Signal timing automatic optimization method and system based on generative model
By using a generative model-based automatic signal timing optimization method, multi-source traffic data and federated learning algorithms are employed to achieve cross-domain collaborative optimization of multiple intersections. This solves the problem of lack of collaborative optimization in modern urban traffic management and improves the real-time performance and efficiency of the traffic system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- AI SUPER EYE TECH CO LTD
- Filing Date
- 2026-01-09
- Publication Date
- 2026-04-24
AI Technical Summary
Existing technologies lack multi-intersection cross-domain collaborative optimization mechanisms, making it difficult to adapt to the refined and real-time management needs of modern urban traffic. This results in low traffic flow prediction accuracy, a lack of forward-looking adaptation of signal timing schemes, low utilization of intersection resources, and a tendency to cause congestion or excessively long waiting times.
An automatic signal timing optimization method based on generative models is adopted. By extracting the temporal, spatial and dynamic impact features of multi-source traffic data at intersections, a mapping and adaptation relationship between intersection traffic scenarios and signal timing parameters is established. Then, federated learning algorithms are used to perform cross-domain collaborative optimization of multiple intersections in the cloud. Combined with visual configuration and distributed storage architecture, dynamic adjustment and optimization of signal timing are realized.
It improves the adaptability of the transportation system to dynamic scenarios, enhances the traffic efficiency of intersections and road networks, reduces vehicle waiting time and the probability of congestion, and supports the refined management of intelligent transportation.
Smart Images
Figure CN121921982A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of traffic signal control technology, specifically to a method and system for automatic optimization of signal timing based on a generative model. Background Technology
[0002] In urban traffic management, traffic flow prediction and signal timing optimization are core components for improving traffic efficiency and alleviating traffic congestion. Traditional traffic flow prediction technologies have significant limitations: on the one hand, relying on statistical models or traditional machine learning models, they can only use historical traffic flow time-series data for single-dimensional prediction, failing to effectively integrate dynamic influencing factors such as real-time road conditions, weather conditions, and emergencies, resulting in low prediction accuracy and poor timeliness. For example, it is difficult to accurately predict sudden changes in traffic flow during the evening rush hour in rainy weather. On the other hand, existing signal timing schemes are mostly fixed timings or semi-dynamic adjustments based on human experience, lacking deep linkage with traffic flow prediction results. They cannot adaptively optimize according to real-time traffic changes and differences in intersection traffic demand, resulting in low intersection resource utilization and easily causing local congestion or excessively long waiting times.
[0003] With the improvement of urban traffic monitoring networks and breakthroughs in generative AI technology, massive amounts of multi-source traffic data have been accumulated. However, current technologies have failed to fully explore the collaborative value of multi-source data. Traffic flow prediction and signal timing optimization are disconnected from each other. Prediction results cannot directly guide timing adjustments, and timing schemes lack forward-looking adaptation to future traffic changes. They cannot comprehensively improve the overall traffic efficiency of the road network and are unable to meet the needs of modern urban traffic for refined and real-time management.
[0004] In summary, existing technologies lack a multi-intersection cross-domain collaborative optimization mechanism, making it difficult to adapt to the technical needs of modern urban traffic's refined and real-time management. Summary of the Invention
[0005] This application provides a signal timing automatic optimization method and system based on a generative model, aiming to solve the technical problem in the existing technology of lacking a multi-intersection cross-domain collaborative optimization mechanism, which makes it difficult to adapt to the needs of modern urban traffic refined and real-time management.
[0006] In view of the above problems, the technical solution to achieve the present application is as follows:
[0007] In a first aspect, this application provides an automatic signal timing optimization method based on a generative model. The method includes: extracting traffic flow features from multi-source traffic data at an intersection using a timing optimization generative model, wherein the traffic flow features include temporal flow characteristics, spatial correlation characteristics, and dynamic impact characteristics; determining a mapping and adaptation relationship between the intersection traffic scenario and signal timing parameters based on the traffic flow features; setting the signal timing process in a traffic signal control interface using the mapping and adaptation relationship and a visual configuration function; automatically filling initial timing parameters in a parameter preset unit based on the mapping and adaptation relationship; transmitting the signal timing process to a cloud-based timing optimization center; and performing cross-domain collaborative optimization of the signal timing process across multiple intersections using a federated learning algorithm in the cloud-based timing optimization center to obtain the optimal timing strategy and optimal timing parameters.
[0008] Preferably, the execution effect of the signal timing process is monitored in real time by traffic flow detectors deployed at multiple intersections to obtain monitoring data; the monitoring data of the traffic flow detectors is then uploaded to the cloud-based timing optimization center.
[0009] Preferably, the monitoring data is stored in a traffic operation log database; a backend traffic data cluster based on a distributed storage architecture is established, and the traffic operation log database is associated with a storage node in the backend traffic data cluster.
[0010] Preferably, a distributed consensus algorithm is used to synchronously store the optimal timing strategy to each storage node and generate a strategy version identifier and blockchain record.
[0011] Preferably, the signal timing task is classified and marked according to its service priority and impact range to divide the release branch types; the release branch types are sorted and entered into the release queue in sequence according to the sorting results, and a conditional branch release strategy is configured.
[0012] Preferably, the conditional branch release strategy is used to release the optimal timing strategy corresponding to the high-priority signal timing task under the business priority mark to the core business branch under the influence range mark for use by the traffic control system.
[0013] Preferably, the monitoring data includes the execution status and traffic quality data of the signal timing process, wherein the traffic quality data includes average delay time, queue length, and traffic efficiency improvement rate; the monitoring data is input into the timing optimization generation model, which analyzes the performance bottleneck of the signal timing process based on preset traffic rules and reinforcement learning reward function, generates the optimal timing strategy for the performance bottleneck, and pushes it to the traffic control terminal.
[0014] In a second aspect, this application provides an automatic signal timing optimization system based on a generative model. The system includes: a traffic flow feature extraction module, used to extract traffic flow features from multi-source traffic data at intersections using a timing optimization generative model, wherein the traffic flow features include temporal flow features, spatial correlation features, and dynamic impact features; a mapping adaptation relationship determination module, used to determine the mapping adaptation relationship between intersection traffic scenarios and signal timing parameters based on the traffic flow features; a signal timing process setting module, used to set the signal timing process in a traffic signal control interface using the mapping adaptation relationship and a visual configuration function; an automatic parameter filling module, used by a parameter preset unit to automatically fill initial timing parameters according to the mapping adaptation relationship and transmit the signal timing process to a cloud-based timing optimization center; and a collaborative optimization module, used in the cloud-based timing optimization center to perform cross-domain collaborative optimization of the signal timing process across multiple intersections using a federated learning algorithm to obtain the optimal timing strategy and optimal timing parameters.
[0015] In summary, one or more technical solutions provided in this application extract multi-dimensional traffic flow features through a timing optimization generation model, construct a mapping relationship between intersection traffic scenarios and timing parameters, and use a federated learning algorithm to achieve cross-domain collaborative optimization of multiple intersections, thereby balancing the traffic demand of each intersection, enhancing the adaptability of the traffic system to dynamic scenarios, and providing technical support for the refined management of intelligent transportation. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0017] Figure 1 This application provides a flowchart illustrating the automatic signal timing optimization method based on a generative model.
[0018] Figure 2 This application provides a schematic diagram of the structure of an automatic signal timing optimization system based on a generative model.
[0019] Figure labeling: Traffic flow feature extraction module 11, mapping and adaptation relationship determination module 12, signal timing process setting module 13, parameter auto-fill module 14, collaborative optimization module 15. Detailed Implementation
[0020] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0021] Example 1: The present application will be described in detail below with reference to the accompanying drawings, as follows... Figure 1 As shown, this application provides an automatic signal timing optimization method based on a generative model, the method comprising:
[0022] Using a timing optimization generation model, traffic flow features of multi-source traffic data at intersections are extracted. These traffic flow features include temporal flow features, spatial correlation features, and dynamic impact features. Based on these traffic flow features, the mapping and adaptation relationship between intersection traffic scenarios and signal timing parameters is determined.
[0023] In one embodiment, the timing optimization generation model is a model built based on deep learning or machine learning technology. Its core function is to extract valuable features from massive traffic data and generate signal timing parameters that match these features. By learning the patterns in historical traffic data, the timing optimization generation model can identify the changing patterns of traffic flow under different time, space and dynamic conditions. Multi-source traffic data refers to a comprehensive data set from different sensors, monitoring equipment and external environmental data, including meteorological conditions and event information.
[0024] Temporal flow characteristics reflect the changing patterns of traffic flow over time, such as traffic fluctuations during morning and evening rush hours; spatial correlation characteristics describe the mutual influence of traffic flow between different intersections or road segments, such as the overflow effect of traffic flow at adjacent intersections; dynamic impact characteristics consider the immediate impact of real-time road conditions, weather conditions, and emergencies on traffic flow, such as the decrease in vehicle speed and changes in traffic flow caused by rainfall; mapping and adaptation relationships refer to the correspondence established between the extracted traffic flow characteristics and signal timing parameters. Through this relationship, signal timing parameters can be dynamically adjusted according to real-time traffic scenarios to optimize traffic flow.
[0025] Optionally, the timing optimization generation model first processes multi-source traffic data at the intersection, including traffic flow sensor data, video surveillance data, meteorological data, and surrounding event information. The timing optimization generation model uses complex algorithms to extract temporal flow characteristics, spatial correlation characteristics, and dynamic impact characteristics from this data. For example, by analyzing historical traffic flow data, the model can identify the periodic changes in traffic flow during morning and evening peak hours, i.e., temporal flow characteristics; by analyzing the flow overflow and transfer between adjacent intersections, it determines the mutual influence between different intersections, i.e., spatial correlation characteristics; and simultaneously, by combining real-time meteorological data and emergency event information, it assesses the immediate impact of these dynamic factors on traffic flow, i.e., dynamic impact characteristics.
[0026] Based on these characteristics, the timing optimization generation model determines the mapping and adaptation relationship between intersection traffic scenarios and signal timing parameters. This makes signal timing no longer a fixed adjustment or one based on human experience, but rather dynamically optimized according to real-time traffic scenarios. For example, if a significant increase in traffic flow in a certain direction is detected during the morning rush hour, and traffic overflow is found at adjacent intersections, along with rainy weather, the timing optimization generation model will calculate an optimal signal timing scheme based on these characteristics. This will increase the green light time for that direction and adjust the signal timing for other directions, improving the intersection's traffic efficiency and reducing vehicle waiting time and congestion probability. This dynamic optimization mechanism not only improves the traffic efficiency of a single intersection but also provides basic data support for subsequent multi-intersection collaborative optimization, thereby achieving refined and real-time management of the entire road network.
[0027] In the traffic signal control interface, the signal timing process is set through the mapping and adaptation relationship and the visual configuration function; the parameter preset unit automatically fills in the initial timing parameters according to the mapping and adaptation relationship and transmits the signal timing process to the cloud timing optimization center; in the cloud timing optimization center, the signal timing process is optimized across multiple intersections using a federated learning algorithm to obtain the optimal timing strategy and optimal timing parameters.
[0028] In one embodiment, the traffic signal control interface refers to a visual operating platform for managing and adjusting traffic signal timings. It typically provides graphical tools that enable traffic management personnel to intuitively set and adjust signal timing parameters. The visual configuration function refers to the ability to set signal timing parameters through a graphical interface (such as drag, slider, chart, etc.), which allows non-technical users to operate easily. The parameter preset unit is responsible for automatically generating initial signal timing parameters based on the mapping adaptation relationship, providing a basic configuration for subsequent optimization.
[0029] The cloud-based timing optimization center receives and processes signal timing data from multiple intersections and optimizes it using advanced algorithms. Federated learning is a distributed machine learning method that allows multiple data sources (such as signal timing data from different intersections) to collaboratively train models without sharing the original data, thereby achieving cross-domain collaborative optimization. The optimal timing strategy and optimal timing parameters refer to the signal timing scheme and specific parameters that maximize the traffic efficiency of the intersection after optimization by the federated learning algorithm.
[0030] Optionally, in the traffic signal control interface, operators can use the visual configuration function to set the signal timing process according to the mapping adaptation relationship. For example, operators can intuitively select the intersection direction, adjust the green light duration, and set the phase sequence through the interface. The interface will provide real-time feedback on the effect of the adjustment, such as the expected improvement in traffic efficiency or reduction in delay time. The parameter preset unit automatically fills in the initial timing parameters according to the mapping adaptation relationship, providing a reasonable starting point for the signal timing process. Specifically, if the traffic scenario of a certain intersection is identified as high traffic volume during the morning rush hour, the parameter preset unit will automatically fill in the initial timing parameters that match the scenario, such as increasing the green light time of the main direction and decreasing the green light time of the secondary direction. These initial parameters are then transmitted to the cloud timing optimization center.
[0031] At the cloud-based signal timing optimization center, federated learning algorithms perform cross-domain collaborative optimization of the signal timing process. Specific signal timing data for multiple intersections are uploaded to the cloud. Through distributed computing, the federated learning algorithm analyzes the mutual influence between intersections, such as traffic overflow and demand shifts, thereby adjusting the signal timing parameters of each intersection to achieve optimal traffic efficiency for the entire road network. For example, in an area containing multiple intersections, if the traffic flow at one intersection suddenly increases, the federated learning algorithm will automatically adjust the signal timing of adjacent intersections to guide traffic dispersion, avoid congestion, achieve road network-level optimization, and improve the overall operational efficiency of the traffic system. At the same time, the federated learning algorithm protects data privacy by avoiding direct sharing of raw data.
[0032] Furthermore, this application provides a method for cross-domain collaborative optimization of the signal timing process across multiple intersections using a federated learning algorithm at the cloud-based timing optimization center. The method includes:
[0033] The execution effect of the signal timing process is monitored in real time by traffic flow detectors deployed at multiple intersections to obtain monitoring data; the monitoring data of the traffic flow detectors is then uploaded to the cloud-based timing optimization center.
[0034] In one embodiment, a traffic flow detector refers to a sensor device installed at an intersection or road segment to monitor traffic parameters such as traffic flow, vehicle speed, and queue length in real time. The traffic flow detector integrates a geomagnetic induction coil, a video surveillance camera, a radar sensor, and a microwave detector, and can provide high-precision real-time traffic data. Real-time monitoring refers to the continuous collection of traffic data through these detectors and the transmission of the data to a processing system in order to reflect changes in traffic flow in a timely manner.
[0035] Monitoring data refers to traffic parameter data collected by traffic flow detectors, including flow rate, vehicle speed, queue length, and delay time. This data reflects the actual implementation effect of the signal timing scheme. The cloud-based timing optimization center receives and processes signal timing data from multiple intersections and optimizes it using advanced algorithms. This platform can handle large-scale data and supports complex computational tasks, such as the operation of federated learning algorithms.
[0036] Optionally, traffic flow detectors are deployed at multiple intersections. These detectors monitor the execution effect of the signal timing process in real time. For example, geomagnetic induction coils can detect the frequency and speed of vehicles passing through intersections, video surveillance cameras can analyze queue length and vehicle type distribution, and radar sensors can measure vehicle speed and vehicle spacing. The collected monitoring data is transmitted to the cloud-based timing optimization center in real time. After receiving this data, the cloud-based timing optimization center stores it in a distributed database for subsequent analysis and optimization. Through real-time monitoring and data uploading, the signal timing scheme can be dynamically adjusted to adapt to real-time changes in traffic flow.
[0037] Furthermore, the method described in this application includes:
[0038] The monitoring data is stored in a traffic operation log database; a backend traffic data cluster based on a distributed storage architecture is established, and the traffic operation log database is associated with a storage node in the backend traffic data cluster.
[0039] In one embodiment, the traffic operation log database refers to a database used to store monitoring data collected by traffic flow detectors and signal timing-related log information, recording the operating status and historical data of the traffic system; the distributed storage architecture is an architecture that distributes data across multiple storage nodes, achieving high availability, scalability, and fault tolerance through distributed computing and storage technologies; the backend traffic data cluster is a distributed system composed of multiple storage nodes, used for centralized management and storage of traffic data, supporting efficient processing and querying of large-scale data; storage node association refers to connecting the traffic operation log database with one or more storage nodes in the backend traffic data cluster, enabling data to be distributed for storage and management while ensuring data consistency and integrity.
[0040] Optionally, monitoring data is first stored in a traffic operation log database, including traffic flow, vehicle speed, queue length, and signal timing parameters, recording the real-time operating status and historical changes of the traffic system. For example, data collected by traffic flow detectors at each intersection is written to the traffic operation log database in real time, forming a complete traffic operation log. A backend traffic data cluster based on a distributed storage architecture is established, and the traffic operation log database is associated with a storage node in this cluster. Specifically, the backend traffic data cluster distributes data across multiple storage nodes through a distributed file system or distributed database. Each storage node is responsible for storing a portion of the data, and data consistency is ensured through a distributed consensus algorithm. For example, in an urban traffic system, the traffic operation log database is connected to a storage node in the backend cluster, and monitoring data is synchronized to that storage node in real time and automatically backed up to other nodes through the mechanism of the distributed storage architecture, ensuring high availability and fault tolerance of the data.
[0041] Preferably, through a distributed storage architecture, the traffic operation log database can efficiently process large-scale data and support high-concurrency read and write operations. In addition, distributed storage also improves the fault tolerance of data. Even if a storage node fails, other nodes can still ensure the integrity and availability of data, which significantly enhances the data management capabilities of the traffic system and provides a data foundation for signal timing optimization and traffic analysis.
[0042] Furthermore, this application provides a method for associating the traffic operation log database with a storage node in the backend traffic data cluster, the method comprising:
[0043] Using a distributed consensus algorithm, the optimal timing strategy is synchronously stored on each storage node, and a strategy version identifier and blockchain record are generated.
[0044] In one embodiment, a distributed consensus algorithm is an algorithm used in a distributed system to ensure data consistency among multiple nodes. Examples include the RAFT algorithm, the Paxos algorithm, or consensus mechanisms in blockchains. Through communication and voting mechanisms between nodes, it guarantees data consistency and integrity in a distributed environment. An optimal timing strategy refers to a signal timing scheme calculated by an optimization algorithm (such as federated learning) that maximizes intersection traffic efficiency. A strategy version identifier is a unique identifier assigned to each updated optimal timing strategy to distinguish different versions of the strategy, facilitating tracking and management. A blockchain record refers to storing the update record of the optimal timing strategy on the blockchain using blockchain technology, ensuring the immutability and traceability of the record. The distributed ledger characteristics of blockchain guarantee data security and transparency.
[0045] Optionally, after the cloud-based timing optimization center calculates the optimal timing strategy using a federated learning algorithm, the system uses a distributed consensus algorithm to synchronously store the strategy on each storage node of the backend traffic data cluster. The specific process is as follows: The optimal timing strategy is first generated on the primary storage node and broadcast within the cluster using a distributed consensus algorithm. Each storage node receives and verifies the validity of the strategy. Once consensus is reached, all nodes synchronously update the strategy. For example, in a cluster with 10 storage nodes, after the primary node generates the optimal timing strategy, it uses the Raft algorithm for multiple rounds of voting to ensure that more than half of the nodes confirm the strategy before all nodes synchronously update it.
[0046] To facilitate management and tracking, the system assigns a unique version identifier to each updated optimal timing strategy. For example, the version identifier can be a timestamp combined with a strategy number. Update records of the optimal timing strategy are stored on the blockchain. Each update record contains information such as the strategy version identifier, update time, and strategy summary. The distributed ledger nature of the blockchain ensures the immutability and traceability of these records. For instance, when the optimal timing strategy is updated, the system generates a blockchain transaction containing the version identifier and strategy summary, broadcasts and verifies it through the blockchain network, and ultimately stores it in the blockchain ledger. This ensures the high availability and data consistency of the optimal timing strategy, while the immutable records provided by blockchain technology enhance the system's security and transparency.
[0047] Furthermore, the method described in this application includes:
[0048] The signal timing task is classified and marked according to its service priority and impact range, and the release branch type is divided. The release branch types are sorted and entered into the release queue in sequence according to the sorting results, and the conditional branch release strategy is configured.
[0049] In one embodiment, the service priority of a signal timing task refers to the classification based on the importance and urgency of the signal timing task. For example, a timing task at a critical intersection may be marked as high priority, while a task at a secondary intersection may be marked as low priority. The scope of influence refers to the geographical area or the degree of influence of the signal timing task on traffic flow. For example, a timing task at an intersection may only affect a local area, while a regional timing task may affect the traffic flow of the entire city.
[0050] Classification tagging refers to classifying signal timing tasks according to business priority and scope of impact, and assigning corresponding tags; release branch type refers to different task processing paths divided according to classification tags. For example, high-priority tasks may enter the fast processing branch, while low-priority tasks enter the regular processing branch; conditional branch release strategy refers to formulating different release strategies according to the priority and scope of impact of tasks to ensure that tasks can be released and executed in a predetermined order and under predetermined conditions.
[0051] Optionally, signal timing tasks are first categorized and labeled according to business priority and impact scope. For example, tasks are graded based on their urgency and impact scope, such as whether they involve traffic congestion, emergencies, or affect major traffic arteries. High-priority tasks include emergency response or optimization tasks at critical intersections, while low-priority tasks include routine maintenance or adjustments at minor intersections. The tasks are then divided into release branch types based on the categorization labels and sorted. For example, high-priority tasks are assigned to the fast processing branch and will be prioritized for release. Low-priority tasks enter the regular processing branch and are queued for processing in normal order. The sorting process is typically based on the urgency and impact scope of the tasks to ensure rapid response to critical tasks. Tasks are then added to the release queue sequentially according to the sorting results, and conditional branch release strategies are configured. For example, high-priority tasks may be released immediately when specific conditions are met, while low-priority tasks are released during regular maintenance periods, ensuring rapid response to critical tasks. This ensures the flexibility and adaptability of task processing and improves the overall management efficiency of the traffic signal timing system.
[0052] Furthermore, the method described in this application includes:
[0053] The conditional branching release strategy is used to prioritize the release of the optimal timing strategy corresponding to the high-priority signal timing task under the business priority mark to the core business branch under the influence range mark for use by the traffic control system.
[0054] In one embodiment, the conditional branching release strategy is a strategy that dynamically adjusts the release order and target of tasks based on preset conditions. It determines the release path and execution order of tasks based on the business priority and scope of influence of signal timing tasks. The business priority label refers to the hierarchical identification of signal timing tasks according to their importance and urgency. For example, high-priority tasks usually involve the optimization of key intersections and main roads.
[0055] The optimal signal timing strategy refers to the best signal timing scheme calculated through federated learning, which effectively improves traffic efficiency and reduces congestion. Impact range labeling refers to the classification and identification of the geographical area or traffic impact involved in the signal timing task. For example, core business branches may involve urban arterial roads or key transportation hubs. Core business branches refer to modules or paths in the traffic control system that are responsible for handling critical traffic tasks, usually used to handle high-priority and high-impact tasks. The traffic control system refers to a comprehensive system for managing and controlling urban traffic signals, including functions such as traffic light control, traffic flow monitoring, and optimization strategy execution.
[0056] Optionally, the conditional branching deployment strategy ensures that high-priority signal timing tasks can be quickly and effectively deployed to core business branches for use by the traffic control system. Specifically, signal timing tasks are evaluated based on business priority and impact scope markers. For example, high-priority tasks involving urban arterial roads are marked as core tasks requiring immediate processing. According to the conditional branching deployment strategy, the system prioritizes deploying the optimal timing strategy corresponding to high-priority tasks to the core business branches. Core business branches typically have higher processing capacity and faster response times, enabling them to execute these strategies rapidly. For instance, the system might send the optimal timing strategy to the core business branches via a dedicated fast track, rather than through the regular deployment process. Upon receiving the optimal timing strategy, the core business branches immediately notify the traffic control system to execute these strategies. Further, this involves adjusting the signal light duration at key intersections, increasing green light time to alleviate congestion, or changing signal phase sequence to optimize traffic flow, ensuring rapid response to critical tasks and reducing traffic congestion and delays.
[0057] Furthermore, this application provides a method for obtaining the optimal timing strategy and optimal timing parameters, the method comprising:
[0058] The monitoring data includes the execution status and traffic quality data of the signal timing process, wherein the traffic quality data includes average delay time, queue length, and traffic efficiency improvement rate; the monitoring data is input into the timing optimization generation model, which analyzes the performance bottleneck of the signal timing process based on preset traffic rules and reinforcement learning reward function, generates the optimal timing strategy for the performance bottleneck, and pushes it to the traffic control terminal.
[0059] In one embodiment, monitoring data refers to data collected by traffic flow detectors regarding the execution status and traffic quality of the signal timing process, used to evaluate the effectiveness of the current traffic signal timing scheme; execution status refers to the actual operation of the signal timing scheme, such as green light duration and phase switching; traffic quality data refers to data reflecting traffic flow efficiency and congestion, including average delay time, queue length, and traffic efficiency improvement rate; the timing optimization generation model is a machine learning or deep learning-based model that can generate optimized signal timing strategies based on the input monitoring data.
[0060] Preset traffic rules refer to traffic management rules that are pre-defined in the model, such as minimum green light time and maximum queue length, to ensure that the signal timing scheme meets actual traffic management needs; reinforcement learning reward function is a mechanism used to evaluate and optimize signal timing schemes, guiding the model to learn the optimal strategy by defining rewards and penalties; performance bottleneck refers to the key link in the signal timing process that affects traffic efficiency, such as insufficient green light time in a certain direction leading to excessively long queues; traffic control terminal refers to the equipment or system used to actually implement the signal timing strategy, such as a traffic signal controller.
[0061] Optionally, the monitoring data includes the execution status of the signal timing process and traffic quality data. Traffic quality data specifically includes average delay time, queue length, and traffic efficiency improvement rate. This data is collected in real time by traffic flow detectors and uploaded to the cloud-based timing optimization center. Specifically, the traffic flow detectors collect data on the execution status of the signal timing process and traffic quality data in real time; for example, the detectors collect data every minute, including the current average delay time, queue length, and traffic efficiency improvement rate, and upload this data to the cloud-based timing optimization center. The monitoring data is then input into the timing optimization generation model. Based on preset traffic rules and a reinforcement learning reward function, the model analyzes the performance bottlenecks in the signal timing process. For example, the model might analyze and find that the queue length in a certain direction at a certain intersection is too long during peak hours, leading to a decrease in traffic efficiency.
[0062] Based on the analysis results, an optimal timing strategy is generated for performance bottlenecks and pushed to the traffic control terminal. For example, the model suggests increasing the green light time for a particular direction and adjusting the timing for other directions to optimize overall traffic efficiency. Upon receiving the strategy, the traffic control terminal immediately adjusts the traffic light settings. Preferably, real-time optimization of signal timing schemes improves the operational efficiency of the traffic system. This data-driven optimization method effectively addresses dynamic changes in traffic flow and enhances the overall operational efficiency of urban traffic.
[0063] In summary, the beneficial effects of the embodiments of this application are:
[0064] This application employs a timing optimization generation model to extract traffic flow features from multi-source traffic data at intersections. These features include temporal flow characteristics, spatial correlation characteristics, and dynamic impact characteristics. Based on these traffic flow features, a mapping and adaptation relationship between intersection traffic scenarios and signal timing parameters is determined. At the traffic signal control interface, the signal timing process is set using this mapping and adaptation relationship, combined with a visual configuration function. The parameter preset unit automatically fills in the initial timing parameters based on the mapping and adaptation relationship and transmits the signal timing process to the cloud-based timing optimization center. At the cloud-based timing optimization center, a federated learning algorithm is used to perform cross-domain collaborative optimization of the signal timing process across multiple intersections, obtaining the optimal timing strategy and optimal timing parameters. This application provides a signal timing automatic optimization method and system based on a generative model. By extracting multi-dimensional traffic flow features through a timing optimization generation model, constructing a mapping relationship between intersection traffic scenarios and timing parameters, and using a federated learning algorithm to achieve cross-domain collaborative optimization across multiple intersections, this application achieves a balanced approach to traffic demand at each intersection, enhances the adaptability of the traffic system to dynamic scenarios, and provides technical support for the refined management of intelligent transportation.
[0065] Example 2, based on the same inventive concept as the signal timing automatic optimization method based on the generative model in the foregoing examples, such as... Figure 2 As shown, this application provides an automatic signal timing optimization system based on a generative model, the system comprising:
[0066] The traffic flow feature extraction module 11 is used to extract traffic flow features from multi-source traffic data at intersections using a timing optimization generation model. The traffic flow features include time-series flow features, spatial correlation features, and dynamic impact features.
[0067] The mapping and adaptation relationship determination module 12 is used to determine the mapping and adaptation relationship between the intersection traffic scenario and the signal timing parameters based on the traffic flow characteristics.
[0068] The signal timing process setting module 13 is used to set the signal timing process in the traffic signal control interface by means of the mapping adaptation relationship and in combination with the visual configuration function.
[0069] The parameter auto-fill module 14 is used by the parameter preset unit to automatically fill the initial timing parameters according to the mapping adaptation relationship and transmit the signal timing process to the cloud timing optimization center.
[0070] The collaborative optimization module 15 is used to perform cross-domain collaborative optimization of the signal timing process across multiple intersections using a federated learning algorithm in the cloud-based timing optimization center, so as to obtain the optimal timing strategy and optimal timing parameters.
[0071] Furthermore, the collaborative optimization module 15 is used to perform the following method:
[0072] The execution effect of the signal timing process is monitored in real time by traffic flow detectors deployed at multiple intersections to obtain monitoring data; the monitoring data of the traffic flow detectors is then uploaded to the cloud-based timing optimization center.
[0073] Furthermore, the collaborative optimization module 15 is also used to perform the following method:
[0074] The monitoring data is stored in a traffic operation log database; a backend traffic data cluster based on a distributed storage architecture is established, and the traffic operation log database is associated with a storage node in the backend traffic data cluster.
[0075] Furthermore, the collaborative optimization module 15 is also used to perform the following method:
[0076] Using a distributed consensus algorithm, the optimal timing strategy is synchronously stored on each storage node, and a strategy version identifier and blockchain record are generated.
[0077] Furthermore, the collaborative optimization module 15 is also used to perform the following method:
[0078] The signal timing task is classified and marked according to its service priority and impact range, and the release branch type is divided. The release branch types are sorted and entered into the release queue in sequence according to the sorting results, and the conditional branch release strategy is configured.
[0079] Furthermore, the collaborative optimization module 15 is also used to perform the following method:
[0080] The conditional branching release strategy is used to prioritize the release of the optimal timing strategy corresponding to the high-priority signal timing task under the business priority mark to the core business branch under the influence range mark for use by the traffic control system.
[0081] Furthermore, the collaborative optimization module 15 is also used to perform the following method:
[0082] The monitoring data includes the execution status and traffic quality data of the signal timing process, wherein the traffic quality data includes average delay time, queue length, and traffic efficiency improvement rate; the monitoring data is input into the timing optimization generation model, which analyzes the performance bottleneck of the signal timing process based on preset traffic rules and reinforcement learning reward function, generates the optimal timing strategy for the performance bottleneck, and pushes it to the traffic control terminal.
[0083] It should be noted that the order of the embodiments described above is for descriptive purposes only and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.
[0084] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0085] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.
Claims
1. A method for automatic signal timing optimization based on a timing optimization generation model, characterized in that, The method includes: Using a timing optimization generation model, traffic flow features of multi-source traffic data at intersections are extracted, wherein the traffic flow features include temporal flow features, spatial correlation features, and dynamic impact features. Based on the traffic flow characteristics, determine the mapping and adaptation relationship between the intersection traffic scenario and the signal timing parameters; In the traffic signal control interface, the signal timing process is set through the mapping and adaptation relationship and the visual configuration function. The parameter preset unit automatically fills in the initial timing parameters according to the mapping adaptation relationship, and transmits the signal timing process to the cloud timing optimization center; In the cloud-based timing optimization center, a federated learning algorithm is used to perform cross-domain collaborative optimization of the signal timing process across multiple intersections to obtain the optimal timing strategy and optimal timing parameters.
2. The method as described in claim 1, characterized in that, In the cloud-based timing optimization center, a federated learning algorithm is used to perform cross-domain collaborative optimization of the signal timing process across multiple intersections. The method includes: The execution effect of the signal timing process is monitored in real time by traffic flow detectors deployed at multiple intersections, and monitoring data is obtained. The monitoring data of the traffic flow detector is uploaded to the cloud-based timing optimization center.
3. The method as described in claim 2, characterized in that, The method includes: The monitoring data is stored in a traffic operation log database; A backend traffic data cluster based on a distributed storage architecture is established, and the traffic operation log database is associated with a storage node in the backend traffic data cluster.
4. The method as described in claim 3, characterized in that, The traffic operation log database is associated with a storage node in the backend traffic data cluster, and the method includes: Using a distributed consensus algorithm, the optimal timing strategy is synchronously stored on each storage node, and a strategy version identifier and blockchain record are generated.
5. The method as described in claim 1, characterized in that, The method includes: The signal timing tasks are categorized and marked according to their business priority and scope of impact, and the release branch types are divided accordingly. Sort the release branches according to their types and enter them into the release queue in order of sorting, and configure the conditional branch release strategy.
6. The method as described in claim 5, characterized in that, The method includes: The conditional branching release strategy is used to prioritize the release of the optimal timing strategy corresponding to the high-priority signal timing task under the business priority mark to the core business branch under the influence range mark for use by the traffic control system.
7. The method as described in claim 2, characterized in that, The method for obtaining the optimal timing strategy and optimal timing parameters includes: The monitoring data includes the execution status of the signal timing process and traffic quality data, wherein the traffic quality data includes average delay time, queue length, and traffic efficiency improvement rate. The monitoring data is input into the timing optimization generation model. The timing optimization generation model analyzes the performance bottleneck of the signal timing process based on preset traffic rules and reinforcement learning reward function, generates the optimal timing strategy for the performance bottleneck, and pushes it to the traffic control terminal.
8. A signal timing automatic optimization system based on a generative model, characterized in that, The system is used to implement the automatic signal timing optimization method based on a generative model as described in any one of claims 1-7, the system comprising: The traffic flow feature extraction module is used to extract traffic flow features from multi-source traffic data at intersections using a timing optimization generation model. The traffic flow features include time-series flow features, spatial correlation features, and dynamic impact features. The mapping and adaptation relationship determination module is used to determine the mapping and adaptation relationship between the intersection traffic scenario and the signal timing parameters based on the traffic flow characteristics. The signal timing process setting module is used to set the signal timing process in the traffic signal control interface through the mapping adaptation relationship and in combination with the visual configuration function. The parameter auto-fill module is used by the parameter preset unit to automatically fill the initial timing parameters according to the mapping adaptation relationship, and transmit the signal timing process to the cloud timing optimization center. The collaborative optimization module is used in the cloud-based timing optimization center to perform cross-domain collaborative optimization of the signal timing process across multiple intersections using a federated learning algorithm, so as to obtain the optimal timing strategy and optimal timing parameters.