Intelligent traffic light adjustment method and system fusing car networking information

Abstract

The application discloses a traffic light intelligent adjustment method and system fusing Internet of Vehicles information, and relates to the technical field of intelligent traffic control.The method comprises the following steps: collecting multi-source traffic data through a roadside device, performing time-space alignment and filtering of abnormal values; constructing a deep learning prediction model based on the filtered data, combining historical and real-time Internet of Vehicles information to predict traffic demand; generating a dynamic timing strategy according to the prediction, synchronizing to associated intersections through V2X; deploying an edge computing node to adjust traffic light parameters in real time and sending a recommended speed to vehicles; establishing a closed-loop feedback mechanism to correct the strategy and model online based on trajectory data and efficiency indicators.The technical problem that the existing traffic light timing system cannot fuse multi-source Internet of Vehicles data in real time to perform accurate demand prediction and adaptive adjustment under a dynamic traffic environment is solved, and the technical effect of significantly improving traffic efficiency and shortening emergency response delay in a complex traffic scenario is achieved.

Description

Technical Field

[0001] This application relates to the field of intelligent traffic control technology, and in particular to a method and system for intelligent adjustment of traffic lights that integrates vehicle network information. Background Technology

[0002] In traditional urban traffic management, traffic light control systems often employ fixed timing or static optimization strategies based on wired detection equipment (such as geomagnetic coils and cameras), making it difficult to perceive complex dynamic changes in traffic flow in real time. Existing technologies rely on historical traffic flow statistics for periodic adjustments, failing to effectively integrate multi-dimensional dynamic information such as real-time vehicle location, pedestrian density, and emergency vehicle passage demands. Especially in high-density vehicle-to-everything (V2X) environments, the massive amounts of data generated by heterogeneous data sources (such as roadside units, vehicle terminals, and mobile terminals) suffer from spatiotemporal asynchrony and inconsistent standards, limiting the accuracy of traffic condition predictions. Furthermore, sudden traffic events (such as accidents or large-scale events) causing localized congestion or emergency vehicle passage demands often result in a sharp drop in regional road network efficiency due to system response delays and a lack of multi-intersection coordination. Therefore, achieving real-time fusion of multi-source V2X data and accurate prediction of dynamic traffic demand, and constructing an adaptive traffic light coordination control mechanism accordingly, has become a key technological bottleneck for improving the intelligence of urban traffic management.

[0003] At present, there is a technical problem in the relevant technologies that traffic light timing systems cannot integrate multi-source vehicle network data in real time to make accurate demand predictions and adaptive adjustments in dynamic traffic environments. Summary of the Invention

[0004] This application provides a method and system for intelligent adjustment of traffic lights that integrates vehicle-to-everything (V2X) information, thereby solving the technical problem that existing traffic light timing systems cannot integrate multi-source V2X data in real time for accurate demand prediction and adaptive adjustment in dynamic traffic environments.

[0005] This application provides a method for intelligent adjustment of traffic lights that integrates vehicle-to-everything (V2X) information, including:

[0006] Multi-source traffic data is collected in real time by roadside sensing devices. This data undergoes spatiotemporal alignment and outlier filtering to obtain filtered multi-source traffic data. A deep learning-based multimodal traffic state prediction model is constructed using this filtered data, combining historical traffic data and real-time vehicle-to-everything (V2X) information to predict traffic demand in each direction within a future variable time window. A dynamic timing strategy is generated based on the prediction results, and strategy parameters are synchronized to associated intersections via V2X communication protocols. Edge computing nodes are deployed to dynamically adjust traffic light controllers in real time, generating traffic light cycle parameters, phase difference parameters, and green light ratio parameters based on the dynamic timing strategy. Suggested driving speed commands are sent to vehicle terminals via a vehicle-road cooperative system. A closed-loop feedback mechanism is established to perform online parameter correction and model iterative updates of the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators.

[0007] This application provides a traffic light intelligent adjustment system that integrates vehicle network information, including:

[0008] The system includes: a multi-source traffic data acquisition module, which collects multi-source traffic data in real time through roadside sensing devices, performs spatiotemporal alignment and outlier filtering on the multi-source traffic data, and obtains filtered multi-source traffic data; a traffic demand prediction module, which constructs a deep learning-based multimodal traffic state prediction model using the filtered multi-source traffic data, and combines historical traffic data and real-time vehicle network information to predict traffic demand in each direction within a future variable time window; and a dynamic timing strategy generation module, which generates dynamic timing based on the prediction results. The system includes a strategy module that synchronizes strategy parameters with associated intersections via V2X communication protocol; an edge computing node deployment module that deploys edge computing nodes to dynamically adjust the traffic light controller in real time, generating traffic light cycle parameters, phase difference parameters, and green light ratio parameters based on the dynamic timing strategy, and sending suggested driving speed commands to vehicle terminals through the vehicle-road cooperative system; and a closed-loop feedback mechanism construction module that establishes a closed-loop feedback mechanism to perform online parameter correction and model iterative updates of the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators.

[0009] The proposed intelligent traffic light adjustment method and system, which integrates vehicle-to-everything (V2X) information, firstly collects multi-source traffic data through roadside equipment, performs spatiotemporal alignment and outlier filtering; based on the filtered data, a deep learning prediction model is constructed, combining historical and real-time V2X information to predict traffic demand; a dynamic timing strategy is generated based on the prediction and synchronized to related intersections via V2X; edge computing nodes are deployed to adjust traffic light parameters in real time and send suggested speeds to vehicles; a closed-loop feedback mechanism is established to correct the strategy and model online based on trajectory data and efficiency indicators, achieving the technical effect of significantly improving traffic efficiency and shortening emergency response delays in complex traffic scenarios. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0011] Figure 1 A flowchart illustrating the intelligent traffic light adjustment method integrating vehicle network information provided in this application embodiment;

[0012] Figure 2 A schematic diagram of the structure of the intelligent traffic light adjustment system that integrates vehicle network information provided in the embodiments of this application.

[0013] Figure labeling: 10 Multi-source traffic data acquisition module, 20 Traffic demand prediction module, 30 Dynamic timing strategy generation module, 40 Edge computing node deployment module, 50 Closed-loop feedback mechanism construction module. Detailed Implementation

[0014] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0017] This application provides a method for intelligent adjustment of traffic lights that integrates vehicle network information, such as... Figure 1 As shown, the method includes:

[0018] Step S100 involves real-time collection of multi-source traffic data using roadside sensing devices. This multi-source traffic data undergoes spatiotemporal alignment and outlier filtering to obtain filtered multi-source traffic data. Specifically, obtaining the filtered multi-source traffic data requires careful selection and deployment of roadside sensing devices, including cameras, radar, sensors, and V2X communication equipment. The locations of these devices should be rationally arranged based on road layout and monitoring needs, ensuring smooth communication. Next, multi-source traffic data collection is conducted, using different devices to collect vehicle flow, pedestrian density, vehicle speed, and V2X communication data. Then, spatiotemporal alignment is performed. Temporally, the clocks of each device are synchronously calibrated using a GPS clock as a reference, and data from different sampling frequencies is interpolated or resampled. Spatially, a unified geographic coordinate system is used, and GIS technology is employed to match the data with the actual road location. Finally, outlier filtering is performed, using statistical analysis, machine learning methods, and data consistency checks to process numerical data, identify abnormal patterns, and verify conflicting data, ultimately obtaining reliable data suitable for subsequent traffic analysis and decision-making.

[0019] In one possible implementation, multi-source traffic data is collected in real time by roadside sensing devices. This multi-source traffic data undergoes spatiotemporal alignment and outlier filtering to obtain filtered multi-source traffic data. Step S100 further includes step S110, which involves grouping the raw vehicle network data according to a time series, calculating the interquartile range and standard deviation of each group, and obtaining a grouped data position difference sequence. Specifically, obtaining the grouped data position difference sequence requires first clarifying the purpose of the scheme: by grouping the raw vehicle network data according to a time series and calculating the interquartile range and standard deviation, the dispersion and distribution characteristics of the data in different time periods can be analyzed to support subsequent data processing. Next, data preparation is performed by collecting raw data from various vehicle network data sources, cleaning and removing erroneous data, and processing missing values. Then, a suitable time interval is determined according to the analysis requirements, and the raw data is grouped according to a time series using a data processing library. Afterwards, for each group, the data is sorted, and the lower quartile Q1 and upper quartile Q3 are calculated to obtain the interquartile range. Simultaneously, the standard deviation is calculated according to the standard deviation formula. Finally, the interquartile range and standard deviation of each group are combined and stored in a new DataFrame to form a sequence of grouped data position differences for further in-depth analysis.

[0020] Step S120 involves identifying outlier data points in the grouped data position difference sequence based on a sliding window mechanism, and compensating for missing data using cubic spline interpolation. Specifically, the grouped data position difference sequence is processed to identify outlier data points and compensate for missing data. First, outliers are identified based on a sliding window mechanism, determining the window size and step size. Statistical features are calculated according to the set sliding window, and outlier judgment criteria are determined based on the threshold method or combined with the interquartile range, marking the outliers. Next, cubic spline interpolation is used to compensate for missing data, locating the missing data points, selecting adjacent known data points, constructing a cubic spline interpolation function, and substituting the x-coordinate of the missing points to calculate the compensation value. Finally, the results are verified and optimized. The reasonableness of the outliers is verified, parameters are adjusted, and re-identification is performed. The interpolation effect is evaluated; if unsatisfactory, the range of known data points is adjusted or other interpolation methods are used to improve data quality and usability, providing a reliable basis for subsequent analysis and decision-making.

[0021] Step S130 involves fusing and verifying the conflicting data collected by radar and cameras using a Kalman filter algorithm to obtain filtered multi-source traffic data. Specifically, the conflicting data from radar and cameras is processed. First, data preprocessing is performed, synchronizing the time of both sets of data and removing noise and outliers. Next, a state model and an observation model are established, defining the traffic target state vector and establishing state transition equations based on physical laws, constructing observation equations for both radar and camera. Then, the Kalman filter algorithm is implemented, initializing the state vector, covariance matrix, and noise covariance matrix, performing prediction and update steps, and updating the state vector and covariance matrix based on the observed values. Afterward, a reliability verification is performed, evaluating the observation reliability based on the observation noise covariance matrix and the fusion result reliability based on the state covariance matrix, setting a threshold to determine the reliability of the result. Finally, the fused and verified state vector is output as filtered data and stored, continuously updated with new data acquisition. Furthermore, the algorithm parameters can be adaptively adjusted, and data from other sensors can be fused to optimize and improve the solution.

[0022] Step S200 involves constructing a deep learning-based multimodal traffic state prediction model using filtered multi-source traffic data. This model combines historical traffic data and real-time vehicle-to-everything (V2X) information to predict traffic demand in all directions within a future variable time window. Specifically, data preparation and integration are performed first. Filtered multi-source traffic data, historical traffic data, and real-time V2X information are prepared and merged into a complete dataset after format conversion and standardization. Next, a deep learning-based multimodal traffic state prediction model is constructed, determining the architecture to include CNN, RNN, or LSTM with an attention mechanism, and designing input, hidden, and output layers. The dataset is then partitioned, a loss function is defined, and an optimization algorithm is selected. The model is trained, with hyperparameters adjusted using a validation set during training. After training, the model is deployed to a real-world system, using real-time data for prediction. The results are post-processed to aid decision-making. The model is then evaluated using evaluation metrics. If performance is unsatisfactory, the reasons are analyzed and optimizations are implemented. Finally, continuous improvement is achieved by updating the data and iterating the model based on actual feedback and new demands to adapt to traffic changes and accurately predict traffic demand in all directions within a future variable time window.

[0023] In one possible implementation, a deep learning-based multimodal traffic state prediction model is constructed using filtered multi-source traffic data. This model combines historical traffic data and real-time vehicle-to-everything (V2X) information to predict traffic demand in all directions within a future variable time window. Step S200 further includes step S210, establishing a spatiotemporal graph convolutional network to process road topology features. Specifically, establishing this network requires data preparation. Road topology and traffic state data are collected from GIS, traffic sensors, etc., and preprocessed through cleaning and normalization. Next, a graph is constructed, abstracting intersections and road segments as nodes and representing connections with edges, which can be represented by an adjacency matrix or edge list. When designing the spatiotemporal graph convolutional network, spatially, graph convolution operations are used and multiple layers are stacked to learn complex spatial dependencies. Temporally, a temporal dimension is introduced and fused with spatial features. During model training, the dataset is divided, a loss function is defined according to the task, an optimization algorithm is selected, parameters are adjusted through batch training, and hyperparameters are adjusted using a validation set. The model is then evaluated using a test set. If it meets the requirements, it is applied to scenarios such as traffic flow prediction. Finally, based on evaluation and application feedback, the model was improved through structural adjustments, feature engineering, and algorithm optimization to enhance its performance and prediction accuracy.

[0024] Step S220 involves modeling traffic flow temporal dependencies using a Long Short-Term Memory (LSTM) network. Specifically, this involves collecting traffic flow data from multiple sources, including road sensors and surveillance cameras, covering indicators such as traffic volume and speed, and recording timestamps. Data preprocessing is then performed, including cleaning noise and outliers, handling missing values ​​using interpolation, and normalizing the data volume. The data is then divided into training, validation, and test sets proportionally. When constructing the LSTM model, the input layer is designed to receive the preprocessed data, and the LSTM layer is used as the core to handle long-term dependencies through a gating mechanism. The output layer is designed according to the task objective. Parameters such as the number of hidden units and the time step are set, and an appropriate activation function is selected. During training, loss functions such as mean squared error are defined, and optimization algorithms such as Adam are selected. Forward propagation is performed on batches of input data, loss is calculated, and backpropagation updates the parameters. Hyperparameters are adjusted using the validation set. When evaluating the model, metrics such as MAE and RMSE are used. The metric values ​​are calculated using the test set and compared with thresholds or benchmark models. If the performance is unsatisfactory, it can be optimized by adjusting the model structure, improving data preprocessing, and optimizing hyperparameters. Finally, the model that meets the requirements can be applied to real-world scenarios such as traffic flow prediction to provide decision support.

[0025] Step S230 involves integrating an attention mechanism to dynamically weight vehicle intent information from V2X communication data. Specifically, this involves first collecting and storing V2X communication data containing basic vehicle information and intent information from the vehicle-to-everything (V2X) system, followed by preprocessing including noise and outlier removal, intent information encoding, and normalization of numerical data. Then, an attention mechanism is designed, representing the preprocessed information as a sequence of feature vectors, defining queries, keys, and values, calculating attention scores and weights, and weighted summing of the value vectors. Next, basic models such as RNNs and LSTMs are selected, and the attention mechanism is integrated into them. The dataset is divided, a loss function is defined according to the task, an optimization algorithm is selected, and training is performed with batch input data. Hyperparameters are adjusted using a validation set. The model is evaluated using a test set, employing metrics such as accuracy and MSE, and compared with thresholds or benchmark models. If performance is unsatisfactory, optimizations are made by adjusting attention mechanism parameters, improving the basic model structure, and optimizing data preprocessing. Finally, the satisfactory model is applied to real-world scenarios such as intelligent traffic signal control, dynamically adjusting traffic management strategies based on real-time V2X data.

[0026] Step S240 involves optimizing the multi-objective loss function using a reinforcement learning framework to balance traffic efficiency and safety indicators, thereby generating a multimodal traffic state prediction model. Specifically, optimizing the multi-objective loss function using a reinforcement learning framework to balance traffic efficiency and safety indicators to generate the multimodal traffic state prediction model first requires clarifying the quantification methods for traffic efficiency and safety indicators, collecting and preprocessing multi-source data from traffic sensors, cameras, V2X communication, etc. Next, a multimodal traffic state prediction model is constructed, combining CNNs and RNNs to process different modalities of data and fuse features, and performing initial training using an initial loss function. Then, a reinforcement learning framework is built, abstracting the traffic system as an environment, defining a multi-objective reward function that includes traffic efficiency and safety indicators, and selecting a suitable reinforcement learning algorithm. Afterwards, the reinforcement learning and prediction model are combined, and the multi-objective loss function is calculated based on the prediction results and actual conditions. Model parameters are updated through policy optimization. The model is then evaluated using a test dataset, and performance is judged based on evaluation indicators. If requirements are not met, the model structure and algorithm parameters are fine-tuned. Finally, the optimized model is deployed to an actual traffic management system to predict traffic conditions and provide decision support based on real-time data, while continuously collecting data to further optimize the model.

[0027] Step S300: Generate a dynamic timing strategy based on the prediction results and synchronize the strategy parameters to the associated intersections via the V2X communication protocol. Specifically, first, acquire and preprocess the multimodal traffic state prediction results, extracting key indicators related to the dynamic timing strategy. Then, based on these indicators, generate schemes for arterial road green wave coordination, emergency vehicle priority passage, and pedestrian crossing protection, integrating and optimizing the strategy parameters. Next, select and configure a suitable V2X communication protocol, encapsulate the optimized strategy parameters into a data format conforming to the protocol, and send them to the traffic signal control equipment at the associated intersections via the protocol. The equipment receives, parses, and verifies the data; if successful, it updates the local signal timing parameters. During strategy implementation, real-time monitoring of traffic state changes is conducted, and the strategy is adjusted based on the monitoring results. If the actual situation deviates significantly from the prediction or the effect is unsatisfactory, the strategy is regenerated and synchronized. Simultaneously, a long-term evaluation mechanism is established to continuously optimize the strategy generation and adjustment methods.

[0028] In one possible implementation, a dynamic timing strategy is generated based on the prediction results, and the strategy parameters are synchronized to the associated intersections via the V2X communication protocol. Step S300 further includes step S310, which classifies traffic scenarios into three modes: normal traffic flow, peak congestion, and emergency events. Specifically, classifying traffic scenarios into these three modes requires data collection and analysis. Data from various sources, including traffic flow, video surveillance, special events, and weather, is collected, preprocessed (cleaned and normalized), and then subjected to time series and cluster analysis. Next, three modes are defined: the normal traffic flow mode is characterized by stable traffic flow and smooth driving, and judgment criteria can be set based on indicators such as vehicle flow and speed; the peak congestion mode occurs during specific time periods, with high traffic flow and slow driving, and also has corresponding judgment thresholds; the emergency event mode requires special scheduling due to special circumstances and is determined based on alarm information from relevant departments. Then, a real-time monitoring system is built, a pattern recognition algorithm is designed, data is collected and analyzed in real time, and updates are provided promptly when the mode changes. Finally, the accuracy of pattern segmentation is verified through historical data and practical applications. Based on the results, the pattern definition, judgment criteria, and recognition algorithm are optimized to adapt to the development and changes of the transportation system.

[0029] Step S320 involves employing a green wave optimization method for normal traffic patterns. Specifically, a green wave optimization method based on a genetic algorithm is used for normal traffic patterns. First, data on road infrastructure, traffic flow, and traffic light settings are collected and preprocessed to remove noise, fill missing values, and normalize the data. Next, a green wave model is constructed, with the goal of improving vehicle throughput. Decision variables such as traffic light cycle duration are defined, and constraints such as traffic flow, time, and safety are set to form a mathematical model. Then, a genetic algorithm is designed, encoding the decision variables into chromosomes, randomly generating an initial population, defining a fitness function to evaluate the merits of different schemes, and designing genetic operations such as selection, crossover, and mutation, with termination conditions set. The algorithm is run to solve the problem, and the output scheme is evaluated. If the scheme is unsatisfactory, the algorithm parameters are adjusted and re-optimized. Finally, the optimized scheme is applied to an actual traffic light control system, a monitoring system is established, and the effectiveness is evaluated based on real-time traffic data. Adjustments and optimizations are made promptly when deviations are found.

[0030] Step S330 involves implementing a multi-intersection collaborative control mechanism for peak congestion patterns. Specifically, this mechanism, based on game theory, first collects and preprocesses data such as traffic flow, intersection topology, and traffic light information, removing noise and outliers and normalizing the data. Next, it models the multi-intersection collaborative control problem, treating each intersection's traffic light control strategy as a participant, defining a payoff function that integrates traffic efficiency, fairness, and stability, and formulating interaction rules for each participant. Then, it selects a suitable game theory model and designs analytical or numerical algorithms to solve the strategy. The solved strategy is applied to the actual traffic light control system, and a real-time monitoring system is established to evaluate its effectiveness, adjusting the strategy promptly based on feedback. The mechanism's effectiveness is periodically evaluated using indicators such as average delay time, and the game theory model, algorithm, and strategy adjustment methods are optimized accordingly. Finally, the mechanism is integrated with the existing traffic management system for compatibility testing to ensure stable operation in coordination with other traffic control measures.

[0031] Step S340 involves constructing a dynamic priority right-of-way allocation model for emergency events. Specifically, this involves first collecting information such as the type and location of the emergency, surrounding traffic conditions, and the location of emergency vehicles through alarm systems, traffic monitoring equipment, and V2X communication technology, and assessing the severity of the event and analyzing traffic conditions. Next, a model is constructed, setting the goal of ensuring rapid arrival of emergency vehicles while minimizing disruption to normal traffic. Route selection and traffic light timing adjustments are defined as decision variables, and constraints such as traffic flow, safety, and legal requirements are set. Mathematical programming or intelligent algorithms are then used to solve the problem. After inputting the relevant information into the model, specific allocation strategies are generated, including navigation for emergency vehicles and adjusting traffic light timings. During strategy implementation, the movement of emergency vehicles and traffic light execution are monitored in real time, and the strategy is dynamically adjusted based on the results. After the event, the effectiveness is evaluated using indicators such as emergency vehicle arrival time, and the model is optimized accordingly. Finally, the model is integrated with existing systems, and relevant personnel are trained to improve collaborative work efficiency and personnel's ability to respond to emergencies.

[0032] Step S400 involves deploying edge computing nodes to dynamically adjust the traffic light controller in real time. Based on the dynamic timing strategy, traffic light cycle parameters, phase difference parameters, and green ratio parameters are generated, and suggested driving speed commands are sent to vehicle terminals via the vehicle-road cooperative system. Specifically, traffic flow, geographic information, and other data are collected and preprocessed. Edge computing nodes are deployed appropriately to build the vehicle-road cooperative system. Next, traffic scenarios are identified based on the collected data, and appropriate strategies are selected and optimized from the strategy library. Based on this, traffic light cycle, phase difference, and green ratio parameters are generated. The edge computing nodes send these parameters to the traffic light controller, which updates its control strategy. Simultaneously, the nodes monitor and provide feedback in real time for dynamic adjustments. Afterward, the nodes integrate and analyze traffic, traffic light, and vehicle location information to calculate suggested vehicle speeds, generate commands, and send them to vehicle terminals. Vehicles respond and provide feedback. Finally, system performance is periodically evaluated using metrics such as average delay time. Based on the results, the dynamic timing strategy is optimized to improve system performance and traffic efficiency.

[0033] In one possible implementation, edge computing nodes are deployed to dynamically adjust the traffic light controller in real time. Based on the dynamic timing strategy, traffic light cycle parameters, phase difference parameters, and green light ratio parameters are generated. Suggested driving speed commands are sent to vehicle terminals via a vehicle-road cooperative system. Step S400 further includes step S410, designing a lightweight model compression algorithm to reduce the number of prediction model parameters. Specifically, designing a lightweight model compression algorithm to reduce the number of prediction model parameters to less than 30% of the original model requires preliminary preparation, including collecting and preprocessing the prediction model and its structural information and representative datasets, and determining evaluation metrics such as parameter quantity and accuracy. Next, common algorithms such as pruning, quantization, and knowledge distillation are investigated, and selected based on model characteristics, application requirements, and hardware limitations. Then, customized design is performed, determining pruning strategies, quantization schemes, knowledge distillation forms, etc., and optimizing the algorithm through hyperparameter search. Model compression is implemented, and the compressed model is then trained or fine-tuned using a dataset. Performance is evaluated using a test dataset to check if the parameter quantity meets the standard; if not, parameters are adjusted or the algorithm is redesigned. The model is optimized based on the evaluation results, and different algorithm combinations can be tried. Finally, the optimized compression model is deployed to the actual application environment, with appropriate hardware and software frameworks selected. Performance and parameter counts are continuously monitored during the application, and regular evaluation and optimization are performed to maintain performance stability.

[0034] Step S420: Data interaction between the local node and the cloud platform is achieved through the API interface. The local node processes millisecond-level real-time control commands, while the cloud platform performs hourly-level strategy optimization. Specifically, to achieve data interaction between the local node and the cloud platform through the API interface, enabling the local node to process millisecond-level real-time control commands and the cloud platform to perform hourly-level strategy optimization, the following steps are required: First, conduct requirements analysis and planning, clarifying the data types and functional requirements of the interaction between the two parties, and formulate API interface specifications, covering URL structure, request methods, data formats, and parameter definitions. Next, develop the local node and the cloud platform separately. The local node implements the API client and optimizes real-time processing capabilities, while the cloud platform builds an API server and implements hourly-level strategy optimization. Then, conduct data interaction testing, including unit testing, integration testing, and security testing, to ensure that the system's functionality, performance, and security meet the standards. Next, deploy the system to the production environment, establish a monitoring and maintenance mechanism, and record logs for problem investigation and auditing. Finally, continuously optimize the system, improve it through data analysis and user feedback, and pay attention to industry trends to introduce new technologies for upgrades and innovations.

[0035] Step S430: Establish a communication quality assessment module to dynamically switch between LTE-V2X direct communication and 5G cellular communication transmission channels. Specifically, establishing a communication quality assessment module to dynamically switch between LTE-V2X direct communication and 5G cellular communication transmission channels first requires demand analysis and planning to clarify the communication quality requirements of different application scenarios, consider equipment characteristics, and determine assessment indicators and thresholds. Next, design and develop the communication quality assessment module, including a data acquisition unit to accurately collect indicator data in real time, a data analysis and assessment unit to preprocess data and comprehensively assess communication quality, and then integrate and optimize each unit. Then, formulate a dynamic switching mechanism, develop a reasonable switching strategy based on the assessment results, and design a switching process including triggering, requesting, accessing, and migrating steps to ensure smooth and reliable switching. Afterwards, conduct system testing and optimization, first in a laboratory to simulate scenarios, then in real-world scenarios, adjusting the system based on the results. Finally, complete deployment and maintenance, deploying the system to actual equipment and networks, training personnel, establishing a monitoring mechanism, promptly handling faults, and upgrading and optimizing the system to adapt to new requirements.

[0036] Step S500: Establish a closed-loop feedback mechanism to perform online parameter correction and model iterative updates for the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators. Specifically, establishing a closed-loop feedback mechanism for online parameter correction and model iterative updates of the dynamic timing strategy requires first collecting high-precision trajectory data and traffic efficiency indicator data, which are then collected and stored using sensor devices and communication networks, and preprocessed by cleaning and normalization. Next, a dynamic timing strategy model is established, its structure is determined, and parameters are initialized. A closed-loop feedback mechanism is designed, using traffic efficiency indicators as feedback indicators, periodically comparing actual values ​​with target values, and triggering adjustments if thresholds are exceeded. Then, online parameter correction and model iterative updates are performed, adjusting parameters based on feedback results, updating the model with new data, and evaluating performance. The system is periodically evaluated using indicators such as the degree of improvement in traffic efficiency, and the feedback mechanism, model, and parameter correction methods are optimized accordingly. Finally, the system is deployed to an actual traffic management system, personnel are trained, a monitoring mechanism is established, and faults are handled promptly, and upgrades and optimizations are made to adapt to traffic changes.

[0037] In one possible implementation, a closed-loop feedback mechanism is established. Based on high-precision trajectory data and traffic efficiency indicators, the dynamic timing strategy is adjusted online and the model is iteratively updated. Step S500 further includes step S510, defining a comprehensive traffic efficiency evaluation index, which includes three dimensions: average delay time, number of stops, and fuel consumption. Specifically, defining a comprehensive traffic efficiency evaluation index that includes three dimensions—average delay time, number of stops, and fuel consumption—first requires clarifying the basis for index determination. These three dimensions can comprehensively evaluate traffic efficiency from the perspectives of traffic flow, congestion, and energy consumption. Next, data is collected through multiple channels, using sensors and video surveillance to obtain average delay time and number of stops data, and obtaining fuel consumption data from the OBD interface or through vehicle dynamics models. Then, the index is quantified, with the three indicators measured in seconds, integers, and liters respectively, and standardized using min-max normalization or Z-score normalization. Finally, the weights of each indicator are reasonably determined using expert scoring or analytic hierarchy process. Finally, the standardized parameters of each dimension are multiplied and summed with their corresponding weights to construct a comprehensive evaluation index for traffic efficiency, with the formula E = w1×x1 + w2×x2 + w3×x3, which provides a strong basis for the evaluation and optimization of the traffic system.

[0038] Step S520: Develop a digital twin-based simulation verification platform to conduct virtual intersection stress tests on the new strategy. Specifically, developing a digital twin-based simulation verification platform for virtual intersection stress testing of the new strategy first requires preliminary planning and requirements analysis to clarify the test objectives, requirements, simulation scope, and accuracy. Next, collect and pre-process data on the geometry, traffic facilities, and traffic flow of the physical intersection. Then, construct a digital twin model, including physical entity modeling and data mapping and synchronization. Following this, develop the simulation verification platform, designing the architecture and developing functional modules such as new strategy import, simulation scenario setting, calculation, result analysis, and visualization. Implement virtual intersection stress tests, designing different scenarios, importing the new strategy for testing, and comparing and analyzing it with a benchmark strategy. Then, verify the platform by comparing it with actual historical data; if there are significant differences, adjust and optimize accordingly. Further optimize the platform based on testing and feedback. Finally, compile a user manual, test report, and technical documentation, deliver the platform and documentation to users, and provide training and technical support.

[0039] Step S530: Construct an abnormal operating condition self-healing system that automatically rolls back to the previous stable version when a policy execution deviation exceeds a threshold. Specifically, constructing this system requires the following steps: First, conduct system requirements analysis and planning, clarifying application scenarios and objectives, defining policies and deviation indicators, and setting thresholds. Next, build a data collection and processing system, deploying acquisition equipment, establishing a transmission network, preprocessing and storing data. Then, develop a deviation detection module, select and optimize the detection algorithm, and perform real-time monitoring and alarms. Construct a policy version management system, defining identifier versions, determining stable versions, and establishing backup and recovery mechanisms. Implement an automatic rollback mechanism, setting trigger conditions, designing the rollback process, and testing and verifying it. Afterward, perform system integration and testing, covering functional and performance testing. Finally, deploy the system to the production environment, establish a monitoring and maintenance mechanism, and continuously improve the system to enhance its self-healing capabilities and reliability.

[0040] In one possible implementation, a closed-loop feedback mechanism is established. Based on high-precision trajectory data and traffic efficiency indicators, the dynamic timing strategy is adjusted online and the model is iteratively updated. Step S500 further includes step S540, establishing a hierarchical response mechanism to classify traffic conditions into four levels of warnings: red, orange, yellow, and blue. Specifically, establishing a hierarchical response mechanism to classify traffic conditions into four levels of warnings requires the following steps: First, collect and integrate multi-source data, using devices such as geomagnetic sensors and video surveillance to collect traffic flow data, while also collecting relevant information such as weather, special events, and road construction, and building a data storage platform for management. Next, determine the warning thresholds for each level. By mining historical data and conducting simulations and predictions, the blue warning is set at 30% below the normal daily average traffic flow, the yellow warning at 30%-60%, the orange warning at 60%-90%, and the red warning at over 90%. Then, develop a warning monitoring system, design the architecture, implement the algorithm, and conduct testing and optimization. Next, a response mechanism should be established, clearly defining the measures corresponding to different levels of early warning, such as enhanced monitoring for blue alerts and activation of a full emergency response for red alerts. The division of responsibilities and coordination mechanisms among various departments should also be determined. Then, standardized early warning information should be disseminated through diverse channels such as traffic radio and electronic displays, and a public feedback and communication mechanism should be established. Finally, the system's operational effectiveness should be regularly evaluated, and improvement measures should be developed and implemented based on the results to continuously enhance the scientific validity and effectiveness of the mechanism.

[0041] Step S550: Initiate regional coordinated control for a red alert, coordinating all traffic lights within a predefined range to enter emergency mode. Specifically, initiating coordinated control within a 5-kilometer radius of a red alert and coordinating traffic lights to enter emergency mode requires the following steps: Preliminary data collection and analysis, including traffic and geographic information data, evaluation and upgrading of the traffic light system, and establishment of a communication network; setting a red alert trigger mechanism, determining the alert threshold, and real-time judgment by the intelligent monitoring system; designing an emergency mode, optimizing traffic light timing, and establishing regional coordinated control and priority control mechanisms for special vehicles; integrating the emergency strategy into the control system and testing it; initiating coordinated control upon alert triggering, and simultaneously dispatching traffic police to the scene for command; establishing an evaluation index system to evaluate the effect in real time and provide feedback, continuously optimizing the control plan; conducting training for relevant personnel to improve emergency response capabilities, and publicizing the information to the public through the media to increase cooperation and jointly address traffic congestion.

[0042] Step S560: Develop a driver behavior prediction module to predict risky behaviors such as cutting in and jaywalking by combining historical trajectory data and adjust the traffic phase in advance. Specifically, developing a driver behavior prediction module to predict risky behaviors such as cutting in and jaywalking by combining historical trajectory data and adjust the traffic phase in advance requires the following steps: First, conduct requirements analysis and planning to clarify business objectives, functional requirements, and formulate a project plan; then, collect historical trajectory and risk behavior labeled data, clean and extract features, and divide the dataset; then, select a suitable machine learning or deep learning model, train it using the training set, and optimize the model using the validation set; integrate the trained module into the traffic light control system and conduct functional and performance tests; finally, deploy the system, establish a real-time monitoring and feedback mechanism, and continuously optimize the model and system based on actual operation to adapt to the constantly changing traffic environment and driver behavior patterns.

[0043] This application's embodiments employ the following methods: collecting multi-source traffic data via roadside equipment, performing spatiotemporal alignment and outlier filtering; constructing a deep learning prediction model based on the filtered data, and combining historical and real-time vehicle-to-everything (V2X) information to predict traffic demand; generating dynamic timing strategies based on the predictions, and synchronizing them to associated intersections via V2X; deploying edge computing nodes to adjust traffic light parameters in real time and sending suggested speeds to vehicles; and establishing a closed-loop feedback mechanism to correct strategies and models online based on trajectory data and efficiency indicators. This achieves the technical effect of significantly improving traffic efficiency and shortening emergency response delays in complex traffic scenarios.

[0044] In the above text, refer to Figure 1 This paper describes in detail a method for intelligent adjustment of traffic lights that integrates vehicle network information according to an embodiment of the present invention. Next, we will refer to... Figure 2 A traffic light intelligent adjustment system integrating vehicle network information is described according to an embodiment of the present invention.

[0045] The intelligent traffic light adjustment system integrating vehicle-to-everything (V2X) information according to embodiments of the present invention addresses the technical problem that existing traffic light timing systems cannot accurately predict and adaptively adjust demand in real time by integrating multi-source V2X data under dynamic traffic environments. This achieves the technical effect of significantly improving traffic efficiency and shortening emergency response delays in complex traffic scenarios. The intelligent traffic light adjustment system integrating V2X information includes: (See attached figures for reference numerals: 10 for multi-source traffic data acquisition, 20 for traffic demand prediction, 30 for dynamic timing strategy generation, 40 for edge computing node deployment, and 50 for closed-loop feedback mechanism construction.)

[0046] The multi-source traffic data acquisition module 10 is used to collect multi-source traffic data in real time through roadside sensing devices, perform spatiotemporal alignment and outlier filtering on the multi-source traffic data, and obtain filtered multi-source traffic data.

[0047] The traffic demand forecasting module 20 is used to build a deep learning-based multimodal traffic state forecasting model using filtered multi-source traffic data, and combine historical traffic data and real-time vehicle network information to predict traffic demand in each direction within a future variable time window.

[0048] The dynamic timing strategy generation module 30 is used to generate a dynamic timing strategy based on the prediction results and synchronize the strategy parameters to the associated intersections through the V2X communication protocol.

[0049] The edge computing node deployment module 40 is used to deploy edge computing nodes to make real-time dynamic adjustments to the traffic light controller. Based on the dynamic timing strategy, it generates traffic light cycle parameters, phase difference parameters, and green light ratio parameters, and sends suggested driving speed commands to vehicle terminals through the vehicle-road cooperative system.

[0050] The closed-loop feedback mechanism construction module 50 is used to establish a closed-loop feedback mechanism, which performs online parameter correction and model iterative update on the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators.

[0051] The specific configuration of the multi-source traffic data acquisition module 10 will be described in detail below. As mentioned above, multi-source traffic data is collected in real time through roadside sensing devices. The multi-source traffic data is then spatiotemporally aligned and outlier filtered to obtain filtered multi-source traffic data. The multi-source traffic data acquisition module 10 further includes: a grouped data position difference sequence acquisition unit, which is used to group the original vehicle network data according to the time series, calculate the interquartile range and standard deviation of each group of data, and obtain the grouped data position difference sequence; an outlier data point identification unit, which is used to identify outlier data points in the grouped data position difference sequence based on a sliding window mechanism and to compensate for missing data using cubic spline interpolation; and a data verification unit, which is used to perform data fusion and credibility verification on conflicting data collected by radar and cameras using a Kalman filter algorithm to obtain filtered multi-source traffic data.

[0052] The specific configuration of the traffic demand prediction module 20 will be described in detail below. As mentioned above, a deep learning-based multimodal traffic state prediction model is constructed using filtered multi-source traffic data. Combining historical traffic data and real-time vehicle network information, the model predicts traffic demand in each direction within a future variable time window. The traffic demand prediction module 20 further includes: a spatiotemporal graph convolutional network construction unit, which is used to establish a spatiotemporal graph convolutional network to process road topology features; a temporal dependency construction unit, which is used to model traffic flow temporal dependencies using a long short-term memory network; a vehicle intent information integration unit, which is used to integrate vehicle intent information from dynamically weighted V2X communication data using an attention mechanism; and a multimodal traffic state prediction model generation unit, which is used to optimize a multi-objective loss function through a reinforcement learning framework, balance traffic efficiency and safety indicators, and generate a multimodal traffic state prediction model.

[0053] The specific configuration of the dynamic timing strategy generation module 30 will be described in detail below. As mentioned above, the dynamic timing strategy is generated based on the prediction results, and the strategy parameters are synchronized to the associated intersections via the V2X communication protocol. The dynamic timing strategy generation module 30 further includes: a mode division unit, which is used to divide traffic scenarios into three modes: normal traffic, peak congestion, and emergency events; a green wave optimization unit, which is used to apply a green wave optimization method for the normal traffic mode; a cooperative control mechanism activation unit, which is used to activate a multi-intersection cooperative control mechanism for the peak congestion mode; and a right-of-way dynamic allocation model construction unit, which is used to construct a priority right-of-way dynamic allocation model for the emergency event mode.

[0054] The following section will describe in detail the specific configuration of the edge computing node deployment module 40. As mentioned above, the deployed edge computing nodes dynamically adjust the traffic light controller in real time, generate traffic light cycle parameters, phase difference parameters, and green light ratio parameters based on the dynamic timing strategy, and send suggested driving speed commands to vehicle terminals through the vehicle-road cooperative system. The edge computing node deployment module 40 further includes: a lightweight model compression algorithm design unit, which is used to design a lightweight model compression algorithm to reduce the number of prediction model parameters; a data interaction unit, which is used to realize data interaction between the local node and the cloud platform through the API interface, wherein the local node processes millisecond-level real-time control commands, and the cloud platform performs hourly-level strategy optimization; and a communication quality assessment module construction unit, which is used to establish a communication quality assessment module and dynamically switch between LTE-V2X direct communication and 5G cellular communication transmission channels.

[0055] The specific configuration of the closed-loop feedback mechanism construction module 50 will be described in detail below. As mentioned above, a closed-loop feedback mechanism is established based on high-precision trajectory data and traffic efficiency indicators to perform online parameter correction and model iterative updates on the dynamic timing strategy. The closed-loop feedback mechanism construction module 50 further includes: an evaluation index definition unit, which is used to define a comprehensive traffic efficiency evaluation index, including three-dimensional parameters: average delay time, number of stops, and fuel consumption; a simulation verification platform development unit, which is used to develop a simulation verification platform based on digital twins to conduct virtual intersection stress tests on the new strategy; and an abnormal working condition self-healing system construction unit, which is used to construct an abnormal working condition self-healing system that automatically rolls back to the previous stable version when a strategy execution deviation is detected to exceed a threshold.

[0056] The closed-loop feedback mechanism construction module 50 further includes: a graded response mechanism construction unit, which is used to establish a graded response mechanism to divide traffic conditions into four levels of warning: red, orange, yellow, and blue; a regional linkage control activation unit, which is used to activate regional linkage control for red warning conditions, coordinating all traffic lights within a predefined range to enter emergency mode; and a behavior prediction module development unit, which is used to develop a driver behavior prediction module to predict risky behaviors such as cutting in and jaywalking based on historical trajectory data and adjust the phase in advance.

[0057] The traffic light intelligent adjustment system integrating vehicle network information provided in this embodiment of the invention can execute the traffic light intelligent adjustment method integrating vehicle network information provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0058] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0059] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for intelligent adjustment of traffic lights integrating vehicle network information, characterized in that, The method includes: Multi-source traffic data is collected in real time by roadside sensing devices, and the multi-source traffic data is spatiotemporally aligned and outlier filtered to obtain filtered multi-source traffic data. A deep learning-based multimodal traffic state prediction model is constructed using filtered multi-source traffic data. By combining historical traffic data and real-time vehicle network information, the traffic demand in each direction within a future variable time window is predicted. Dynamic timing strategies are generated based on the prediction results, and strategy parameters are synchronized to associated intersections via V2X communication protocol. Edge computing nodes are deployed to dynamically adjust the traffic light controller in real time. Based on the dynamic timing strategy, traffic light cycle parameters, phase difference parameters, and green ratio parameters are generated, and suggested driving speed commands are sent to vehicle terminals through the vehicle-road cooperative system. A closed-loop feedback mechanism is established to perform online parameter correction and model iterative update of the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators. The generation of the dynamic timing strategy includes: classifying traffic scenarios into three modes: normal traffic, peak congestion, and emergency events; For normal traffic patterns, a green wave optimization method is adopted; for peak congestion patterns, a multi-intersection collaborative control mechanism is activated; and for emergency events, a dynamic priority right-of-way allocation model is constructed. Specifically, the normal traffic flow mode employs a green wave optimization method based on a genetic algorithm. First, data on road infrastructure, traffic flow, and traffic light settings are collected and preprocessed to remove noise, fill missing values, and normalize. Next, a green wave model is constructed, aiming to improve vehicle throughput. Decision variables for traffic light cycle duration are defined, and traffic flow, time, and safety constraints are set to form a mathematical model. Then, a genetic algorithm is designed, encoding the decision variables into chromosomes, randomly generating an initial population, defining a fitness function to evaluate the merits of different schemes, and designing selection, crossover, and mutation genetic operations with termination conditions. The algorithm is run to solve the problem, and the output scheme is evaluated. If unsatisfactory, the algorithm parameters are adjusted and re-optimized. Finally, the optimized scheme is applied to the actual traffic light control system, a monitoring system is established, and the effectiveness is evaluated based on real-time traffic data. Adjustments and optimizations are made promptly when deviations are found. The peak congestion mode employs a game theory-based multi-intersection collaborative control mechanism. First, traffic flow, intersection topology, and traffic light information data are collected and preprocessed, including noise and outlier removal and data normalization. Next, the multi-intersection collaborative control problem is modeled, with each intersection's traffic light control strategy as a participant. A payoff function considering overall traffic efficiency, fairness, and stability is defined, and interaction rules for each participant are established. Then, a suitable game theory model is selected, and analytical or numerical algorithms are designed to solve the strategies. The solved strategies are applied to the actual traffic light control system, and a real-time monitoring system is established to evaluate the effectiveness, adjusting the strategies promptly based on feedback. The mechanism's effectiveness is periodically evaluated using the average delay time index, and the game theory model, algorithm, and strategy adjustment methods are optimized accordingly. Finally, the mechanism is integrated with the existing traffic management system, and compatibility testing is conducted to ensure stable operation in coordination with other traffic control measures. The emergency event model constructs a dynamic priority right-of-way allocation model. First, it collects information on the emergency event type, location, surrounding traffic conditions, and emergency vehicle locations through alarm systems, traffic monitoring equipment, and V2X communication technology, assessing the severity of the event and analyzing traffic conditions. Next, it builds a model, setting the goal of ensuring rapid arrival of emergency vehicles while minimizing disruption to normal traffic. Route selection and traffic light timing adjustments are defined as decision variables, and traffic flow, safety, and legal constraints are set. Mathematical programming or intelligent algorithms are used to solve the problem. After inputting relevant information into the model, specific allocation strategies are generated, including navigation for emergency vehicles and adjusting traffic light timings. During strategy implementation, the movement of emergency vehicles and traffic light execution are monitored in real time, and the strategies are dynamically adjusted based on the results. After the event, the effectiveness is evaluated using emergency vehicle arrival time as an indicator, and the model is optimized accordingly. Finally, the model is integrated with existing systems, and relevant personnel are trained to improve collaborative work efficiency and personnel's ability to respond to emergencies.

2. The intelligent traffic light adjustment method integrating vehicle network information according to claim 1, characterized in that, The outlier filtering includes: The raw data of the vehicle network is grouped according to the time series, and the interquartile range and standard deviation of each group are calculated to obtain the group data position difference sequence. Outlier data points in the grouped data position difference sequence are identified based on the sliding window mechanism, and missing data are compensated by cubic spline interpolation. Conflicting data collected by radar and cameras are fused and verified using a Kalman filter algorithm to obtain filtered multi-source traffic data.

3. The intelligent traffic light adjustment method integrating vehicle network information according to claim 1, characterized in that, The construction of the multimodal traffic state prediction model includes: Establish a spatiotemporal graph convolutional network to process road topology features; Long Short-Term Memory (LSTM) networks are used to model the temporal dependencies of traffic flow; Integrate attention mechanisms to dynamically weight vehicle intent information in V2X communication data; By optimizing the multi-objective loss function through a reinforcement learning framework, a multimodal traffic state prediction model is generated, balancing traffic efficiency and safety indicators.

4. The intelligent traffic light adjustment method integrating vehicle network information according to claim 1, characterized in that, The deployment of the edge computing nodes includes: Design a lightweight model compression algorithm to reduce the number of parameters in the prediction model; Data interaction between local nodes and the cloud platform is achieved through API interfaces. The local nodes process millisecond-level real-time control commands, while the cloud platform performs hourly-level strategy optimization. Establish a communication quality assessment module to dynamically switch between LTE-V2X direct communication and 5G cellular communication transmission channels.

5. The intelligent traffic light adjustment method integrating vehicle network information according to claim 1, characterized in that, The closed-loop feedback mechanism includes: Define a comprehensive evaluation index for traffic efficiency, which includes three dimensions: average delay time, number of stops, and fuel consumption. Develop a simulation verification platform based on digital twins to conduct virtual intersection stress tests on the new strategy; Build a self-healing system for abnormal operating conditions. When a policy execution deviation is detected to exceed a threshold, it will automatically roll back to the previous stable version.

6. The intelligent traffic light adjustment method integrating vehicle network information according to claim 1, characterized in that, The method further includes: Establish a tiered response mechanism, classifying traffic conditions into four levels of warning: red, orange, yellow, and blue. In response to a red alert, regional coordinated control is activated to coordinate all traffic lights within a predefined area to enter emergency mode. Develop a driver behavior prediction module that combines historical trajectory data to predict risky behaviors such as cutting in and jaywalking, and adjust the phase in advance.

7. A traffic light intelligent adjustment system integrating vehicle network information, characterized in that: The system is used to implement the intelligent traffic light adjustment method integrating vehicle network information as described in any one of claims 1-6, the system comprising: A multi-source traffic data acquisition module is used to collect multi-source traffic data in real time through roadside sensing devices, perform spatiotemporal alignment and outlier filtering on the multi-source traffic data, and obtain filtered multi-source traffic data. The traffic demand forecasting module is used to construct a deep learning-based multimodal traffic state prediction model using filtered multi-source traffic data, and combine historical traffic data and real-time vehicle network information to predict traffic demand in each direction within a future variable time window. A dynamic timing strategy generation module is used to generate a dynamic timing strategy based on the prediction results and synchronize the strategy parameters to the associated intersections via the V2X communication protocol. An edge computing node deployment module is used to deploy edge computing nodes to dynamically adjust the traffic light controller in real time, generate traffic light cycle parameters, phase difference parameters, and green ratio parameters based on the dynamic timing strategy, and send suggested driving speed commands to vehicle terminals through the vehicle-road cooperative system. A closed-loop feedback mechanism construction module is used to establish a closed-loop feedback mechanism, which performs online parameter correction and model iterative update of the dynamic timing strategy based on high-precision trajectory data and traffic efficiency indicators.

Images