Digital twinborn system for low-orbit giant satellite base system and construction method of digital twinborn system

By designing a digital twin system for low-orbit giant constellation systems, multi-dimensional dynamic construction of data and models and information network reconstruction are realized, the problem of insufficient operational efficiency and management level of low-orbit giant constellation systems is solved, and operational efficiency and overall performance are improved .

CN118939974AActive Publication Date: 2024-11-12NANJING UNIV OF AERONAUTICS & ASTRONAUTICS +1
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411121649.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-12
Estimated Expiration
2044-08-15

AI Technical Summary

Technical Problem

The existing technology is unable to effectively manage and optimize the complex network and topological dynamics of low-rail giant constellation systems, resulting in insufficient operational efficiency and management level and lack of an integrated system comprehensive management model.

Method used

Design a digital twin system for low-rail giant constellation systems, including data management module, model management module and information management module. Through data interconnection, model interoperability and information interoperability, it realizes multi-dimensional dynamic construction and information between physical systems and virtual systems. Network reconstruction, using multi-scale and multi-level feature information for local to global network reconstruction.

Benefits of technology

The operational efficiency and management level of the low-rail giant constellation system have been improved, and potential problems have been discovered through real-time monitoring, virtual testing and system design optimization, and overall performance and collaborative operation capabilities have been improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939974A_ABST
    Figure CN118939974A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of aerospace, and relates to a digital twin system for a low-orbit giant satellite base and a construction method thereof, and the system comprises a data management module which carries out the comprehensive management of real-time data, longitudinal data, transverse data, simulation data and fusion data of a giant satellite base system; the model management module is used for performing multi-dimensional dynamic construction and comprehensive description on geometric, physical, behavior, rule, constraint and other part characteristics of the complex physical system from different dimensions, different spatial scales and different time scales; the information management module performs local-to-global network reconstruction on the low-orbit giant satellite base system by using the multi-scale and multi-level feature information to restore a complete information network; a collaborative coupling mechanism of data interconnection, information intercommunication and model interoperation is adopted, efficient and real-time monitoring and management are achieved, and a solution is provided for digital, visual and intelligent supervision.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of low-orbit giant star constellation control technology, and in particular to a digital twin system for a low-orbit giant star constellation system and a construction method thereof. Background Art

[0002] In recent years, low-orbit space has become the focus of global aerospace development. Since low-orbit satellite constellations have obvious advantages in launch costs and global coverage, and have the characteristics of strong anti-destruction, low transmission delay and low power consumption link, the low-orbit giant star system characterized by low orbit and large scale is showing a booming development trend worldwide.

[0003] Unlike traditional single satellites, the Giant Star system has a large number of systems, dynamic and time-varying links and topologies, and complex network spatiotemporal behaviors. Traditional satellite management models alone cannot meet its operational needs. Therefore, an integrated system comprehensive management model is urgently needed to conduct comprehensive analysis and virtual verification of the problems faced by the Giant Star throughout its life cycle, so as to improve the basic capabilities and management level of the constellation system. Digital twin technology is the key to digital construction in the constellation field.

[0004] As a means of digital construction in the constellation field, digital twins closely connect satellite constellations and ground base stations in theoretical architecture, with digital expressions of physical objects, physical and virtual interfaces to achieve interactive connections, data integration, fusion, analysis, mining, and service-oriented packaging and application of models, data, functions, etc. Therefore, research on the combination of digital twins and constellation industries will effectively promote digitalization, networking, intelligence, and service-oriented transformation and upgrading, and play a key role in the engineering application of giant star constellations. Summary of the invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a digital twin system for a low-orbit giant star system and a construction method thereof.

[0006] In order to achieve the purpose of the present invention, the present invention will be implemented by adopting the following technical solutions.

[0007] A digital twin system for a low-orbit giant star system, comprising a data management module, a model management module and an information management module, wherein the data management module and the model management module interact with each other in a data interconnection and model interoperation manner; the data management module and the information management module interact with each other in a data interconnection and information intercommunication manner; the model management module and the information management module interact with each other in a model interoperation and information intercommunication manner; wherein:

[0008] The data management module uses data processing, data classification, data association and data integration to comprehensively manage the real-time data, vertical data, horizontal data, simulation data and fusion data of the low-orbit giant star system, so as to realize the data connection and interaction between the physical system and the virtual system;

[0009] The model management module uses geometric models, simulation models, mechanism models and data models to dynamically construct and comprehensively characterize the geometry, physics, behavior, rules and constraint characteristics of the LEO giant star system from different dimensions, different spatial scales and different time scales;

[0010] The information management module relies on twin data and twin models, and utilizes the multi-scale and multi-level feature information input from the data management module and the model management module to reconstruct the network of the digital twin system from local to global. Through global feature aggregation, local feature extraction, cross-scale refinement of spatial features, and feature adaptive fusion technology, the complete information network is restored to build a digital twin system.

[0011] As a preferred solution of the present invention, the real-time data is derived from measured data from on-orbit satellite sensors, providing key real-time insights for real-time monitoring and decision-making;

[0012] The longitudinal data covers the historical records, periodic data and performance evolution trends of the LEO Giant Star system, which are helpful for understanding the performance changes, problem trends and long-term patterns of the LEO Giant Star system evolution in different time periods;

[0013] The horizontal data focuses on the relationship between satellites or subsystems, including communication, mutual influence and performance differences, and uses the data records of the same or similar constellation of a certain low-orbit giant star constellation as a reference and comparison to optimize the coordinated operation of the entire low-orbit giant star constellation system and improve the overall performance;

[0014] The simulation data refers to the theoretical analysis data obtained by simulation of the digital twin system at the digital level, which supports virtual testing of the response of the digital twin system, helps to discover potential problems before actual operation, and improves system design and performance;

[0015] The fusion data integrates and reconstructs real-time data, longitudinal data, transverse data and simulation data to generate a more comprehensive and accurate data set that covers theory, simulation and reality.

[0016] As a preferred solution of the present invention, the geometric model is a comprehensive expert knowledge of Kepler's law, quaternions, orbital perturbation theory, spherical harmonics, and orbital roots to describe the positional relationship of each satellite in the low-orbit giant star system in space;

[0017] The simulation model simulates the motion trajectory of the satellite constellation in space using a dynamics and control model and an atmospheric disturbance model;

[0018] The mechanism model is based on expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, providing a physical explanation of the behavior of the entire LEO giant star system and a description of the rules of internal logic;

[0019] The data model uses statistical methods and machine learning techniques to capture the complex relationships between satellite orbit information from the data management module.

[0020] A method for constructing a digital twin system comprises the following steps:

[0021] S1. Build a data management module through data processing, data classification, data association and data integration to comprehensively manage the collected data of the physical system to achieve data connection and interaction between the physical system and the virtual system;

[0022] S2. Construct a model management module through geometric models, simulation models, mechanism models and data models to dynamically construct and comprehensively characterize the geometry, physics, behavior, rules and constraint characteristics of the physical system in multiple dimensions;

[0023] S3. Build an information management module through global feature aggregation, local feature extraction, cross-scale refinement of spatial features, and adaptive fusion of features. Reconstruct the network of the digital twin system from local to global from different dimensions, different spatial scales, and different time scales to restore the complete information network.

[0024] As a preferred embodiment of the present invention, step S1 comprises the following steps:

[0025] S11. Process the acquired physical system data by cleaning, filtering, denoising, and filling missing values ​​to ensure the quality and accuracy of the original data;

[0026] S12, decoding the data processed in step S11 and converting it into a format that the system can understand and process. At the same time, organizing and marking the data and setting a unified language format to ensure that data from different sources can be understood and used within a unified framework;

[0027] S13, for the data in a unified language format decoded, organized and marked in step S12, multiple groups of data points are connected through data association to explore the internal connections and mutual influences to ensure the compatibility and operability between different data sets;

[0028] S14, integrating the data associated in step S14 from different sources and with different association relationships into a unified data set through data integration to support system modeling and analysis;

[0029] S15. For the data integrated in step S14, ensure that the data can work together through data conversion and standardization to support model management and information management, ensure that the data can be effectively applied and analyzed, and ensure the interoperability of data between different models and modules.

[0030] As a preferred embodiment of the present invention, step S2 comprises the following steps:

[0031] S21. Integrate satellite orbit parameters, attitude, reference coordinate system, position and velocity vectors, coordinate system transformation matrix, attitude quaternion, relative position vector parameter information to form a geometric model to accurately describe the position relationship of each satellite in the LEO Giant Star system in space;

[0032] S22. Design satellite control systems based on linear and nonlinear control theories to ensure the stability of the attitude and position of each satellite within the constellation and establish simulation models;

[0033] S23. Based on the expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, construct orbital mechanics models, communication protocol models, and inter-satellite-to-earth communication models, and integrate the above models into a mechanism model to provide a physical explanation of the system behavior and a rule description of the internal logic;

[0034] S24. Use statistical methods and machine learning techniques to establish data models to accurately capture the complex relationships between satellite orbit information from the data management module;

[0035] S25. Comprehensively manage the functional characteristics of different models to build an accurate and complete digital twin model of the GSTAR system.

[0036] As a preferred embodiment of the present invention, step S3 comprises the following steps:

[0037] S31. Use the Transformer model as a tool for feature extraction and representation learning to extract coarse-grained features of data information and model information at the token level. Aggregate all token information at the output of the Encoder by weighted average and summary. Further encode the input token-level position information into a global representation through position embedding to form a global view.

[0038] S32, introduce the local window attention branch, learn the short-range interaction within the local window, extract the fine-grained feature representation of data information and model information, and capture and extract token-level local features;

[0039] S33, introduce the multi-scale attention mechanism to explore the integration of two features, refine the spatial features in a cross-scale manner through Multi-Head Attention, and capture the feature relationship at different scales;

[0040] S34. Use the Reinforcement Learning algorithm to optimize the adaptive fusion process of features, define the state, action and reward function, where the state represents the feature representation of the current data, the action represents the way of feature fusion, and the reward function evaluates the quality of the fusion result; based on the multi-scale attention fusion (MAF) scheme, enhance the interactive connection between global-local token pairs, train multiple token-level agents at different stages together in a unified framework, and deploy both local and global features using linear operations to ensure the consistency of feature distribution;

[0041] S35. According to the learning process and data feedback of the model, the optimal feature fusion strategy is determined through the reinforcement learning algorithm, and the weights and fusion strategies of the global features and local features are dynamically adjusted so that the feature fusion process can maximize the performance of the information management module;

[0042] S36. Use cloud computing platform and distributed training technology to provide an efficient, flexible and scalable computing environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the structure of the digital twin system of the present invention;

[0044] Figure 2 This is a schematic diagram of the structure of the data management module of the present invention;

[0045] Figure 3 This is a flow chart of the model management module of the present invention;

[0046] Figure 4 This is a flow chart of the information management module of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0048] As an embodiment of the present invention, Figure 1As shown in the figure, the connection between the physical system and the virtual system emphasizes data interconnection, information exchange, and model interoperability. Therefore, a digital twin system for the LEO system is divided into three modules: data management, model management, and information management. As the driver of the digital twin system, data has the characteristics of massiveness, strong correlation, and high coupling. Due to the limited transit time of the satellite, data transmission has a certain timeliness. The data of the LEO also presents fragmented discrete characteristics. Therefore, the data management module is designed to comprehensively manage the real-time data, vertical data, horizontal data, simulation data, and fusion data of the physical system, including data processing, data classification, data association, and data integration, to achieve data connection and interaction between the physical system and the virtual system. As the carrier of the digital twin system, the model is used to simulate and simulate the operation process of the physical world. Therefore, the model management module is designed to dynamically construct and comprehensively characterize the various features of the complex physical system, such as geometry, physics, behavior, rules, and constraints, from different dimensions, different spatial scales, and different time scales, including geometric models, simulation models, mechanism models, and data models. Relying on twin data and twin models, the information management module uses these multi-scale and multi-level feature information to reconstruct the network of the digital twin system from local to global. Through global feature aggregation, local feature extraction, cross-scale refinement of spatial features and adaptive feature fusion, the complete information network is restored to build a digital twin system.

[0049] As an embodiment of the present invention, Figure 2 As shown, the real-time data is derived from the measured data of on-orbit satellite sensors, providing key real-time insights for real-time monitoring and decision-making;

[0050] The longitudinal data covers the historical records, periodic data and performance evolution trends of the LEO Giant Star system, which are helpful for understanding the performance changes, problem trends and long-term patterns of the LEO Giant Star system evolution in different time periods;

[0051] The horizontal data focuses on the relationship between satellites or subsystems, including communication, mutual influence and performance differences, and uses the data records of the same or similar constellation of a certain low-orbit giant star constellation as a reference and comparison to optimize the coordinated operation of the entire low-orbit giant star constellation system and improve the overall performance;

[0052] The simulation data refers to the theoretical analysis data obtained by simulation of the digital twin system at the digital level, which supports virtual testing of the response of the digital twin system, helps to discover potential problems before actual operation, and improves system design and performance;

[0053] The fusion data integrates and reconstructs real-time data, longitudinal data, transverse data and simulation data to generate a more comprehensive and accurate data set that covers theory, simulation and reality.

[0054] As an embodiment of the present invention, Figure 3 As shown, the geometric model is a comprehensive expert knowledge of Kepler's law, quaternions, orbital perturbation theory, spherical harmonics, and orbital roots to describe the positional relationship of each satellite in the low-orbit giant star system in space;

[0055] The simulation model simulates the motion trajectory of the satellite constellation in space using a dynamics and control model and an atmospheric disturbance model;

[0056] The mechanism model is based on expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, providing a physical explanation of the behavior of the entire LEO giant star system and a description of the rules of internal logic;

[0057] The data model uses statistical methods and machine learning techniques to capture the complex relationships between satellite orbit information from the data management module.

[0058] As an embodiment of the present invention, Figure 1 As shown, a method for constructing a digital twin system includes the following steps:

[0059] S1. Build a data management module through data processing, data classification, data association and data integration to comprehensively manage the collected data of the physical system to achieve data connection and interaction between the physical system and the virtual system;

[0060] S2. Construct a model management module through geometric models, simulation models, mechanism models and data models to dynamically construct and comprehensively characterize the geometry, physics, behavior, rules and constraint characteristics of the physical system in multiple dimensions;

[0061] S3. Build an information management module through global feature aggregation, local feature extraction, cross-scale refinement of spatial features, and adaptive fusion of features. Reconstruct the network of the digital twin system from local to global from different dimensions, different spatial scales, and different time scales to restore the complete information network.

[0062] As an embodiment of the present invention, Figure 2 As shown, step S1 includes the following steps:

[0063] S11. Process the acquired physical system data by cleaning, filtering, denoising, and filling missing values ​​to ensure the quality and accuracy of the original data;

[0064] S12, decoding the data processed in step S11 and converting it into a format that the system can understand and process. At the same time, organizing and marking the data and setting a unified language format to ensure that data from different sources can be understood and used within a unified framework;

[0065] S13, for the data in a unified language format decoded, organized and marked in step S12, multiple groups of data points are connected through data association to explore the internal connections and mutual influences to ensure the compatibility and operability between different data sets;

[0066] S14, integrating the data associated in step S14 from different sources and with different association relationships into a unified data set through data integration to support system modeling and analysis;

[0067] S15. For the data integrated in step S14, ensure that the data can work together through data conversion and standardization to support model management and information management, ensure that the data can be effectively applied and analyzed, and ensure the interoperability of data between different models and modules.

[0068] As an embodiment of the present invention, Figure 3 As shown, step S2 includes the following steps:

[0069] S21. Integrate satellite orbit parameters, attitude, reference coordinate system, position and velocity vectors, coordinate system transformation matrix, attitude quaternion, relative position vector parameter information to form a geometric model to accurately describe the position relationship of each satellite in the LEO Giant Star system in space;

[0070] S22. Design satellite control systems based on linear and nonlinear control theories to ensure the stability of the attitude and position of each satellite within the constellation and establish simulation models;

[0071] S23. Based on the expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, construct orbital mechanics models, communication protocol models, and inter-satellite-to-earth communication models, and integrate the above models into a mechanism model to provide a physical explanation of the system behavior and a rule description of the internal logic;

[0072] S24. Use statistical methods and machine learning techniques to establish data models to accurately capture the complex relationships between satellite orbit information from the data management module;

[0073] S25. Comprehensively manage the functional characteristics of different models to build an accurate and complete digital twin model of the GSTAR system.

[0074] As an embodiment of the present invention, Figure 4 As shown, step S3 includes the following steps:

[0075] S31. Use the Transformer model as a tool for feature extraction and representation learning to extract coarse-grained features of data information and model information at the token level. Aggregate all token information at the output of the Encoder by weighted average and summary. Further encode the input token-level position information into a global representation through position embedding to form a global view.

[0076] S32, introduce the local window attention branch, learn the short-range interaction within the local window, extract the fine-grained feature representation of data information and model information, and capture and extract token-level local features;

[0077] S33, introduce the multi-scale attention mechanism to explore the integration of two features, refine the spatial features in a cross-scale manner through Multi-Head Attention, and capture the feature relationship at different scales;

[0078] S34. Use the Reinforcement Learning algorithm to optimize the adaptive fusion process of features, define the state, action and reward function, where the state represents the feature representation of the current data, the action represents the way of feature fusion, and the reward function evaluates the quality of the fusion result; based on the multi-scale attention fusion (MAF) scheme, enhance the interactive connection between global-local token pairs, train multiple token-level agents at different stages together in a unified framework, and deploy both local and global features using linear operations to ensure the consistency of feature distribution;

[0079] S35. According to the learning process and data feedback of the model, the optimal feature fusion strategy is determined through the reinforcement learning algorithm, and the weights and fusion strategies of the global features and local features are dynamically adjusted so that the feature fusion process can maximize the performance of the information management module;

[0080] S36. Use cloud computing platform and distributed training technology to provide an efficient, flexible and scalable computing environment.

[0081] The technical solution of the present invention is described in detail above in conjunction with the embodiments / drawings, but the present invention is not limited to the above technical solution. For ordinary technicians in this technical field, after knowing the contents recorded in the present invention, they can make several equivalent changes and substitutions without departing from the principle of the present invention. These equivalent changes and substitutions should also be regarded as belonging to the protection scope of the present invention.

Claims

1. A digital twin system for a low-orbit giant star system, characterized in that: The digital twin system includes a data management module, a model management module and an information management module. The data management module interacts with the model management module by means of data interconnection and model interoperation; the data management module interacts with the information management module by means of data interconnection and information intercommunication; the model management module interacts with the information management module by means of model interoperation and information intercommunication; wherein: The data management module uses data processing, data classification, data association and data integration to comprehensively manage the real-time data, vertical data, horizontal data, simulation data and fusion data of the low-orbit giant star system, so as to realize the data connection and interaction between the physical system and the virtual system; The model management module uses geometric models, simulation models, mechanism models and data models to dynamically construct and comprehensively characterize the geometry, physics, behavior, rules and constraint characteristics of the LEO giant star system from different dimensions, different spatial scales and different time scales; The information management module relies on twin data and twin models, and utilizes the multi-scale and multi-level feature information input from the data management module and the model management module to reconstruct the network of the digital twin system from local to global. Through global feature aggregation, local feature extraction, cross-scale refinement of spatial features, and feature adaptive fusion technology, the complete information network is restored to build a digital twin system.

2. The digital twin system for the LEO giant star system according to claim 1, characterized in that: The real-time data is derived from the measured data of on-orbit satellite sensors, providing key real-time insights for real-time monitoring and decision-making; The longitudinal data covers the historical records, periodic data and performance evolution trends of the LEO Giant Star system, which are helpful for understanding the performance changes, problem trends and long-term patterns of the LEO Giant Star system evolution in different time periods; The horizontal data focuses on the relationship between satellites or subsystems, including communication, mutual influence and performance differences, and uses the data records of the same or similar constellation of a certain low-orbit giant star constellation as a reference and comparison to optimize the coordinated operation of the entire low-orbit giant star constellation system and improve the overall performance; The simulation data refers to the theoretical analysis data obtained by simulation of the digital twin system at the digital level, which supports virtual testing of the response of the digital twin system, helps to discover potential problems before actual operation, and improves system design and performance; The fusion data integrates and reconstructs real-time data, longitudinal data, transverse data and simulation data to generate a more comprehensive and accurate data set that covers theory, simulation and reality.

3. The digital twin system for the LEO giant star system according to claim 1, characterized in that: The geometric model is a comprehensive description of the positional relationship of each satellite in the LEO giant star system in space by integrating Kepler's laws, quaternions, orbital perturbation theory, spherical harmonics, and orbital element expert knowledge; The simulation model simulates the motion trajectory of the satellite constellation in space using a dynamics and control model and an atmospheric disturbance model; The mechanism model is based on expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, providing a physical explanation of the behavior of the entire LEO giant star system and a description of the rules of internal logic; The data model uses statistical methods and machine learning techniques to capture the complex relationships between satellite orbit information from the data management module.

4. A method for constructing a digital twin system, characterized in that: The following steps are involved: S1. Build a data management module through data processing, data classification, data association and data integration to comprehensively manage the collected data of the physical system to achieve data connection and interaction between the physical system and the virtual system; S2. Construct a model management module through geometric models, simulation models, mechanism models and data models to dynamically construct and comprehensively characterize the geometry, physics, behavior, rules and constraint characteristics of the physical system in multiple dimensions; S3. Build an information management module through global feature aggregation, local feature extraction, cross-scale refinement of spatial features, and adaptive fusion of features. Reconstruct the network of the digital twin system from local to global from different dimensions, different spatial scales, and different time scales to restore the complete information network.

5. The method for constructing a digital twin system according to claim 4, characterized in that: The step S1 comprises the following steps: S11. Process the acquired physical system data by cleaning, filtering, denoising, and filling missing values ​​to ensure the quality and accuracy of the original data; S12, decoding the data processed in step S11 and converting it into a format that the system can understand and process. At the same time, organizing and marking the data and setting a unified language format to ensure that data from different sources can be understood and used within a unified framework; S13, for the data in a unified language format decoded, organized and marked in step S12, multiple groups of data points are connected through data association to explore the internal connections and mutual influences to ensure the compatibility and operability between different data sets; S14, integrating the data associated in step S14 from different sources and with different association relationships into a unified data set through data integration to support system modeling and analysis; S15. For the data integrated in step S14, ensure that the data can work together through data conversion and standardization to support model management and information management, ensure that the data can be effectively applied and analyzed, and ensure the interoperability of data between different models and modules.

6. The method for constructing a digital twin system according to claim 4, characterized in that: The step S2 comprises the following steps: S21. Integrate satellite orbit parameters, attitude, reference coordinate system, position and velocity vectors, coordinate system transformation matrix, attitude quaternion, relative position vector parameter information to form a geometric model to accurately describe the position relationship of each satellite in the LEO Giant Star system in space; S22. Design satellite control systems based on linear and nonlinear control theories to ensure the stability of the attitude and position of each satellite within the constellation and establish simulation models; S23. Based on the expert knowledge of orbital mechanics, communication protocols, laser communication theory, electromagnetic wave propagation theory, and system engineering theory, construct orbital mechanics models, communication protocol models, and inter-satellite-to-earth communication models, and integrate the above models into a mechanism model to provide a physical explanation of the system behavior and a rule description of the internal logic; S24. Use statistical methods and machine learning techniques to establish data models to accurately capture the complex relationships between satellite orbit information from the data management module; S25. Comprehensively manage the functional characteristics of different models to build an accurate and complete digital twin model of the GSTAR system.

7. The method for constructing a digital twin system according to claim 4, characterized in that: Step S3 includes the following steps: S31. Use the Transformer model as a tool for feature extraction and representation learning to extract coarse-grained features of data information and model information at the token level. Aggregate all token information at the output of the Encoder by weighted average and summary. Further encode the input token-level position information into a global representation through position embedding to form a global view. S32, introduce the local window attention branch, learn the short-range interaction within the local window, extract the fine-grained feature representation of data information and model information, and capture and extract token-level local features; S33, introduce the multi-scale attention mechanism to explore the integration of two features, refine the spatial features in a cross-scale manner through Multi-Head Attention, and capture the feature relationship at different scales; S34. Use the Reinforcement Learning algorithm to optimize the adaptive fusion process of features, define the state, action and reward function, where the state represents the feature representation of the current data, the action represents the way of feature fusion, and the reward function evaluates the quality of the fusion result; based on the multi-scale attention fusion (MAF) scheme, enhance the interactive connection between global-local token pairs, train multiple token-level agents at different stages together in a unified framework, and deploy both local and global features using linear operations to ensure the consistency of feature distribution; S35. According to the learning process and data feedback of the model, the optimal feature fusion strategy is determined through the reinforcement learning algorithm, and the weights and fusion strategies of the global features and local features are dynamically adjusted so that the feature fusion process can maximize the performance of the information management module; S36. Use cloud computing platform and distributed training technology to provide an efficient, flexible and scalable computing environment.

Citation Information

Patent Citations

  • Low-orbit constellation satellite simulation modeling method

    CN115392057A

  • Digital twin micro-nano satellite system

    CN115828593A

  • Space target intelligent detection system and method based on digital twin technology

    CN117911886A

  • Data and knowledge collaborative twin modeling method and system for highway tunnel fire environment

    CN117994416A

  • Vulnerability-driven cyberattack protection system and method for industrial assets

    US20230058974A1