Model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency

Abstract

The invention relates to a model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency. The method comprises the following steps: acquiring quaternary knowledge driven architecture data of a spacecraft; performing knowledge extension on the quaternary knowledge-driven architecture data to obtain seven-element knowledge extension architecture data; according to the seven-element knowledge extension architecture data, performing mapping analysis on the operation behavior of the spacecraft to obtain spacecraft behavior mapping analysis data; according to the spacecraft behavior mapping analysis data, carrying out correction calculation on the behavior difference of the spacecraft to obtain spacecraft behavior adjustment data; and according to the spacecraft behavior adjustment data, performing data correction on the quaternary knowledge-driven architecture data to obtain corrected knowledge-driven architecture data. By adopting the method, data missing, drifting or delay possibly existing in part of key sensors can be avoided, and the accuracy of the model and the consistency of virtual-real mapping are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of automatic control, in particular to a model adaptive correction method based on virtual-real mapping consistency of spacecraft multi-domain information. BACKGROUND

[0002] In the prior art, a spacecraft realizes dynamic perception, prediction evaluation and cooperative control of a whole-system state of the spacecraft through multi-source sensor arrangement, cross-domain data standardization, time-space synchronization processing, feature extraction and correlation analysis, which is a basic support capability for constructing a high-precision digital twin model and carrying out system-level intelligent decision-making. However, different data of the traditional spacecraft multi-domain information technology have heterogeneity and time-space inconsistency, data sources of different physical domains are various, sampling frequencies are different, and precision differences are large, which easily leads to deviation or distortion in the information fusion process, and further affects the accuracy of the model and the virtual-real mapping consistency. SUMMARY

[0003] In order to overcome the defects of the prior art, the application provides a model adaptive correction method based on virtual-real mapping consistency of spacecraft multi-domain information. The application provides a model adaptive correction method, device and computer equipment based on virtual-real mapping consistency of spacecraft multi-domain information, which can avoid data missing, drift or delay of part of key sensors, and improve the accuracy of the model and the virtual-real mapping consistency.

[0004] The technical scheme adopted by the application to solve the technical problems is:

[0005] In a first aspect, the application provides a model adaptive correction method based on virtual-real mapping consistency of spacecraft multi-domain information, comprising:

[0006] obtaining four-element knowledge-driven architecture data of a spacecraft;

[0007] performing knowledge extension on the four-element knowledge-driven architecture data to obtain seven-element knowledge extension architecture data;

[0008] performing mapping analysis on the running behavior of the spacecraft according to the seven-element knowledge extension architecture data to obtain spacecraft behavior mapping analysis data;

[0009] performing correction calculation on the behavior difference of the spacecraft according to the spacecraft behavior mapping analysis data to obtain spacecraft behavior adjustment data;

[0010] performing data correction on the four-element knowledge-driven architecture data according to the spacecraft behavior adjustment data to obtain corrected knowledge-driven architecture data.

[0011] In a second aspect, the application further provides a model adaptive correction device based on spacecraft multi-domain information virtual-real mapping consistency, comprising:

[0012] An architecture data acquisition module is configured to acquire four-element knowledge-driven architecture data of a spacecraft.

[0013] An architecture data dimension upgrading module is configured to perform knowledge expansion on the four-element knowledge-driven architecture data to obtain seven-element knowledge expansion architecture data.

[0014] An architecture data analysis module is configured to perform mapping analysis on the operation behavior of the spacecraft according to the seven-element knowledge expansion architecture data to obtain spacecraft behavior mapping analysis data.

[0015] A correction data calculation module is configured to perform correction calculation on the behavior difference of the spacecraft according to the spacecraft behavior mapping analysis data to obtain spacecraft behavior adjustment data.

[0016] An architecture data correction module is configured to perform data correction on the four-element knowledge-driven architecture data according to the spacecraft behavior adjustment data to obtain corrected knowledge-driven architecture data.

[0017] In a third aspect, the application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements any step of the model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency when executing the computer program.

[0018] The model adaptive correction method, device and computer device based on spacecraft multi-domain information virtual-real mapping consistency described above are based on four-element knowledge-driven architecture data of a spacecraft, construct and expand to form seven-element knowledge expansion architecture data with higher semantic dimensions and context association capabilities, realize multi-angle and full-link modeling and analysis of the operation behavior of the spacecraft, not only enhance the representation capability of complex task logic relations, but also provide more abundant knowledge support for behavior mapping analysis. Then, the mapping analysis results of the seven-element knowledge expansion architecture data are subjected to difference identification and quantitative correction to generate targeted spacecraft behavior adjustment data, and then the original knowledge architecture is subjected to high-precision closed-loop correction and update to form an adaptive cycle from knowledge generation, behavior feedback to knowledge evolution. The fusion processing capability and knowledge iteration capability of the spacecraft system for multi-source dynamic information are significantly improved, the spacecraft has stronger intelligent perception, behavior reasoning and strategy optimization capability when facing uncertain task environment, data loss, drift or delay of some key sensors are avoided, and the accuracy of the model and the consistency of virtual-real mapping are improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1An application environment diagram of a model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency in an embodiment;

[0020] Figure 2 A flowchart of a model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency in an embodiment;

[0021] Figure 3 A flowchart of a spacecraft behavior mapping analysis data obtaining method in an embodiment;

[0022] Figure 4 A flowchart of a spacecraft behavior mapping analysis data obtaining method in another embodiment;

[0023] Figure 5 A flowchart of a spacecraft behavior adjustment data obtaining method in an embodiment;

[0024] Figure 6 A flowchart of a path difference analysis data obtaining method in an embodiment;

[0025] Figure 7 A flowchart of a spacecraft behavior adjustment data obtaining method in another embodiment;

[0026] Figure 8 A flowchart of a seven-element knowledge extension architecture data obtaining method in an embodiment;

[0027] Figure 9 A structural block diagram of a model adaptive correction device based on spacecraft multi-domain information virtual-real mapping consistency in an embodiment;

[0028] Figure 10 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0029] The application will be further described below in conjunction with the drawings and embodiments.

[0030] An application environment in which a model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency provided by an embodiment of the present application can be applied is shown in FIG. 1. Figure 1 The terminal 102 communicates with the server 104 through a network. A data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0031] In an exemplary embodiment, as shown in FIG. 2, the model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency in an embodiment of the present application includes the following steps. Figure 2As shown, a model adaptive correction method based on spacecraft multi-domain information virtual-real mapping consistency is provided, and the method is applied to Figure 1 The server in the method is taken as an example for description, including the following steps 202 to 210.

[0032] Step 202, acquiring four-knowledge-driven architecture data of the spacecraft.

[0033] Among them, the four-knowledge-driven architecture data can be a four-core module containing structure object (OB), structure knowledge (KB), data resource (DB) and process model (DP). Among them, the structure object is used to describe the physical composition and system topology of the spacecraft; the structure knowledge expresses various design rules, control constraints and operation logic; the data resource integrates historical, simulation and real-time observation data; and the process model defines the stage division and switching conditions of the spacecraft in the task execution process.

[0034] Specifically, through the digital twin modeling platform, the structure model, control system, functional components and task process of the spacecraft are systematically analyzed, and the four-knowledge-driven architecture data composed of four core knowledge units is constructed: first, the structure object model (OB) is used to describe the component composition, system topology, interface relationship and physical characteristics of the spacecraft, which is the basis for aligning the virtual model with the real object; second, the structure knowledge base (KB) contains expert knowledge and engineering specifications such as design rules, control constraints, fault logic and operation boundaries, which is the logical basis for control and judgment; third, the data resource base (DB) integrates telemetry data, simulation output, ground test records and flight history data to form a multi-source data base for state recognition and trend evaluation; fourth, the process model base (DP) defines the switching conditions between each task stage, system running state and stage, which is used to drive the behavior evolution path and scheduling logic. The four-knowledge-driven architecture data as the core knowledge skeleton of the digital twin supports modeling, simulation, prediction and decision-making of the spacecraft running state.

[0035] Step 204, knowledge extension is performed on the four-knowledge-driven architecture data to obtain seven-knowledge-extended architecture data.

[0036] Among them, the knowledge extension can be the introduction of new knowledge dimensions and modeling mechanisms on the basis of the four-knowledge-driven architecture to enhance the cognitive ability of the system to the spacecraft behavior, control intention and state evolution.

[0037] Among them, the seven-knowledge-extended architecture data can be a new generation of multi-dimensional knowledge system generated by knowledge extension on the basis of the four-knowledge-extended architecture data, containing structure object (OB), structure knowledge (KB_S), behavior knowledge (KB_B), data resource (DB), process model (DP), structure-behavior bridging engine (BE) and time dynamic evolution model (TD).

[0038] Specifically, based on the information in the structure object (OB), the structure knowledge (KB), the data resource (DB), and the process model (DP), the behavior elements related to the spacecraft task execution are extracted, the behavior knowledge identification is completed, and the preliminary four-element behavior knowledge identification feature data are obtained; then, according to the functional attributes, control intentions, state response rules and other contents of the behavior features, the behavior features are classified and grouped in terms of semantics and functions, and the classification knowledge identification feature data with category labels are generated; in combination with the task context and the behavior evolution law, the behavior enhancement is performed on each type of behavior feature data, such as completing the boundary conditions, reasoning the implicit action chain or introducing the typical state evolution mode, so that the enhanced knowledge identification feature data with better predictability and interpretability are obtained; finally, based on the enhanced features, the behavior semantic model, the structure-behavior bridging relationship and the time dynamic evolution modeling are added, the behavior expansion of each type of behavior is completed, and finally the seven-element knowledge expansion architecture data containing the structure knowledge, the behavior knowledge, the semantic mapping and the evolution ability are constructed.

[0039] That is, on the basis of the original structure object, the structure knowledge, the data layer and the process model, three key knowledge dimensions are added, which are: first, the behavior knowledge base (KB_B), which is used to represent the behavior state nodes, the control intentions, the action sequences and the behavior transition rules in each task phase of the spacecraft, so that the system can perform semantic modeling and path planning on the operation logic of the spacecraft; second, the structure-behavior bridging engine (BE), which realizes cross-domain knowledge reasoning by establishing the semantic coupling relationship between the structure parameters and the behavior states, and can infer the behavior deviation from the structure anomaly or infer the potential structure problem from the behavior deviation; third, the time dynamic model (TD), which models the evolution of the behavior state in the time dimension, and realizes the inferability and deviation trend judgment of the future behavior of the spacecraft through the continuous state transition function or evolution trajectory prediction.

[0040] Step 206, according to the seven-element knowledge expansion architecture data, the running behavior of the spacecraft is mapped and analyzed, and the spacecraft behavior mapping analysis data are obtained.

[0041] Among them, the running behavior can be the functional action sequence and state change trajectory exhibited by the system composed modules of the spacecraft in the task execution process, which reflects the state transition, control response and function execution process experienced by the spacecraft from starting to completing the task.

[0042] Among them, the mapping analysis can be the process of aligning and comparing the actual running behavior of the spacecraft with the expected behavior path defined in the seven-element knowledge architecture, which specifically includes state matching, path comparison, semantic deviation identification and confidence score calculation.

[0043] The spacecraft behavior mapping analysis data can be a structured result of the mapping analysis, and contains state node offset scores, path similarity, confidence domain consistency scores, state fuzziness indicators, and other dimensional information.

[0044] Specifically, the predefined task behavior path and state transition logic in the behavior knowledge base (KB_B) are called, and the time sequence expression of the expected behavior trajectory is derived in combination with the time dynamic model (TD); then the multi-source observation data in the database (DB) are accessed in real time, and the actual data are mapped to the corresponding behavior state dimension through the structure-behavior bridge engine (BE) to form the current actual behavior path. On this basis, the offset degree between the actual behavior path and the expected trajectory is compared by using path-level semantic similarity calculation, state node matching, confidence domain consistency scoring, etc. In particular, state importance weight and fuzziness score are introduced to weight and adjust the mapping credibility and fuzziness risk of key behavior states, thereby enhancing the sensitivity and explanatory power of behavior deviation. Finally, the spacecraft behavior mapping analysis data containing behavior path similarity, state node offset tensor, trust source consistency score, and fuzziness risk indicator are output.

[0045] Step 208, according to the spacecraft behavior mapping analysis data, the behavior difference of the spacecraft is calculated to obtain the spacecraft behavior adjustment data.

[0046] The behavior difference can be the deviation of the current actual operation behavior of the spacecraft from the expected behavior modeled in the seven-element knowledge architecture in terms of state sequence, action logic, or time process, including behavior not achieved, state lag, path misplacement, or execution logic error.

[0047] The spacecraft behavior adjustment data can be the control parameter adjustment result generated by joint reasoning of the behavior offset tensor, confidence score, and fuzziness indicator after identifying the behavior difference. The data includes state correction amount, control variable adjustment value, trust domain weight update strategy, and behavior priority adjustment suggestion.

[0048] Specifically, the key state nodes with significant deviation in the current behavior path are identified from the spacecraft behavior mapping analysis data, and the state deviation score, confidence consistency score and fuzziness index corresponding to the key state nodes are extracted to construct a behavior difference tensor; combined with the behavior intention information and control target expectation of the state in the behavior knowledge base (KB_B), a control intention semantic vector modeling is introduced to generate a semantic-driven behavior correction candidate path or control variable deviation direction; the time dynamic model (TD) is used to predict and evaluate the behavior evolution trend to determine whether the deviation has persistent or phase abnormal characteristics; and on this basis, the control parameter mapping engine is called to input the behavior difference tensor, semantic vector and structure-behavior bridge relationship to generate structured control parameter correction data, and convert it into spacecraft behavior adjustment data, which contains state-level correction amount, control variable fine-tuning value, trust domain weight update strategy and correction priority suggestion, etc., forming an integrated correction result with multi-source input, behavior understanding, dynamic reasoning and executable parameter output capability.

[0049] In step 210, the four-element knowledge-driven architecture data is corrected according to the spacecraft behavior adjustment data to obtain corrected knowledge-driven architecture data.

[0050] The corrected knowledge-driven architecture data can be a knowledge structure generated by dynamically updating the original knowledge-driven architecture (such as KB, DB, DP, etc.) under the action of the spacecraft behavior adjustment data, reflecting the adaptive learning results of the system to the deviation, anomaly or environmental change, which may include parameter range reset, process node replacement, data source trust weight redistribution, etc.

[0051] Specifically, in the process of correcting the four-element knowledge-driven architecture data according to the spacecraft behavior adjustment data, first, in the structure knowledge base (KB), the constraint boundary, logical rule or control condition is updated at the parameter level according to the control parameter adjustment suggestion to ensure that the structure model is consistent with the actual running state; secondly, in the data layer (DB), the confidence score of each data source is updated, and the low consistency or abnormal signal source is marked for subsequent weight compression or trust scheduling; meanwhile, in the process model (DP), the state switching condition or path priority in the process is dynamically adjusted according to the task state deviation involved in the behavior correction, for example, the conversion threshold of a certain type of behavior state is changed from "based on temperature" to "based on attitude stability"; in addition, for serious deviation that cannot be reconciled, part of the process nodes or action sequences can be directly replaced based on the behavior intention model, so as to realize the structural correction of the execution logic. Finally, the corrected knowledge-driven architecture data containing updated structure knowledge, data weight and process logic is output, that is, the updated DTB architecture, so that the digital twin model has higher consistency between virtual and real, dynamic adaptability and multi-source fault tolerance capability in subsequent operation.

[0052] The aforementioned adaptive correction method for a model based on the consistency of virtual-real mapping of multi-domain information in spacecraft constructs and expands a seven-element knowledge extension architecture based on the spacecraft's four-element knowledge-driven architecture data, which has higher semantic dimensions and contextual association capabilities. This enables multi-angle, full-link modeling and analysis of spacecraft operational behavior, enhancing the representation of complex mission logical relationships and providing richer knowledge support for behavior mapping analysis. Then, the mapping analysis results of the seven-element knowledge extension architecture data are subjected to difference identification and quantitative correction to generate targeted spacecraft behavior adjustment data. This, in turn, performs high-precision closed-loop correction and update of the original knowledge architecture, forming an adaptive loop from knowledge generation, behavior feedback to knowledge evolution. This significantly improves the spacecraft system's ability to fuse and process multi-source dynamic information and its knowledge iteration capabilities, enabling the spacecraft to possess stronger intelligent perception, behavior reasoning, and strategy optimization capabilities when facing uncertain mission environments. It also avoids potential data loss, drift, or delays in some key sensors, improving model accuracy and the consistency of virtual-real mapping.

[0053] In one exemplary embodiment, such as Figure 3 As shown, based on the seven-element knowledge extension architecture data, the operational behavior of the spacecraft is mapped and analyzed to obtain spacecraft behavior mapping analysis data, including steps 302 to 304.

[0054] in:

[0055] Step 302: Based on the seven-element knowledge extension architecture data, analyze the spacecraft's behavioral intent to obtain spacecraft behavioral intent analysis data.

[0056] Behavioral intent refers to the underlying goal orientation and control objectives of a spacecraft's operational behavior during a specific mission phase. It is not merely a description of the current action, but also expresses "what the system is trying to achieve." For example, in "attitude adjustment" behavior, the intent might be "aligning with the communication direction" or "entering an observation attitude." Behavioral intent is typically determined by a combination of mission planning, process logic, control strategies, and target parameters.

[0057] The spacecraft behavior intent analysis data can be a set of structured information generated after analyzing the matching relationship between the current operational state sequence and the mission objective. It is used to express the understanding results of the spacecraft's behavior intent at different time points or state nodes. This data not only includes the intent label corresponding to each behavior state (such as "stable orbit", "observation completed", "ready to return", etc.), but also includes key fields such as target completion score, intent offset direction, mission path category, and behavior-target matching confidence.

[0058] Specifically, the system invokes the pre-defined task behavior model in the behavior knowledge base (KB_B) to perform semantic matching and path comparison on the current spacecraft's operational state sequence, identifying the control target and mission intent corresponding to its behavioral evolution trajectory. Simultaneously, it combines a time dynamic model (TD) to infer the evolution trend of the current state over time, determining whether it is steadily progressing towards the predetermined goal or exhibiting abnormal behavioral signs such as deviation, lag, or jumps. Using the structure-behavior bridging engine (BE), the system derives the causal relationship between the current state and the target intent based on key control parameters (such as thrust, attitude, and heat flux density), further enhancing the physical rationality and contextual relevance of intent understanding, generating spacecraft behavior intent analysis data.

[0059] Step 304: Perform trust domain weighted consistency mapping on the spacecraft behavior intent analysis data to obtain spacecraft behavior mapping analysis data.

[0060] Among them, the Trust Domain Weighted Consistency Mapping can be a technical mechanism for verifying the accuracy of behavioral intentions by integrating data from multiple sources. Its core is that, for a certain behavioral intention, the system extracts the state performance from multiple trust domains (such as telemetry, simulation, expert rules, historical behavior, etc.) and scores it through similarity calculation and fuzzy analysis; then, based on the preset weights or dynamic credibility of each trust domain, the scores are weighted and integrated to evaluate the consistency of the behavioral intention under the support of multi-dimensional data.

[0061] Specifically, for the intent expression of each behavioral state, data source inputs from multiple trust domains are collected, including telemetry data, simulation model outputs, historical mission behavior data, and expert rule bases. Then, based on the credibility weight configuration of each trust domain, similarity calculations (e.g., vector distance, semantic matching, state structure comparison) and ambiguity assessments (e.g., information entropy, volatility, deviation degree) are performed on the performance of the corresponding behavioral states in each data source, forming a state-level multi-trust scoring matrix. A weighted fusion mechanism is employed, combining the consistency scores and ambiguity factors of each trust domain into the behavioral intent analysis results, generating a consistency mapping tensor containing semantic matching scores, confidence fusion scores, and ambiguity adjustment factors. By integrating the weighted consistency scores of all states, structured spacecraft behavior mapping analysis data is output, which comprehensively reflects the consistency, deviation trends, and credibility and executability of actual behavioral intents across each trust domain.

[0062] In this embodiment, spacecraft behavior intent analysis is conducted based on a seven-element knowledge extension architecture. Behavior mapping analysis data is generated by combining trust domain weighted consistency mapping. This approach not only integrates structural knowledge, behavioral semantics, temporal dynamics, and multi-source data support, but also enables the system to "understand what the spacecraft wants to do" through behavior intent recognition. Furthermore, a trust domain scoring mechanism dynamically balances the credibility of various data sources (such as telemetry, simulation, and expert rules), effectively suppressing abnormal data interference and improving the accuracy and reliability of behavior deviation identification. The final output behavior mapping analysis data provides a highly reliable and interpretable cognitive foundation for subsequent behavior correction and control adjustments, effectively enhancing the intelligent perception, adaptive response, and virtual-real consistency management capabilities of the spacecraft digital twin system.

[0063] In one exemplary embodiment, such as Figure 4 As shown, trust domain weighted consistency mapping is performed on the spacecraft behavior intent analysis data to obtain spacecraft behavior mapping analysis data, including steps 402 to 408. Wherein:

[0064] Step 402: According to the spacecraft's operational function information, classify the spacecraft's behavioral intent analysis data to obtain the classification data of each spacecraft's behavioral intent.

[0065] Among them, the spacecraft behavior intent classification data can be a structured dataset formed after the system performs semantic classification of the identified behavior intents according to their corresponding functional domains (such as orbit control, attitude control, communication, thermal control, etc.) based on the functional division rules of spacecraft mission operation after the behavior intent analysis is completed. Each type of data contains multiple behavior intent items with similar control objectives, execution logic or state paths.

[0066] Specifically, the functional labels, control objectives, and corresponding system module information associated with each behavioral intent in the spacecraft behavioral intent analysis are extracted, such as functional domains like attitude adjustment, orbit control, communication management, energy distribution, and thermal control regulation. Based on the functional mapping rules defined in the Behavior Knowledge Base (KB_B) and combined with the component-behavior association information provided by the Structure-Behavior Bridging Engine (BE), the behavioral intent data is semantically categorized and aggregated according to the spacecraft's operational functional information to form multiple behavioral intent functional classification sets. During the classification process, behavioral path context information of state nodes is also introduced to ensure that behavioral intents in the same category are consistent not only in functional attributes but also in temporal evolution trends and control objectives, resulting in the output spacecraft behavioral intent classification data.

[0067] Step 404: Calculate the similarity between the behavioral intent classification data of each spacecraft and the corresponding data source configuration trust parameters to obtain the data similarity calculation information of each spacecraft.

[0068] The trust parameters configured for the data source can be a set of pre-defined credibility indicators for each type of data source (such as telemetry data, simulation models, historical mission data, expert rule bases, etc.), used to reflect the stability, accuracy, and decision reliability of the data source in a specific context. Common trust parameters include data source weight, anomaly rate, update time frequency, historical consistency score, and fault tolerance level.

[0069] The spacecraft data similarity calculation information can be the semantic matching or state consistency score obtained by comparing each type of behavioral intent classification data with data sources from different trust domains. This information is usually represented in the form of a similarity matrix or vector, where each element reflects the degree of correspondence between a certain behavioral intent and the actual data or model output in a specific trust domain. The higher the similarity, the stronger the support of the data in that trust domain for the behavioral intent.

[0070] Specifically, for each categorized spacecraft behavior intent classification data, state performance data under the corresponding functional dimension is extracted based on the data source type configured in the trust architecture (such as telemetry data, simulation output, expert rule models, or historical mission samples). This data is then compared with the current behavior intent in terms of semantics, state parameters, and execution logic (similarity calculation). The similarity calculation employs a multi-modal matching method, including vector space cosine similarity, structural nesting matching, and semantic graph alignment techniques. Furthermore, the comparison results are weighted and processed in conjunction with trust parameters from the data source (such as confidence score, update time, and anomaly rate) to improve reliability. The final output spacecraft data similarity calculation information includes a matching score matrix for each behavior intent category under each trust domain, reflecting whether the current behavior has received effective support from multi-source data at the "intent-execution" level.

[0071] Step 406: Perform fuzziness calculation on the behavior intent classification data of each spacecraft and the corresponding data source configuration trust parameters to obtain the fuzziness calculation information of each spacecraft data.

[0072] The spacecraft data ambiguity calculation information can be a quantitative result calculated by the system based on uncertainty indicators such as data volatility, entropy, and anomaly rate when assessing the expression stability between various behavioral intentions and corresponding trust domain data sources. This information is used to indicate whether the data has ambiguity, discontinuity, or unclear expression when expressing a certain behavioral intention; the higher the value, the stronger the uncertainty of the information source under that behavior.

[0073] Specifically, in the process of calculating fuzziness by configuring trust parameters for each spacecraft behavior intent classification data and its corresponding data source, the clarity and stability of the data source input associated with each type of behavior intent are evaluated to express that behavior intent. Specifically, corresponding data segments are extracted from the spacecraft behavior intent classification data, and indicators such as instantaneous volatility of state variables, time series variance, outlier distribution density, and information entropy are calculated to quantify whether the data expresses the intent clearly and whether there is ambiguity or uncertainty. Simultaneously, the fuzziness results are weighted and normalized by combining the fuzziness tolerance parameter and historical consistency score configured for each data source, resulting in a fuzziness score with relative credibility. A higher score indicates a more fuzzy expression of the current behavior intent by the data source, and vice versa. The final fuzziness calculation information is a quantitative description of the uncertainty of each behavior intent category under multiple trust domains, serving as a key supplementary indicator in the consistency assessment along with the similarity score.

[0074] Step 408: Perform consistency scoring on the similarity calculation information and ambiguity calculation information of each spacecraft data to obtain spacecraft behavior mapping analysis data.

[0075] Among them, the consistency score can be a comprehensive evaluation result obtained by fusing the similarity score and ambiguity score of each type of behavioral intention under multiple trust domains. It is used to measure whether the behavioral intention has sufficient data support and semantic stability in the current actual state.

[0076] Specifically, based on spacecraft behavior intent classification data and data source configuration trust parameters, weighting factors are set for the similarity calculation information and ambiguity calculation information of each spacecraft data, taking into account parameters such as long-term stability, real-time performance, and credibility level. Subsequently, a nonlinear scoring function is constructed to assign higher mapping consistency scores to combinations of high similarity and low ambiguity, while low similarity or high ambiguity weakens their scoring contribution. For data with conflicting or outlier characteristics, the system introduces a redundancy suppression mechanism to reduce their interference with the overall score. The final spacecraft behavior mapping analysis data includes not only the fusion consistency score of a single behavior intent under multiple trust domains, but also the overall path-level consistency trend, key node offset indication, and confidence interval.

[0077] In this embodiment, behavioral intent analysis data is classified based on spacecraft operational function information. Similarity and ambiguity calculations are then performed using configured trust parameters to generate a consistency score. This not only ensures that the classification granularity aligns with the actual control function partitions but also avoids the misjudgment and overfitting problems that may occur with traditional single-matching mechanisms by combining semantic similarity and data ambiguity as dual measures. Simultaneously, the introduction of trust domain configuration information dynamically weights the reliability of the data source, making the final generated behavioral mapping analysis data more credible and discriminative. This provides a more stable and interpretable scoring basis for behavioral difference detection, adaptive control adjustment, and spacecraft virtual-real consistency management, significantly improving the system's intelligent cognition and response capabilities to complex operational states.

[0078] In one exemplary embodiment, such as Figure 5 As shown, based on the spacecraft behavior mapping analysis data, the behavior differences of the spacecraft are corrected and calculated to obtain the spacecraft behavior adjustment data, including steps 502 to 504.

[0079] in:

[0080] Step 502: Based on the spacecraft behavior mapping analysis data, perform a difference analysis on the actual behavior path and the expected behavior path of the spacecraft to obtain path difference analysis data.

[0081] The path difference analysis data can be a set of structured analysis results generated through multi-dimensional alignment, state comparison, and offset measurement during the comparison of the actual and expected behavior paths of a spacecraft. This data includes not only the misalignment of the behavior path on the time axis, the offset of state nodes, and the difference in transition order, but also behavioral intention deviation scores, response lag indicators, control chain breakpoints, and overall path similarity indicators.

[0082] Specifically, behavioral state sequences extracted from spacecraft behavior mapping analysis data are used to construct actual behavior paths, and predefined mission target paths are invoked from the seven-element knowledge extension architecture to construct expected behavior paths. Then, based on multi-modal features such as semantic labels, timestamps, behavioral intentions, and control responses of behavioral state nodes, temporal alignment and structural comparison are performed on the two paths to identify differences between the paths in dimensions such as state transition order, control target offset, response delay, and path breakage. Furthermore, a state importance weighting mechanism is introduced to perform weighted enhancement analysis on the offsets of key mission nodes, and the error magnitude, evolution trend, and influence of each offset segment in the entire path are calculated. The generated path difference analysis data not only includes path-level structural offset scores, behavioral trajectory alignment errors, and stage offset indicators, but also identifies potential influencing factors causing the differences.

[0083] Step 504: Perform data control transformation on the path difference analysis data to obtain spacecraft behavior adjustment data.

[0084] Data control transformation can be the process of converting offset features (such as state deviation, path breakage, response delay, etc.) identified in path difference analysis data into specific executable control command adjustment information through a control parameter mapping mechanism. This transformation can be based on control law models, mapping rules from behavioral intentions to parameters, neural network regulators, or fuzzy inference modules, achieving accurate projection from high-dimensional cognitive layer difference information to low-level control variables.

[0085] Specifically, key offset indicators in the path difference analysis data are feature-extracted, including the offset magnitude of state nodes, control response lag, behavioral path break location, and degree of behavioral intent mismatch. Based on a control law mapping model (such as controller gain function, behavioral intent to control parameter conversion network, or fuzzy inference engine), these difference features are mapped into executable control parameter adjustments, such as attitude angle fine-tuning, thrust compensation, and stage switching threshold adjustment. At the same time, state importance weights, confidence region credibility scores, and fuzziness factors are introduced to dynamically weight the generated adjustments to ensure that the control output has stability, fault tolerance, and specificity. The resulting spacecraft behavior adjustment data is a set of structured control correction instructions, including parameter fine-tuning values, control priority labels, and constraint boundary adjustment suggestions. This data can be directly input into the spacecraft control system to achieve rapid response and closed-loop adaptive correction of path offsets.

[0086] In this embodiment, difference analysis between the actual and expected behavioral paths is conducted based on spacecraft behavior mapping analysis data. This difference analysis data is then transformed using data control to generate spacecraft behavior adjustment data. This not only accurately identifies key state deviations, timing misalignments, and behavioral logic anomalies within the path but also, through a control intent-driven transformation mechanism, converts high-dimensional behavioral deviation information into structured, executable control parameter adjustment results. This effectively supports the system's real-time correction of abnormal behavior, adaptive regulation, and mission objective recovery. This mechanism strengthens the collaborative capabilities between the model perception and execution systems, significantly improving the spacecraft's dynamic adaptability, control accuracy, and ability to maintain consistency between virtual and real systems during operation.

[0087] In one exemplary embodiment, such as Figure 6 As shown, based on spacecraft behavior mapping analysis data, a difference analysis is performed on the actual behavior path and the expected behavior path of the spacecraft to obtain path difference analysis data, including steps 602 to 606. Wherein:

[0088] Step 602: Adjust the state importance of the spacecraft behavior mapping analysis data to obtain state adjustment mapping analysis data.

[0089] Among them, state importance adjustment can be the process of assigning different importance weights to each behavioral state based on its functional status, control criticality, time sensitivity, and scope of influence in the mission when processing spacecraft behavior mapping analysis data, and then using these weights to reconstruct the state offset score.

[0090] Among them, the state adjustment mapping analysis data can be the intermediate processing result after state importance adjustment. It is the enhanced expression data formed by weighting the offset value, confidence score and fuzzy index of each state node according to the importance factor based on the original behavior mapping analysis data.

[0091] Specifically, key features of each behavioral state node are extracted from spacecraft behavior mapping analysis data, including state offset scores, confidence consistency scores, ambiguity indices, and mission stage information. Then, based on the mission model and behavior knowledge base, high-weight states with critical control or logical dominance over mission execution are identified, such as path start points, mission turning points, and critical control nodes. The importance weight factor for each state is calculated by considering its temporal sensitivity and influence range. Subsequently, the offset of each state is further weighted with its relevant scores to increase the offset weight of critical states while suppressing the interference effect of low-importance states. Finally, the resulting state adjustment mapping analysis data not only retains the numerical information of state offsets but also incorporates semantic layer weights related to mission execution logic and control sensitivity.

[0092] Step 604: Perform local offset aggregation on the state adjustment mapping analysis data to obtain local offset mapping analysis data.

[0093] Local offset aggregation can be achieved by windowing the state-adjusted behavior path data and combining the offsets of adjacent state nodes in the behavior time series to identify continuous offset trends, abrupt changes, or semantically abnormal segments within local areas of the behavior path.

[0094] Local migration mapping analysis data can be the output of the local migration aggregation process. It records the migration patterns, anomaly types, and trend changes of the spacecraft in each behavioral path segment, using a sliding window or path segment as the unit. This data not only includes local average migration intensity, fluctuation patterns, and migration density, but may also include structural features such as control chain breaks and behavioral intention jumps.

[0095] Specifically, using a certain time or state span in the state adjustment mapping analysis data as a window, the offset values, weight information, and behavioral consistency scores of adjacent state nodes are combined and calculated to identify offset patterns within continuous regions, such as local behavioral features like "continuous upward trend," "intermittent offset," and "drastic jumps." Simultaneously, the semantic relationships between states within the window are analyzed, such as the rationality of state transition logic and the integrity of local control chains, further enhancing the semantic interpretability of the offset aggregation results. The output local offset mapping analysis data is presented in units of behavioral segments, identifying which path regions exhibit anomalous evolution, offset clustering, or control fluctuations.

[0096] Step 606: Perform full path tensor offset processing on the local offset mapping analysis data to obtain path difference analysis data.

[0097] The full-path tensor migration process involves reassembling all local behavioral fragments into a complete behavioral path tensor after completing local migration analysis, and then performing a multi-dimensional comparative analysis with the expected path tensor in the knowledge model. By using tensor difference functions (such as tensor distance, trajectory similarity, dynamic time warping, etc.) to perform structural alignment and error measurement between the actual and expected paths, the degree of migration of the entire path in space, time, and semantics can be quantified.

[0098] Specifically, all aggregated local offset mapping analysis data are reorganized into a multidimensional behavioral trajectory tensor according to the temporal and structural order of the behavioral path. Each tensor segment represents a local offset feature within a time window. Then, an alignment mapping is constructed between this actual path tensor and the expected behavioral path tensor defined in the seven-element knowledge extension architecture. Tensor difference calculations (such as Frobenius norm, Dynamic Time Warping (DTW), and behavioral similarity kernel functions) are used to comprehensively measure the overall offset intensity, trend consistency, and key segment misalignment at the path level. At the same time, path universality constraints and behavioral semantic similarity are introduced as adjustment factors to prevent the amplification of local anomalies from interfering with the global score. The resulting path difference analysis data includes not only numerical full-path offset scores but also key segment offset indicators, offset heat distribution maps, and behavioral path deviation pattern labels. This constitutes the core foundational data for assessing the degree of deviation in spacecraft operational status, triggering correction mechanisms, and decision-making interventions.

[0099] In one embodiment, the formula for calculating path difference analysis data is:

[0100]

[0101] Where ΔB represents path difference analysis data, i represents the number of states, and α i For the importance weight of the state, For spacecraft behavior mapping analysis data, As a penalty index, Penalty for trust consistency Let i be the actual state vector. The i-th expected state vector For multi-mode state variability, For the actual state behavior distribution, For the expected state behavior distribution, Let λ be the Körbeck-Leibler divergence, and λ1 be the Körbeck-Leibler divergence adjustment coefficient. For multimodal behavioral state shift, μ i The state ambiguity score is given, where γ is the ambiguity adjustment coefficient, γ·μ i For ambiguity penalty, P a Let P be the tensor of the actual behavior path. e For the expected behavior path tensor, Let β be the squared Frobenius norm, and β be the global offset adjustment coefficient. This represents the offset of the overall path structure.

[0102] In this embodiment, by sequentially performing state importance adjustment, local offset aggregation, and full-path tensor offset processing on the spacecraft behavior mapping analysis data, the influence of key control nodes can be highlighted at the state level through importance weighting, avoiding interference from non-critical states in the overall judgment. At the local path level, sliding window aggregation is used to identify behavior offset trends and segment anomalies. Finally, tensor-level comparison is constructed at the overall path level to achieve multi-dimensional, temporal offset trajectory measurement and behavior anomaly structure modeling. This hierarchical processing strategy significantly improves the robustness and discriminative power of behavior difference identification, providing accurate and interpretable behavioral deviation data for subsequent control correction generation, and effectively enhancing the behavioral cognition depth and model adaptive adjustment capability of the spacecraft digital twin system.

[0103] In one exemplary embodiment, such as Figure 7 As shown, data control transformation is performed on the path difference analysis data to obtain spacecraft behavior adjustment data, including steps 702 to 706. Wherein:

[0104] Step 702: Map the path difference analysis data to the control intent space to obtain the control intent offset semantic vector.

[0105] The control intent space can be a multi-dimensional representation space used to express the semantic mapping relationship between spacecraft mission objectives, control strategies, and behavioral objectives. In this space, each point or vector represents a specific control intent, such as "entering attitude stabilization," "completing orbit control acceleration," or "maintaining communication attitude."

[0106] Among them, the control intention offset semantic vector can be a vector expression generated by mapping the behavioral offset information in the path difference analysis data to the control intention space. It is used to represent the degree and direction of the current spacecraft's operating state deviating from the preset control intention at the semantic layer.

[0107] Specifically, the offset magnitude, direction, duration, and corresponding behavior labels and control response types of each behavioral state segment in the path difference analysis data are extracted to construct a structured set of offset features. Then, the control intent knowledge base is invoked to align these difference features with the semantic relationships between them and the predetermined control objectives. Using an embedding model or graph mapping mechanism, each type of offset feature is embedded into the control intent semantic space to obtain its deviation description in the control objective dimension. The resulting data not only identifies "what the current behavior has deviated from" and "which type of erroneous execution path it has deviated from" relative to the target task, but also forms a set of control intent offset semantic vectors with semantic interpretation capabilities. These vectors locate the direction, intensity, and type of the offset trajectory in the control intent space.

[0108] Step 704: Compress the uncertain state distribution of the behavioral trajectory in the control intention offset semantic vector to obtain the adjusted intention offset semantic vector.

[0109] Among them, the uncertain state distribution of behavioral trajectory can be a set of state nodes with low confidence, high data noise, high ambiguity, or multiple possible interpretations in the actual spacecraft operation path. These states are usually caused by sensor anomalies, fuzzy matching of behavior, or unclear control conditions, which will introduce instability and ambiguity in control decision-making. Therefore, it is necessary to identify and avoid them in the expression of control intention through compression, weight suppression, or fuzzy information processing.

[0110] The adjusted intention offset semantic vector can be the result of uncertain state distribution compression and semantic cleansing based on the control intention offset semantic vector. It removes ambiguous or semantically unclear interference segments in the behavior trajectory, retaining only semantically clear and stable offset features, representing a highly reliable and structurally compact control correction intention input.

[0111] Specifically, the confidence, ambiguity, and stability indices of the behavioral state nodes corresponding to each component element in the control intent offset semantic vector are analyzed to identify "weakly expressed" state regions with high volatility, incomplete information, or semantic ambiguity. Using a fuzzy compression function, these uncertain state parts undergo information compression and feature dimensionality reduction processing. Methods such as entropy-weighted attenuation, confidence-gated suppression, or ambiguity masking are employed to weaken their interference with the overall control intent expression, preserving and strengthening high-confidence, clearly expressed state segment information, forming a set of control intent offset semantic vectors. This vector represents a "cleaned version" of the original offset intent, effectively filtering out the impact of data-level uncertainty on control intent parsing.

[0112] Step 706: Trust domain redundancy squeezing and fusion are performed on the adjustment intention offset semantic vector to obtain spacecraft behavior adjustment data.

[0113] Among them, the redundancy squeezing fusion of the trust domain can be used to evaluate the support and consistency of multiple data trust sources (such as telemetry, simulation, historical behavior, and expert knowledge) for control correction decisions. By compressing conflicting data, suppressing low-confidence data, and normalizing weights, a set of the most credible and consistent control correction quantities can be fused together.

[0114] Specifically, the semantic vector of the adjusted intention offset is aligned with the corresponding control parameter support data in multiple trust domains (such as telemetry data, simulation model output, expert rules, and historical mission behavior databases). The consistency and contribution strength of each source in expressing the current control intention offset are analyzed. Then, based on the preset weight parameters of each trust domain, historical stability scores, and real-time credibility under the current state, the outputs of each source are differentially adjusted. A redundancy suppression mechanism is introduced by constructing a fusion function to explicitly suppress conflicting or repetitive data contributions, thereby strengthening the control influence of high-confidence and high-consistency data. Residual compression and fuzzy weighting factors are also introduced to prevent abnormal data amplification from interfering with system judgment. The final spacecraft behavior adjustment data is a set of structured control correction outputs, including cleaned control parameter adjustments, execution confidence scores, source credibility identifiers, and behavior correction priority labels. It has the ability to directly drive the control system to perform adaptive corrections, while also preserving the semantic interpretation of behavior and the transparency of data sources.

[0115] In this embodiment, path difference analysis data is mapped to the control intent space to generate a control intent offset semantic vector. Further compression of the uncertain state distribution and implementation of trust domain redundancy squeezing fusion not only clarifies the relationship between the offset and the control objective through semantic vector mapping but also effectively eliminates ambiguity and low-confidence states in the trajectory, improving the clarity of the control intent expression. Simultaneously, multi-source fusion and redundancy suppression mechanisms are introduced to ensure that the final behavior adjustment data possesses high reliability, low noise, and a clear execution direction. This processing flow significantly enhances the system's fault-tolerant identification and adaptive correction capabilities for abnormal behavior in complex and variable operating scenarios, significantly strengthening the practical application value of the digital twin model in intelligent decision-making and control closed loops.

[0116] In one exemplary embodiment, such as Figure 8 As shown, the quaternary knowledge-driven architecture data is expanded to obtain a seventh-element knowledge-extended architecture data, including steps 802 to 806. Wherein:

[0117] Step 802: Perform behavioral knowledge recognition on the quaternary knowledge-driven architecture data to obtain quaternary behavioral knowledge recognition feature data.

[0118] Among them, behavioral knowledge recognition can extract key elements related to behavior execution from the original structural knowledge, data flow and control logic, including state transition conditions, control objectives, task intent, execution path and temporal relationship.

[0119] The four-element behavioral knowledge recognition feature data can be a set of structured information related to behavior extracted from the four-element knowledge-driven architecture. This data includes the behavioral label, control intent, data triggering conditions, stage affiliation, and temporal relationship of each state node, which is an explicit expression of the "implicit behavioral logic" in the original knowledge system.

[0120] Specifically, key information related to spacecraft operational behavior is extracted from the structured objects (OB), structured knowledge (KB), data resources (DB), and process models (DP) in the quaternary knowledge-driven architecture data. This includes control parameters, system component functions, mission phase state nodes, state transition conditions, and real-time feedback signal characteristics. Using semantic extraction algorithms, rule matching engines, and state flowchart analysis methods, this information is transformed into a structured representation for behavior modeling, such as the temporal logical relationship between behavior triggering conditions, behavior target intent, execution path labels, and behavior state nodes. These behavioral features are further associated with corresponding control targets and data response mechanisms to form a preliminary set of behavioral cognition vectors as quaternary behavioral knowledge recognition feature data.

[0121] Step 804: Based on the four-element behavioral knowledge recognition feature data, classify the behavioral features of the spacecraft to obtain the knowledge recognition feature data for each category.

[0122] Among them, the classification knowledge recognition feature data can be the structured result generated after functional division and semantic classification of the four-element behavioral knowledge recognition feature data. The system clusters and classifies behavioral features according to dimensions such as control objectives, behavioral categories, and acting systems. Each data category represents a type of task fragment with similar functions, execution patterns, and behavioral intentions.

[0123] Specifically, the task intent, control objective, acting system (such as attitude, orbit control, thermal control, and communication), and state transition structure associated with each quaternary behavioral knowledge recognition feature data are analyzed to construct a behavioral description vector containing multi-dimensional attributes such as functional semantics, execution context, and control type. Based on a predefined spacecraft operation functional domain model, semantic matching, vector clustering, or rule mapping are used to classify behavioral features into categories with clear semantic boundaries and mission functional consistency, such as "attitude adjustment," "orbit correction," and "energy management." During the classification process, boundary adjustments and anomaly identification are performed by comparing historical mission patterns with expert behavioral templates to ensure that behaviors within the same category are not only highly related in functional logic but also have structural consistency and execution compatibility, thus obtaining knowledge recognition feature data for each category.

[0124] Step 806: Perform behavior enhancement on the knowledge recognition feature data of each category to obtain the enhanced knowledge recognition feature data.

[0125] Among them, behavior enhancement can be based on existing classification behavior features and supplement missing state nodes, strengthen semantic expression, and fill control path structure through context modeling, task template matching, and historical trajectory introduction, thereby forming a more complete behavioral knowledge unit with better execution logic and task semantic consistency.

[0126] Among them, the enhanced knowledge recognition feature data can be the output of the behavior enhancement process. It contains more complete state chains, behavior paths, control conditions and target intent information than the original classification features, and has stronger semantic clarity, behavior continuity and execution rationality.

[0127] Specifically, for each category of classified knowledge recognition feature data, expert rule templates or historical task behavior trajectories are invoked to supplement the missing intermediate state nodes, control preconditions, typical action sequences, etc. in the classification features. Then, a context enhancement mechanism is introduced, and the implicit logical relationships between behaviors are inferred by combining the process model (DP), such as state transition dependencies, resource mutual exclusion, or stage synchronization. At the same time, potential control targets are identified through the behavior intent reverse inference mechanism, and their intent labels and parameter requirements are supplemented to generate enhanced knowledge recognition feature data. This data not only has a more complete behavior path, control chain, and intent expression, but also has stronger logical coherence and semantic clarity.

[0128] Step 808: Perform behavioral extensions on each enhanced knowledge recognition feature data to obtain the seven-element knowledge extension architecture data.

[0129] Among them, behavior extension can be to further transform the enhanced behavior knowledge into a form that can be embedded in the system knowledge structure. By constructing behavior knowledge modules (KB_B), structure-behavior bridging engines (BE), and time dynamic models (TD), behavior information is connected with structure, data, and processes, ultimately forming a complete seven-element knowledge extension architecture.

[0130] Specifically, an independent behavioral knowledge module (KB_B) is constructed based on the enhanced knowledge recognition feature data, clarifying behavioral state nodes, action sequences, control intentions, state transition logic, etc., and semantically mapping and logically connecting it with the structural knowledge (KB_S) and process model (DP) in the four-element knowledge-driven architecture data. Secondly, by constructing a structure-behavior bridging engine (BE), a correspondence is established between key behavioral nodes and specific structural components and control parameters, realizing causal coupling between behavioral logic and the physical system. Simultaneously, a time dynamic model (TD) is introduced, based on historical evolution patterns and control response time windows, to establish a prediction function or evolutionary trajectory of behavioral states in the time dimension, thereby achieving path-level behavioral deduction and trend modeling. This results in a seven-element knowledge extended architecture data, including structural objects (OB), structural knowledge (KB_S), behavioral knowledge (KB_B), data resources (DB), process model (DP), structure-behavior bridging engine (BE), and time dynamic model (TD), constituting a digital twin cognitive framework with semantic completeness, control interpretability, and predictive capabilities.

[0131] In this embodiment, by identifying, classifying, enhancing, and expanding the behavioral knowledge of the quaternary knowledge-driven architecture data, a 7-element knowledge-extended architecture data is ultimately constructed. This enables the explicit extraction and structured expression of implicit behavioral logic in the original knowledge. Semantic classification clarifies the boundaries of behavioral functions, ensuring a reasonable modular organization of the model. Subsequently, contextual modeling and behavioral chain completion mechanisms are introduced to enhance behavioral features, effectively improving the continuity of behavioral expression and the accuracy of control semantics. Finally, by extending and generating behavioral knowledge modules (KB_B), a structure-behavior bridging engine (BE), and a time dynamic model (TD), a knowledge modeling framework with multi-domain linkage of structure, behavior, and time is realized. The overall architecture moves from static modeling to dynamic cognition, significantly enhancing the digital twin system's ability to understand, infer, and adaptively control complex task states, providing a high-precision, multi-dimensional knowledge foundation for subsequent virtual-real consistency analysis and dynamic model correction.

[0132] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order restriction on the execution of these steps, and they may be executed in other orders.

[0133] Based on the same inventive concept, this application also provides a model adaptive correction device for implementing the above-mentioned model adaptive correction method based on the consistency of virtual and real mapping of spacecraft multi-domain information. For example... Figure 9 As shown, a model adaptive correction device based on the consistency of virtual-real mapping of spacecraft multi-domain information is provided, including: an architecture data acquisition module 902, an architecture data dimensionality enhancement module 904, an architecture data analysis module 906, a correction data calculation module 908, and an architecture data correction module 910. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the model adaptive correction device based on the consistency of virtual-real mapping of spacecraft multi-domain information provided below can refer to the limitations of the model adaptive correction method based on the consistency of virtual-real mapping of spacecraft multi-domain information above, and will not be repeated here. Each module in the above-mentioned model adaptive correction device based on the consistency of virtual-real mapping of spacecraft multi-domain information can be implemented entirely or partially by software, hardware, or a combination thereof. Each module can be embedded in the processor of a computer device in hardware form or independent of the processor, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0134] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 10 As shown. This computer device includes a processor, memory, input / output interfaces (I / O), and communication interfaces. Those skilled in the art will understand that... Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0135] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0136] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0137] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0138] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0139] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0140] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0141] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A model adaptive correction method based on the consistency of virtual-real mapping of multi-domain information in spacecraft, characterized in that, The method includes: Acquire quaternary knowledge-driven architecture data for spacecraft; The quaternary knowledge-driven architecture data is expanded to obtain seven-element knowledge-extended architecture data. Based on the seven-element knowledge extension architecture data, the operational behavior of the spacecraft is mapped and analyzed to obtain spacecraft behavior mapping analysis data. Based on the spacecraft behavior mapping analysis data, the behavior differences of the spacecraft are corrected and calculated to obtain spacecraft behavior adjustment data. Based on the spacecraft behavior adjustment data, the quaternary knowledge-driven architecture data is corrected to obtain corrected knowledge-driven architecture data.

2. The method according to claim 1, characterized in that, The process of mapping and analyzing the spacecraft's operational behavior based on the seven-element knowledge extension architecture data to obtain spacecraft behavior mapping and analysis data includes: Based on the seven-element knowledge extension architecture data, the behavioral intent of the spacecraft is analyzed to obtain spacecraft behavioral intent analysis data; Trust domain weighted consistency mapping is performed on the spacecraft behavior intent analysis data to obtain the spacecraft behavior mapping analysis data.

3. The method according to claim 2, characterized in that, The process of performing trust-domain weighted consistency mapping on the spacecraft behavior intent analysis data to obtain the spacecraft behavior mapping analysis data includes: Based on the operational function information of the spacecraft, the behavioral intent analysis data of the spacecraft is classified to obtain behavioral intent classification data for each spacecraft. The similarity calculation of the spacecraft's behavioral intent classification data is performed between the corresponding data source configuration trust parameters and the data of each spacecraft to obtain the data similarity calculation information of each spacecraft. Furthermore, the ambiguity calculation of each spacecraft behavior intent classification data is performed with the corresponding data source configuration trust parameters to obtain the ambiguity calculation information of each spacecraft data. Consistency scores are calculated for the similarity and ambiguity of the data of each spacecraft to obtain the spacecraft behavior mapping analysis data.

4. The method according to claim 1, characterized in that, The step of performing correction calculations on the spacecraft's behavior differences based on the spacecraft behavior mapping analysis data to obtain spacecraft behavior adjustment data includes: Based on the spacecraft behavior mapping analysis data, a difference analysis is performed on the actual behavior path and the expected behavior path of the spacecraft to obtain path difference analysis data. The path difference analysis data is subjected to data control transformation to obtain the spacecraft behavior adjustment data.

5. The method according to claim 4, characterized in that, The step of performing a difference analysis on the actual behavior path and the expected behavior path of the spacecraft based on the spacecraft behavior mapping analysis data to obtain path difference analysis data includes: The spacecraft behavior mapping analysis data is adjusted for state importance to obtain state-adjusted mapping analysis data; Local offset aggregation is performed on the state adjustment mapping analysis data to obtain local offset mapping analysis data; The local offset mapping analysis data is subjected to full path tensor offset processing to obtain the path difference analysis data.

6. The method according to claim 5, characterized in that, The formula for calculating the path difference analysis data is: Where ΔB represents path difference analysis data, i represents the number of states, and α i For the importance weight of the state, For spacecraft behavior mapping analysis data, As a penalty index, Penalty for trust consistency Let i be the actual state vector. The i-th expected state vector For multi-mode state variability, For the actual state behavior distribution, For the expected state behavior distribution, Let λ be the Körbeck-Leibler divergence, and λ1 be the Körbeck-Leibler divergence adjustment coefficient. For multimodal behavioral state shift, μ i The state ambiguity score is given, where γ is the ambiguity adjustment coefficient, γ·μ i For ambiguity penalty, P a Let P be the tensor of the actual behavior path. e For the expected behavior path tensor, Let β be the squared Frobenius norm, and β be the global offset adjustment coefficient. This represents the offset of the overall path structure.

7. The method according to claim 4, characterized in that, The process of performing data control transformation on the path difference analysis data to obtain the spacecraft behavior adjustment data includes: The path difference analysis data is mapped to the control intent space to obtain the control intent offset semantic vector; The uncertain state distribution of the behavioral trajectory in the control intention offset semantic vector is compressed to obtain the adjustment intention offset semantic vector; The spacecraft behavior adjustment data is obtained by performing trust domain redundancy squeezing and fusion on the adjusted intention offset semantic vector.

8. The method according to claim 1, characterized in that, The process of extending the quaternary knowledge-driven architecture data to obtain seven-element knowledge-extended architecture data includes: Behavioral knowledge recognition is performed on the data of the four-element knowledge-driven architecture to obtain four-element behavioral knowledge recognition feature data; Based on the four-element behavioral knowledge recognition feature data, the behavioral features of the spacecraft are classified to obtain knowledge recognition feature data for each category. Behavior enhancement is performed on each of the aforementioned classification knowledge recognition feature data to obtain each enhanced knowledge recognition feature data; Behavioral extensions are performed on each of the enhanced knowledge recognition feature data to obtain the seven-element knowledge extension architecture data.

9. A model adaptive correction device based on the consistency of virtual-real mapping of multi-domain information in spacecraft, characterized in that, The device includes: The architecture data acquisition module is used to acquire the quaternary knowledge-driven architecture data of the spacecraft; The architecture data dimensionality enhancement module is used to expand the knowledge of the four-element knowledge-driven architecture data to obtain seven-element knowledge-extended architecture data. The architecture data analysis module is used to expand the architecture data based on the seven-element knowledge, perform mapping analysis on the operational behavior of the spacecraft, and obtain spacecraft behavior mapping analysis data. The correction data calculation module is used to perform correction calculations on the behavior differences of the spacecraft based on the spacecraft behavior mapping analysis data, and obtain spacecraft behavior adjustment data; The architecture data correction module is used to correct the quaternary knowledge-driven architecture data based on the spacecraft behavior adjustment data to obtain corrected knowledge-driven architecture data.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images