A digital twin-based comprehensive transportation hub emergency scenario simulation system

Abstract

The present invention discloses a digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system, which relates to the field of digital simulation technology and includes: a data processing unit, a simulation and deduction unit, and a visualization display unit; the data processing unit is used to obtain the basic emergency scenario data of the target digital twin comprehensive transportation hub, and to integrate the basic data from multiple sources to build a model data environment; the simulation and deduction unit calls the simulation and deduction algorithm according to the emergency scenario model data to build a simulation and deduction algorithm model, and performs emergency situation deduction and dispatch disposal based on the simulation and deduction algorithm model; the visualization display unit allocates and manages user permissions, plots deduction data based on GIS map basic information, and is also used to query and visualize deduction data according to user query instructions. The present invention builds a digital twin platform based on the simulation and deduction algorithm model and dynamic data, making the deduction process more accurate and realizing visualization display in combination with GIS map data.

Description

Technical Field

[0001] The present invention relates to the field of digital simulation technology, and in particular to a digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system. Background Art

[0002] Public transportation hubs are characterized by large scale, extensive facilities, and compact layouts. They are densely populated, complexly organized, and feature a multitude of transportation options. In the event of an emergency, delayed evacuation can lead to serious accidents. Existing research on public transportation scheduling within public transportation hubs has primarily focused on the daily operational decision-making needs of public transportation hubs. However, research on emergency coordinated scheduling in unusual scenarios, such as emergencies, technical failures, service disruptions, natural disasters, and accidents, remains insufficient.

[0003] The emergency scenario simulation of a comprehensive transportation hub is a simulation system for specific emergency scenarios and their handling processes within the hub. Although it has the advantages of strong flexibility and high intervention, the current transportation emergency scenario simulation system generally has problems such as non-standard element representation and insufficient realism of the simulation environment, which makes it difficult to accurately deduce the evolution of the situation and cannot effectively reflect the evolution mechanism of emergencies and the impact of handling decisions on the evolution process of the emergency situation. Summary of the Invention

[0004] In light of this, the present invention provides a comprehensive transportation hub emergency scenario simulation system based on digital twins. This system constructs an information model of all the elements of the transportation hub, creates a digital twin simulation platform, and implements dynamic matching, call-up, and correlation analysis of the simulation model, thereby completing emergency situation simulation and dispatching. This system overcomes two major bottlenecks in existing transportation emergency simulation technology: first, the lack of standardized element representation and a high-fidelity simulation environment; second, the difficulty in accurately simulating the evolution of the situation and effectively quantifying the evolution mechanism of emergencies and the real-time impact of response decisions.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A digital twin-based emergency scenario simulation system for integrated transportation hubs aims to improve emergency response efficiency through high-precision simulation in virtual space. The system consists of three core modules: a data processing unit, a simulation unit, and a visualization unit. These modules work together to provide data-driven intelligent decision support.

[0007] The data processing unit is responsible for building the data base of the digital twin, which is used to obtain the basic emergency scenario data of the target digital twin integrated transportation hub, and integrate the basic data from multiple sources to build a model data environment. During this process, the strategy training module set up in the data processing unit uses historical sample data and simulation feedback to generate a strategy model based on reinforcement learning algorithm training, providing intelligent scheduling support for subsequent simulation and deduction processes. The simulation and deduction unit is used to call the simulation and deduction algorithm based on the emergency scenario model data to construct a simulation and deduction algorithm model, and perform emergency situation deduction and scheduling based on the model. At the same time, the system combines the strategy model output by the strategy training module to dynamically optimize the response strategy during the deduction process, thereby improving the system response efficiency and decision-making quality. Specifically, the strategy model is integrated into the deduction process as an auxiliary decision-making module. The system inputs the current scenario status into the strategy model in each deduction process, and the model outputs recommended response strategies, such as evacuation path plans and resource scheduling sequences. These strategies are input into the scheduling module together with the simulation deduction results as the basis for decision-making, thereby improving the rationality of strategy selection and system response results. The strategy model can output multi-granularity response strategies, including resource scheduling sequences at the tactical level, evacuation path guidance at the operational level, regional lockdown plans, and information broadcast priority rankings. The system automatically screens and executes the most suitable strategy action combination based on the response target. The visualization display unit is used to allocate and manage user permissions, and to plot deduction data based on GIS map basic information. It is also used to query and visualize deduction data based on user query instructions and conditions. The visualization display unit also supports the visualization of the response plan results output by the strategy model, displaying recommended paths, scheduling priorities, and comparative analysis diagrams with actual scheduling plans, to assist users in understanding the basis for model recommendations and improve the efficiency of human-computer collaborative decision-making.

[0008] Furthermore, the data processing unit includes a basic data module, a data integration module, a factor modeling module and a strategy training module;

[0009] The basic data module is used to collect multi-source data from the transportation hub infrastructure layer based on the established data model and build a basic database based on this data resource. This multi-source data includes basic map data (GIS vector data, satellite imagery), IoT sensor data (equipment status, environmental parameters), and interaction data (passenger flow statistics). Based on this data, a basic database covering all elements, including buildings, roads, and equipment, is constructed.

[0010] The data integration module is connected to the basic data module. It is used to integrate multi-source data and implement an eight-dimensional data processing process, including cleaning, classification, coding, mapping, labeling, fusion, pattern mining, and spatiotemporal analysis. These are used to eliminate noise data, divide data by traffic elements, unify semantic identification, spatial coordinate conversion, metadata annotation, multi-source data alignment, and passenger flow spatiotemporal patterns. The integrated processed data is then stored in the core database.

[0011] The element modeling module is used to represent all element entities of a transportation hub as an information model, with the goal of realizing business functions and building a hub digital twin information model library.

[0012] The strategy training module is based on the multi-source data environment output by the data integration module and the factor modeling module in the data processing unit, and uses the reinforcement learning algorithm to perform training tasks in the constructed simulation environment. Specifically, in each round of training, the system first sets the current scene state, including state parameters such as passenger flow density, channel occupancy, and warning level; the strategy model selects a scheduling action (such as evacuation path adjustment, traffic resource allocation plan, etc.) based on the current state; the simulation environment executes the action and generates feedback results, such as handling time, congestion index changes, etc.; the system uses the feedback information as a reward signal to update the strategy model parameters, and iteratively executes multiple rounds of training processes to continuously improve the response capability and strategy effect of the strategy model in complex emergency scenarios. The current scene input parameters further include quantifiable structured input features, such as spatial density distribution, equipment operating status, channel accessibility, and regional risk level. The system continuously updates the state vector during the deduction cycle to drive the strategy model to generate corresponding response strategies. The policy model can be constructed based on a value function approximation method (such as a deep Q network) or a policy gradient algorithm (such as PPO). By cumulatively learning state transition samples, the policy behavior function is iteratively optimized to achieve dynamic response capabilities in complex simulation environments.

[0013] Furthermore, the hub digital twin information model library includes a model database, a rule relationship database and an algorithm database; the model database is used to store the model number, model type, model name, scene event or disposal event number to be matched, algorithm name and storage location of the information model, the rule relationship database is used to store the relationship rule data between models, and the algorithm database is used to store the scene event evolution model and the scheduling disposal event model and the strategy model obtained by the strategy training module.

[0014] Furthermore, the simulation and deduction unit includes an emergency situation awareness module, a model matching module and a deduction and simulation module;

[0015] The emergency situation awareness module is used to analyze the integrated processing data and extract scene events and event object parameters (such as fire location and passenger flow density) to obtain emergency scene model data;

[0016] The model matching module is connected to the emergency situation awareness module. The model matching module is used to call the simulation deduction algorithm model according to the emergency scenario model data matching. The simulation deduction algorithm model responds and processes according to the data processing demand signal. In addition to adopting the improved KNN algorithm dynamic matching algorithm model, a strategy matching mechanism based on the reinforcement learning model is also introduced. By constructing a state-strategy mapping relationship, the strategy category of the current scenario is judged according to real-time scenario parameters (such as time type, impact range, resource status, etc.), and the strategy model outputs the optimal matching solution and selects the corresponding simulation deduction algorithm model combination; after the deduction is executed, the system records feedback indicators such as deduction deviation and resource scheduling efficiency, and updates the strategy model based on feedback to achieve continuous optimization of strategy matching behavior, thereby improving the accuracy of model selection and the response adaptability of combined deduction.

[0017] The simulation module is connected to the model matching module. It is used to input event data from different emergency scenarios as simulation node conditions into the corresponding simulation algorithm model, generating event evolution model files and event handling model files. An evolution sequence is then established based on a chronological order. During each simulation cycle, all simulation algorithm models in the evolution sequence are run, and the file information is analyzed and processed. Based on the analysis and processing results, the simulation algorithm model is optimized based on the event data from different scenarios and the pre-set simulation result data. Simultaneously, the policy training module in the data processing unit continuously optimizes the policy model based on feedback data from the simulation process, improving the accuracy and stability of its policy output and establishing an adaptive evolutionary optimization system. To ensure system stability, if the policy model output fails to meet pre-set safety constraints or clearly conflicts with the manually configured expert policy, the system automatically switches to a default scheduling solution or manual review mode, ensuring that key emergency actions are controllable and fault-tolerant.

[0018] Furthermore, the simulation unit also includes a simulation control module and a message management module. The simulation control module is used to control the simulation time and step size, and schedule the model database (loading / unloading models) based on feedback messages. The message management module is implemented using a message queue mechanism to record the data results of emergency scenario simulation and exchange information with the visualization unit and transportation department to ensure accurate transmission and sharing of information. It also uses reinforcement learning algorithms to evaluate and learn strategies for key decision points in the simulation process, including:

[0019] During the data processing phase, a policy training model is constructed. By leveraging historical simulation data and interacting with multiple rounds of simulation, reinforcement learning training is performed in the simulation environment to generate a policy model with adaptive scheduling capabilities. This model is iteratively optimized based on trial-and-error feedback in the simulation environment and trained using reinforcement learning algorithms such as deep Q networks or proximal policy optimization, aiming to improve the ability to select strategies for complex emergency scenarios.

[0020] During the simulation, the system uses the strategy model to evaluate strategy selection based on the current environmental status (such as channel occupancy, equipment availability, passenger flow density, etc.), and selects the optimal response plan from among alternative response plans such as evacuation route optimization, resource scheduling sequence, and traffic control measures.

[0021] After the simulation is executed, the system records the feedback results, including indicators such as response efficiency and risk suppression level, and uses them as update samples to continuously update the strategy model parameters;

[0022] By building a self-learning closed loop of "strategy training-strategy application-feedback optimization," the system's adaptability to dynamic scenarios and intelligent emergency decision-making capabilities are enhanced. The strategy model, running through data training, simulation, and optimization feedback, serves as the core intelligent agent dynamically invoked by each system module. This not only improves the real-time and accuracy of strategy selection, but also enhances the system's overall closed-loop intelligent control capabilities.

[0023] Furthermore, a situation assessment is performed based on the deduction model of the optimized model. The situation assessment process includes:

[0024] The event data in the emergency scenario is used as the simulation node conditions to input into the corresponding simulation algorithm model. At the same time, some scenarios use simulation models generated based on reinforcement learning strategies to achieve more flexible and intelligent response scheduling, improve the ability to respond to complex emergency situations, and establish an evolution sequence according to the time sequence. In each simulation cycle, all simulation algorithm models in the evolution sequence are calculated;

[0025] When resources arrive after the incident is handled, the emergency scenario information is fed back to the simulation deduction algorithm model for processing to determine whether there is evolution. If so, the corresponding new scenario event is generated according to the relationship rules between the deduction results and the model and added to the evolution sequence to conduct the current emergency situation assessment. Otherwise, the current emergency situation assessment is directly carried out, and the situation of the organization, resources and environment is updated.

[0026] Furthermore, the visualization display unit includes a map basic information management module, a query statistics module and a data plotting display module;

[0027] The basic map information management module is used to manage basic map data and pre-process it according to the actual needs of emergency scenario simulation. Pre-processing includes map clipping, adding and editing layer attributes, performing topological operations based on the roads and spatial locations of transportation hubs and building a topological network based on the operation results, publishing map services, classifying layers according to the set categories and layer names to form a tree diagram, storing the tree diagram in the core database, and adding deduced elements to the map layer according to feature attributes. The query statistics module supports querying deduced data by time range, geographic location, data type, and other conditions, and performs statistical analysis on the query results to generate charts and reports. The data plotting and display module plots the deduced data on the map in real time, supports multi-layer overlay display, and sorts and displays data according to information priority (such as the urgency of the event).

[0028] Furthermore, the visual display unit also includes a user management module, which is used to match corresponding role permissions according to user basic information.

[0029] The basic map information management module implements topological operations to construct a three-dimensional road network and publishes WMTS map services, supporting dynamic layer rendering (heat map / trajectory flow). The data plotting and display module uses a priority algorithm to display key information (red warnings are automatically pinned to the top) and supports three-dimensional model overlays. The query statistics module provides a natural language query interface. The user management module can configure role permissions based on the RBAC model, supporting: commanders (situation overview), operators (equipment control), etc.

[0030] It can be seen from the above technical solution that the advantages of the present invention are:

[0031] 1. The present invention builds a digital twin platform based on simulation deduction algorithm models and dynamic data, integrates and processes data and models elements, making element representation more standardized and creating a more realistic simulation environment. It also realizes a more accurate emergency simulation situation deduction process based on the model connection relationship and evolution relationship. It also realizes the visualization of data in combination with GIS map data, intuitively reflecting the evolution mechanism of emergencies and the impact of disposal decisions on the evolution process of emergency situations. At the same time, by introducing a reinforcement learning model, the system can be continuously trained through historical data and simulation feedback, optimize the emergency response strategy within each deduction cycle, and further improve the decision-making accuracy and flexibility in the emergency simulation process.

[0032] 2. Automatic model combination and evolution are achieved through a rule engine, breaking through the static model limitations of traditional systems. A closed loop of "perception-deduction-feedback-optimization" is established, dynamically correcting model parameters based on real-time data to improve prediction accuracy. Reinforcement learning models play a key role in this closed loop. Through continuous feedback learning, they intelligently adjust emergency response strategies, enabling the system to better cope with different emergency scenarios and automatically optimize the simulation and deduction process. By integrating BIM and GIS, bidirectional mapping between physical space and virtual models, as well as multi-physics coupled simulation of pedestrian flow, vehicle flow, and structural stress fields, is achieved, providing comprehensive transportation hubs with full-lifecycle emergency management capabilities, significantly improving the efficiency and accuracy of emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0034] Figure 1 It is a structural schematic diagram of the present invention.

[0035] Figure 2 Schematic diagram of the overall framework structure of this embodiment.

[0036] Figure 3 Schematic diagram of the structure of the data processing unit of the present invention.

[0037] Figure 4 It is a structural diagram of the simulation deduction unit of the present invention.

[0038] Figure 5 It is a structural schematic diagram of the visual display unit of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0040] This system adopts a layered architecture, such as Figure 2As shown, the bottom layer is the data perception layer (data processing unit), which integrates infrastructure IoT data, operation management system data, and environmental perception data. The middle layer is the simulation application layer (simulation deduction unit), which includes the emergency situation awareness engine, dynamic model matching library, and cross-temporal and spatial deduction kernel. The top layer is the interactive service layer (visualization display unit), which provides a GIS spatiotemporal visualization interface, a multi-role permission management system, and a situation assessment dashboard. The various units communicate through standardized interfaces to ensure the flexibility of module replacement and horizontal expansion. The overall system follows the concept of digital twins and improves the efficiency and accuracy of emergency response through simulation in virtual space. Figures 1 to 5 This embodiment provides a comprehensive transportation hub emergency scenario simulation and deduction system based on digital twins. Among them, the application of digital twin technology in the comprehensive transportation hub emergency scenario simulation and deduction is mainly reflected in a highly integrated and intelligent simulation method, which improves the efficiency and accuracy of emergency response through simulation in virtual space. Figure 1 and Figure 2 As shown in the figure, the system specifically includes: a data processing unit, a simulation and deduction unit, and a visualization unit. The data processing unit is used to obtain the basic emergency scenario data of the target digital twin integrated transportation hub, integrate the basic data from multiple sources, and build a model data environment.

[0041] Specifically, if Figure 3 As shown in the figure, the data processing unit includes a basic data module, a data integration module, and a feature modeling module. The basic data module is used to build a digital twin model based on the hub's physical entities (such as buildings, roads, and equipment). Next, data collection is performed, including basic map data (GIS geospatial data, terrain elevation data, etc.); interactive data (such as pedestrian and vehicle flow dynamics); and IoT perception data (such as real-time monitoring data from sensors, cameras, and firefighting equipment). Finally, a basic database is established to store the raw data, providing a foundation for subsequent processing.

[0042] The data integration module is connected to the basic data module. It integrates multi-source data, integrating heterogeneous data such as BIM models, GIS data, video images, and IoT sensor data. It also cleans the data to remove noise and errors. Classification and coding: categorize data by data type (e.g., structured vs. unstructured) and standardize the coding format. Mapping and tagging: map data to a unified coordinate system and add metadata tags. Fusion: integrates data from different sources (e.g., BIM and GIS) to form a complete data environment. Pattern mining and spatiotemporal analysis: use machine learning algorithms to discover data patterns and analyze spatiotemporal correlations. Based on the complete data environment constructed by the data integration module, the system further extracts key data dimensions, such as basic map data, IoT sensor data, and interaction data, and constructs the data input set required by the strategy training module. Based on this input set, the strategy training module sets initial state parameters, such as passenger flow density, equipment operating status, and regional risk level, to generate a preliminary strategy model for the simulation module to generate and verify emergency response strategies.

[0043] The element modeling module abstracts the physical entities of a transportation hub (such as buildings, roads, and equipment) into information models. This module aims to implement business functions and construct a hub digital twin information model library. This library includes a model database, which stores model numbers, types, names, and association algorithms; a rule-relationship database, which stores rule data linking models (such as model combinations and evolution rules); and an algorithm database, which stores scenario-event evolution models and dispatch and disposal model algorithms. The reinforcement learning module automatically optimizes the rules and relationships in the model library based on real-time scenario feedback. Through reinforcement learning, the system continuously adjusts event-related model parameters, enabling more precise responses to emergencies and dynamically adjusting models related to traffic flow, equipment scheduling, and other aspects, thereby improving the accuracy and efficiency of emergency management. Data resources at the infrastructure layer include basic map data, interaction data, and IoT perception data. The hub digital twin information model library comprises a model database, a rule-relationship database, and an algorithm database. The model database stores the information model's model number, model type, model name, the scenario event or disposal event number to be matched, the algorithm name, and the storage location.

[0044] Since it is necessary to obtain multiple types of data of the target integrated transportation hub when simulating the emergency scenario of the target integrated transportation hub, it is necessary to first perform data processing tasks and build a data environment. Specifically, it is necessary to load the building information data model (Building Information Modeling, BIM) used for the simulation of emergency scenarios of the integrated transportation hub, multi-regional geographic information system GIS data and basic data files. Among them, the basic data files include video image data, building element models, road traffic element models, terrain site models, vegetation element models, spatial facility element models and identifications, production information and IoT perception data, as well as data model connection relationships.

[0045] The simulation unit is connected to the data processing unit. It uses simulation algorithms based on emergency scenario model data to construct a simulation model. Emergency scenario simulation and dispatch response are then conducted based on this model. The strategy training module, a core component of the simulation unit, utilizes the emergency scenario model data generated by the data processing unit to construct a preliminary strategy model. It continuously optimizes strategy parameters through feedback from the simulation process, enhancing the model's dynamic response capabilities. Emergency scenarios include abnormal situations such as technical failures, service disruptions, natural disasters, accidents, calamities, and emergencies. Emergencies are sudden events that cause or may cause serious social harm, necessitating emergency response measures for public health, social security, and other public health and safety issues. Based on hub management needs, analysis algorithms are provided for scenarios such as daily passenger capacity assurance, extreme weather, fire incidents, large passenger evacuations, and abnormal data analysis. A "simulation mode" function is provided, allowing all relevant data to be input through a combination of automatic and manual methods. The system then generates simulated images and corresponding data through the simulation algorithm.

[0046] Specifically, if Figure 4 As shown, the simulation and deduction unit includes an emergency situation awareness module, a model matching module and a deduction and simulation module.

[0047] The Emergency Situation Awareness module is primarily responsible for data parsing, analyzing integrated and processed data, extracting scenario events and event object parameters, and converting the parsed data into emergency scenario model data suitable for simulation and deduction. The Model Matching module, connected to the Emergency Situation Awareness module, primarily involves algorithm model matching and response processing. Algorithm model matching involves matching and calling the corresponding simulation and deduction algorithm model from the algorithm database based on the emergency scenario model data. Response processing involves the simulation and deduction algorithm model responding to data processing demand signals in preparation for deduction and calculation.

[0048] The simulation module is connected to the model matching module and involves event data input, simulation calculations, and model optimization. Event data input primarily involves inputting event data from different emergency scenarios into the algorithm model as simulation node conditions. The simulation calculations include event evolution models and event handling models. The event evolution model simulates the development of events over time (e.g., fire spread, passenger flow accumulation). The event handling model simulates the effects of emergency measures on the events (e.g., evacuation plans, traffic control). Finally, the algorithm model is optimized based on the simulation results and preset data. The simulation process involves inputting event data from different emergency scenarios into the corresponding simulation algorithm model as simulation node conditions. Event evolution model files and event handling model files are generated, and an evolution sequence is established in chronological order. During each simulation cycle, all simulation algorithm models in the evolution sequence are run, and the file information is analyzed and processed. Based on the analysis results, the simulation results of the event data in different scenarios relative to the simulation algorithm model are determined. The simulation algorithm model is then optimized based on the simulation results and the preset simulation result data.

[0049] During the optimization process of the simulation module, the system uses an algorithm framework based on the combination of Policy Gradient and Proximal Policy Optimization (PPO / Proximal Policy Optimization) to optimize the policy model. The method is as follows:

[0050] Policy Optimization Algorithm: During each simulation cycle, the system first inputs the policy model based on the current scenario state (S), which then outputs the current scheduling policy (A). The system calculates the reward value (R) based on the simulation feedback and performs a gradient update on the parameter θ in the policy model based on the policy-provided formula ∇θJ(θ)=E[∇θlogπθ(A|S)*R].

[0051] PPO Optimization: To further improve optimization stability and prevent policy failure caused by overly rapid updates, the system introduces the PPO algorithm. The PPO algorithm effectively controls the policy update amplitude by introducing a clipping loss function: L_CLIP(θ)=min(r(θ)*A,clip(r(θ),1-ε,1+ε)*A, where r(θ)=πθ(A|S) / πθ_old(A|S) is the policy ratio and ε is the control threshold.

[0052] Strategy Optimization Process: After each simulation cycle, the system uses feedback (such as evacuation efficiency and resource allocation effectiveness) as a reward signal to update the parameters of the strategy model through PPO optimization. The optimized strategy model will be applied in the next simulation cycle to achieve adaptive optimization and rapid response to emergency scenarios.

[0053] Algorithm database: During the strategy optimization process, the system stores the optimized strategy model in the algorithm database for quick call in subsequent deduction scenarios, and dynamically adjusts the strategy model according to real-time scenario data to achieve closed-loop optimization of the reinforcement learning strategy model and simulation deduction module.

[0054] In this embodiment, various simulation tools are integrated through corresponding application programming interfaces (APIs), and corresponding simulation algorithms are called. These algorithms are used to construct simulation models. These include mathematical algorithms such as trend analysis, statistical analysis, and data fitting, as well as business algorithms such as environmental simulation, security simulation, traffic simulation, and equipment simulation. The algorithm database contains a comprehensive set of simulation algorithms, facilitating their use within the digital twin integrated transportation hub emergency scenario simulation system.

[0055] More specifically, the simulation unit also includes a simulation control module and a message management module. The simulation control module primarily controls the duration and step size of the simulation, simulating emergency scenarios at different timescales. It also dynamically loads and unloads models and updates parameters based on feedback (such as real-time data and user input). The message management module records all data results (such as indicators, parameters, and events) during the emergency scenario simulation. It also interacts with the visualization unit, sending simulation results for display and communicating with the transportation department to transmit key information for decision-making (via a message queue or event bus).

[0056] The simulation control module can set the duration of the simulation and the time interval between each step according to actual needs. This helps simulate emergency scenarios at different time scales, thereby better evaluating the effectiveness of response strategies. During the simulation process, the simulation control module schedules the model database based on feedback messages, such as real-time data and user input. This includes operations such as loading or unloading models and updating model parameters to ensure the accuracy and real-time performance of the simulation. The message management module is responsible for recording all data results generated during the simulation, including various indicators, parameters, events, etc. The message management module also needs to exchange information with other systems or departments, such as the visualization unit and the transportation department. It can send simulation results to the visualization unit for display and send key information to relevant departments such as the transportation department to enable them to make timely decisions. Specifically, messages are transmitted between different systems or departments through a message queue or event bus mechanism.

[0057] In this embodiment, a situation assessment is performed based on the deduction model of the optimized model. The situation assessment process includes:

[0058] Step A: Input the event data in the emergency scenario as the simulation node conditions into the corresponding simulation algorithm model, and establish an evolution sequence according to the time sequence. In each simulation cycle, all simulation algorithm models in the evolution sequence are operated;

[0059] Step B: When resources arrive after the incident is handled, the emergency scenario information is fed back to the simulation deduction algorithm model for processing to determine whether there is evolution. If so, the corresponding new scenario event is generated according to the relationship rules between the deduction results and the model and added to the evolution sequence to conduct the current emergency situation assessment. Otherwise, the current emergency situation assessment is directly conducted and the situation of the organization, resources and environment is updated.

[0060] In traditional information evolution systems, the relationships between models and between models and data are fixed. However, for the different types and properties of entity models within an integrated transportation hub, dynamic organization is required for computational scheduling. In this embodiment, the hub digital twin information model library organizes numerous models according to a specific structure and is subdivided into a model database, a rule relationship database, and an algorithm database to enable model retrieval and matching, as well as the evolutionary calculation of scenario events. The evolutionary calculation of scenario events involves the combination and evolution of models. The combination of models involves combining scenario event models with corresponding scheduling and handling event models to generate new scenario event model algorithms. For example, when a road is congested, traffic control is implemented and a congestion index is output. The connection rule is the initial congestion index minus the traffic control impact to obtain the current congestion index. Model evolution refers to the generation of new scenario events when specific scenario events occur at the same entity. For example, when passenger flow congestion reaches the second-level warning level and there is a conflict or fire at the scene location, resulting in casualties, this scenario event will trigger the corresponding model and evolve.

[0061] Digital twin technology for integrated transportation hubs, centered around digital twin 3D and dynamic scene simulations, aims to dynamically interact between physical and digital objects, enabling assessment of current status and diagnosis of past issues, thereby forming an intelligent decision-making support system based on a "perception-prediction-action" model. This technology utilizes internal modeling and rendering of transportation hubs, as well as simulations of train distribution, passenger flow, and vehicle flow at stations, to comprehensively display early warning and forecast information in real time. By building a dynamic scene simulation module and integrating BIM and GIS, integrating multi-source intelligent monitoring sensor information in real time, and utilizing 3D interactive digital twin visualization technology, it enables the coordinated management of the digital and physical worlds, creating a fully integrated closed-loop system for the entire hub lifecycle and achieving one-stop operational management for future hub safety and intelligent control.

[0062] In this embodiment, based on MECE (Mutually Exclusive Collectively Excessive

[0063] Based on the principle of emergency situation assessment (MECE), an emergency situation assessment model is constructed. Real-time sensor data such as weather, fire sensors, cameras, videos, and vehicles passing through checkpoints are collected. A secondary monitoring indicator system is constructed for dimensions such as extreme weather, fire accidents, large passenger flow events, temperature, humidity, smoke sensor equipment alarms, and video equipment alarms. Based on the secondary monitoring dimension indicators, a primary monitoring indicator system is established for emergency events and equipment alarms. The weight of each indicator is calculated using the hierarchical entropy analysis method, and a comprehensive emergency situation scoring model is constructed based on the probability and severity of risk events. The specific form is as follows:

[0064] ,

[0065] ,

[0066] ,

[0067] in, The information entropy of the indicator is is the proportion of the jth indicator in the i-th scheme, n is the number of indicators, The weight of the j-th indicator, is the i-th value of the j-th indicator, For the comprehensive score of the i-th plan, four risk status levels of event emergency, namely, very high, high, medium, and low, are designed based on the comprehensive score of emergency pressure state discrimination, so as to realize real-time monitoring and early warning of emergency events in comprehensive transportation hubs.

[0068] The visualization unit is primarily used to manage user permissions, plot inference data on a GIS map, and support query and visualization. Users can query inference data based on query instructions and conditions, which may include time range, geographic location, data type, etc. Query results are visualized in the form of graphs and charts, helping users more intuitively understand the inference data and identify patterns, trends, and anomalies.

[0069] Specifically, if Figure 5As shown, the visualization display unit includes a map basic information management module, a query statistics module, a data plotting and display module, and a user management module. The map basic information management module involves map data management, map service publishing, and deduction element overlay processes. Map data management mainly involves preprocessing GIS map basic data (map clipping, layer attribute addition and editing, topological operations based on the roads and spatial locations of transportation hubs, and building a topological network based on the operation results). Map service publishing is to publish the processed map data as a service for other modules to call. Deduction element overlay is to overlay the deduction data (such as event location, impact range) onto the map layer, that is, to publish the map service, classify the layers according to the set categories and layer names to form a tree diagram, store the tree diagram in the core database, and add the deduction elements to the map layer according to the element attributes, thereby realizing the query, editing and configuration of the emergency scenario situation elements.

[0070] The query statistics module supports querying deduced data by time range, geographic location, data type and other conditions, and performs statistical analysis on the query results to generate charts and reports. The data plotting and display module plots the deduced data on the map in real time, supports multi-layer overlay display, and sorts and displays data according to information priority (such as the urgency of the event). The user management module is used to match the corresponding role permissions (such as management command personnel, fire operation department, health department) according to the user's basic information, and perform user authentication to ensure the legitimacy of the user's identity and protect the security of system data. Users of this system include managers who maintain basic information and configure solutions, as well as departments involved in emergency simulation, including fire, health and public security departments. The emergency scenario simulation service is designed for high-level emergency command and management personnel. Managers can use functions such as situation information query, past situation query, map browsing, and situation element configuration.

[0071] Digital twins are a technology that digitally creates virtual entities of physical entities, leveraging historical and real-time data, as well as algorithmic models, to simulate, verify, predict, and control the entire lifecycle of physical entities. For integrated transportation hubs, data twin technology can be used to build a cross-screen collaborative mechanism for management units, business collaboration, mobile collaboration, and server-side services. This enables real-time equipment monitoring, proactive accident prevention, rapid fault diagnosis, and optimized maintenance strategies.

[0072] This system combines Building Information Modeling (BIM) and Geographic Information Systems (GIS) to construct a high-precision 3D digital twin model. It also integrates IoT sensor data and video data to achieve real-time synchronization between virtual and physical entities. Through 3D interactive linkage, users can operate and monitor the virtual scene in real time. Based on simulation results, the system assesses the current emergency situation (e.g., risk classification). Finally, it uses time series analysis to predict event trends (e.g., passenger flow growth forecasts). Statistical analysis of historical data reveals event patterns (e.g., accident-prone areas). Event evolution curves are fitted using regression and interpolation methods. Simulations also include environmental, security, transportation, and equipment simulations. These simulations include the impact of environmental disasters like fire on hubs, the impact of security measures (e.g., evacuation route blockades) on crowd flow, the impact of traffic control and road closures on vehicle flow, and emergency response scenarios such as equipment failures and power outages. Feedback from simulations allows for dynamic adjustment of model parameters and optimization of simulation results. This system integrates multi-source data, multiple algorithms, and models to achieve one-stop emergency management. Digital twin technology enables real-time interaction and dynamic simulation between virtual and physical scenes. It provides an intuitive visual interface, supporting emergency commanders in making quick decisions and improving the efficiency and accuracy of emergency response.

[0073] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system, characterized by: include: Data processing unit, simulation unit and visualization display unit; The data processing unit is used to obtain basic emergency scenario data of the target digital twin integrated transportation hub, and integrate the basic data from multiple sources to build a model data environment; The simulation deduction unit is connected to the data processing unit, and the simulation deduction unit calls the simulation deduction algorithm to construct a simulation deduction algorithm model based on the emergency scenario model data, and performs emergency situation deduction and dispatch disposal based on the simulation deduction algorithm model; The visualization display unit is responsible for the allocation and management of user rights, plotting deduction data based on GIS map basic information, and querying and visualizing the deduction data according to the user's query instructions and conditions; The data processing unit includes a basic data module, a data integration module and an element modeling module; The basic data module is used to obtain data resources of the infrastructure layer according to the established data model, and to build a basic database based on the data resources; The data integration module is connected to the basic data module. The data integration module is responsible for integrating data from multiple sources, cleaning, classifying, encoding, mapping, labeling, fusing, and performing pattern mining and spatiotemporal analysis on the data, and storing the integrated data in the core database. The element modeling module is responsible for converting all element entities of the transportation hub into information models, and building a digital twin information model library of the hub with the goal of realizing business functions. The simulation deduction unit includes an emergency situation awareness module, a model matching module and a deduction simulation module; the emergency situation awareness module is used to parse the integrated processing data and extract the scene events and event object parameters to obtain emergency scene model data; the model matching module is connected to the emergency situation awareness module, and the model matching module is used to call the simulation deduction algorithm model according to the emergency scene model data matching, and the simulation deduction algorithm model responds and processes according to the data processing demand signal; the deduction simulation module is connected to the model matching module, and the deduction simulation module is used to input the event data under different emergency scenarios as simulation deduction node conditions into the corresponding simulation deduction algorithm model to obtain event evolution model files and event handling model files, and establish an evolution sequence according to the time sequence. In each deduction cycle, all simulation deduction algorithm models in the evolution sequence are calculated, the file information is analyzed and processed, and the deduction results of the event data under different scenarios relative to the simulation deduction algorithm model are determined according to the analysis and processing results, and the simulation deduction algorithm model is optimized based on the deduction results and preset deduction result data; The simulation deduction unit also includes a deduction control module and a message management module; The simulation control module is used to control the simulation time and step size settings, and to schedule and manage the model database according to the feedback messages; the message management module is used to record the data results of the emergency scenario simulation and to exchange information with the visualization display unit and the transportation department; The situation assessment is performed based on the deduction model of the optimized model. The situation assessment process includes: The event data in the emergency scenario is used as the node conditions of the simulation deduction and input into the corresponding simulation deduction algorithm model. An evolution sequence is established according to the time sequence. In each deduction cycle, all simulation deduction algorithm models in the evolution sequence are operated. When resources arrive after the incident is handled, the emergency scenario information will be fed back to the simulation deduction algorithm model for processing to determine whether there is evolution. If so, the corresponding new scenario event will be generated according to the relationship rules between the deduction results and the model and added to the evolution sequence to conduct the current emergency situation assessment. Otherwise, the current emergency situation assessment will be conducted directly, and the situation of the organization, resources and environment will be updated. The situation assessment and model optimization will be combined with the reinforcement learning model to achieve model adaptive adjustment and intelligent strategy optimization through continuous learning and reinforcement of the deduction feedback results.

2. The digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system according to claim 1 is characterized in that: The hub digital twin information model library includes a model database, a rule relationship database and an algorithm database; the model database is responsible for storing various attributes of the information model, including model number, model type, model name, scene event or disposal event number to be matched, and related algorithm name and storage location; the rule relationship database is used to store the relationship rule data that connects the models; the algorithm database is used to store the evolution model of the scene event and the event model of scheduling and disposal.

3. The digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system according to claim 1 is characterized in that: The visualization display unit includes a map basic information management module, a query statistics module and a data plotting display module; The map basic information management module is used to manage the basic map data and pre-process the basic map data according to the actual needs of emergency scenario simulation. The pre-processing includes map clipping, adding and editing layer attributes, performing topological operations based on the roads and spatial locations of transportation hubs and building a topological network based on the operation results, publishing map services, classifying layers according to set categories and layer names to form a tree diagram, storing the tree diagram in the core database, and adding deduced elements to map layers according to element attributes; The query statistics module is used for users to query and browse the corresponding deduction and situation evolution information; the data plotting and display module is used to plot the relevant evolution data on the map, and display the data according to the information priority, and display messages of the same level in the order in which they are received.

4. The digital twin-based comprehensive transportation hub emergency scenario simulation and deduction system according to claim 3 is characterized in that: The visual display unit further includes a user management module, which is used to match corresponding role permissions according to user basic information.

Images