An urban traffic network intelligent scheduling method and system based on digital twinning

Abstract

This invention relates to intelligent traffic scheduling technology, and discloses a method and system for intelligent scheduling of urban traffic networks based on digital twins. The method includes: acquiring multi-source traffic data from multiple intersections in a preset target area, performing precision alignment and fusion to obtain fused traffic data; constructing digital twins of two adjacent intersections in the target area based on the fused traffic data to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body; generating an initial scheduling strategy based on the multi-source traffic data using the first and second dynamic virtual traffic bodies; performing traffic simulation on the first and second dynamic virtual traffic bodies using the initial scheduling strategy; performing game-theoretic collaborative scheduling decisions based on the obtained simulation data to obtain an optimal scheduling strategy; generating a control instruction set based on the optimal scheduling strategy; and realizing intelligent scheduling of the traffic network based on the control instruction set. This invention can improve the effectiveness of traffic network scheduling.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic scheduling technology, and in particular to an intelligent scheduling method and system for urban traffic networks based on digital twins. Background Technology

[0002] In urban traffic network management, coordinated signal control at adjacent intersections is a key technology for improving regional traffic efficiency. Existing coordinated control methods can be mainly divided into the following categories: First, offline optimization schemes based on fixed timing. These methods rely on historical traffic flow data to formulate static plans, which cannot adapt to real-time dynamic changes in traffic flow. Their effectiveness decreases significantly when sudden events or atypical traffic patterns occur. Second, online schemes based on simple inductive control, such as using geomagnetic coils or video detectors to trigger phase switching, have a certain degree of real-time performance, but due to their limited sensing range and independent decision-making at each intersection, they lack a global perspective and are prone to causing discontinuous "green waves" or even secondary congestion. Third, centralized optimization models that have emerged in recent years collect data and uniformly calculate the optimal strategy through a regional control center. This method improves the coordination effect to some extent, but its core bottleneck is that the computational complexity increases exponentially with the number of intersections, making it difficult to meet the millisecond-level latency requirements of real-time response in large-scale road networks. Furthermore, the highly centralized architecture will cause the entire system to fail if communication is interrupted or the central server fails.

[0003] Existing methods generally treat the traffic system as a static model that can be completely controlled by a single entity, ignoring the inherent spatial connections and competitive interests between intersections. Each intersection has its own independent traffic demand and efficiency goals, and the optimization strategy of one intersection may come at the expense of the interests of downstream intersections. Traditional centralized optimization aims to find the global optimum, but often sacrifices local efficiency and cannot effectively quantify and balance the conflicts of interest between different intersections. As a result, the optimization strategy may be theoretically feasible, but its performance is poor in practice due to a lack of flexibility. Summary of the Invention

[0004] This invention provides a digital twin-based intelligent scheduling method and system for urban traffic networks, the main purpose of which is to solve the problem of poor performance of existing intelligent scheduling methods for urban traffic networks.

[0005] To achieve the above objectives, the present invention provides an intelligent scheduling method for urban traffic networks based on digital twins, comprising:

[0006] Real-time acquisition of multi-source traffic data from multiple intersections in a preset target area; precision alignment and fusion of the multi-source traffic data to obtain fused traffic data;

[0007] Based on the fused traffic data, digital twins of two adjacent intersections in the target area are constructed to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body;

[0008] An initial scheduling strategy is generated based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body.

[0009] The initial scheduling strategy is used to perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain simulation data.

[0010] The first and second dynamic virtual traffic bodies are used to make game-like collaborative scheduling decisions based on the inference data to obtain the optimal scheduling strategy.

[0011] A control instruction set is generated based on the optimal scheduling strategy, and intelligent scheduling of the traffic network is achieved by issuing the control instruction set to traffic signal equipment in the target area.

[0012] Optionally, the step of precision-aligning and fusing the multi-source traffic data to obtain fused traffic data includes:

[0013] The multi-source traffic data is decoded and subjected to spatiotemporal reference unification processing to obtain spatiotemporally aligned data;

[0014] Construct a spatiotemporal data cube based on the spatiotemporal alignment data;

[0015] Data conflict detection and credit assessment are performed on the spatiotemporal data cube to obtain a list of conflicting data units with credit scores;

[0016] Based on the list of conflicting data units, identify the conflict-free units in the spatiotemporal data cube;

[0017] The conflict-free data units are fused using a weighted average method, and the conflict data units are further fused using a weighted average method based on the credit scores of the conflict data unit list to obtain fused traffic data.

[0018] Optionally, the step of performing data conflict detection and credit assessment on the spatiotemporal data cube to obtain a list of conflicting data units with credit scores includes:

[0019] The data list of each cell in the spatiotemporal data cube is read sequentially to obtain a set of data lists;

[0020] Remove data lists containing only a single data item from the data list set to obtain the filtered data list set;

[0021] Calculate the variance, range, and interquartile range of each data list in the filtered data list set to obtain the statistical distribution characteristics;

[0022] By comparing the statistical distribution characteristics with a preset threshold, conflicting units in the spatiotemporal data cube are identified, and a list of conflicting units is obtained.

[0023] A multidimensional credit assessment is performed on each conflicting unit in the conflicting unit list to obtain a comprehensive credit score;

[0024] The comprehensive credit score is associated with the list of conflicting units to obtain a list of conflicting data units with credit scores.

[0025] Optionally, the step of performing a multi-dimensional credit assessment on each conflicting unit in the conflicting unit list to obtain a comprehensive credit score includes:

[0026] Obtain the historical credit score for each conflict unit in the conflict unit list;

[0027] Real-time data confidence is calculated for each conflict unit in the conflict unit list based on real-time acquired multi-source traffic data to obtain a confidence score;

[0028] Obtain the fusion value of each conflicting unit and its neighboring units in the list of conflicting units;

[0029] The consistency of the fusion value between each conflicting unit in the conflicting unit list and its neighboring units is calculated to obtain a spatiotemporal proximity consistency score;

[0030] The historical credit score, the confidence score, and the spatiotemporal proximity consistency score are weighted and fused based on preset weight parameters to obtain a comprehensive credit score.

[0031] Optionally, generating an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body includes:

[0032] Feature extraction is performed on the multi-source traffic data corresponding to the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain a first feature vector and a second feature vector;

[0033] The first feature vector and the second feature vector are used to perform policy matching with a preset policy library to obtain a first matching policy and a second matching policy.

[0034] The first matching strategy and the second matching strategy are conflict-resolved and merged to obtain the initial scheduling strategy.

[0035] Optionally, the step of using the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decisions based on the inference data to obtain the optimal scheduling strategy includes:

[0036] The traffic performance index of the first dynamic virtual traffic body is calculated based on the simulation data to obtain the first index;

[0037] The traffic performance index of the second dynamic virtual traffic body is calculated based on the simulation data to obtain the second index;

[0038] Based on the dynamic fine-tuning identification of the adjustment directions of the first dynamic virtual traffic body and the second dynamic virtual traffic body, the first adjustment direction and the second adjustment direction are obtained;

[0039] Determine whether the first adjustment direction and the second adjustment direction are the same;

[0040] If they are the same, the initial scheduling strategy is iteratively adjusted according to the first adjustment direction to obtain the optimal scheduling strategy;

[0041] If they are not the same, then determine whether the first indicator is greater than the second indicator;

[0042] If it is less than or equal to, then the initial scheduling strategy is adjusted in a coordinated manner according to the first adjustment direction to obtain the optimal scheduling strategy;

[0043] If the value is greater than the initial scheduling strategy, then the initial scheduling strategy is adjusted in a coordinated manner according to the second adjustment direction to obtain the optimal scheduling strategy.

[0044] Optionally, the step of identifying the adjustment directions of the first and second dynamic virtual traffic bodies based on dynamic fine-tuning to obtain the first and second adjustment directions includes:

[0045] The initial scheduling strategy is adjusted in the positive direction based on a preset fine-tuning step size to obtain a positive adjustment scheduling strategy.

[0046] Based on the positive adjustment scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are calculated to obtain the adjusted first indicator and the adjusted second indicator.

[0047] If the adjusted first indicator is greater than the first indicator, then the first adjustment direction of the first dynamic virtual traffic body is determined to be a positive direction; otherwise, the first adjustment direction is determined to be a negative direction.

[0048] If the adjusted second indicator is greater than the second indicator, then the second adjustment direction of the second dynamic virtual traffic body is determined to be a positive direction; otherwise, the second adjustment direction is determined to be a negative direction.

[0049] Optionally, the step of coordinating the adjustment of the initial scheduling strategy according to the first adjustment direction to obtain the optimal scheduling strategy includes:

[0050] The initial scheduling strategy is adjusted according to the first adjustment direction and the preset adjustment step size to obtain the adjusted scheduling strategy.

[0051] Based on the adjusted scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are recalculated to obtain the third and fourth indicators.

[0052] Calculate the absolute value of the difference between the third indicator and the fourth indicator;

[0053] Determine whether the absolute value of the difference is greater than a preset difference threshold;

[0054] If the absolute value of the difference is greater than the difference threshold, the adjusted scheduling strategy is adjusted according to the first adjustment direction and the adjustment step size. After obtaining a new adjusted scheduling strategy, the process returns to the step of recalculating the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body according to the adjusted scheduling strategy to obtain the third and fourth indicators.

[0055] If the absolute value of the difference is less than or equal to the difference threshold, then the adjusted scheduling strategy is confirmed as the optimal scheduling strategy.

[0056] Optionally, the step of calculating the traffic performance index of the first dynamic virtual traffic body based on the simulation data to obtain the first index includes:

[0057] The total delay, total throughput, maximum queue length, and coordinated throughput of the intersection of the first dynamic virtual traffic body are obtained from the simulation data to obtain the first total delay, first throughput, first queue length, and first throughput.

[0058] Calculate the ratio of the first throughput to the first total delay, and multiply the calculated ratio by the reciprocal of the first queue length to obtain an independent performance index;

[0059] Calculate the difference between the first queue length and the preset baseline queue length, and divide the calculated difference by the baseline queue length to obtain the collaborative penalty factor;

[0060] Calculate the ratio of the first throughput to the first throughput to obtain the collaborative reward factor;

[0061] The independent performance index, the collaborative penalty factor, and the collaborative reward factor are weighted and fused to obtain the first index.

[0062] To address the above problems, the present invention also provides an intelligent dispatching system for urban traffic networks based on digital twins, the system comprising:

[0063] The multi-source data fusion module is used to acquire multi-source traffic data from multiple intersections in a preset target area in real time, perform precision alignment on the multi-source traffic data and fuse it to obtain fused traffic data.

[0064] The traffic body construction module is used to construct digital twins of two adjacent intersections in the target area based on the fused traffic data, thereby obtaining a first dynamic virtual traffic body and a second dynamic virtual traffic body.

[0065] The initial strategy generation module is used to generate an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body;

[0066] The optimal strategy generation module is used to perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body using the initial scheduling strategy to obtain simulation data, and to use the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decision based on the simulation data to obtain the optimal scheduling strategy.

[0067] The intelligent scheduling module is used to generate a set of control instructions based on the optimal scheduling strategy, and to realize intelligent scheduling of the traffic network by issuing the set of control instructions to traffic signal equipment in the target area.

[0068] This invention acquires multi-source traffic data from multiple intersections in a preset target area in real time, performs precision alignment and fusion of the multi-source traffic data to obtain fused traffic data, and constructs digital twins of two adjacent intersections in the target area based on the fused traffic data, resulting in a first dynamic virtual traffic body and a second dynamic virtual traffic body. This achieves a precise mapping from the physical world to the virtual space, providing a core carrier for subsequent simulation and optimization decisions. The first and second dynamic virtual traffic bodies are used to generate an initial scheduling strategy based on the multi-source traffic data, ensuring the basic effectiveness of the method and significantly shortening the optimization computation time. The convergence time of the method is reduced, improving decision-making efficiency. Traffic simulation is performed on the first and second dynamic virtual traffic bodies using the initial scheduling strategy to obtain simulation data. Based on this simulation data, the first and second dynamic virtual traffic bodies perform game-like collaborative scheduling decisions to obtain the optimal scheduling strategy. This achieves synergistic efficiency between intersections and balanced optimization of the overall network. A control command set is generated based on the optimal scheduling strategy, and intelligent scheduling of the traffic network is achieved by issuing the control command set to traffic signal equipment in the target area. This solves the problem of poor performance of existing intelligent scheduling methods for urban traffic networks. Attached Figure Description

[0069] Figure 1 A flowchart illustrating an intelligent scheduling method for urban traffic networks based on digital twins, provided in an embodiment of the present invention;

[0070] Figure 2 This is a functional block diagram of an intelligent urban traffic network scheduling system based on digital twins, provided as an embodiment of the present invention.

[0071] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0072] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0073] This application provides a digital twin-based intelligent scheduling method for urban traffic networks. The executing entity of this digital twin-based intelligent scheduling method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the digital twin-based intelligent scheduling method for urban traffic networks can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0074] Reference Figure 1 The diagram shown is a flowchart illustrating an intelligent urban traffic network scheduling method based on digital twins according to an embodiment of the present invention. In this embodiment, the intelligent urban traffic network scheduling method based on digital twins includes:

[0075] S1. Real-time acquisition of multi-source traffic data from multiple intersections in a preset target area, precision alignment and fusion of the multi-source traffic data to obtain fused traffic data.

[0076] In this embodiment of the invention, the multi-source traffic data may include traffic light data and real-time traffic flow at intersections, as well as real-time speed, location, and queue length of vehicles.

[0077] In this embodiment of the invention, the step of precision alignment and fusion of the multi-source traffic data to obtain fused traffic data includes:

[0078] The multi-source traffic data is decoded and subjected to spatiotemporal reference unification processing to obtain spatiotemporally aligned data;

[0079] Construct a spatiotemporal data cube based on the spatiotemporal alignment data;

[0080] Data conflict detection and credit assessment are performed on the spatiotemporal data cube to obtain a list of conflicting data units with credit scores;

[0081] Based on the list of conflicting data units, identify the conflict-free units in the spatiotemporal data cube;

[0082] The conflict-free data units are fused using a weighted average method, and the conflict data units are further fused using a weighted average method based on the credit scores of the conflict data unit list to obtain fused traffic data.

[0083] In detail, the process of decoding the multi-source traffic data and performing spatiotemporal benchmark unification processing to obtain spatiotemporally aligned data involves parsing and extracting raw data streams from different sources (such as protocol messages from traffic lights, video streams from cameras, and NMEA statements from GPS) according to their specific communication protocols and data formats, converting them into standardized data objects that the system can process. Spatiotemporal benchmark unification is a key innovation that ensures all data can be jointly analyzed: in the time dimension, the timestamps of all data are unified to the same time benchmark, and the delay caused by different transmission links is compensated to ensure that the data are within the same time slice; in the spatial dimension, the geographical coordinates (such as latitude and longitude) of all data are mapped to a unified high-precision map lane-level coordinate system.

[0084] In detail, the construction of the spatiotemporal data cube based on the spatiotemporal aligned data is a three-dimensional mathematical model with three dimensions: time, space, and traffic indicators. The time dimension is sliced ​​at fixed intervals (e.g., 1 minute); the space dimension is represented by lanes, intersections, and road segments; and the traffic indicator dimension includes flow rate, speed, queue length, etc. This step fills the corresponding cells of the cube with the discrete spatiotemporal aligned data output from the previous step according to their time and space attributes. For example, all speed values ​​detected in a "certain lane" within the past minute are aggregated into the "speed" cell corresponding to that lane.

[0085] In this embodiment of the invention, the step of performing data conflict detection and credit assessment on the spatiotemporal data cube to obtain a list of conflicting data units with credit scores includes:

[0086] The data list of each cell in the spatiotemporal data cube is read sequentially to obtain a set of data lists;

[0087] Remove data lists containing only a single data item from the data list set to obtain the filtered data list set;

[0088] Calculate the variance, range, and interquartile range of each data list in the filtered data list set to obtain the statistical distribution characteristics;

[0089] By comparing the statistical distribution characteristics with a preset threshold, conflicting units in the spatiotemporal data cube are identified, and a list of conflicting units is obtained.

[0090] A multidimensional credit assessment is performed on each conflicting unit in the conflicting unit list to obtain a comprehensive credit score;

[0091] The comprehensive credit score is associated with the list of conflicting units to obtain a list of conflicting data units with credit scores.

[0092] In detail, the range refers to the difference between the maximum and minimum values ​​in the calculated data list.

[0093] In detail, the interquartile range is the difference between the 75th percentile and the 25th percentile of a set of data, which measures the dispersion of the middle 50% of the data.

[0094] In detail, the step of identifying conflicting units in the spatiotemporal data cube by comparing the statistical distribution characteristics with preset thresholds involves comparing the statistical distribution characteristics (variance, range, interquartile range) of each unit with the preset thresholds. For example, if the range of a unit's data exceeds a threshold (e.g., 30 km / h), or the variance is too large, then the unit is determined to be a conflicting unit.

[0095] In this embodiment of the invention, the step of performing a multi-dimensional credit assessment on each conflicting unit in the conflicting unit list to obtain a comprehensive credit score includes:

[0096] Obtain the historical credit score for each conflict unit in the conflict unit list;

[0097] Real-time data confidence is calculated for each conflict unit in the conflict unit list based on real-time acquired multi-source traffic data to obtain a confidence score;

[0098] Obtain the fusion value of each conflicting unit and its neighboring units in the list of conflicting units;

[0099] The consistency of the fusion value between each conflicting unit in the conflicting unit list and its neighboring units is calculated to obtain a spatiotemporal proximity consistency score;

[0100] The historical credit score, the confidence score, and the spatiotemporal proximity consistency score are weighted and fused based on preset weight parameters to obtain a comprehensive credit score.

[0101] In detail, the historical credit score is a dynamically updated metric that quantifies the long-term reliability performance of each data source. This score is continuously calibrated based on the error between the data provided by the data source over a past period and the final verified "true value" (usually calculated from later system fusion results or higher-precision data). A high score indicates that the data source has historically performed stably and accurately; a low score suggests it is error-prone or has unstable performance.

[0102] In detail, the real-time data confidence calculation for each conflict unit in the conflict unit list based on real-time acquired multi-source traffic data focuses on the generation quality and contextual rationality of the data itself. For example, for GPS data, the Horizontal Accuracy Factor (HDOP) is extracted from its message information. The larger the HDOP value, the greater the positioning error, and the lower the confidence score of the data. It also checks whether the data value is within the physically possible range (e.g., vehicle speed on urban roads should not exceed 200 km / h). If unreasonable values ​​are found, the confidence score is reduced.

[0103] In detail, calculating the consistency of the fusion value of each conflicting unit in the conflicting unit list with its neighboring units involves calculating the difference (such as absolute difference or relative error) between the fusion values ​​of the conflicting unit and its neighboring units. The smaller the difference, the more the data value matches the known spatiotemporal trend of the system, and the higher its spatiotemporal proximity consistency score; conversely, a data value that is out of step with the surrounding environment (for example, the data in this unit shows a flow of 0, while the flow of upstream and downstream units is very high) will have a very low consistency score. This score can extremely effectively identify "outliers" caused by transient interference, local sensor failure, or transmission errors.

[0104] In detail, the weighted fusion of the historical credit score, the confidence score, and the spatiotemporal proximity consistency score based on preset weight parameters can be achieved by using weights of 0.3, 0.5, and 0.2 respectively for the confidence score and the spatiotemporal proximity consistency score.

[0105] S2. Construct digital twins of two adjacent intersections in the target area based on the fused traffic data to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body.

[0106] In this embodiment of the invention, the construction of digital twins of two adjacent intersections in the target area based on the fused traffic data to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body refers to using a high-precision map as the skeleton and fused traffic data as the nerves. First, the static infrastructure models of the two intersections (including lane topology, traffic light positions, etc.) are reproduced one-to-one in the virtual space. Then, real-time traffic conditions (such as vehicle positions, speeds, and signal phases) are injected into the models in a data-driven manner to make them operational. Finally, each intersection entity is encapsulated with an intelligent agent (i.e., a dynamic virtual traffic body) with perception and decision-making capabilities. This agent has a built-in reinforcement learning model and can interact and coordinate with other agents with the goal of optimizing the traffic efficiency of the intersection.

[0107] S3. Generate an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body.

[0108] In this embodiment of the invention, the step of generating an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body includes:

[0109] Feature extraction is performed on the multi-source traffic data corresponding to the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain a first feature vector and a second feature vector;

[0110] The first feature vector and the second feature vector are used to perform policy matching with a preset policy library to obtain a first matching policy and a second matching policy.

[0111] The first matching strategy and the second matching strategy are conflict-resolved and merged to obtain the initial scheduling strategy.

[0112] In detail, the feature extraction of the multi-source traffic data corresponding to the first and second dynamic virtual traffic bodies is a crucial step in data preprocessing and information condensation. While the original multi-source traffic data (such as flow rate, speed, and queue length) is rich in information, its high dimensionality and complex format prevent it from being directly used by decision-making logic. Therefore, the first dynamic virtual traffic body (representing intersection one) and the second dynamic virtual traffic body (representing intersection two) need to independently extract key features from their respective data that highly summarize the essence of their current traffic state. These features may include, but are not limited to, traffic flow intensity, saturation of queued vehicles, urgency of vehicle delays, and pattern deviations from historical periods. Through mathematical methods (such as statistical aggregation and principal component analysis), these features are constructed into a structured, low-dimensional numerical set, i.e., a feature vector.

[0113] In detail, the process of matching the first and second feature vectors with a preset strategy library to obtain a first matching strategy and a second matching strategy involves each dynamic virtual traffic entity having a preset strategy library. This strategy library is essentially a large "situation-countermeasure" mapping table, storing scheduling strategies that have been verified as effective under various historical or simulated traffic scenarios. The strategy library can be constructed based on expert experience rules or generated by training machine learning models (such as decision trees or neural networks). Here, the first dynamic virtual traffic entity uses the first feature vector as a query condition to perform matching or reasoning in its strategy library, quickly retrieving one or more strategies most suitable for the current intersection state, and combining them into a first matching strategy (e.g., "suggest increasing the phase difference by 10 seconds"). Similarly, the second dynamic virtual traffic entity independently performs the same operation to obtain a second matching strategy that serves its own intersection interests (e.g., "suggest decreasing the phase difference by 5 seconds" or "suggest increasing the red light time by 5 seconds"). The output of this step is two potentially conflicting locally optimal strategies, each aiming to optimize its own traffic efficiency.

[0114] In detail, the conflict resolution and fusion of the first and second matching strategies to obtain the initial scheduling strategy can be based on preset rules that prioritize the overall system benefits. For example, it can compare the severity of traffic conditions at two intersections and prioritize the strategy of the one with more severe traffic congestion; or it can calculate a weighted average of the two strategies, with the weights dynamically allocated according to the importance of the current traffic load at each intersection. After resolving the conflict, the agreed-upon adjustments are applied to the current scheduling parameters, fusing them to generate a unified and executable initial scheduling strategy.

[0115] S4. Using the initial scheduling strategy, perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain simulation data.

[0116] In this embodiment of the invention, the step of using the initial scheduling strategy to perform traffic simulation on the first and second dynamic virtual traffic bodies to obtain simulation data refers to inputting the initial scheduling strategy as a control parameter into the digital twin ultra-real-time simulation engine, driving the collaborative simulation environment composed of the two virtual traffic bodies to simulate micro-behavioral interactions such as vehicle following, lane changing, and traffic light response over a future period of time (e.g., 15 minutes) at speeds far exceeding reality, and finally collecting and recording macro-performance indicators (such as average delay, queue length, and throughput at the two intersections) and micro-vehicle trajectory data that can comprehensively reflect the simulation results. This dataset is the simulation data used for subsequent game decision-making.

[0117] In this embodiment of the invention, the extrapolated data may include the total delay of the intersection, the total throughput of the intersection, and the maximum queue length of the intersection.

[0118] S5. Using the first dynamic virtual traffic body and the second dynamic virtual traffic body, a game-like collaborative scheduling decision is made based on the simulation data to obtain the optimal scheduling strategy.

[0119] In this embodiment of the invention, the step of using the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decisions based on the inference data to obtain the optimal scheduling strategy includes:

[0120] The traffic performance index of the first dynamic virtual traffic body is calculated based on the simulation data to obtain the first index;

[0121] The traffic performance index of the second dynamic virtual traffic body is calculated based on the simulation data to obtain the second index;

[0122] Based on the dynamic fine-tuning identification of the adjustment directions of the first dynamic virtual traffic body and the second dynamic virtual traffic body, the first adjustment direction and the second adjustment direction are obtained;

[0123] Determine whether the first adjustment direction and the second adjustment direction are the same;

[0124] If they are the same, the initial scheduling strategy is iteratively adjusted according to the first adjustment direction to obtain the optimal scheduling strategy;

[0125] If they are not the same, then determine whether the first indicator is greater than the second indicator;

[0126] If it is less than or equal to, then the initial scheduling strategy is adjusted in a coordinated manner according to the first adjustment direction to obtain the optimal scheduling strategy;

[0127] If the value is greater than the initial scheduling strategy, then the initial scheduling strategy is adjusted in a coordinated manner according to the second adjustment direction to obtain the optimal scheduling strategy.

[0128] In detail, the calculation of traffic performance indicators for the first and second dynamic virtual traffic entities based on the extrapolated data requires extracting quantifiable traffic performance indicators that represent the core interests of each dynamic virtual traffic entity. The performance of the first dynamic virtual traffic entity can be measured by a first indicator (such as total queue length or average delay), while the performance of the second dynamic virtual traffic entity can be measured by a second indicator (such as throughput or vehicle throughput rate). This transformation process abstracts the complex extrapolation results into comparable and competitive numerical targets between the two agents.

[0129] In detail, the determination of whether the first adjustment direction and the second adjustment direction are the same is a conflict detection node in the game, which determines the branch of the subsequent decision path. The system compares the two adjustment directions obtained in the first step. If the directions are the same (e.g., both sides suggest "increasing" the phase difference), it means that the interests of the two intersections are the same in the current situation, and there is an adjustment direction that can improve the performance of both sides at the same time, making cooperation simple. If the directions are different (e.g., one wants to "increase" and the other wants to "decrease"), it means that there is a direct conflict of interests between the two intersections, and the gain of one side's interests must come at the cost of the loss of the other side's interests. In this case, the decision-making mechanism needs to be more complex to resolve this conflict.

[0130] In detail, phase difference refers to the time difference between the start of the traffic light cycle (or the start of the green light of a critical phase) at one intersection (downstream intersection) and the start of the traffic light cycle (or the start of the green light of a critical phase) at a reference intersection (upstream intersection) in signal coordination control.

[0131] In detail, the determination of whether the first indicator is greater than the second indicator requires an arbitration mechanism to decide whose solution to adopt when the adjustment directions are opposite. The arbitration basis for this step is the urgency of the current performance. The system compares the values ​​of the first and second indicators (for example, comparing the queue length at intersection one with the throughput loss value converted from intersection two). If the first indicator is greater than the second indicator, it indicates that the performance problem of the intersection represented by the first DVTA is more serious and more urgent; conversely, the problem of the second DVTA deserves priority. This step transforms the question of "whose interests should be prioritized" into an objective and quantifiable comparison.

[0132] In this embodiment of the invention, the step of obtaining a first adjustment direction and a second adjustment direction based on dynamically fine-tuning the identification of the adjustment directions of the first dynamic virtual traffic body and the second dynamic virtual traffic body includes:

[0133] The initial scheduling strategy is adjusted in the positive direction based on a preset fine-tuning step size to obtain a positive adjustment scheduling strategy.

[0134] Based on the positive adjustment scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are calculated to obtain the adjusted first indicator and the adjusted second indicator.

[0135] If the adjusted first indicator is greater than the first indicator, then the first adjustment direction of the first dynamic virtual traffic body is determined to be a positive direction; otherwise, the first adjustment direction is determined to be a negative direction.

[0136] If the adjusted second indicator is greater than the second indicator, then the second adjustment direction of the second dynamic virtual traffic body is determined to be a positive direction; otherwise, the second adjustment direction is determined to be a negative direction.

[0137] In this embodiment of the invention, the positive adjustment of the initial scheduling strategy based on a preset fine-tuning step size to obtain a positive adjustment scheduling strategy is an initialization operation for dynamic fine-tuning. Its purpose is to generate an exploratory strategy for detecting the direction of changes in system performance. The "initial scheduling strategy" is the currently executing baseline scheme (e.g., the phase difference between intersection A and intersection B is 20 seconds). The system adjusts a parameter (such as the phase difference) of this strategy in a positive direction (i.e., by increasing its value) with a preset, small fine-tuning step size (e.g., increasing it by 2 seconds), thereby obtaining a novel, "perturbed" positive adjustment scheduling strategy (e.g., the phase difference becomes 22 seconds). This micro-adjustment ensures that the detection process does not cause drastic impacts on the system, while accurately testing the trend of the impact of parameter changes on system performance.

[0138] In detail, the calculation of the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body based on the positive adjustment scheduling strategy involves first performing traffic simulation based on the positive adjustment scheduling strategy, and then calculating the traffic performance indicators based on the simulation results.

[0139] In this embodiment of the invention, the step of coordinating the initial scheduling strategy according to the first adjustment direction to obtain the optimal scheduling strategy includes:

[0140] The initial scheduling strategy is adjusted according to the first adjustment direction and the preset adjustment step size to obtain the adjusted scheduling strategy.

[0141] Based on the adjusted scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are recalculated to obtain the third and fourth indicators.

[0142] Calculate the absolute value of the difference between the third indicator and the fourth indicator;

[0143] Determine whether the absolute value of the difference is greater than a preset difference threshold;

[0144] If the absolute value of the difference is greater than the difference threshold, the adjusted scheduling strategy is adjusted according to the first adjustment direction and the adjustment step size. After obtaining a new adjusted scheduling strategy, the process returns to the step of recalculating the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body according to the adjusted scheduling strategy to obtain the third and fourth indicators.

[0145] If the absolute value of the difference is less than or equal to the difference threshold, then the adjusted scheduling strategy is confirmed as the optimal scheduling strategy.

[0146] In detail, adjusting the initial scheduling strategy according to the first adjustment direction and the preset adjustment step size to obtain the adjusted scheduling strategy is the starting operation of the collaborative adjustment loop. Its function is to perform a specific, quantitative strategy adjustment based on the result of the game decision. The first adjustment direction is the output of the previous game process (e.g., "increase the phase difference"), which indicates the direction of optimization. The preset adjustment step size (e.g., increasing or decreasing by 2 seconds each time) is a hyperparameter that controls the adjustment magnitude, and its value is small to ensure the stability and refinement of the optimization process. The system applies these two inputs to the current initial scheduling strategy (e.g., the phase difference is 20 seconds) to generate a brand new, adjusted scheduling strategy (e.g., the phase difference becomes 22 seconds). This step transforms the abstract "direction" instruction into a concrete scheme that can be verified by simulation.

[0147] Specifically, if the absolute value of the difference is greater than the difference threshold, it indicates that the performance of the two traffic bodies is in an unacceptable imbalance (e.g., one intersection's efficiency has greatly improved while the performance of the other intersection has deteriorated sharply), and the optimization process must continue to iterate to seek a more balanced solution. If the absolute value of the difference is less than or equal to the threshold, it means that the system has reached a satisfactory equilibrium state, and the optimization objective has been achieved.

[0148] In this embodiment of the invention, the step of calculating the traffic performance index of the first dynamic virtual traffic body based on the inference data to obtain the first index includes:

[0149] The total delay, total throughput, maximum queue length, and coordinated throughput of the intersection of the first dynamic virtual traffic body are obtained from the simulation data to obtain the first total delay, first throughput, first queue length, and first throughput.

[0150] Calculate the ratio of the first throughput to the first total delay, and multiply the calculated ratio by the reciprocal of the first queue length to obtain an independent performance index;

[0151] Calculate the difference between the first queue length and the preset baseline queue length, and divide the calculated difference by the baseline queue length to obtain the collaborative penalty factor;

[0152] Calculate the ratio of the first throughput to the first throughput to obtain the collaborative reward factor;

[0153] The independent performance index, the collaborative penalty factor, and the collaborative reward factor are weighted and fused to obtain the first index.

[0154] In this embodiment of the invention, the step of calculating the traffic performance index of the second dynamic virtual traffic body based on the simulation data to obtain the second index is the same as the step of calculating the traffic performance index of the first dynamic virtual traffic body based on the simulation data to obtain the first index, and will not be described again here.

[0155] S6. Generate a control instruction set according to the optimal scheduling strategy, and realize intelligent scheduling of the traffic network by issuing the control instruction set to traffic signal equipment in the target area.

[0156] In this embodiment of the invention, generating a control instruction set according to the optimal scheduling strategy refers to converting the abstract strategy (such as the coordinated phase difference and green light ratio scheme) obtained by game optimization into standardized instructions that can be recognized and executed by specific traffic control equipment through a preset instruction compilation engine. This includes generating phase timing parameters (such as green light start time and duration) for the signal controller, generating guidance information text and release timing for the variable message sign, and encapsulating them into an instruction set through a secure communication protocol.

[0157] In this embodiment of the invention, the intelligent scheduling of the traffic network is achieved by issuing the control command set to the traffic signal equipment in the target area. The encapsulated control command set can be sent securely, reliably, and with low latency to the execution queue of the signal controllers at each intersection in the target area via V2X or fiber optic networks. After verifying the command signature, the equipment immediately drives the traffic lights to operate according to the new strategy, thereby achieving intelligent scheduling goals such as coordinated release between intersections and regional congestion relief.

[0158] like Figure 2 The diagram shown is a functional block diagram of an intelligent urban traffic network scheduling system based on digital twins, provided by an embodiment of the present invention.

[0159] The intelligent urban traffic network scheduling system 100 based on digital twins described in this invention can be installed in an electronic device. Depending on the functions implemented, the intelligent urban traffic network scheduling system 100 based on digital twins may include a multi-source data fusion module 101, a traffic body construction module 102, an initial strategy generation module 103, an optimal strategy generation module 104, and an intelligent scheduling module 105. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0160] In this embodiment, the functions of each module / unit are as follows:

[0161] The multi-source data fusion module 101 is used to acquire multi-source traffic data from multiple intersections in a preset target area in real time, perform precision alignment and fusion on the multi-source traffic data, and obtain fused traffic data.

[0162] The traffic body construction module 102 is used to construct digital twins of two adjacent intersections in the target area based on the fused traffic data, to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body.

[0163] The initial strategy generation module 103 is used to generate an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body.

[0164] The optimal strategy generation module 104 is used to perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body using the initial scheduling strategy to obtain simulation data, and to use the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decision based on the simulation data to obtain the optimal scheduling strategy.

[0165] The intelligent scheduling module 105 is used to generate a set of control instructions according to the optimal scheduling strategy, and to realize intelligent scheduling of the traffic network by sending the set of control instructions to traffic signal equipment in the target area.

[0166] In detail, the modules of the digital twin-based intelligent urban traffic network dispatching system 100 described in this embodiment of the invention employ the same methods as described above. Figure 1 The method uses the same technical means as the intelligent scheduling method for urban traffic networks based on digital twins described in the article, and can produce the same technical effects, so it will not be elaborated here.

[0167] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0168] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0169] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0170] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0171] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0172] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0173] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or systems stated in a system claim may also be implemented by a single unit or system through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for intelligent scheduling of urban traffic networks based on digital twins, characterized in that, The method includes: Multi-source traffic data is decoded and subjected to spatiotemporal benchmark unification processing to obtain spatiotemporally aligned data. A spatiotemporal data cube is constructed based on the spatiotemporal aligned data. The data list of each cell in the spatiotemporal data cube is read sequentially to obtain a data list set. Data lists with only single data points are removed from the data list set to obtain a filtered data list set. The variance, range, and interquartile range of each data list in the filtered data list set are calculated to obtain statistical distribution characteristics. Conflicting cells in the spatiotemporal data cube are identified by comparing the statistical distribution characteristics with preset thresholds to obtain a conflict cell list. The historical credit score of each conflict cell in the conflict cell list is obtained. Based on real-time acquired multi-source traffic data, each conflict cell in the conflict cell list is analyzed in real-time. Based on the confidence level, a confidence score is obtained. The fusion value of each conflicting unit in the conflicting unit list and its neighboring units is obtained. The consistency of the fusion value of each conflicting unit in the conflicting unit list and its neighboring units is calculated to obtain a spatiotemporal proximity consistency score. The historical credit score, the confidence score, and the spatiotemporal proximity consistency score are weighted and fused based on preset weight parameters to obtain a comprehensive credit score. The comprehensive credit score is associated with the conflicting unit list to obtain a list of conflicting data units with credit scores. Based on the list of conflicting data units, non-conflicting units in the spatiotemporal data cube are identified. The non-conflicting units are fused using a weighted average method. The conflicting data units are then weighted and fused based on the credit scores of the list of conflicting data units to obtain fused traffic data. Based on the fused traffic data, digital twins of two adjacent intersections in the target area are constructed to obtain a first dynamic virtual traffic body and a second dynamic virtual traffic body; An initial scheduling strategy is generated based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body. The initial scheduling strategy is used to perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain simulation data. The first and second dynamic virtual traffic bodies are used to make game-like collaborative scheduling decisions based on the inference data to obtain the optimal scheduling strategy. A control instruction set is generated based on the optimal scheduling strategy, and intelligent scheduling of the traffic network is achieved by issuing the control instruction set to traffic signal equipment in the target area.

2. The intelligent scheduling method for urban traffic networks based on digital twins as described in claim 1, characterized in that, The step of generating an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body includes: Feature extraction is performed on the multi-source traffic data corresponding to the first dynamic virtual traffic body and the second dynamic virtual traffic body to obtain a first feature vector and a second feature vector; The first feature vector and the second feature vector are used to perform policy matching with a preset policy library to obtain a first matching policy and a second matching policy. The first matching strategy and the second matching strategy are conflict-resolved and merged to obtain the initial scheduling strategy.

3. The intelligent scheduling method for urban traffic networks based on digital twins as described in claim 1, characterized in that, The step of using the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decisions based on the inferred data to obtain the optimal scheduling strategy includes: The traffic performance index of the first dynamic virtual traffic body is calculated based on the simulation data to obtain the first index; The traffic performance index of the second dynamic virtual traffic body is calculated based on the simulation data to obtain the second index; Based on the dynamic fine-tuning identification of the adjustment directions of the first dynamic virtual traffic body and the second dynamic virtual traffic body, the first adjustment direction and the second adjustment direction are obtained; Determine whether the first adjustment direction and the second adjustment direction are the same; If they are the same, the initial scheduling strategy is iteratively adjusted according to the first adjustment direction to obtain the optimal scheduling strategy; If they are not the same, then determine whether the first indicator is greater than the second indicator; If it is less than or equal to, then the initial scheduling strategy is adjusted in a coordinated manner according to the first adjustment direction to obtain the optimal scheduling strategy; If the value is greater than the initial scheduling strategy, then the initial scheduling strategy is adjusted in a coordinated manner according to the second adjustment direction to obtain the optimal scheduling strategy.

4. The intelligent scheduling method for urban traffic networks based on digital twins as described in claim 3, characterized in that, The step of identifying the adjustment directions of the first and second dynamic virtual traffic bodies based on dynamic fine-tuning to obtain the first and second adjustment directions includes: The initial scheduling strategy is adjusted in the positive direction based on a preset fine-tuning step size to obtain a positive adjustment scheduling strategy. Based on the positive adjustment scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are calculated to obtain the adjusted first indicator and the adjusted second indicator. If the adjusted first indicator is greater than the first indicator, then the first adjustment direction of the first dynamic virtual traffic body is determined to be a positive direction; otherwise, the first adjustment direction is determined to be a negative direction. If the adjusted second indicator is greater than the second indicator, then the second adjustment direction of the second dynamic virtual traffic body is determined to be a positive direction; otherwise, the second adjustment direction is determined to be a negative direction.

5. The intelligent scheduling method for urban traffic networks based on digital twins as described in claim 3, characterized in that, The step of coordinating the initial scheduling strategy according to the first adjustment direction to obtain the optimal scheduling strategy includes: The initial scheduling strategy is adjusted according to the first adjustment direction and the preset adjustment step size to obtain the adjusted scheduling strategy. Based on the adjusted scheduling strategy, the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body are recalculated to obtain the third and fourth indicators. Calculate the absolute value of the difference between the third indicator and the fourth indicator; Determine whether the absolute value of the difference is greater than a preset difference threshold; If the absolute value of the difference is greater than the difference threshold, the adjusted scheduling strategy is adjusted according to the first adjustment direction and the adjustment step size. After obtaining a new adjusted scheduling strategy, the process returns to the step of recalculating the traffic performance indicators of the first dynamic virtual traffic body and the second dynamic virtual traffic body according to the adjusted scheduling strategy to obtain the third and fourth indicators. If the absolute value of the difference is less than or equal to the difference threshold, then the adjusted scheduling strategy is confirmed as the optimal scheduling strategy.

6. The intelligent scheduling method for urban traffic networks based on digital twins as described in claim 3, characterized in that, The step of calculating the traffic performance index of the first dynamic virtual traffic body based on the simulation data to obtain the first index includes: The total delay, total throughput, maximum queue length, and coordinated throughput of the intersection of the first dynamic virtual traffic body are obtained from the simulation data to obtain the first total delay, first throughput, first queue length, and first throughput. Calculate the ratio of the first throughput to the first total delay, and multiply the calculated ratio by the reciprocal of the first queue length to obtain an independent performance index; Calculate the difference between the first queue length and the preset baseline queue length, and divide the calculated difference by the baseline queue length to obtain the collaborative penalty factor; Calculate the ratio of the first throughput to the first throughput to obtain the collaborative reward factor; The independent performance index, the collaborative penalty factor, and the collaborative reward factor are weighted and fused to obtain the first index.

7. A digital twin-based intelligent dispatching system for urban traffic networks, characterized in that, The system includes: The multi-source data fusion module decodes multi-source traffic data and performs spatiotemporal benchmark unification processing to obtain spatiotemporally aligned data. A spatiotemporal data cube is constructed based on this aligned data. The data list of each unit in the spatiotemporal data cube is read sequentially to obtain a data list set. Data lists containing only a single data point are removed from the data list set to obtain a filtered data list set. The variance, range, and interquartile range of each data list in the filtered data list set are calculated to obtain statistical distribution characteristics. These statistical distribution characteristics are compared with preset thresholds to identify conflicting units in the spatiotemporal data cube, resulting in a conflicting unit list. The historical credit score of each conflicting unit in the conflicting unit list is obtained. Based on real-time acquired multi-source traffic data, each conflicting unit in the conflicting unit list is evaluated... The system performs real-time data confidence calculation to obtain a confidence score, acquires the fusion value of each conflicting unit in the conflicting unit list with its neighboring units, calculates the consistency of the fusion value of each conflicting unit in the conflicting unit list with its neighboring units, and obtains a spatiotemporal proximity consistency score. Based on preset weight parameters, the system performs weighted fusion of the historical credit score, the confidence score, and the spatiotemporal proximity consistency score to obtain a comprehensive credit score. The system associates the comprehensive credit score with the conflicting unit list to obtain a list of conflicting data units with credit scores. Based on the list of conflicting data units, the system identifies non-conflicting units in the spatiotemporal data cube, fuses the non-conflicting units using a weighted average method, and performs weighted fusion of the conflicting data units based on the credit scores of the list of conflicting data units to obtain fused traffic data. The traffic body construction module is used to construct digital twins of two adjacent intersections in the target area based on the fused traffic data, thereby obtaining a first dynamic virtual traffic body and a second dynamic virtual traffic body. The initial strategy generation module is used to generate an initial scheduling strategy based on the multi-source traffic data using the first dynamic virtual traffic body and the second dynamic virtual traffic body; The optimal strategy generation module is used to perform traffic simulation on the first dynamic virtual traffic body and the second dynamic virtual traffic body using the initial scheduling strategy to obtain simulation data, and to use the first dynamic virtual traffic body and the second dynamic virtual traffic body to perform game-like collaborative scheduling decision based on the simulation data to obtain the optimal scheduling strategy. The intelligent scheduling module is used to generate a set of control instructions based on the optimal scheduling strategy, and to realize intelligent scheduling of the traffic network by issuing the set of control instructions to traffic signal equipment in the target area.

Images