SUMO-based high-concurrency distributed simulation system and method

Abstract

The invention discloses a SUMO-based high-concurrency distributed simulation system and method, and relates to the technical field of traffic simulation, and the system comprises a road network division module which is used for receiving an SUMO road network file, dividing a region-level road network into a plurality of independent sub-regions, and generating a sub net.xml configuration file containing a boundary interface; the simulation scheduler serves as a master control node and is responsible for scheduling simulation examples of all sub-areas and maintaining boundary vehicle state synchronization. The simulation execution modules serve as simulation nodes, each simulation execution module is bound with a single CPU core, and SUMO sub-network simulation is independently operated; the WebSocket service module supports real-time control, data push and state display of a browser client; the load balancing module monitors CPU load, memory occupancy rate and simulation progress difference, and dynamically adjusts task allocation; and the data fusion and visualization module fuses the data of each sub-region to generate a continuous and uniform front-end visual traffic state picture. According to the method, high-performance simulation in a multi-core and multi-node environment is realized through road network division and a distributed scheduling mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of traffic simulation, and in particular to a high-concurrency distributed simulation system and method based on SUMO. BACKGROUND

[0002] Urban traffic simulation is an important tool for smart city, autonomous driving simulation, traffic management decision, etc. SUMO (Simulation of Urban Mobility) is a widely used open-source microscopic traffic simulation tool with high configurability and powerful simulation functions.

[0003] However, SUMO is originally single-threaded and cannot fully utilize multi-core CPU resources. When facing large-scale road networks, high-frequency vehicle interactions, and real-time visualization requirements, there are obvious performance bottlenecks. Some existing solutions attempt to improve single-instance computing power through high-performance servers, but have limited scalability and cannot achieve online real-time interaction or dynamic load distribution. SUMMARY

[0004] Therefore, the present application provides a high-concurrency distributed simulation system and method based on SUMO to solve the deficiencies in the prior art.

[0005] The first aspect of the present application provides a high-concurrency distributed simulation system based on SUMO, comprising: a road network division module that receives a SUMO road network file, divides a regional road network into multiple independent sub-regions according to road network topology complexity, traffic flow density, and CPU core number, and generates a sub-net.xml configuration file containing boundary interface definitions for each sub-region; a simulation scheduler as a master node that assigns simulation tasks for each sub-region, maintains boundary vehicle state synchronization for each sub-region, controls synchronization mechanisms based on time steps, uniformly advances simulation processes for all simulation nodes, and receives and integrates simulation data uploaded by each simulation execution module; a simulation execution module as a simulation node, the number of which is at least two; each simulation execution module is bound to a single CPU core, executes SUMO simulation tasks for the corresponding sub-region in the form of an independent sub-process, and collects simulation data in the sub-region in real time; a WebSocket service module that establishes a long connection between a browser client and the simulation scheduler, supports real-time simulation state pushing, remote parameter adjustment, and interactive instruction transmission; a load balancing module that detects the computing load, memory occupancy rate, and simulation progress difference of each CPU core in real time, and dynamically adjusts the allocation of sub-region simulation tasks; A data fusion and visualization module is configured to perform space-time alignment and consistency check on the simulation data of each sub-region, and generate a continuous and unified traffic state visualization picture at the front end.

[0006] In a possible implementation of the first aspect, the road network division module divides the regional road network into a plurality of independent sub-regions by using a preset algorithm, and ensures that the topologies of the boundary roads of adjacent sub-regions are matched during the sub-region division process, and the calculation complexity of each sub-region corresponds to the CPU core computing power.

[0007] In a possible implementation of the first aspect, the preset algorithm is a METIS algorithm, or a graph partitioning algorithm, or a fusion algorithm of the METIS algorithm and the graph partitioning algorithm.

[0008] In a possible implementation of the first aspect, the sub-net.xml configuration file contains the in-out rules of the boundary vehicles, the data interaction format, and the synchronization trigger condition.

[0009] In a possible implementation of the first aspect, the simulation scheduler collects the boundary vehicle states of each simulation execution module in a timely manner through a preset boundary vehicle data synchronization protocol, performs path connection and state consistency correction on the vehicles traveling across the sub-regions, and dynamically configures the step length of the time step control synchronization mechanism based on the simulation accuracy requirement.

[0010] In a possible implementation of the first aspect, after the simulation execution module is bound to the CPU core, the simulation kernel is invoked through the TraCI interface of SUMO; the simulation data uploading frequency of each sub-process is consistent with the step length of the time step control synchronization mechanism, and the dynamic traffic event instructions issued by the simulation scheduler are received during the simulation process.

[0011] In a possible implementation of the first aspect, the simulation data of the sub-region collected by the simulation execution module in real time includes the real-time collection of the vehicle position, vehicle speed, traffic flow, and road network congestion state data in the sub-region.

[0012] In a possible implementation of the first aspect, the monitoring indicators of the load balancing module include the CPU usage, the simulation progress difference, and the memory occupancy rate; when the CPU usage, the simulation progress difference, and the memory occupancy rate exceed the corresponding preset threshold values, the simulation scheduler is triggered to perform sub-region migration, and an incremental data synchronization method is used to ensure the simulation continuity during the migration process.

[0013] In a possible implementation manner of the first aspect, the data fusion and visualization module display screen includes a global road network traffic flow heat map, real-time vehicle trajectory, road section passing efficiency, boundary vehicle cross-region flow state, and simulation statistical data; the update frequency of the visualization screen is synchronized with the simulation time step, and the user can select one or more specified sub-regions for viewing and selecting all corresponding simulation data.

[0014] The second aspect of the application provides a high-concurrency distributed simulation method based on SUMO, comprising: The road network division module receives a SUMO road network file and simulation parameters, combines the number of CPU cores and the distribution of computing power, uses a preset segmentation algorithm to segment the road network into multiple sub-regions, generates a sub-net.xml configuration file containing boundary interaction rules, and establishes an initial binding relationship between the sub-regions and the CPU cores; Based on the initial binding relationship, the simulation tasks of each sub-region are allocated to the corresponding simulation execution module, the simulation execution module binds the CPU core and starts a SUMO sub-simulation process through a TraCI interface; Each simulation execution module runs the sub-simulation according to the time step set by the simulation scheduler, collects the vehicle position, vehicle speed, traffic flow, and road network congestion state data in real time, and uploads them to the simulation scheduler at a synchronization period; the simulation scheduler performs consistency checking and path connection on the boundary vehicle data of each sub-region, and generates global simulation data by merging; The simulation scheduler pushes the global simulation data to the browser client through a WebSocket service module, realizes real-time visualization display through a data fusion and visualization module, receives the interaction instructions issued by the browser client, and synchronizes them to the corresponding simulation execution module to perform parameter adjustment or event triggering; The load balancing module monitors the computing load, memory occupancy rate, and simulation progress of each CPU core in real time, and when it is detected that a monitoring index exceeds a corresponding preset threshold, the simulation scheduler reassigns the road network sub-regions during the synchronization interval of adjacent time steps, adjusts the task binding relationship of the simulation execution module, and completes the simulation initialization of the migrated region through an incremental data synchronization method to ensure the continuity of the simulation process. After the simulation is completed, the simulation scheduler integrates the simulation data of all sub-regions to generate a complete simulation report, and supports client export or online viewing.

[0015] The beneficial effects are that the application provides a high-concurrency distributed simulation system and method based on SUMO, which comprises a road network division module, receives a SUMO road network file, divides a regional road network into multiple independent sub-regions, and generates a sub-net.xml configuration file containing a boundary interface; a simulation scheduler serves as a master node, allocates sub-regional simulation tasks, synchronizes boundary vehicle states, and uniformly promotes the simulation process based on a time step control synchronization mechanism; a simulation execution module serves as a simulation node, is bound to a single CPU core, and executes corresponding sub-regional simulation and data collection in an independent sub-process; a WebSocket service module establishes a long connection between a browser client and the master node, supports simulation state pushing, parameter adjustment, and instruction transmission; a load balancing module monitors CPU load, memory occupancy, and simulation progress difference, and dynamically adjusts task allocation; and a data fusion and visualization module performs space-time alignment and consistency check on sub-regional data, and generates a continuous and unified front-end visual traffic state picture. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only part of the embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0017] Figure 1 is a high-concurrency distributed simulation system based on SUMO provided by the embodiment of the present application; Figure 2 is a high-concurrency distributed simulation method flowchart based on SUMO provided by the embodiment of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0019] In this application, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0020] Embodiment one The existing urban traffic simulation system has a computing performance bottleneck in large-scale and high-precision scenarios, which is specifically manifested as: Single-threading leads to inability to handle large-scale road networks or multi-region simulation in parallel, resulting in low simulation efficiency; Traditional simulation lacks real-time interaction and visualization capabilities, and can only preset inputs and output results after pre-computation, and cannot dynamically adjust parameters; The existing architecture does not support dynamic load scheduling, making it difficult to balance resources according to actual computing pressure, and has poor scalability; It is difficult to integrate with modern Web service systems, hindering remote management and visualization applications.

[0021] Therefore, the present application provides a high-concurrency distributed simulation system based on SUMO, as shown in Figure 1 which comprises: A road network division module receives a SUMO road network file, divides a regional road network into multiple independent sub-regions in combination with road network topology complexity, traffic flow density and CPU core number, and generates a sub-net.xml configuration file for each sub-region, which includes boundary interface definition; A simulation scheduler, as a master node, is used to allocate simulation tasks for each sub-region, maintain synchronization of vehicle states at the boundaries of each sub-region, and based on a time step control synchronization mechanism, uniformly advance the simulation process of all simulation nodes, while receiving and integrating simulation data uploaded by each simulation execution module; A simulation execution module, as a simulation node, is at least two in number; and each simulation execution module is bound to a single CPU core to execute a SUMO simulation task for a corresponding sub-region in the form of an independent sub-process, and to collect simulation data in the sub-region in real time; A WebSocket service module is used to establish a long connection between a browser client and the simulation scheduler, support real-time simulation state pushing, remote parameter adjustment and interactive instruction transmission; Load balancing module, for real-time detection of each CPU core computing load, memory occupancy and simulation progress difference, and dynamic adjustment of sub-region simulation task allocation; Data fusion and visualization module, for spatio-temporal alignment and consistency check of each sub-region simulation data, and generation of continuous and unified traffic state visualization picture at the front end.

[0022] The embodiment provides a high-concurrency distributed simulation system based on SUMO, which comprises a road network division module, a simulation scheduler, a simulation execution module, a WebSocket service module, a load balancing module and a data fusion and visualization module, the modules are coordinated with each other to form the high-concurrency distributed simulation scheme of the embodiment, and the scheme is as follows: Road network division module: The module is a pre-basic module of distributed simulation, and the core role is to split a large-scale road network into sub-regions which can be calculated in parallel.

[0023] Input data, SUMO standard road network file, road network topology complexity evaluation data, historical traffic flow density distribution data, CPU core number and single-core computing power parameters.

[0024] Core processing logic, METIS algorithm, graph partitioning algorithm or fusion algorithm of the two are adopted to intelligently segment the whole road network, and the segmentation process needs to meet two hard constraints: one is that the boundary road topologies of adjacent sub-regions are completely matched, so as to ensure that the paths of vehicles crossing regions are continuous and have no breakpoints; the other is that the computing complexity of each sub-region is accurately adapted to the CPU core computing power, so as to avoid waste of computing power or overload of cores. An independent sub-net.xml configuration file is generated for each sub-region, and in addition to containing the sub-region road network topology information, the file also needs to clearly define the boundary vehicle entry and exit rules, data interaction format, synchronization trigger condition. For example, it is stipulated that when a vehicle travels to within a 50-meter threshold of the boundary road, the vehicle is determined to be a vehicle to be crossed, and data synchronization needs to be triggered; it is stipulated that the data field to be synchronized includes vehicle ID, position coordinates, driving speed, target sub-region identifier, etc. According to the matching result of the computing complexity of the sub-region and the CPU core computing power, a corresponding relationship table of the sub-region-CPU core-simulation execution module is generated, which is directly used as the basis for task allocation by the simulation scheduler.

[0025] Output result, multiple independent sub-region road network configuration files, initial binding relationship between sub-regions and hardware resources.

[0026] Simulation scheduler: It is the central module of the whole system, and is responsible for global simulation task allocation, synchronization control and data integration.

[0027] Task allocation and start control, read the initial binding relationship output by the road network division module, and issue sub-regional simulation tasks to the corresponding simulation execution module, while passing the sub-net.xml configuration file of the sub-region to trigger the start of the sub-simulation process of each simulation execution module.

[0028] Time step control synchronization, adopt time step control synchronization mechanism, set dynamically configurable time step (range 10ms-1000ms, can be adjusted according to simulation accuracy requirements). All simulation execution modules must report completion signal to simulation scheduler after completing simulation calculation of current time step; simulation scheduler will only issue instruction to enter next time step after receiving completion signal of all nodes, to ensure global simulation progress completely consistent.

[0029] Boundary vehicle state synchronization and consistency correction, based on preset boundary vehicle data synchronization protocol, collect boundary vehicle state data uploaded by each simulation execution module at the end of each time step; for vehicles driving across sub-regions, simulation scheduler needs to accurately transfer final state data of the vehicle in sub-region A to simulation execution module in sub-region B, and correct initial state of the vehicle in simulation logic of sub-region B, to avoid problems such as vehicle disappearance or duplication; for vehicle data with state conflict, simulation scheduler will automatically calibrate to unique consistent state based on historical trajectory data and traffic flow rules.

[0030] Data integration and instruction forwarding, aggregate sub-regional simulation data (vehicle position, speed, traffic flow, congestion state, etc.) uploaded by each simulation execution module to generate global simulation data set; at the same time, receive user interaction instructions forwarded by WebSocket service module and issue to corresponding simulation execution module for execution.

[0031] Simulation execution module: is the computing power carrier of simulation calculation, the number is at least two, and is one-to-one bound with a single CPU core.

[0032] Resource binding mechanism, each simulation execution module exclusively occupies a CPU core and runs in the form of independent sub-process, to avoid multi-core resource preemption and maximize the use of hardware parallel computing capability.

[0033] Sub-simulation running process, call SUMO simulation kernel through SUMO's TraCI interface, load sub-net.xml configuration file issued by simulation scheduler, initialize sub-regional road network simulation environment; perform micro-traffic simulation in sub-region according to time step set by simulation scheduler, and real-time collect four types of core data: vehicle position and speed, road traffic flow and road network congestion state.

[0034] A key feature is that the module's simulation data upload frequency is completely consistent with the global time step, ensuring synchronization with the master control node.

[0035] WebSocket service module: This module serves as a two-way interactive bridge between the system and the user, enabling real-time visualization of the simulation status and rapid transmission of user commands.

[0036] Long-lived connections are established between the browser client and the emulation scheduler. Compared with traditional HTTP short connections, long-lived connections can control the interaction latency to within 50ms, meeting the needs of real-time interaction.

[0037] The bidirectional data transmission logic receives the global simulation dataset integrated by the simulation scheduler and pushes it to the browser client in real time, providing a data source for the data fusion and visualization module. The pushed content includes vehicle trajectory data, traffic flow heat map data, congestion status data, etc. at each time step. It also receives interactive commands issued by users on the client, including simulation start and stop control, modification of traffic flow density parameters, addition of temporary road network events (accidents, road construction), simulation speed adjustment, and export of data for specific sub-areas, and forwards the commands to the simulation scheduler in real time.

[0038] Load balancing module: This module is a dynamic regulator that ensures stable system operation. Its core function is to balance the computing load of each CPU core in real time.

[0039] The core monitoring metrics collect three key metrics for each CPU core in real time, and each metric has a preset threshold: CPU utilization rate, threshold range 70%~90%; memory utilization rate, threshold 80%; simulation progress difference, threshold 5-10 time steps (i.e., if the difference between a certain node and the global average progress exceeds this range, it is judged as progress lag).

[0040] The load adjustment trigger logic determines that the simulation execution module corresponding to any core is overloaded when any indicator of any core exceeds the preset threshold. It then sends a sub-region migration trigger signal to the simulation scheduler and simultaneously reports key data such as the sub-region distribution of overloaded nodes, the number of boundary vehicles, and the current simulation progress.

[0041] The dynamic adjustment strategy includes the following steps: For migration target selection, the simulation scheduler prioritizes sub-regions with fewer boundary vehicles and moderate computational complexity as migration targets, thereby reducing data synchronization costs during the migration process. The migration time window is locked, and the migration operation is performed during the synchronization gap between adjacent time steps. This window is the gap between all simulation nodes completing the calculation of the current time step and waiting to enter the next time step, so as to avoid the migration operation interfering with the real-time simulation process. Incremental data synchronization only transmits the latest incremental data (vehicle status and road network status data at the current time step) of the sub-region to be migrated, rather than the full historical data, which greatly reduces the amount of data transmission and migration time, and ensures the continuity of simulation. Task rebinding removes the binding relationship between the overloaded core and the sub-region to be migrated, reallocates the sub-region tasks to the simulation execution modules corresponding to the idle cores with lower loads, and completes the simulation initialization of the migrated region.

[0042] Data fusion and visualization module: This module is responsible for transforming distributed simulation data into intuitive visualizations. Its working principle is as follows: Data preprocessing involves receiving the global simulation dataset uploaded by the simulation scheduler and performing two key operations: first, spatiotemporal alignment, which unifies the simulation data from different sub-regions to the same time axis and spatial coordinate system; and second, consistency verification, which eliminates data deviations caused by calculation differences and ensures the integrity and accuracy of the data.

[0043] The visualization rendering logic generates visual images based on preprocessed data, and the image update frequency is completely synchronized with the global simulation time step. A global road network traffic flow heatmap visually displays the congestion level of each road segment; a real-time vehicle trajectory map presents the travel paths of single or multiple vehicles; a road segment traffic efficiency statistics chart displays the average vehicle speed and traffic volume of each road segment; a boundary vehicle cross-regional flow status map tracks the status changes of vehicles crossing regions; and simulation statistics reports summarize key indicators such as average vehicle speed, congestion duration, and traffic volume.

[0044] The interactive display function allows users to select one or more sub-regions to view detailed simulation data for that region; it also supports data export to meet subsequent analysis needs.

[0045] In this embodiment, the overall engineering process of a high-concurrency distributed simulation system based on SUMO is divided into four stages: initialization, simulation operation, dynamic load adjustment, and simulation end. The modules work closely together to complete the distributed simulation task of a large-scale road network.

[0046] 1. Initialization Phase The road network segmentation module receives the SUMO road network file and simulation parameters, completes intelligent road network segmentation, and generates sub-region configuration files and initial binding relationship tables. The simulation scheduler starts, reads the initial binding relationship table, and distributes sub-region simulation tasks and configuration files to each simulation execution module. Each simulation execution module is bound to a CPU core, and the SUMO sub-simulation kernel is started through the TraCI interface to initialize the sub-region simulation environment. The WebSocket service module establishes a long connection between the browser client and the emulation scheduler, and waits for data transmission. The load balancing module starts and begins collecting load metrics for each CPU core in real time.

[0047] 2. Simulation Execution Phase The simulation scheduler issues a time step start command, sets the current time step size, and all simulation execution modules synchronously start sub-region simulation calculations. Each simulation execution module performs the simulation and collects core data according to the time step, and uploads it to the simulation scheduler after the time step is completed; The simulation scheduler performs consistency checks and path connections on the vehicle data at the boundaries of each sub-region, and merges them to generate a global simulation dataset. The WebSocket service module pushes global simulation data to the browser client, while the data fusion and visualization module completes the screen rendering and display. Users send interactive commands through the client. The commands are forwarded to the simulation scheduler via the WebSocket service module, and then sent to the corresponding simulation execution module by the simulation scheduler to adjust the simulation parameters in real time.

[0048] 3. Dynamic load adjustment phase The load balancing module continuously monitors the load metrics of each CPU core. When it detects that a metric exceeds a threshold, it triggers a sub-region migration signal. The simulation scheduler filters migration objects, locks the time step interval as the migration window, and performs sub-region task unbinding and rebinding operations. Incremental data synchronization is used to transmit the latest data of the sub-region to be migrated to the target node, and the target node completes the simulation initialization. The simulation scheduler re-verifies the boundary vehicle data between the migration area and adjacent sub-regions, calibrates the global simulation progress, and ensures that the simulation process is continuous and uninterrupted. The load balancing module continuously monitors the load status of each node after migration. If the load imbalance problem still exists, the above adjustment process is repeated.

[0049] 4. Simulation End Phase The simulation scheduler receives the end command issued by the user, or after the preset simulation duration is reached, it issues a stop command to all simulation execution modules. Each simulation execution module stops its sub-simulation calculations and uploads the final sub-region simulation data to the simulation scheduler; The simulation dispatcher integrates data from all sub-regions to generate a complete simulation report, which includes traffic statistics, congestion analysis, and event impact assessment. The WebSocket service module pushes the simulation report to the client, allowing users to view or export data online, thus ending the entire simulation process.

[0050] In some embodiments, the road network segmentation module uses a preset algorithm to segment the regional road network into multiple independent sub-regions. During the sub-region segmentation process, it ensures that the boundary road topology of adjacent sub-regions matches, and the computational complexity of each sub-region corresponds to the computing power of the bound CPU core.

[0051] In some embodiments, the preset algorithm is the METIS algorithm, or a graph partitioning algorithm, or a fusion algorithm of the METIS algorithm and the graph partitioning algorithm.

[0052] In some embodiments, the subnet.xml configuration file contains the entry and exit rules for boundary vehicles, the data interaction format, and the synchronization triggering conditions.

[0053] In some embodiments, the simulation scheduler periodically collects the boundary vehicle status of each simulation execution module through a preset boundary vehicle data synchronization protocol, and performs path connection and status consistency correction for vehicles traveling across sub-regions; the step size of the time step control synchronization mechanism is dynamically configured based on the simulation accuracy requirements.

[0054] In some embodiments, after the simulation execution module is bound to the CPU core, it calls the simulation kernel through the SUMO TraCI interface; the simulation data upload frequency of each subprocess is consistent with the step size of the time step control synchronization mechanism, and it receives dynamic traffic event instructions issued by the simulation scheduler during the simulation process.

[0055] In some embodiments, the simulation data collected in real time within the sub-region by the simulation execution module includes: real-time collection of vehicle location, vehicle speed, traffic flow, and road network congestion status data within the sub-region.

[0056] In some embodiments, the monitoring metrics of the load balancing module include CPU utilization, simulation progress difference, and memory usage. When the CPU utilization, simulation progress difference, and memory usage exceed the corresponding preset thresholds, the simulation scheduler is triggered to perform sub-region migration, and incremental data synchronization is used during the migration process to ensure simulation continuity.

[0057] In some embodiments, the data fusion and visualization module displays a global road network traffic flow heatmap, real-time vehicle trajectories, traffic efficiency of each road segment, cross-regional vehicle flow status at the boundary, and simulation statistics. The update frequency of the visualization is synchronized with the simulation time step, and users can customize the viewing of one or more specified sub-regions to view all the corresponding simulation data.

[0058] Example 2 Based on the SUMO-based high-concurrency distributed simulation system provided in Embodiment 1 of this application, correspondingly, Embodiment 2 of this application also provides a SUMO-based high-concurrency distributed simulation method, such as... Figure 2 As shown, it includes: The road network segmentation module receives the SUMO road network file and simulation parameters, and, in combination with the number of CPU cores and computing power distribution, uses a preset segmentation algorithm to divide the road network into multiple sub-regions, generating a sub-net.xml configuration file containing boundary interaction rules, and simultaneously establishing the initial binding relationship between the sub-regions and CPU cores; Based on the initial binding relationship, the simulation tasks of each sub-region are assigned to the corresponding simulation execution modules. The simulation execution modules are bound to the CPU core and start the SUMO sub-simulation process through the TraCI interface. Each simulation execution module runs a sub-simulation according to the time step set by the simulation scheduler, and collects data on vehicle location, vehicle speed, traffic flow and road network congestion status in real time within the sub-region, and uploads it to the simulation scheduler according to the synchronization cycle; the simulation scheduler performs consistency verification and path connection on the vehicle data at the boundaries of each sub-region, and merges them to generate global simulation data. The simulation scheduler pushes the global simulation data to the browser client through the WebSocket service module, and realizes real-time visualization display through the data fusion and visualization module. At the same time, it receives the interactive instructions sent by the browser client and synchronizes them to the corresponding simulation execution module to adjust the parameters or trigger events. The load balancing module monitors the computing load, memory usage, and simulation progress of each CPU core in real time. When it detects that a monitored indicator exceeds the corresponding preset threshold, the simulation scheduler reallocates the road network sub-regions during the synchronization interval between adjacent time steps, adjusts the task binding relationship of the simulation execution module, and completes the simulation initialization of the migration region through incremental data synchronization to ensure the continuity of the simulation process. After the simulation is completed, the simulation scheduler integrates the simulation data from all sub-regions, generates a complete simulation report, and supports client export or online viewing.

[0059] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computing software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0060] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0061] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A high-concurrency distributed simulation system based on SUMO, characterized in that, include: The road network segmentation module receives the SUMO road network file, combines the road network topology complexity, traffic flow density and the number of CPU cores, divides the regional road network into multiple independent sub-regions, and generates a subnet.xml configuration file containing boundary interface definitions for each sub-region; The simulation scheduler, as the master control node, is used to allocate simulation tasks for each sub-region and maintain the synchronization of vehicle status at the boundaries of each sub-region. Based on the time step control synchronization mechanism, it uniformly promotes the simulation process of all simulation nodes, and at the same time receives and integrates the simulation data uploaded by each simulation execution module. The simulation execution module, as a simulation node, shall be at least two in number; and each simulation execution module shall be bound to a single CPU core, execute the SUMO simulation task of the corresponding sub-region as an independent sub-process, and collect simulation data of the sub-region in real time. The WebSocket service module is used to establish a long connection between the browser client and the simulation scheduler, supporting real-time push of simulation status, remote parameter adjustment and interactive command transmission. The load balancing module is used to detect the computing load, memory usage, and simulation progress difference of each CPU core in real time, and dynamically adjust the allocation of simulation tasks in sub-regions. The data fusion and visualization module is used to perform spatiotemporal alignment and consistency verification of simulation data from each sub-region, and to generate a continuous and unified traffic status visualization on the front end.

2. The high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The road network segmentation module uses a preset algorithm to divide the regional road network into multiple independent sub-regions. During the sub-region segmentation process, it ensures that the boundary road topology of adjacent sub-regions matches, and the computational complexity of each sub-region corresponds to the computing power of the bound CPU core.

3. The high-concurrency distributed simulation system based on SUMO according to claim 2, characterized in that, The preset algorithm is the METIS algorithm, or a graph partitioning algorithm, or a fusion algorithm of the METIS algorithm and the graph partitioning algorithm.

4. The high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The subnet.xml configuration file contains the entry and exit rules for boundary vehicles, the data interaction format, and the synchronization trigger conditions.

5. A high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The simulation scheduler periodically collects the boundary vehicle status of each simulation execution module through a preset boundary vehicle data synchronization protocol, and performs path connection and status consistency correction for vehicles traveling across sub-regions; the step size of the time step control synchronization mechanism is dynamically configured based on the simulation accuracy requirements.

6. The high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, After the simulation execution module is bound to the CPU core, it calls the simulation kernel through the SUMO TraCI interface; the simulation data upload frequency of each subprocess is consistent with the step size of the time step control synchronization mechanism, and it receives dynamic traffic event instructions issued by the simulation scheduler during the simulation process.

7. A high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The simulation execution module collects simulation data within the sub-region in real time, including: real-time data on vehicle location, vehicle speed, traffic flow, and road network congestion status within the sub-region.

8. A high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The monitoring metrics of the load balancing module include CPU utilization, simulation progress difference, and memory usage. When the CPU utilization, simulation progress difference, and memory usage exceed the corresponding preset thresholds, the simulation scheduler is triggered to perform sub-region migration, and incremental data synchronization is used during the migration process to ensure simulation continuity.

9. A high-concurrency distributed simulation system based on SUMO according to claim 1, characterized in that, The data fusion and visualization module displays a global road network traffic flow heatmap, real-time vehicle trajectories, traffic efficiency of each road segment, cross-regional vehicle flow status at boundaries, and simulation statistics. The update frequency of the visualization is synchronized with the simulation time step, and users can customize the viewing of one or more specified sub-regions to view all the corresponding simulation data.

10. A high-concurrency distributed simulation method based on SUMO, characterized in that, include: The road network segmentation module receives the SUMO road network file and simulation parameters, and, in combination with the number of CPU cores and computing power distribution, uses a preset segmentation algorithm to divide the road network into multiple sub-regions, generating a sub-net.xml configuration file containing boundary interaction rules, and simultaneously establishing the initial binding relationship between the sub-regions and CPU cores; Based on the initial binding relationship, the simulation tasks of each sub-region are assigned to the corresponding simulation execution modules. The simulation execution modules are bound to the CPU core and start the SUMO sub-simulation process through the TraCI interface. Each simulation execution module runs a sub-simulation according to the time step set by the simulation scheduler, and collects data on vehicle location, vehicle speed, traffic flow and road network congestion status in real time within the sub-region, and uploads it to the simulation scheduler according to the synchronization cycle; the simulation scheduler performs consistency verification and path connection on the vehicle data at the boundaries of each sub-region, and merges them to generate global simulation data. The simulation scheduler pushes the global simulation data to the browser client through the WebSocket service module, and realizes real-time visualization display through the data fusion and visualization module. At the same time, it receives the interactive instructions sent by the browser client and synchronizes them to the corresponding simulation execution module to adjust the parameters or trigger events. The load balancing module monitors the computing load, memory usage, and simulation progress of each CPU core in real time. When it detects that a monitored indicator exceeds the corresponding preset threshold, the simulation scheduler reallocates the road network sub-regions during the synchronization interval between adjacent time steps, adjusts the task binding relationship of the simulation execution module, and completes the simulation initialization of the migration region through incremental data synchronization to ensure the continuity of the simulation process. After the simulation is completed, the simulation scheduler integrates the simulation data from all sub-regions, generates a complete simulation report, and supports client export or online viewing.

Images