intelligently understand control intentions and automatically correlate and match the most relevant data resources from massive datasets to support the generation of more scientific and accurate traffic signal control strategies.

[0004] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure provide a traffic signal control method, apparatus, electronic device, and computer-readable medium based on data matching to address one or more of the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a traffic signal control method based on data matching. The method includes: constructing a semantic association network, wherein the semantic association network includes various traffic control domain nodes and various keyword nodes; generating a data supply label set based on acquired traffic datasets, and linking each data supply label in the data supply label set to the semantic association network; generating a data demand label set based on acquired traffic control demand information, and linking each data demand label in the data demand label set to the semantic association network; generating a target traffic dataset based on the data supply label set and the data demand label set; generating a traffic signal control parameter set based on the target traffic dataset; and controlling various traffic lights based on the traffic signal control parameter set.

[0008] Secondly, some embodiments of this disclosure provide a traffic signal control device based on data matching, comprising: a construction unit configured to construct a semantic association network, wherein the semantic association network includes various traffic control domain nodes and various keyword nodes; a first association unit configured to generate a data supply tag set based on acquired traffic datasets, and to link various data supply tags in the data supply tag set to the semantic association network; a second association unit configured to generate a data demand tag set based on acquired traffic control demand information, and to link various data demand tags in the data demand tag set to the semantic association network; a first generation unit configured to generate a target traffic dataset based on the data supply tag set and the data demand tag set; a second generation unit configured to generate a traffic signal control parameter set based on the target traffic dataset; and a control unit configured to control various traffic lights based on the traffic signal control parameter set.

[0009] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0010] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0011] Fifthly, some embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0012] The above embodiments of this disclosure have the following beneficial effects: The above embodiments of this application have the following beneficial effects: Through the traffic signal control method based on data matching of this application, intelligent matching and precise application of multi-source data in complex traffic scenarios are realized, effectively improving the intelligence level and scientific nature of traffic signal control. Specifically, traditional data retrieval methods rely on direct keyword matching or limited document retrieval, which may lead to insufficient semantic understanding, difficulties in cross-domain association, and omission of valuable data, making it difficult to meet the needs of intelligent transportation systems for in-depth data mining and divergent applications. Based on this, the traffic signal control method based on data matching of this application embodiments: First, by constructing a semantic association network containing traffic control domain nodes and keyword nodes, a cross-disciplinary, multi-level semantic understanding framework is established. This step breaks down the domain barriers in traditional retrieval at the knowledge system level, placing traffic data in a broader knowledge context, laying a solid foundation for subsequent intelligent association and divergent matching, and overcoming the problem of single-domain knowledge limitations. Then, based on the acquired traffic datasets, a data supply label set is generated and linked to the semantic network, realizing deep semantic annotation and knowledge integration of the original data resources. This process automatically mines the connotation of data through pre-trained large models, transforming simple dataset names and fields into rich semantic labels and making them an organic component of the knowledge network. This solves the bottleneck of traditional methods, which rely on shallow semantic expression and manual annotation. Next, a data demand label set is generated based on the acquired traffic control demand information and linked to the semantic network, achieving accurate parsing and structured expression of complex user intentions. This mechanism transforms control demands described in natural language into semantic nodes recognizable by the network, enabling precise understanding and location of ambiguous and divergent user needs, overcoming the failure of traditional keyword matching when terms are inconsistent. Then, a target traffic dataset is generated based on the data supply label set and the data demand label set, achieving intelligent data discovery based on semantic association. This step, through graph search and path calculation in the knowledge network, can discover potentially related datasets with semantically similar user needs but different keywords, achieving intelligent retrieval "from point to network," significantly improving the recall and novelty of data discovery. Subsequently, a set of traffic signal control parameters is generated based on the target traffic dataset, completing the value transformation from data resources to control strategies. This process transforms the matched high-quality data resources into specific, executable signal control parameters, establishing a crucial bridge for data-driven decision-making and enabling the value of data to be concretely realized in traffic control practice. Finally, based on the set of traffic signal control parameters, each traffic light is controlled, achieving data-driven precision traffic management. This step directly applies the data processing results to physical traffic facilities, forming a complete technical closed loop of "data perception - intelligent matching - strategy generation - precise execution," providing an innovative path for building intelligent traffic control systems.In summary, the embodiments of this application achieve in-depth mining and efficient utilization of traffic data resources through a collaborative technical route of "semantic network construction - supply and demand tag generation - intelligent matching - parameter transformation - precise control". This method not only improves the intelligence of data retrieval and solves the shortcomings of traditional methods in semantic understanding and cross-domain correlation, but also effectively transforms data value into control benefits through a complete technical closed loop, providing an innovative technical paradigm for the construction of intelligent transportation systems. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the traffic signal control method based on data matching according to the present disclosure; Figure 2 This is a schematic diagram of the structure of some embodiments of a traffic signal control device based on data matching according to the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0015] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0016] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0017] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0018] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0019] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0020] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] Figure 1 A flow 100 of some embodiments of a data-matching-based traffic signal control method according to the present disclosure is shown. This data-matching-based traffic signal control method includes the following steps: Step 101: Construct a semantic association network.

[0022] In some embodiments, the execution entity (e.g., a computing device) of the data matching-based traffic signal control method can construct a semantic association network. This semantic association network is a graph-structured data structure used to characterize the complex relationships between traffic control research fields and traffic technology keywords. Its nodes include traffic control field nodes representing subject classifications and keyword nodes representing specific technical concepts. Nodes are connected by weighted edges, with the weights representing the strength of the association.

[0023] In some optional implementations of certain embodiments, the aforementioned execution entity may construct a semantic association network through the following steps: Step one involves constructing a framework encompassing nodes across various traffic control domains. These traffic control domain nodes represent the classification systems of related disciplines such as intelligent transportation and control science, with each node corresponding to a unique discipline classification identifier and name. This framework is a pre-constructed hierarchical discipline classification system based on an authoritative classification system (such as the *Chinese Library Classification*). In practice, the implementing entity first needs to establish a structured framework covering traffic control and related fields. This step, by importing a standardized classification system, provides a stable and standardized top-level skeleton for the entire semantic network. This framework ensures that all subsequent data linking and retrieval operations can be performed within a unified discipline system, laying the foundation for cross-domain knowledge association. For example, the implementing entity parses and imports classification structures strongly related to traffic control from the electronic version of the *Chinese Library Classification* (ZTFLH). For instance, it might create a node "U Transportation," under which a sub-node "U49 Traffic Engineering and Traffic Management," and further, under "U491 Traffic Technology Management," a sub-node "U491.5 Traffic Signal Control." Each node stores information including: a classification number (e.g., "U491.5"), a classification name (e.g., "traffic signal control"), and its hierarchical path within the entire tree structure.

[0024] Step two involves generating keyword nodes based on the constructed framework. Keyword nodes are technical terms, concepts, or keywords extracted from knowledge data sources (such as academic papers, patents, and technical reports) in fields like traffic engineering, control science, and computer applications. These nodes represent the core content of the literature. In practice, after establishing the disciplinary framework, the implementing entity needs to extract specific technical concepts from massive knowledge data sources in the transportation field to populate the framework. This step uses automated information extraction technology to transform unstructured document content into structured knowledge units (i.e., keyword nodes), enabling the semantic network to possess fine-grained knowledge content related to traffic signal control. For example, the implementing entity obtains metadata from academic papers in the fields of intelligent transportation, autonomous driving, and signal control optimization from databases such as CNKI and IEEE. For a paper titled "Research on Real-time Optimization Method of Intersection Signals Based on Reinforcement Learning," its built-in keywords might be "signal timing," "reinforcement learning," and "traffic congestion." If a paper lacks keywords, its abstract is input into a pre-trained large model, which might output "Q-learning," "delay minimization," etc. Ultimately, terms such as "signal timing," "reinforcement learning," and "traffic congestion" will be created as keyword nodes in the semantic network.

[0025] Step three involves linking each keyword node to each traffic control domain node and assigning weights to all connecting edges to obtain a semantic association network. Here, "linking" refers to establishing edges between keyword nodes and traffic control domain nodes in the graph structure. The initial weight is the initial association strength value assigned when the edge is created, usually set to a uniform constant. In practice, this step is the core of network construction, aiming to establish semantic associations between the disciplinary framework and specific knowledge content. By linking keywords to their respective disciplinary domains, a binary association structure of "traffic domain - technical concept" is formed. Assigning initial weights to the edges ultimately forms a semantic association network serving intelligent matching for traffic signal control. This, through processing a large number of papers and data in the traffic field, provides a quantitative foundation for subsequent graph search-based relevance calculations oriented towards traffic data needs. For example, continuing the previous example, the disciplinary classification of the paper on signal optimization might belong to "U491.5 Traffic Signal Control". The executing entity will perform the following operations: search for the keyword node "reinforcement learning" in the semantic network, and create it if it does not exist. Search for the traffic control domain node "U491.5" in the semantic network. Establish a connection between the "Reinforcement Learning" node and the "U491.5" node. Assign an initial weight to this connection, for example, set the initial weight value to 1. Similarly, link other keyword nodes of this paper, such as "signal timing" and "traffic congestion," to the "U491.5" node or other related domain nodes (such as "Machine Learning" under "TP18 Artificial Intelligence Theory"), and assign an initial weight of 1 to each.

[0026] Step 102: Based on the acquired traffic datasets, generate a data supply label set and link each data supply label in the data supply label set to the semantic association network.

[0027] In some embodiments, the aforementioned executing entity may generate a data supply label set based on the acquired traffic datasets, and link each data supply label in the data supply label set to the aforementioned semantic association network.

[0028] In some optional implementations of certain embodiments, the aforementioned execution entity may generate a data supply label set based on the acquired traffic datasets through the following steps, and link each data supply label in the data supply label set to the aforementioned semantic association network: Step one involves acquiring various traffic datasets. A traffic dataset refers to a structured collection of data related to the transportation sector registered on a data trading platform. Each dataset includes basic information such as its name, content description, and data field structure. In practice, the executing entity accesses the data trading platform's backend database to query all registered and tradable traffic datasets and extract their complete metadata. For example, the executing entity reads all dataset records categorized as "Transportation" from the platform's database table, obtaining information such as the names, descriptions, and field lists of multiple datasets, including "Dataset A: Traffic Flow Data at the South Entrance of Zhongguancun Street" and "Dataset B: Signal Timing Scheme for the CBD Area of ​​Chaoyang District."

[0029] Step two, for each traffic dataset obtained, perform the following steps: The first sub-step involves organizing the attribute information of the acquired traffic dataset into structured text information. This attribute information includes name information, a data description, and a list of data fields. In practice, the executing entity uses a predefined text template to combine the various attribute fields of the dataset into a uniformly formatted and complete descriptive text, ensuring consistency in the description format across all datasets. For example, for "Dataset A: Traffic Flow Data at the South Entrance of Zhongguancun Street," the executing entity combines its name, description (real-time traffic flow data collected via geomagnetic sensors), and field list (timestamp, lane number, traffic flow value) into a complete structured text in the format {Dataset Name, Dataset Description, Data Fields}.

[0030] Sub-step two involves inputting the structured text information into a pre-trained large model to obtain data supply labels corresponding to the aforementioned traffic dataset. The pre-trained large model refers to an artificial intelligence model (such as GPT-4 or Ernie) trained on massive amounts of text data and possessing natural language understanding capabilities. Data supply labels are keywords or phrases that accurately characterize the dataset's content features, technical characteristics, or application scenarios. In practice, the executing entity sends the prepared structured descriptive text to the large model processing service, requesting the model to generate a set of keywords representing the dataset's features through specific instructions. For example, inputting the descriptive text of the aforementioned intersection traffic flow dataset into the large model, along with the instruction "Please generate 5 keywords that best represent the technical characteristics and application areas of this dataset based on the dataset information," the large model returns data supply labels such as "traffic flow monitoring, geomagnetic sensing technology, real-time data acquisition, intelligent transportation systems, and urban traffic management."

[0031] Step three involves integrating the data supply tags corresponding to the various traffic datasets into a data supply tag set. This set refers to the complete tag collection formed after merging and deduplicating the data supply tags generated from all datasets. In practice, the executing entity collects all tags generated from each dataset, removes duplicate content using a string matching algorithm, and forms a unified tag library for the platform. For example, after processing 50 traffic datasets on the platform, the 215 initial tags generated, such as "traffic flow monitoring," "signal control optimization," "reinforcement learning," and "computer vision," are deduplicated, ultimately resulting in a data supply tag set containing 198 unique keywords.

[0032] Step four involves matching each data supply label in the aforementioned data supply label set with each keyword node or traffic control domain node in the aforementioned semantic association network, and linking each data supply label with each matching node. In practice, the executing entity searches for nodes in the semantic association network that are exactly the same as the data supply label name. For each matching node found, an association is established between that node and the corresponding dataset. For example, for the data supply label "signal control optimization," after finding a keyword node with the same name in the semantic network, a connection is established between that node and all datasets containing data related to signal control optimization, indicating that these datasets are all related to the technical concept of "signal control optimization."

[0033] Optionally, the aforementioned implementing entity may also perform the following steps: Step one: In response to any data supply label failing to match any node in the semantic association network, a corresponding keyword node is created in the semantic association network. In practice, during the matching process, when the executing entity finds that a data supply label does not have a corresponding node in the existing semantic network, it automatically adds a new keyword node named after that label to the network. For example, when the data supply label "vehicle-road cooperation" has no corresponding node in the existing semantic network, the system automatically creates a new keyword node named "vehicle-road cooperation," assigns a unique identifier to the node, and records metadata information such as the creation time.

[0034] Step two involves linking the aforementioned data to the labels and corresponding keyword nodes, and updating the semantic association network. In practice, after creating a new keyword node, the executing entity immediately establishes the association between that node and the corresponding dataset, and persistently saves all node and link changes to the semantic network database. For example, after creating a new node for "intelligent roadside equipment," it immediately establishes links between that node and all datasets containing roadside equipment-related data, and saves the updated complete semantic network structure to the graph database, ensuring that subsequent data retrieval and recommendation operations can use the latest network structure.

[0035] Step 103: Based on the acquired traffic control demand information, generate a data demand tag set and link each data demand tag in the data demand tag set to a semantic association network.

[0036] In some embodiments, the aforementioned implementing entity may generate a data demand tag set based on the acquired traffic control demand information, and link each data demand tag in the aforementioned data demand tag set to the aforementioned semantic association network.

[0037] In some optional implementations of certain embodiments, the aforementioned executing entity may generate a data demand tag set based on the acquired traffic control demand information, and link each data demand tag in the aforementioned data demand tag set to the aforementioned semantic association network through the following steps: Step one: Obtain traffic control requirement information. This traffic control requirement information refers to natural language text input by the user describing the traffic signal control needs, including the traffic problems to be solved, the expected control objectives, or a description of the required data types. In practice, the implementing entity receives the user-submitted requirement description text through the requirement input interface of the data trading platform. For example, if the user inputs "needs to optimize intersection queue length data during evening rush hour," the implementing entity obtains this text as the traffic control requirement information.

[0038] Step two involves inputting the aforementioned traffic control demand information into a pre-trained large-scale model to obtain a data demand label set. This set includes various data demand labels. Data demand labels are keywords or phrases extracted from the demand information that represent the core demand content. The pre-trained large-scale model refers to an artificial intelligence model with natural language understanding capabilities (such as GPT-4, Ernie, etc.). In practice, the executing entity sends the user-input demand description text to the large-scale model processing service, prompting the model to extract key information from the demand and generate the corresponding label set through specific instructions. For example, inputting the demand "need to optimize intersection queue length data during evening rush hour" into the large-scale model, along with the instruction "please extract the key technical keywords from this demand," will result in the large-scale model returning data demand labels such as "queue length analysis," "evening rush hour traffic optimization," and "intersection traffic management."

[0039] Step three involves matching each of the aforementioned data requirement tags with the keyword nodes in the semantic association network. In practice, the executing entity searches for keyword nodes in the semantic association network that perfectly match the data requirement tag name to determine whether each tag exists in the existing semantic network. For example, for the data requirement tag "queue length analysis," a search is conducted among the keyword nodes in the semantic network to confirm whether the concept has been included in the network.

[0040] Step four: In response to each data requirement tag successfully matching any keyword in the semantic association network, the system links each data requirement tag to its matching keyword node. In practice, the execution entity establishes an association between all data requirement tags that find matching nodes and their corresponding keyword nodes, forming a connection between requirements and knowledge concepts. For example, when the tag "evening rush hour traffic optimization" finds a keyword node with the same name in the semantic network, the system establishes a link between the requirement and the node, allowing searches based on this requirement to start from this keyword node.

[0041] Step 5: In response to any data requirement tag failing to match any keyword in the semantic association network, the system links the data requirement tag to a preset general node or removes it from the data requirement tag set. The preset general node refers to a general category node in the semantic network specifically used to include concepts without clear classification. In practice, the executing entity handles data requirement tags that fail to match a specific keyword node in two ways: either links them to a general concept node in the semantic network, or removes the tag directly from the requirement tag set. For example, when the requirement tag "dynamic traffic allocation" has no corresponding node in the semantic network, the system links it to a general node named "other traffic control concepts," or directly removes the tag from the current search tag set to ensure the feasibility of subsequent graph search processes.

[0042] Step 104: Generate the target traffic dataset based on the data supply label set and the data demand label set.

[0043] In some embodiments, the aforementioned executing entity may generate a target traffic dataset based on the aforementioned data supply label set and the aforementioned data demand label set.

[0044] In some optional implementations of certain embodiments, the execution entity can generate the target traffic dataset by following these steps based on the data supply label set and the data demand label set:

[0045] Step one involves identifying the keyword nodes linked to each data requirement tag in the aforementioned data requirement tag set as the search starting node set. This search starting node set refers to the collection of semantic network nodes corresponding to the user's requirement, serving as the starting point for the graph search algorithm. In practice, the executing entity extracts the corresponding semantic network node identifiers from the linked data requirement tags to form the search starting node set. For example, if the user's generated data requirement tag "signal timing optimization" links to a keyword node with the same name in the semantic network, the system adds this node to the search starting node set as one of the starting points for subsequent path searches.

[0046] Step two involves identifying the nodes linked to each data supply label in the aforementioned data supply label set as the candidate target node set. This candidate target node set refers to the collection of semantic network nodes corresponding to all traffic datasets on the platform, serving as the target points for the graph search algorithm. In practice, the execution entity traverses the data supply labels of all traffic datasets on the platform, collecting all nodes linked to these labels in the semantic network to form the candidate target node set. For example, the traffic datasets registered on the platform include various types such as "traffic flow data," "signal timing schemes," and "vehicle trajectory data," and the semantic network nodes corresponding to these datasets are collected to form the candidate target node set.

[0047] Step 3: Based on the aforementioned set of starting search nodes and set of candidate target nodes, generate the path distance from each starting search node to each candidate target node. The path distance refers to the sum of the weights of all edges on the shortest path between two nodes in the semantic association network, representing the strength of the semantic association between the two concepts. In practice, the executing entity uses a graph search algorithm to calculate the shortest path distance from each starting search node to each candidate target node. For example, Dijkstra's algorithm is used to calculate the shortest path from the "signal timing optimization" node to the "real-time traffic flow" node, and the path distance between the two nodes is obtained by summing the weights of all edges on the path.

[0048] Step 4: For each traffic dataset in the acquired traffic datasets, perform the following operations: Sub-step one involves obtaining all candidate target nodes linked to the aforementioned traffic dataset. In practice, the execution entity queries the semantic network to find all candidate target nodes directly connected to the current traffic dataset. For example, for a signal timing dataset at a certain intersection, it identifies all nodes linked to it in the semantic network, including candidate target nodes such as "signal control," "optimization algorithm," and "traffic simulation."

[0049] Sub-step two involves generating a comprehensive association distance for the aforementioned traffic dataset based on the acquired candidate target nodes and their corresponding path distances. In practice, the executing entity extracts the path distances between each candidate target node and each search starting node from the calculation results of step three, and uses a minimum value strategy to determine the comprehensive association distance of the dataset. For example, if a traffic dataset links three candidate target nodes with minimum path distances of 2, 3, and 5 to the search starting node, the system takes the minimum value of 2 as the comprehensive association distance for that dataset.

[0050] Step 5: Based on the comprehensive association distances corresponding to the aforementioned traffic datasets, select traffic datasets that meet the preset comprehensive association distance criteria, and integrate the selected traffic datasets into a candidate result set. The preset comprehensive association distance criteria refer to the maximum association distance threshold set by the system, used to select datasets with sufficiently high relevance to user needs. In practice, the executing entity compares the comprehensive association distance of each dataset with the preset threshold, retaining datasets with a distance less than or equal to the threshold. For example, setting the comprehensive association distance threshold to 4, the system selects all traffic datasets with a comprehensive association distance less than or equal to 4, forming a candidate result set containing 15 datasets.

[0051] Step six involves sorting and organizing the candidate result set, and determining the sorted candidate result set as the target traffic dataset. In practice, the executing entity sorts the datasets in the candidate result set according to the comprehensive correlation distance from smallest to largest; the smaller the distance, the higher the relevance to the user's needs. For example, the 15 datasets in the candidate result set are sorted by comprehensive correlation distance, with the dataset with a distance of 1 ranked first and the dataset with a distance of 4 ranked last, forming the final list of target traffic datasets recommended to the user.

[0052] Step 105: Generate a set of traffic signal control parameters based on the target traffic dataset.

[0053] In some embodiments, the aforementioned executing entity may generate a set of traffic signal control parameters based on the aforementioned target traffic dataset.

[0054] In some optional implementations of certain embodiments, the aforementioned execution entity may generate a set of traffic signal control parameters based on the aforementioned target traffic dataset through the following steps: Step 1: For each traffic dataset in the target traffic dataset above, perform the following operations: The first sub-step involves reading the raw data content from the aforementioned traffic dataset. The raw data content refers to the unprocessed initial data records stored in the traffic dataset, including time-series data, spatial location data, and traffic state parameters. In practice, the executing entity extracts complete raw data records from the target traffic dataset through a data access interface. For example, it reads a complete data record table containing timestamps, lane numbers, and traffic flow values ​​from the "Zhongguancun Street South Entrance Traffic Flow Dataset".

[0055] Sub-step two involves preprocessing the raw data to generate standardized data. Data preprocessing refers to cleaning, denoising, filling in missing values, and standardizing the format of the raw data. Standardized data refers to preprocessed data that meets the model's input requirements and is in a standardized format. In practice, the execution entity uses data cleaning algorithms to process the raw data, eliminating outliers and noise interference to ensure data quality meets the model's computational requirements. For example, abnormally high values ​​caused by sensor malfunctions in traffic flow data are removed, and missing time period data is interpolated to complete the data, ultimately generating complete and usable standardized data.

[0056] Step three involves inputting the standardized data into a pre-defined signal control optimization model to obtain preliminary control parameters. The signal control optimization model is a computational model based on traffic flow theory used to generate optimal signal control parameters; the preliminary control parameters are the basic control parameter values ​​calculated by the model based on the input data. In practice, the executing entity inputs the preprocessed standardized data into the signal control optimization model, and the model calculates control parameters applicable to the current traffic conditions. For example, historical traffic flow data and real-time traffic status data can be input into a reinforcement learning-based signal control model, and the model outputs preliminary control parameters including green light duration, signal cycle, and phase difference.

[0057] Sub-step four involves generating a subset of signal control parameters corresponding to the aforementioned traffic dataset, based on the preliminary control parameters described above. This subset of signal control parameters refers to the complete set of control parameters generated for a single traffic dataset, containing all control parameters and their metadata for that dataset. In practice, the executing entity encapsulates the preliminary control parameters output by the model according to a predetermined format, adding metadata such as dataset identifier, generation time, and parameter applicability range. For example, a subset of control parameters containing parameters such as phase scheme, green light duration, and signal cycle might be generated for a specific intersection dataset, and this parameter set might be labeled as applicable to weekday morning rush hour.

[0058] Step two involves processing the various subsets of signal control parameters to obtain the traffic signal control parameter set. This processing includes performing consistency checks, conflict resolution, and priority ranking on multiple parameter subsets. The traffic signal control parameter set is the final complete set of parameters that can be directly issued to the signal controller. In practice, the executing entity comprehensively analyzes multiple parameter subsets from different datasets to eliminate conflicts and determine the final parameter combination. For example, when control parameters generated from two datasets conflict, the system prioritizes the datasets based on their timeliness, data quality, and matching degree, selecting the optimal parameter combination to form the final traffic signal control parameter set, ensuring parameter consistency and executability.

[0059] Step 106: Control each traffic light based on the set of traffic signal control parameters.

[0060] In some embodiments, the aforementioned execution entity may control each traffic light based on the aforementioned set of traffic signal control parameters.

[0061] In some optional implementations of certain embodiments, the aforementioned executing entity may control each traffic light based on the aforementioned set of traffic signal control parameters through the following steps: The first step involves parsing the aforementioned traffic signal control parameter set within the edge computing gateway to obtain the identifiers for each target intersection. These target intersection identifiers are unique codes that identify road intersections, typically using geolocation coding or a unified numbering system adopted by traffic management departments. In practice, the executing entity extracts the intersection identification information contained in the traffic signal control parameter set and parses it into a standard device identification format. For example, intersection identifiers such as "ZGC-001" and "ZGC-002" are parsed from the parameter set, and these identifiers correspond one-to-one with the actual physical location of the intersection.

[0062] The second step involves querying the local device registry for the communication protocol specifications of the corresponding signal controllers for each target intersection identifier, based on the aforementioned target intersection identifiers. The communication protocol specifications refer to the data communication formats and interaction standards supported by the signal controller, including technical parameters such as communication ports, data frame structures, and instruction set definitions. In practice, the executing entity performs a matching query in the device registration database based on the intersection identifier to obtain the detailed communication parameters of the corresponding signal controller. For example, the query might find that the signal controller corresponding to intersection "ZGC-001" uses the TCP / IP communication protocol, uses port 502, and supports data exchange in Modbus protocol format.

[0063] The third step involves establishing communication links with each target signal controller according to the aforementioned communication protocol specifications. In practice, the executing entity initializes network connection parameters based on the retrieved communication protocol specifications and establishes a stable data communication channel with each target signal controller. For example, a TCP connection is established based on the controller's IP address and port number to complete the communication link handshake process, ensuring reliable transmission of subsequent control commands.

[0064] The fourth step involves generating various control commands based on the aforementioned set of signal control parameters. These control commands are operational instructions that conform to the signal controller interface specification and can directly drive changes in the state of the traffic lights. In practice, the executing entity converts the optimized signal control parameters into a command format recognizable by the controller. For example, parameters such as green light duration and signal cycle are encoded according to the controller's command encoding rules to generate specific control command data packets.

[0065] The fifth step involves sending the aforementioned control commands to the corresponding target signal controllers via the established communication links. Each target signal controller includes logic control units, and each logic control unit corresponds to a traffic light. In practice, the executing entity accurately sends the control commands to the corresponding signal controllers through the established communication links. For example, a control command containing phase timing parameters is sent to the signal controller at intersection "ZGC-001". After receiving the command, the controller's internal logic control units drive the traffic lights for different phases, such as those in the east-west and north-south directions.

[0066] The sixth step involves, in response to the aforementioned control commands taking effect, connecting the video streams from the video stream processing units corresponding to the target intersection identifiers and the data streams from the radar data interfaces. In practice, after confirming that the control commands have been successfully executed, the executing entity initiates data acquisition from the intersection's sensing devices to obtain real-time traffic operation status information. For example, it connects the video stream from the camera at intersection "ZGC-001" and the microwave radar detection data to provide data support for subsequent effect evaluation and optimization adjustments.

[0067] The seventh step involves acquiring feedback data streams in real time through the various video and data streams accessed. These feedback data streams include traffic flow, queue length, vehicle speed, and lane occupancy rate for each target signal controller. In practice, the aforementioned implementing entities perform real-time analysis and processing of the collected raw perception data to extract key traffic operation indicators. For example, video analytics algorithms are used to calculate traffic flow at each approach lane, detect vehicle queue length, calculate average vehicle speed, and monitor lane occupancy in real time.

[0068] The eighth step involves coordinating and fine-tuning the aforementioned target signal controllers based on the acquired feedback data stream. In practice, these implementing entities dynamically adjust and optimize signal control parameters according to real-time traffic operation data. For example, they automatically adjust phase sequence or green light time based on detected traffic flow changes, achieving coordinated optimization of signal control at multiple intersections.

[0069] The ninth step involves generating an operation log based on the acquired collaborative fine-tuning process data and feedback data stream, and then transmitting the generated operation log back to the central cloud platform. This operation log is used to optimize the collaborative fine-tuning process. In practice, the aforementioned executing entity records the complete control adjustment process and effect data, forming a structured operation log file, which is then transmitted to the central management platform via the network. For example, it records the time of each parameter adjustment, the content of the adjustment, and changes in traffic indicators before and after the adjustment. This log data will be used for subsequent model optimization and algorithm improvement.

[0070] The steps one through nine described above constitute a technical solution of this disclosure, addressing the technical problems of "inaccurate control command generation, poor device communication compatibility, and lack of real-time status feedback and continuous optimization mechanisms in intelligent traffic signal control." The reasons why existing technologies struggle to achieve accurate, reliable, and adaptive traffic signal control are as follows: traditional control methods often employ fixed timing schemes, failing to adapt to dynamically changing traffic flows; different manufacturers' signal controllers use varying communication protocols, lacking a unified command conversion mechanism, leading to difficulties in system integration; the generation and execution of control commands lack effective verification and confirmation mechanisms, posing a risk of misoperation; the signal control effect lacks real-time, quantitative evaluation methods, making effective feedback optimization difficult; and incomplete control process records prevent the formation of a traceable and analyzable data foundation, hindering continuous system improvement. Eliminating these factors would allow for the construction of a complete traffic signal control system, from accurate command generation to reliable device control and continuous optimization. Therefore, this disclosure: In the first step, by parsing the traffic signal control parameter set within an edge computing gateway, standardized target intersection identifiers are extracted, providing an accurate target location basis for precise device addressing and control command distribution. The second step involves querying the local device registry based on the target intersection identifier to obtain the specific communication protocol specifications of each signal controller, resolving the system integration challenges caused by differences in protocols among multiple vendors. The third step establishes independent communication links with each target signal controller according to the communication protocol specifications, ensuring the reliability and stability of the control command transmission channel. The fourth step generates control commands conforming to the controller interface specifications based on the signal control parameter set, achieving accurate conversion from optimization algorithm output to device-executable commands. The fifth step accurately sends the control commands to the corresponding target signal controllers through the established communication links, and drives specific signal light groups via the controller's internal logic control unit, achieving complete transmission and execution of control intent. The sixth step, after the control commands take effect, connects the video streams and radar data streams from each intersection, constructing a data acquisition foundation for real-time traffic status perception, providing multi-source data support for evaluating control effectiveness. The seventh step, through real-time analysis of the connected video and data streams, extracts multi-dimensional traffic parameters such as traffic flow, queue length, and vehicle speed, forming a feedback data stream for quantitatively evaluating control effectiveness. The eighth step, based on the feedback data stream, performs collaborative fine-tuning processing on each target signal controller, achieving dynamic optimization and adjustment of control parameters. The ninth step involves recording complete collaborative fine-tuning process data and feedback data streams to generate structured operation logs, which are then transmitted back to the central cloud platform. This accumulates valuable real-world scenario data for continuous optimization of control strategies. Thus, five key elements—precise equipment control, reliable communication transmission, real-time status awareness, dynamic collaborative optimization, and complete process recording—are organically combined to construct a complete technology chain from control decision-making to execution feedback.Furthermore, the real-time feedback mechanism and dynamic optimization adjustment capability based on multi-source sensing data enable the system to adapt to changes in traffic flow, improving control effectiveness. Thus, a complete traffic signal control closed loop is constructed, from "precise command generation" to "reliable equipment control" and then to "dynamic optimization feedback," achieving a fundamental improvement in traffic signal control from "static timing" to "dynamic optimization." While ensuring control reliability, this enhances the adaptive capability and long-term optimization potential of the traffic signal system.

[0071] In some optional implementations of certain embodiments, the aforementioned execution entity may perform coordinated fine-tuning of the respective target signal controllers based on the acquired feedback data stream through the following steps: The first step is to standardize the acquired feedback data stream to obtain standardized traffic data. Standardization refers to the process of converting traffic data from different sources and with different units into data with a unified standard and specification. In practice, the implementing entities use data normalization algorithms to uniformly process the multi-source heterogeneous data in the feedback data stream, eliminating differences in units between data. For example, traffic flow data is uniformly converted to "vehicles / hour," queue length data to "meters," and vehicle speed data to "kilometers / hour," forming comparable standardized traffic data.

[0072] The second step involves constructing an optimization model based on the obtained standardized traffic data, with traffic efficiency as the objective. This optimization model is a formal description of the traffic signal control problem as a mathematical optimization problem, including an objective function and constraints. In practice, the implementing entity establishes a mathematical optimization model based on the standardized traffic data, with the core objective of minimizing the total vehicle delay in the region. For example, an optimization model might be constructed with the objective function of minimizing the average vehicle delay time at all intersections in the region, using parameters such as green light duration, signal cycle, and phase difference as decision variables, and the parameter value range and traffic flow conservation as constraints.

[0073] The third step involves parallel processing of the aforementioned optimization problem model to obtain the adjustment amounts of each control parameter corresponding to each target signal controller. Parallel processing refers to simultaneously handling the optimization calculation tasks of multiple intersections to improve computational efficiency. In practice, the execution entity employs a distributed computing architecture, distributing the optimization problems of multiple intersections within a region to different computing nodes for simultaneous solution. For example, the regional optimization problem containing 20 intersections can be decomposed into 20 sub-problems, and multi-core processors can be used to calculate the optimal control parameter adjustment amounts for each intersection in parallel, including the green light duration adjustment values ​​and cycle adjustment values ​​for each phase.

[0074] The fourth step involves generating a coordinated fine-tuning instruction set based on the adjustment amounts of the control parameters corresponding to each target signal controller. Each coordinated fine-tuning instruction in this set corresponds to a specific target signal controller. The coordinated fine-tuning instruction set is a complete set of instructions containing multiple intersection coordinated control commands. In practice, the executing entity encapsulates the calculated parameter adjustment amounts according to a predetermined instruction format to generate control commands that can be directly issued. For example, a coordinated fine-tuning instruction containing information such as intersection identifier, adjustment timestamp, phase green light duration adjustment value, and signal cycle adjustment value is generated for each target signal controller. All instructions are combined to form the coordinated fine-tuning instruction set.

[0075] The fifth step involves verifying the generated coordinated fine-tuning instruction set. Verification refers to the process of checking the safety, rationality, and consistency of the control instructions. In practice, the executing entity employs multiple verification mechanisms to validate the instruction set, including parameter range checks, conflict detection, and safety assessments. For example, it checks whether the green light duration in the instructions is within the allowable range, verifies whether the phase difference settings of adjacent intersections are reasonable, ensures conflict-free signal timing schemes, and marks and corrects instructions with potential safety hazards.

[0076] The sixth step involves responding to the verification process of the coordinated fine-tuning instruction set. Based on each coordinated fine-tuning instruction in the set, and through the established communication links, the corresponding target signal controllers are coordinated for control. In practice, the executing entity accurately sends the verified coordinated fine-tuning instructions to the corresponding target signal controllers via the established communication links. For example, instructions containing coordinated control parameters are sent to the signal controllers at each intersection via a TCP / IP network. Upon receiving the instructions, the controllers immediately update their control parameters, achieving coordinated signal control at multiple intersections within the area and ensuring the continuity and coordination of traffic flow.

[0077] The first to sixth steps described above, as a technical solution of this disclosure embodiment, solve the technical problems of "inconsistent data processing, low optimization calculation efficiency, and lack of security verification for control commands" in traffic signal cooperative control. The reasons why existing technologies struggle to achieve efficient, reliable, and secure regional cooperative signal control are as follows: multi-source traffic data exhibits differences in units and dimensions, leading to calculation deviations in the optimization model when used directly; the regional signal optimization problem is highly complex, and serial calculations cannot meet real-time control requirements; the lack of validity verification during control command generation may result in conflicting timing schemes; and the command issuance mechanism lacks security guarantees, posing a risk of misoperation. If these factors can be eliminated, a complete cooperative control system from data standardization to parallel optimization to secure control can be constructed. Therefore, this disclosure: First, by standardizing the feedback data stream, parameters such as traffic flow, queue length, and vehicle speed are unified into standard units, eliminating the dimensional differences of multi-source data and providing an accurate and reliable data foundation for subsequent optimization calculations. Second, based on standardized traffic data, an optimization problem model with traffic efficiency as the objective is constructed, transforming the complex traffic control problem into a computable mathematical optimization problem, providing a theoretical basis for achieving scientific and precise signal control. The third step involves parallel processing of the optimization problem model, distributing the optimization tasks of multiple intersections within the region to significantly improve computational efficiency and ensure real-time control response. The fourth step generates a collaborative fine-tuning instruction set based on the parameter adjustments of each signal controller, transforming the optimization calculation results into executable control commands, achieving a complete conversion from theoretical optimization to practical control. The fifth step verifies the collaborative fine-tuning instruction set, validating its rationality from multiple dimensions such as parameter range, conflict detection, and security, effectively preventing timing conflicts and safety risks. The sixth step sends the verified instructions to each target signal controller via the communication link, achieving coordinated control of multiple intersections within the region and ensuring the continuity and coordination of traffic flow. Thus, by organically combining the four key aspects of data standardization, parallel optimization calculation, instruction security verification, and collaborative control execution, a highly efficient and reliable regional traffic signal collaborative control system is constructed. Furthermore, the optimization solution based on parallel computing and multiple security verification mechanisms improve the system's real-time performance and reliability while ensuring control effectiveness. Thus, a complete closed loop of regional signal collaborative control was constructed, from "data standardization" to "parallel optimization" and then to "safety control," achieving a fundamental improvement in traffic signal control from "single-point optimization" to "regional collaboration." While ensuring control accuracy, it enhanced the collaborative control capability and operational reliability of the regional traffic signal system.

[0078] In addressing the aforementioned data-matching-based traffic signal control technology issues using technical solutions, specific safety and reliability scenarios arise: when a data-intelligent-driven traffic signal control closed loop performs coordinated fine-tuning in a real road environment, control failures or risks may occur due to system component malfunctions, communication anomalies, or data distortion. These problems are characterized by risks stemming from the unreliability of internal system components and the unpredictability of the external environment, rather than malicious attacks. Traditional control logic based on functional correctness verification is ill-suited to effectively address these challenges. This is often accompanied by the following specific technical problems: the control command's effective status lacks an effective confirmation mechanism, and the system may make subsequent decisions based on the erroneous premise of "sent but not executed"; temporary interruptions or high latency of the communication link can disrupt the integrity and timeliness of regional collaborative control; feedback data streams from sensors such as cameras and radar may be severely distorted due to equipment failure or environmental interference, causing the system to "blind" or make misjudgments; optimization algorithms may generate high-frequency oscillations or logically conflicting fine-tuning commands under complex traffic flow, thereby triggering a chain reaction of deterioration in regional traffic conditions; when the above anomalies occur, the system lacks an automated emergency braking and backoff mechanism, cannot prevent the continuous decline in control performance, and is difficult to quickly restore to a safe and controllable state. The following requirements are necessary for this application scenario: real-time health monitoring capabilities that go beyond simple functional logic and focus on system operational status; a comprehensive, multi-dimensional anomaly detection and diagnosis system encompassing data quality, communication connections, and control logic; fault isolation and system self-healing capabilities to immediately pause optimization processes and automatically switch to a preset safety mode upon detecting clear anomalies or potential risks; and a detailed anomaly event context recording mechanism to provide data support for system stability optimization and fault tracing. We have decided to adopt the following solution: Optionally, the aforementioned implementing entity may also perform the following steps: The first step involves real-time monitoring of the data quality status of the aforementioned feedback data stream, the communication connection status of each of the aforementioned communication links, and the logical security status of the aforementioned collaborative fine-tuning instruction set. Real-time monitoring refers to the continuous collection and evaluation of the core operational indicators of the control system to construct a system health profile. Data quality status is an indicator of the reliability and validity of feedback data; communication connection status is an indicator of the availability of control instruction transmission channels; and logical security status is an indicator of the security and rationality of the control instructions to be executed. In practice, the aforementioned execution entities deploy independent monitoring agents that run in parallel with the core control logic, polling and diagnosing the three statuses at a fixed frequency (e.g., 10 times per second). For example, the monitoring agent will: check whether there are consecutive zero values, outliers exceeding physical limits (such as vehicle speeds exceeding 300 km / h), or severe conflicts between different data sources in the feedback data stream, thereby assessing the data quality status; send heartbeat packets to all target signal controllers, and if no response is received within a set timeout window (such as 3 seconds), the communication connection status of the intersection is determined to be abnormal; for the collaborative fine-tuning instruction set that has passed the above verification, further analyze its triggering frequency within a very short time window (such as 1 minute), and if the frequency exceeds the preset safe frequency threshold (such as 5 times / minute), its logical safety status is determined to be a high-frequency oscillation risk.

[0079] The second step involves pausing the current collaborative fine-tuning process and executing the corresponding fault mitigation strategy in response to the detection of any of the following abnormal operating conditions. An abnormal operating condition refers to any monitored state indicator exceeding the system's preset safe operating boundary, potentially leading to a significant decline in driving safety or traffic efficiency. Pausing means immediately interrupting the aforementioned collaborative fine-tuning process to prevent potentially harmful optimization operations from being performed in an unhealthy state. Fault mitigation strategies refer to one or more pre-designed emergency response plans to address specific abnormal operating conditions, aiming to guide the system to a known, controllable, and safe state. In practice, the aforementioned execution entity embeds a state decision-maker, whose input receives the monitoring results from the first step and compares them in real time with preset trust thresholds, timeout limits, and safe frequency thresholds. For example, if the decision-maker detects that the vehicle queue length data from a radar at a certain intersection changes drastically more than 5 times within 10 seconds (data quality status is unreliable), or detects that the heartbeat of the communication link with the "Zhongguancun Street-Haidian South Road" intersection times out (communication connection status is interrupted), or detects that the system generates 4 phase difference adjustment commands for the same intersection within 30 seconds (logic safety status is high frequency), the decision-maker will immediately send an "emergency stop" signal to the core control logic to suspend fine-tuning and index the corresponding fault mitigation strategy according to the anomaly type.

[0080] The third step is to execute the corresponding fault mitigation strategy. This strategy includes at least one of the following: a basic guarantee timing scheme, abnormal status alarms, and recording abnormal event context. The basic guarantee timing scheme refers to a pre-designed signal timing scheme stored locally, which does not aim for optimal efficiency but absolutely guarantees basic safety and orderly traffic flow at the intersection. An abnormal status alarm is an emergency notification issued to system maintenance personnel, containing details and location of the abnormality. Recording abnormal event context involves sealing all system operation data for a period before and after the abnormality occurs to form a complete chain of evidence for post-event analysis. In practice, the executing entity executes a pre-defined combination of strategies based on the type and severity of the abnormal condition. For example, activating the pre-stored basic guarantee timing scheme: When a large-scale failure in data quality is detected at a core intersection in the area, the executing entity immediately retrieves the "rain mode" or "night mode" basic guarantee timing scheme for that area from the read-only memory of the edge computing gateway, overriding the dynamic scheme generated by the optimization model, ensuring that the traffic lights cycle at least at a conservative but safe pace. An abnormal status alarm is sent to the traffic control center, requesting manual intervention: While executing the above-mentioned scheme switch, the executing entity automatically generates an abnormal status alarm message containing "Abnormal type: Data quality deterioration, affected intersections: A, B, C, backup scheme activated," and pushes it to the traffic control center's large screen system via a dedicated network channel, requesting manual intervention for advanced diagnosis and decision-making. The abnormal event context is recorded in the above-mentioned operation log for subsequent diagnosis and analysis: Throughout the process, the executing entity packages all relevant data (including raw feedback data, intermediate calculation results, generated instructions, communication logs, decision-maker status, etc.) from 30 seconds before the abnormality trigger to 10 seconds after the mitigation strategy is implemented into an immutable data packet, which is stored as part of the operation log. For example, this data packet can be used for subsequent analysis to determine whether the root cause of this data distortion is sensor hardware failure or the influence of severe weather.

[0081] The first to third steps described above, as a technical solution of this disclosure embodiment, solve the technical problem of "control failure, performance degradation, and safety risks caused by internal faults or external interference in intelligent traffic signal control closed-loop systems under real and complex environments." The reasons why data-driven intelligent control methods in the prior art struggle to maintain high reliability and resilience in actual deployment are as follows: systems are typically designed based on the assumption that "components are always normal and the environment is always ideal," lacking a normalized response mechanism to abnormal conditions such as data distortion, communication interruptions, and algorithm oscillations; when a local fault occurs, there is a lack of system-level "emergency stop" and "rollback" capabilities, which may lead to the spread and amplification of the fault within the control closed loop; once the optimization process starts, it is difficult to automatically stop, and "negative optimization" may be performed under poor data-driven conditions; at the same time, the system lacks detailed abnormal context records, making fault diagnosis and system optimization lack data support. If the above factors can be eliminated, a resilient control system with fault self-awareness, decision self-constraint, and risk self-avoidance capabilities can be constructed. To this end, this disclosure includes the following steps: First, by real-time monitoring of the data quality status of the feedback data stream, the connection status of the communication link, and the logical security status of the fine-tuning instruction set, a real-time system health profile covering the entire "data-communication-decision" link is constructed. This achieves a fundamental shift from simple functional execution to deep perception of the system's operational status, providing accurate judgment criteria for timely detection of potential risks. Second, by immediately suspending the current collaborative fine-tuning process in response to any detected abnormal operating condition, a crucial "safety braking" mechanism is introduced into the highly automated control closed loop. This step effectively prevents the continuous execution of potentially harmful "optimization" operations in cases of system perception distortion, instruction failure, or decision instability, preventing local anomalies from evolving into systemic loss of control. Third, by executing corresponding fault mitigation strategies, a multi-layered, automated system safety network is constructed. Enabling pre-stored basic guarantee timing schemes ensures that the intersection can still operate in a known safe state when the intelligent algorithm fails, achieving seamless degradation from "optimal control" to "safety assurance." Sending abnormal status alarms to the traffic control center and requesting manual intervention establishes a human-machine collaborative emergency response channel, enhancing the system's ability to handle complex faults. Recording the context of abnormal events in the operational log accumulates valuable real-world data for subsequent root cause analysis, model iteration, and threshold optimization, driving continuous safety evolution of the system. This organically combines the three core components of end-to-end status monitoring, intelligent safety braking, and multi-strategy fault mitigation, adding a crucial safety assurance process. Furthermore, rapid response based on real-time diagnostics and automatic switching of preset safety schemes ensures system functional continuity while enhancing its resilience and survivability against internal faults and external interference.Thus, a complete next-generation intelligent traffic control system architecture has been constructed, encompassing "functional implementation," "state perception," and "safety assurance." This has enabled a leap in traffic signal control from "automation" to "autonomy and resilience," and while pursuing optimal control efficiency, it has built a solid foundation for the stable, reliable, and safe operation of urban traffic.

[0082] The above embodiments of this disclosure have the following beneficial effects: The above embodiments of this application have the following beneficial effects: Through the traffic signal control method based on data matching of this application, intelligent matching and precise application of multi-source data in complex traffic scenarios are realized, effectively improving the intelligence level and scientific nature of traffic signal control. Specifically, traditional data retrieval methods rely on direct keyword matching or limited document retrieval, which may lead to insufficient semantic understanding, difficulties in cross-domain association, and omission of valuable data, making it difficult to meet the needs of intelligent transportation systems for in-depth data mining and divergent applications. Based on this, the traffic signal control method based on data matching of this application embodiments: First, by constructing a semantic association network containing traffic control domain nodes and keyword nodes, a cross-disciplinary, multi-level semantic understanding framework is established. This step breaks down the domain barriers in traditional retrieval at the knowledge system level, placing traffic data in a broader knowledge context, laying a solid foundation for subsequent intelligent association and divergent matching, and overcoming the problem of single-domain knowledge limitations. Then, based on the acquired traffic datasets, a data supply label set is generated and linked to the semantic network, realizing deep semantic annotation and knowledge integration of the original data resources. This process automatically mines the connotation of data through pre-trained large models, transforming simple dataset names and fields into rich semantic labels and making them an organic component of the knowledge network. This solves the bottleneck of traditional methods, which rely on shallow semantic expression and manual annotation. Next, a data demand label set is generated based on the acquired traffic control demand information and linked to the semantic network, achieving accurate parsing and structured expression of complex user intentions. This mechanism transforms control demands described in natural language into semantic nodes recognizable by the network, enabling precise understanding and location of ambiguous and divergent user needs, overcoming the failure of traditional keyword matching when terms are inconsistent. Then, a target traffic dataset is generated based on the data supply label set and the data demand label set, achieving intelligent data discovery based on semantic association. This step, through graph search and path calculation in the knowledge network, can discover potentially related datasets with semantically similar user needs but different keywords, achieving intelligent retrieval "from point to network," significantly improving the recall and novelty of data discovery. Subsequently, a set of traffic signal control parameters is generated based on the target traffic dataset, completing the value transformation from data resources to control strategies. This process transforms the matched high-quality data resources into specific, executable signal control parameters, establishing a crucial bridge for data-driven decision-making and enabling the value of data to be concretely realized in traffic control practice. Finally, based on the set of traffic signal control parameters, each traffic light is controlled, achieving data-driven precision traffic management. This step directly applies the data processing results to physical traffic facilities, forming a complete technical closed loop of "data perception - intelligent matching - strategy generation - precise execution," providing an innovative path for building intelligent traffic control systems.In summary, the embodiments of this application achieve in-depth mining and efficient utilization of traffic data resources through a collaborative technical route of "semantic network construction - supply and demand tag generation - intelligent matching - parameter transformation - precise control". This method not only improves the intelligence of data retrieval and solves the shortcomings of traditional methods in semantic understanding and cross-domain correlation, but also effectively transforms data value into control benefits through a complete technical closed loop, providing an innovative technical paradigm for the construction of intelligent transportation systems.

[0083] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a traffic signal control device based on data matching, which are similar to... Figure 1 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0084] like Figure 2 As shown, a traffic signal control device 200 based on data matching in some embodiments includes: a construction unit 201, a first association unit 202, a second association unit 203, a first generation unit 204, a second generation unit 205, and a control unit 206. The system comprises the following components: a construction unit 201, configured to construct a semantic association network, which includes traffic control domain nodes and keyword nodes; a first association unit 202, configured to generate a data supply tag set based on the acquired traffic datasets, and link the data supply tags in the data supply tag set to the semantic association network; a second association unit 203, configured to generate a data demand tag set based on the acquired traffic control demand information, and link the data demand tags in the data demand tag set to the semantic association network; a first generation unit 204, configured to generate a target traffic dataset based on the data supply tag set and the data demand tag set; a second generation unit 205, configured to generate a traffic signal control parameter set based on the target traffic dataset; and a control unit 206, configured to control each traffic light based on the traffic signal control parameter set.

[0085] It is understandable that the units described in the device 200 are related to the reference. Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 200 and the units contained therein, and will not be repeated here.

[0086] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device 300 suitable for implementing some embodiments of the present disclosure. Figure 3The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0087] like Figure 3 As shown, the electronic device 300 may include a processing unit 301 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0088] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0089] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0090] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0091] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0092] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: construct a semantic association network, wherein the semantic association network includes various traffic control domain nodes and various keyword nodes; generate a data supply label set based on the acquired traffic datasets, and link each data supply label in the data supply label set to the semantic association network; generate a data demand label set based on the acquired traffic control demand information, and link each data demand label in the data demand label set to the semantic association network; generate a target traffic dataset based on the data supply label set and the data demand label set; generate a traffic signal control parameter set based on the target traffic dataset; and control each traffic light based on the traffic signal control parameter set.

[0093] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0094] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0095] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a building unit, a first association unit, a second association unit, a first generation unit, a second generation unit, and a control unit. The names of these units do not necessarily limit the unit itself; for example, a building unit may also be described as a "unit for building a semantic association network."

[0096] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0097] Some embodiments of this disclosure also provide a computer program product, including a computer program that, when executed by a processor, implements any of the above-described data-matching-based traffic signal control methods.

[0098] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A traffic signal control method based on data matching, comprising: Construct a semantic association network, wherein the semantic association network includes nodes of various traffic control domains and nodes of various keywords; Based on the acquired traffic datasets, a data supply label set is generated, and each data supply label in the data supply label set is linked to the semantic association network. Based on the acquired traffic control demand information, a data demand tag set is generated, and each data demand tag in the data demand tag set is linked to the semantic association network. Generate a target traffic dataset based on the data supply label set and the data demand label set; Based on the target traffic dataset, a set of traffic signal control parameters is generated; Based on the set of traffic signal control parameters, each traffic light is controlled.

2. The method according to claim 1, wherein, The construction of the semantic association network includes: Construct a framework that includes nodes from various traffic control domains; Based on the constructed framework, generate each keyword node; By linking each keyword node to each traffic control domain node and assigning weights to all connecting edges, a semantic association network is obtained.

3. The method according to claim 1, wherein, The step of generating a data supply label set based on the acquired traffic datasets, and linking each data supply label in the data supply label set to the semantic association network, includes: Obtain various traffic datasets; For each traffic dataset obtained, perform the following steps: The attribute information of the acquired traffic dataset is organized into structured text information, including name information, data description and data field list; The structured text information is input into a pre-trained large model to obtain the data supply labels corresponding to the traffic dataset. The obtained data supply labels corresponding to each traffic dataset are integrated into a data supply label set; The data supply tags in the data supply tag set are matched with the keyword nodes or traffic control domain nodes in the semantic association network, and the data supply tags are linked with the matched nodes.

4. The method according to claim 3, wherein, The method further includes: In response to any data supply tag among the various data supply tags failing to match any node in the semantic association network, a corresponding keyword node is created in the semantic association network; The data supply tags are linked with corresponding keyword nodes, and the semantic association network is updated.

5. The method according to claim 1, wherein, The step of generating a data demand tag set based on the acquired traffic control demand information, and linking each data demand tag in the data demand tag set to the semantic association network, includes: Obtain traffic control demand information; The traffic control demand information is input into a pre-trained large model to obtain a data demand label set, wherein the data demand label set includes various data demand labels; The data requirement tags are matched with the keyword nodes in the semantic association network; In response to each data requirement tag in the data requirement tags being successfully matched with any keyword in the semantic association network, each data requirement tag is linked with each matching keyword node; In response to any data requirement tag failing to match any keyword in the semantic association network, the data requirement tag is linked to a preset general node, or the data requirement tag is removed from the data requirement tag set.

6. The method according to claim 1, wherein, The step of generating a target traffic dataset based on the data supply label set and the data demand label set includes: Each keyword node linked to each data requirement tag in the data requirement tag set is determined as the search starting node set; Each node linked to each data supply tag in the data supply tag set is determined as a candidate target node set; Based on the set of search starting nodes and the set of candidate target nodes, generate the path distance from each search starting node to each candidate target node; For each traffic dataset in the acquired traffic datasets, perform the following operations: Obtain each candidate target node linked to the traffic dataset; Based on the acquired candidate target nodes and the path distances corresponding to each candidate target node, a comprehensive association distance corresponding to the traffic dataset is generated. Based on the comprehensive association distances corresponding to each traffic dataset, each traffic dataset that meets the preset comprehensive association distance conditions is selected, and the selected traffic datasets are integrated into a candidate result set. The candidate result set is sorted and organized, and the sorted candidate result set is determined as the target traffic dataset.

7. The method according to claim 1, wherein, The step of generating a set of traffic signal control parameters based on the target traffic dataset includes: For each traffic dataset in the target traffic dataset, perform the following operations: Read the raw data content from the traffic dataset; The original data content is preprocessed to generate regularized data; The regularized data is input into a preset signal control optimization model to obtain preliminary control parameters; Based on the preliminary control parameters, a subset of signal control parameters corresponding to the traffic dataset is generated; The various subsets of signal control parameters are detected and processed to obtain the set of traffic signal control parameters.

8. A traffic signal control device based on data matching, comprising: The building unit is configured to build a semantic association network, wherein the semantic association network includes various traffic control domain nodes and various keyword nodes; The first association unit is configured to generate a data supply label set based on the acquired traffic datasets, and to link each data supply label in the data supply label set to the semantic association network. The second association unit is configured to generate a data demand tag set based on the acquired traffic control demand information, and to link each data demand tag in the data demand tag set to the semantic association network. The first generation unit is configured to generate a target traffic dataset based on the data supply label set and the data demand label set; The second generation unit is configured to generate a set of traffic signal control parameters based on the target traffic dataset; The control unit is configured to control each traffic light based on the set of traffic signal control parameters.

9. An electronic device, comprising: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images