A traffic network dynamic simulation and deduction system based on digital twinning technology

Abstract

This invention relates to the field of dynamic simulation and extrapolation technology for traffic networks. Specifically, it relates to a dynamic simulation and extrapolation system for traffic networks based on digital twin technology. It includes a road segmentation module, a twin creation module, a location extrapolation module, a model expansion module, and a scheme delivery module. The road segmentation module acquires static basic data and dynamic sensing data of the traffic network, and divides the traffic network into multiple traffic segments. The twin creation module creates a corresponding virtual twin based on the matched static basic data and dynamic sensing data of each traffic segment. Simultaneously, within the virtual twin, the movement speed of moving objects is analyzed based on the dynamic sensing data. By constructing a dedicated virtual twin for each independent traffic segment through the twin creation module, a 1:1 accurate mapping between physical road segments and virtual models is achieved.

Description

Technical Field

[0001] This invention relates to the field of dynamic simulation and deduction technology of traffic networks, and more specifically, to a dynamic simulation and deduction system for traffic networks based on digital twin technology. Background Technology

[0002] The core of digital twin technology in the transportation sector is to construct a virtual mapping of the physical road network to achieve real-time monitoring of traffic operation status, trend prediction, and simulation of control schemes.

[0003] Existing traffic network simulation and extrapolation technologies mostly operate within a unified global road network modeling scenario. This involves collecting overall static and dynamic data of the traffic network to construct a single virtual simulation model covering the entire area. However, unified global road network modeling easily leads to large data processing volumes, low simulation and extrapolation accuracy, and an inability to accurately capture the differences in traffic operation characteristics across different road segments. Furthermore, congestion control schemes for local road segments often suffer from insufficient adaptability, directly impacting the effectiveness of traffic control strategies and hindering accurate prediction and control of road network conditions. Secondly, existing technologies lack a refined historical and real-time trajectory matching mechanism for predicting the trajectory of moving objects, resulting in significant deviations between predictions and actual driving conditions. This leads to limitations in both the temporal and spatial dimensions of congestion prediction, making it impossible to accurately determine the location, duration, and spread of congestion in advance. To mitigate these issues, a dynamic traffic network simulation and extrapolation system based on digital twin technology is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic simulation and deduction system for traffic networks based on digital twin technology, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, a dynamic simulation and deduction system for traffic networks based on digital twin technology is provided, including a road segment division module, a twin creation module, a location deduction module, a model expansion module, and a scheme delivery module; The road segment division module is used to acquire static basic data and dynamic sensing data of the traffic network, and to divide the traffic network into road segments to acquire multiple traffic segments. The twin creation module is used to create a corresponding virtual twin based on the static basic data and dynamic perception data matched by each traffic segment. At the same time, the moving object's speed is analyzed in the virtual twin based on the dynamic perception data. The location prediction module is used to collect the historical movement trajectory of the moving object, extract the real-time movement trajectory of the moving object, combine the historical movement trajectory with the real-time movement trajectory to select the predicted movement trajectory, combine the predicted movement trajectory with the movement speed to perform spatiotemporal location prediction, and obtain the predicted spatiotemporal location information of the moving object in the future period. The model extension module is used to perform traffic congestion analysis in each virtual twin based on the predicted spatiotemporal location information of the corresponding moving object, determine the traffic congestion location and time, and perform model extension analysis based on the traffic congestion time and the moving speed of the moving object in the virtual twin to obtain the traffic segments that each moving object can reach during the traffic congestion time. Then, based on the reached traffic segments, the virtual twin corresponding to the segment is extracted as the extended model of the virtual twin, and the virtual twin of traffic congestion is extended according to the obtained extended model. The scheme push module is used to generate traffic control schemes in the expanded virtual twin based on the predicted spatiotemporal location information of each moving object, input the traffic control schemes into the virtual twin for simulation, and push the schemes based on the simulation results.

[0006] As a further improvement to this technical solution, the road segment division module establishes a connection with the monitoring terminal of the traffic network, and collects the inherent static basic data of the traffic network and the real-time dynamic perception data through the monitoring terminal, so as to fully cover the physical attributes and real-time operating status of the road network. Static basic data refers to the fixed buildings and facilities along traffic routes; Dynamically sensed data, with pedestrians and vehicles as moving objects; The overall traffic network is then divided into multiple independent traffic segments, achieving unit-based decomposition of the network; each segment is an independent traffic segment.

[0007] As a further improvement to this technical solution, in the twin creation module, for each independent traffic segment that has been divided, the acquired static basic data and dynamic perception data are segmented to obtain static basic data and dynamic perception data specific to that traffic segment, and a virtual twin corresponding to the traffic segment is constructed separately based on the matched static basic data and dynamic perception data. Within the constructed virtual twin, relevant parameters of the moving object are extracted from the dynamic perception data, and the real-time movement speed of the moving object is calculated through parameter parsing.

[0008] As a further improvement to this technical solution, in the location inference module, a dedicated mobile database is established for each mobile object in the dynamic perception data of the traffic network, and then the historical movement trajectory of each mobile object is collected and organized in the dynamic perception data and saved to the corresponding dedicated mobile database. Moving objects are extracted from the virtual twins of each traffic segment to obtain the moving objects that exist in the virtual twin in real time. At the same time, moving objects are traced to obtain the real-time movement trajectory of the moving object in the traffic segment. Then, the corresponding moving database is extracted based on the obtained list of moving objects.

[0009] As a further improvement to this technical solution, in the location inference module, for the same moving object, the historical movement trajectory of the movement database is matched and compared with the real-time movement trajectory, and the predicted movement trajectory that fits the current driving state is selected. Then, the selected predicted movement trajectory is combined with the real-time movement speed of the moving object to carry out the location inference calculation in the spatiotemporal dimension, so as to obtain the predicted spatiotemporal location information of each moving object in the future period. The future time period is set from 0 minutes to 60 minutes.

[0010] As a further improvement to this technical solution, in the model extension module, the predicted spatiotemporal location information of the moving objects contained in the virtual twins of each road segment is summarized and analyzed to obtain a list of moving objects at each road segment location in the corresponding time period. Based on static basic data, traffic congestion flow thresholds and traffic evacuation rates are set for different road segments in the virtual twin. Then, the list of moving objects at the same road segment location is compared with the traffic congestion flow thresholds. When the list of moving objects exceeds the traffic congestion flow threshold, it is determined that traffic congestion will occur at the location of the road segment during the corresponding time period. The traffic congestion time is analyzed in combination with the flow evacuation rate to obtain the traffic congestion time corresponding to the location of the road segment, and the location of the road segment is identified as the location of the traffic congestion. Conversely, monitoring continues as long as the list of moving objects does not exceed the traffic congestion threshold.

[0011] As a further improvement to this technical solution, the model expansion module combines the traffic congestion time with the moving speed of the moving object to reverse-deduce the associated road segments that the moving object can reach during the congestion period. It then extracts the exclusive virtual twins corresponding to these associated road segments as the expansion model and splices and integrates them with the original virtual twins of the congested road segments to complete the expansion of the scope of the traffic congestion virtual twins. Starting with a virtual twin of traffic congestion, the model is expanded by combining the movement trajectory of the moving object to the associated road segment.

[0012] As a further improvement to this technical solution, the solution push module extracts a list of traffic control solutions from the monitoring end of the traffic network, and then, within the virtual twin scene after the scope expansion is completed, combines the list of traffic control solutions with the predicted spatiotemporal location information of moving objects to generate a traffic control solution adaptation, thereby obtaining a list of traffic control solutions adapted to the current traffic congestion location.

[0013] As a further improvement to this technical solution, the solution push module imports a list of traffic control solutions adapted to the current traffic congestion location into the expanded virtual twin to conduct simulation operation and simulation deduction, obtains the actual control effect data of each traffic control solution, and then selects the optimal control solution based on the effect data obtained from the deduction and completes the push.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. In this traffic network dynamic simulation and deduction system based on digital twin technology, a dedicated virtual twin is constructed for each independent traffic segment through a twin creation module, achieving a 1:1 accurate mapping between the physical segment and the virtual model. At the same time, the real-time speed of moving objects is accurately calculated within the twin, and the speed parameters are specifically bound to the moving objects, allowing the virtual twin to realistically reproduce the real-time traffic operation status of the segment. Compared with the global single model of the existing technology, the segment-level dedicated twin of this system has stronger dynamism and accuracy, providing a realistic virtual simulation scenario for subsequent trajectory prediction and congestion prediction.

[0015] 2. In this traffic network dynamic simulation and extrapolation system based on digital twin technology, a location extrapolation module establishes a dedicated mobile database for each moving object, realizing the structured collection of historical trajectories and the accurate extraction of real-time trajectories. At the same time, through multi-dimensional trajectory matching based on driving direction, path overlap, and speed change trend, predicted trajectories that closely match the current driving state are selected. Combined with real-time speed, spatiotemporal location extrapolation is carried out from 0 to 60 minutes. During the extrapolation process, downstream road segment topology information can be automatically connected, effectively improving the accuracy of moving object trajectory prediction. This solves the problems of large prediction deviation and limited spatiotemporal range in existing technologies, and can accurately predict the future location of each moving object in advance, providing a core basis for the early determination of traffic congestion.

[0016] 3. In this traffic network dynamic simulation and deduction system based on digital twin technology, the system accurately quantifies and determines the location and duration of traffic congestion based on the predicted spatiotemporal location of moving objects. Simultaneously, it reverse-engineers the associated road segments where congestion spreads by combining congestion time with the speed of moving objects. Instead of stitching together complete twins of associated road segments, it extracts a local area based on the trajectory of the moving object as an extended model, which is then stitched together with the original twin of the congested road segment. This achieves dynamic and accurate expansion of the virtual twin, covering the range of congestion spread while avoiding the waste of computing power caused by redundant modeling. It solves the problems of fixed model range and lagging deduction in existing technologies, allowing simulation and deduction to closely follow the trend of congestion spread and providing a comprehensive and accurate simulation scenario for the generation of control plans. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the structure of a traffic network dynamic simulation and deduction system based on digital twin technology according to the present invention; Figure 2 This is a flowchart illustrating the road segment division module of the present invention; Figure 3 This is a flowchart illustrating the twin creation module of the present invention; Figure 4 This is a flowchart illustrating the location deduction module of the present invention; Figure 5 This is a flowchart illustrating the model extension module of the present invention; Figure 6 This is a flowchart illustrating the push module of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figures 1-6 As shown, the purpose of this embodiment is to provide a dynamic simulation and deduction system for traffic networks based on digital twin technology, including a road segment division module, a twin creation module, a location deduction module, a model expansion module, and a scheme delivery module; The road segmentation module is used to acquire static basic data and dynamic sensing data of the traffic network, and to divide the traffic network into road segments to obtain multiple traffic segments. In the road segmentation module, a connection is established with the monitoring terminal of the traffic network. The monitoring terminal collects the inherent static basic data of the traffic network as well as the dynamic perception data that changes in real time, so as to fully cover the physical attributes and real-time operating status of the road network. Select existing official monitoring platforms of the traffic network (including traffic cameras, checkpoint equipment, roadside sensing units, and road network management backend), and establish a stable and encrypted two-way communication connection through TCP / IP protocol or platform-specific open interface to ensure the real-time performance and security of data transmission; Static basic data refers to the fixed buildings and facilities along traffic routes; Through the established communication link, static data collection instructions are sent to the monitoring terminal to extract data on fixed buildings and facilities in traffic sections. This type of data is fixed for a long time and does not change in real time. It can be reused for a long time after a single collection. Specifically, it includes the number of lanes in the road section, lane width, road section length, intersection channelization structure, location of traffic signs and markings, guardrails, medians, pedestrian crossings, fixed traffic signal poles, and coordinates of the start and end nodes of the road section. Dynamically perceived data is used to identify pedestrians and vehicles as moving objects. Based on the communication connection, a real-time data monitoring mode is enabled to collect data on moving objects within the road network in real time from the monitoring terminal. The data is dynamically perceived and updated in real time. The collection frequency is adapted to the road network operating speed (1 time / second for regular road sections and 1 time / 0.5 seconds for congested road sections). Only pedestrians and vehicles are included, and they are uniformly classified as moving objects. The collected content includes the real-time spatial coordinates of the moving objects, the type of moving objects (motor vehicles / non-motor vehicles / pedestrians), and the real-time movement status (driving / stationary). The overall traffic network is then divided into multiple independent traffic segments, achieving unit-based decomposition of the network; each segment is an independent traffic segment.

[0020] Based on the verified static basic data, the topology of the entire traffic network is sorted out, the connection relationship between each road segment, the location of intersection nodes, and the physical boundaries of road segments are clarified, and a basic topology map of the road network is drawn. Intersection stop lines, road segment lane change nodes, and fixed isolation facilities are used as natural splitting nodes to avoid arbitrary splitting and ensure that each independent road segment has complete traffic function and conforms to the actual traffic control logic. The entire traffic network is standardized and decomposed into multiple non-overlapping, clearly defined, and independent traffic segments, achieving the unitization of the traffic network. Each decomposed unit is an independent traffic segment.

[0021] The twin creation module is used to create a corresponding virtual twin based on the static basic data and dynamic perception data matched by each traffic segment. At the same time, the moving speed of the moving object is analyzed in the virtual twin according to the dynamic perception data. In the twin creation module, for each independent traffic segment that has been divided, the acquired static basic data and dynamic perception data are separated to obtain static basic data and dynamic perception data specific to that segment. Based on the matched static basic data and dynamic perception data, a virtual twin corresponding to the traffic segment is constructed separately. For each individual traffic segment being processed, spatial filtering is performed on the global static basic data based on its spatial boundary coordinates. Only data on fixed buildings and facilities falling within the boundary of the segment are retained, while static data crossing segments or irrelevant areas are removed, thus achieving dedicated segmentation of static data. At the same time, the segmented segment-specific static data is standardized, unifying the data format and coordinate system, and supplementing missing local detail data to ensure that the static data completely matches the physical structure of the segment, without redundancy or missing data. Then, the processed dedicated static data is bound to the corresponding segment number to form a single segment static data package, which is exclusively used for the construction of the virtual twin of the segment and is not shared with other segments. Based on the spatial boundary range of the same independent road segment, the real-time collected global dynamic perception data is filtered. By comparing the real-time coordinates of moving objects, only the data of moving objects such as pedestrians and vehicles currently in the road segment are retained, while the data of moving objects outside the road segment are removed. The filtered exclusive dynamic data is deduplicated and denoised to filter out invalid data such as abnormal stops and repeated collections. Each moving object is marked with a unique identifier to distinguish different moving objects and ensure that the dynamic data accurately corresponds to the real-time operating status within the road segment. Then, the data collection timestamp is synchronously bound to form a real-time data package of dynamic data for a single road segment, ensuring that the dynamic data accurately corresponds to the road segment and time. Based on the current road segment's dedicated static basic data package, a virtual twin static base is built, restoring the fixed physical structure of the road segment, such as lanes, isolation facilities, signs and markings, and intersection channelization, in a 1:1 ratio. The geometric shape and spatial layout of the real road segment are replicated, and the road segment's dedicated dynamic perception data package is loaded into the static base in real time. The position and type of all moving objects in the road segment are synchronously mapped in the virtual space, completing the fusion of the static base and dynamic elements, and initially generating a road segment's dedicated virtual twin. Within the constructed virtual twin, relevant parameters of the moving object are extracted from the dynamic perception data, and the real-time movement speed of the moving object is calculated through parameter parsing.

[0022] Entering the constructed virtual twin of the road segment, all loaded moving objects are located. The core dynamic parameters of each moving object are extracted one by one according to its unique identifier, including real-time spatial coordinates, coordinates collected in the previous frame, and time interval between adjacent data collections. The extracted parameters are then validated to remove invalid parameters with abnormal coordinates or missing time intervals. Based on the extracted valid spatial coordinates and time interval parameters, the real-time displacement of a single moving object is calculated using the spatial displacement to time ratio method. The displacement and the corresponding time interval are then substituted into the speed formula to calculate the real-time driving speed of the moving object. The forward and reverse driving directions are distinguished to complete the speed analysis. The calculated real-time movement speed is synchronously bound to the corresponding moving object within the virtual twin, the dynamic parameters of the twin are updated, the construction of a dedicated virtual twin for the entire road segment is completed, and the result is output to the subsequent location inference module.

[0023] The location projection module is used to collect the historical movement trajectory of the moving object, extract the real-time movement trajectory of the moving object, combine the historical movement trajectory with the real-time movement trajectory to select the predicted movement trajectory, combine the predicted movement trajectory with the movement speed to perform spatiotemporal location projection, and obtain the predicted spatiotemporal location information of the moving object in the future period. In the location inference module, a dedicated mobile database is established for each mobile object in the dynamic perception data of the traffic network. Then, the historical movement trajectory of each mobile object is collected and organized in the dynamic perception data and saved to the corresponding dedicated mobile database. Traverse all moving objects (pedestrians / vehicles) in the dynamic sensing data of the traffic network, assign a unique identifier ID (such as license plate / pedestrian identification code) to each moving object as the primary key of the database index, and create an independent exclusive mobile database for each moving object based on the unique ID. The database has preset fields, including object ID, trajectory collection timestamp, trajectory coordinates, driving direction, and collection segment number, to ensure that the data is stored in a structured manner. Historical driving / walking trajectory data for each moving object is filtered out from the global dynamic perception data, organized in ascending order by timestamp, and invalid trajectory points with abnormal coordinates and repeated collections are removed. Then, the organized historical movement trajectory (including timestamp, coordinates, and road segment information) is written into the dedicated database of the corresponding moving object one by one, completing the accurate collection of historical trajectories and supporting rapid retrieval later.

[0024] Moving objects are extracted from the virtual twins of each traffic segment to obtain the moving objects that exist in the virtual twin in real time. At the same time, moving objects are traced to obtain the real-time movement trajectory of the moving object in the traffic segment. Then, the corresponding moving database is extracted based on the obtained list of moving objects.

[0025] By accessing the virtual twin data of each traffic segment, extracting all real-time mobile objects existing in the twin at the current moment, generating an object list containing object ID, current road segment, and real-time status, and based on the time series of dynamic perception data, extracting the continuous coordinate points of the object in the current road segment from the entry time to the current time, and stitching them together to form the real-time movement trajectory of the object; Based on the unique ID in the real-time object list, the dedicated mobile database corresponding to each mobile object is retrieved in batches to ensure that historical trajectory data is accurately associated with the current object, without mismatch or omission. In the location extrapolation module, for the same moving object, the historical movement trajectory in the movement database is matched and compared with the real-time movement trajectory. The predicted movement trajectory that fits the current driving state is selected. The selected predicted movement trajectory is then combined with the real-time movement speed of the moving object to carry out the spatiotemporal location extrapolation calculation and obtain the predicted spatiotemporal location information of each moving object in the future period. The future time period is set from 0 minutes to 60 minutes, and the steps are as follows: For the same moving object, extract its historical trajectory from its dedicated database, and perform feature matching with the real-time moving trajectory according to three dimensions: similarity of driving direction, path overlap, and speed change trend. Calculate the feature matching degree, and select historical trajectories with a matching degree ≥80% as candidate predicted trajectories. Then, select the trajectory with the highest matching degree as the final predicted moving trajectory (only retain the trajectory with the highest matching degree to avoid multiple trajectories interfering with the inference accuracy). The selected predicted movement trajectory is combined with the object's real-time movement speed. Starting from the current moment, the coordinates of each time node within the next 0-60 minutes are calculated segment by segment at a preset time granularity (1 minute / step). If the trajectory exceeds the current road segment boundary during the simulation, the downstream road segment topology information is automatically connected to ensure the continuity of the location simulation. Finally, a predicted spatiotemporal location sequence for each moving object within 0-60 minutes is generated, as shown in the following formula: ; in, For trajectory matching degree, The similarity is based on the direction of travel (α=0.3, weight). The path overlap is β=0.5, weighted. Similarity of velocity change trends (γ=0.2, weight); ; in, The predicted coordinates for the next t minutes. , The coordinates at the current time are... For real-time movement speed, For real-time driving direction angle, The time frame is for simulation.

[0026] The model extension module is used to perform traffic congestion analysis in each virtual twin based on the predicted spatiotemporal location information of the corresponding moving object, determine the traffic congestion location and time, and perform model extension analysis based on the traffic congestion time and the moving speed of the moving object in the virtual twin to obtain the traffic segments that each moving object can reach during the traffic congestion time. Then, based on the reached traffic segments, the virtual twin corresponding to the segment is extracted as the extended model of the virtual twin, and the virtual twin of traffic congestion is extended based on the obtained extended model. In the model extension module, the predicted spatiotemporal location information of the moving objects contained in the virtual twins of each road segment is summarized and analyzed to obtain a list of moving objects at each road segment location in the corresponding time period. In the virtual twins corresponding to each road segment, the predicted spatiotemporal location information of moving objects is statistically analyzed by road segment and by time period to obtain a list of moving objects for each road segment at various future times. Based on static basic data, traffic congestion flow thresholds and traffic evacuation rates are set for different road segments in the virtual twin. Then, the list of moving objects at the same road segment location is compared with the traffic congestion flow thresholds. Based on the static basic data of the road segment division module, combined with the number of lanes, road segment length, and road grade, each road segment is individually labeled with exclusive parameters, instead of using uniform universal values; Set a traffic congestion flow threshold, which is the total number of critical moving objects in a road segment that is considered congested when this number is reached; simultaneously set a flow evacuation rate, which is the number of moving objects that can be evacuated from the road segment per unit time. When the list of moving objects exceeds the traffic congestion flow threshold, it is determined that traffic congestion will occur at that road segment location during the corresponding time period. Traffic congestion time analysis is then performed based on the flow evacuation rate to obtain the corresponding traffic congestion time at that road segment location. This road segment location is then designated as the traffic congestion location, using the following formula: ; in, For the duration of traffic congestion, To block the flow threshold, The traffic flow dispersal rate of the road section. The total number of objects to be predicted; Conversely, monitoring continues as long as the list of moving objects does not exceed the traffic congestion threshold.

[0027] In the model extension module, the traffic congestion time is combined with the moving speed of the moving object to reverse the deduction of the related road segments that the moving object can reach during the congestion period. The exclusive virtual twins corresponding to these related road segments are extracted as extension models and spliced ​​and integrated with the original virtual twins of the congested road segments to complete the expansion of the scope of the traffic congestion virtual twins. Starting with a virtual twin of traffic congestion, the model is expanded by combining the movement trajectory of the moving object to the associated road segment.

[0028] The scheme push module is used to generate traffic control schemes in the expanded virtual twin based on the predicted spatiotemporal location information of each moving object, input the traffic control schemes into the virtual twin for simulation, and push the schemes based on the simulation results.

[0029] In the solution push module, a list of traffic control solutions is extracted from the monitoring terminal of the traffic network. Then, in the virtual twin scene after the scope is expanded, the list of traffic control solutions is combined with the predicted spatiotemporal location information of moving objects to generate a traffic control solution adaptation and obtain a list of traffic control solutions adapted to the current traffic congestion location.

[0030] Based on the communication link established in the early stage, a scheme retrieval instruction is sent to the road network monitoring and management platform to obtain the complete set of official standardized traffic control schemes, covering various conventional emergency schemes such as signal timing adjustment, road segment diversion, lane control, pedestrian guidance, and congestion avoidance; The extended virtual twin scene, traffic congestion location, congestion duration, and spatiotemporal location information of moving objects predicted by the location inference module are retrieved from the model extension module. The parameters of the current congestion scene are locked. Then, the general control schemes in the scheme library are accurately matched with the current congestion location, predicted object distribution, and road segment attributes. Schemes that are not suitable for the current scenario are eliminated, and schemes that can be implemented are retained. A list of exclusive control schemes that are only suitable for the current traffic congestion is generated, and redundant schemes are not retained. In the solution push module, the list of traffic control solutions adapted to the current traffic congestion location is imported into the expanded virtual twin to carry out simulation operation and simulation exercise, obtain the actual control effect data of each traffic control solution, and then select the optimal control solution based on the effect data obtained from the simulation and complete the push. The adapted control scheme list is imported one by one into the virtual twin scenario with the expanded scope. The simulation is started to simulate the road network operation status after the actual implementation of the scheme. The simulation duration covers the entire traffic congestion period to ensure that the simulation results are close to reality. At the same time, the simulation process is monitored in real time and various operation indicators are recorded synchronously to ensure that the actual control effect data of each scheme is fully collected and no key parameters are missed. The simulation results of each scheme are sorted and statistically analyzed. The congestion relief efficiency, evacuation speed and traffic restoration time of different schemes are compared to clarify the advantages and disadvantages of each scheme. Then, according to the preset optimal scheme judgment criteria, the single optimal control scheme with the best control effect, the lowest execution cost and the strongest adaptability is selected from the list of suitable schemes to avoid the execution chaos caused by the parallel push of multiple schemes. The selected optimal control scheme is pushed to the traffic network monitoring and control platform through the communication link to generate directly executable control instructions, completing the closed loop from simulation to decision implementation. The execution results can then be fed back for system iteration and optimization.

[0031] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A traffic network dynamic simulation and deduction system based on digital twin technology, characterized in that: It includes a road segmentation module, a twin creation module, a location inference module, a model expansion module, and a solution delivery module; The road segment division module is used to acquire static basic data and dynamic sensing data of the traffic network, and to divide the traffic network into road segments to acquire multiple traffic segments. The twin creation module is used to create a corresponding virtual twin based on the static basic data and dynamic perception data matched by each traffic segment. At the same time, the moving object's speed is analyzed in the virtual twin based on the dynamic perception data. The location prediction module is used to collect the historical movement trajectory of the moving object, extract the real-time movement trajectory of the moving object, combine the historical movement trajectory with the real-time movement trajectory to select the predicted movement trajectory, combine the predicted movement trajectory with the movement speed to perform spatiotemporal location prediction, and obtain the predicted spatiotemporal location information of the moving object in the future period. The model extension module is used to perform traffic congestion analysis in each virtual twin based on the predicted spatiotemporal location information of the corresponding moving object, determine the traffic congestion location and time, and perform model extension analysis based on the traffic congestion time and the moving speed of the moving object in the virtual twin to obtain the traffic segments that each moving object can reach during the traffic congestion time. Then, based on the reached traffic segments, the virtual twin corresponding to the segment is extracted as the extended model of the virtual twin, and the virtual twin of traffic congestion is extended according to the obtained extended model. The scheme push module is used to generate traffic control schemes in the expanded virtual twin based on the predicted spatiotemporal location information of each moving object, input the traffic control schemes into the virtual twin for simulation, and push the schemes based on the simulation results.

2. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the road segmentation module, a connection is established with the monitoring terminal of the traffic network. The monitoring terminal collects the inherent static basic data of the traffic network as well as the dynamic perception data that changes in real time, so as to fully cover the physical attributes and real-time operating status of the road network. Static basic data refers to the fixed buildings and facilities along traffic routes; Dynamically perceived data is used to identify pedestrians and vehicles as moving objects. The overall traffic network is then divided into multiple independent traffic segments, achieving unit-based decomposition of the network; each segment is an independent traffic segment.

3. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the twin creation module, for each independent traffic segment that has been divided, the acquired static basic data and dynamic perception data are segmented to obtain static basic data and dynamic perception data specific to that traffic segment. Based on the matched static basic data and dynamic perception data, a virtual twin corresponding to the traffic segment is constructed separately. Within the constructed virtual twin, relevant parameters of the moving object are extracted from the dynamic perception data, and the real-time movement speed of the moving object is calculated through parameter parsing.

4. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the location inference module, a dedicated mobile database is established for each mobile object in the dynamic perception data of the traffic network. Then, the historical movement trajectory of each mobile object is collected and organized in the dynamic perception data and saved to the corresponding dedicated mobile database. Moving objects are extracted from the virtual twins of each traffic segment to obtain the moving objects that exist in the virtual twins in real time. At the same time, moving objects are traced to obtain the real-time movement trajectory of the moving object in the traffic segment. Then, the corresponding moving database is extracted based on the obtained list of moving objects.

5. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the location extrapolation module, for the same moving object, the historical movement trajectory of the moving object is matched and compared with the real-time movement trajectory of the moving object, and the predicted movement trajectory that fits the current driving state is selected. Then, the selected predicted movement trajectory is combined with the real-time movement speed of the moving object to carry out the location extrapolation calculation in the spatiotemporal dimension, so as to obtain the predicted spatiotemporal location information of each moving object in the future time period. The future time period is set from 0 minutes to 60 minutes.

6. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the model extension module, the predicted spatiotemporal location information of the moving objects contained in the virtual twins of each road segment is summarized and analyzed to obtain a list of moving objects at each road segment location in the corresponding time period. Based on static basic data, traffic congestion flow thresholds and traffic evacuation rates are set for different road segments in the virtual twin, and then the list of moving objects at the same road segment location is compared with the traffic congestion flow thresholds. When the list of moving objects exceeds the traffic congestion flow threshold, it is determined that traffic congestion will occur at the location of the road segment during the corresponding time period. The traffic congestion time is analyzed in conjunction with the flow evacuation rate to obtain the traffic congestion time corresponding to the location of the road segment, and the location of the road segment is identified as the location of the traffic congestion. Conversely, monitoring continues as long as the list of moving objects does not exceed the traffic congestion threshold.

7. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the model extension module, the traffic congestion time is combined with the moving speed of the moving object to reverse the deduction of the related road segments that the moving object can reach during the congestion period. The exclusive virtual twins corresponding to these related road segments are extracted as extension models and spliced ​​and integrated with the original virtual twins of the congested road segments to complete the expansion of the scope of the traffic congestion virtual twins. Starting with a virtual twin of traffic congestion, the model is expanded by combining the movement trajectory of the moving object to the associated road segment.

8. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the proposed solution push module, a list of traffic control solutions is extracted from the monitoring terminal of the traffic network. Then, within the virtual twin scene after the scope expansion is completed, the list of traffic control solutions is combined with the predicted spatiotemporal location information of moving objects to generate a traffic control solution adaptation, thereby obtaining a list of traffic control solutions adapted to the current traffic congestion location.

9. The traffic network dynamic simulation and deduction system based on digital twin technology according to claim 1, characterized in that: In the proposed solution push module, a list of traffic control solutions adapted to the current traffic congestion location is imported into the expanded virtual twin. Simulation and simulation are then conducted to obtain actual control effect data for each traffic control solution. Based on the simulation effect data, the optimal control solution is selected and pushed out.

Images