Satellite constellation state monitoring model recommendation system and method based on dimension matching

Abstract

The invention provides a satellite constellation state monitoring model recommendation system and a satellite constellation state monitoring model recommendation method based on dimension matching, belongs to the technical field of artificial intelligence, and aims to solve the problems that a traditional algorithm recommendation technology has obvious deficiencies in learning ability and adaptability, so that the user experience is poor when facing complex and changeable satellite data processing tasks. The method comprises the following steps of: inputting telemetry data to be processed into an intelligent algorithm recommendation system; according to expert experience, the telemetering code of the to-be-processed telemetering data, the dimension information of the to-be-processed telemetering data and the task scene of the current to-be-processed telemetering data, obtaining an algorithm meeting a condition from an algorithm library; and 2, selecting an algorithm corresponding to the requirement from the algorithms meeting the condition, obtaining algorithm related parameters according to the selected algorithm, the telemetering code of the telemetering data to be processed and the satellite id, and matching the algorithm related parameters according to the model training information module to obtain a satellite constellation state monitoring model corresponding to the requirement.

Description

Technical Field

[0001] The present invention relates to a satellite constellation status monitoring model recommendation system and method based on dimension matching, and belongs to the technical field of artificial intelligence. Background Art

[0002] With the rapid advancement of computer technology, many fields are increasingly integrating it, especially with the widespread application of various computer algorithms in various fields. However, due to the wide variety of algorithms and the significant performance differences in handling different business scenarios, users must have a deep understanding of the underlying principles of the algorithms. This high barrier to entry not only increases the cost of using the algorithms but also, to a certain extent, weakens their effectiveness in practical applications.

[0003] In the satellite field, due to the huge amount of satellite telemetry data and the extremely high real-time requirements of processing tasks, satellite researchers are in urgent need of algorithms with stable performance and excellent efficiency to assist them in quickly analyzing the operating status of satellites. In addition, due to the extremely high requirements for refinement in the satellite field, the demand for refined analysis of satellite data is increasing, which further requires data processing algorithms to meet higher standards in terms of refinement and specialization. However, there is currently a lack of an algorithm recommendation system that combines actual satellite analysis experience, which makes it difficult for researchers to quickly match the most suitable algorithm when faced with complex and changeable satellite data processing tasks. This lack of technology not only limits the in-depth mining of the value of satellite data, but also increases the complexity of researchers' work, affecting the overall research efficiency and quality of results.

[0004] With the continuous advancement of satellite technology, mission scenarios are becoming increasingly complex, placing greater demands on the adaptability of intelligent algorithm recommendation methods. However, current algorithm recommendation technologies have significant shortcomings in learning and adaptability. In particular, when faced with new mission scenarios, the system may be unable to quickly recommend an appropriate algorithm. This limitation not only reduces the accuracy and practicality of algorithm recommendations but also limits the efficiency and quality of satellite data analysis. Summary of the Invention

[0005] In order to solve the problem that traditional algorithm recommendation technology has obvious deficiencies in learning ability and adaptability, which makes it difficult for users to quickly match the most suitable algorithm when facing complex and changeable satellite data processing tasks, the present invention proposes a satellite constellation status monitoring model recommendation system and method based on dimension matching.

[0006] The technical solution adopted by the present invention to solve the above problems is: the structure of the satellite constellation status monitoring model recommendation system based on dimension matching proposed by the present invention includes:

[0007] Algorithmic intelligent recommendation system, algorithm library and database;

[0008] The algorithm library is used to store several algorithms;

[0009] The algorithm intelligent recommendation system is used to obtain qualified algorithms from the algorithm library based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, and match the satellite constellation status monitoring model corresponding to the requirements according to one of the qualified algorithms;

[0010] The database is used to store experience records in the algorithm intelligent recommendation system and algorithm information in the algorithm library.

[0011] Furthermore, the algorithm intelligent recommendation system includes an algorithm screening module, a model training information module and an autonomous learning module;

[0012] The algorithm screening module obtains the algorithm recommendation results and feeds back to the user based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed;

[0013] The model training information module is used to match the algorithm-related parameters with the training information of all types of satellite constellation status monitoring models to obtain the corresponding satellite constellation status monitoring model;

[0014] The autonomous learning module is used to record the optimal information in the analysis process of the algorithm screening module and form an experience record based on the telemetry data of the corresponding scenario. The optimal information includes algorithm selection, parameter configuration, telemetry data processing effect and corresponding telemetry data.

[0015] Furthermore, the algorithm library is connected to the database in communication, and includes a plurality of algorithm modules, each of which is used to store an algorithm, and each of which includes an input parameter interface, an output result interface, and a state feedback interface;

[0016] The input parameter interface is used to receive the output parameters of the algorithm screening module;

[0017] The output result interface is used to output the filtered algorithm;

[0018] The state feedback interface is used to feedback the output parameters of the algorithm screening module and the matching degree of the algorithms in the algorithm library.

[0019] Furthermore, if a new algorithm needs to be added to the algorithm library, it is only necessary to encapsulate the algorithm into a corresponding algorithm module in the same format as the algorithm module and store the algorithm information in the database.

[0020] Furthermore, the algorithm intelligent recommendation system is bidirectionally connected to a database. When the algorithm intelligent recommendation system receives telemetry data to be processed, the database calls the experience records in the database to quickly and accurately generate algorithm recommendation results.

[0021] The satellite constellation status monitoring model recommendation method based on dimension matching includes:

[0022] Step 1: Input the telemetry data to be processed into the algorithm intelligent recommendation system. Based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, the qualified algorithm is obtained from the algorithm library;

[0023] Step 2: Select the algorithm corresponding to the requirements from the qualified algorithms, obtain the algorithm-related parameters based on the selected algorithm, the telemetry code and satellite ID of the telemetry data to be processed, match the algorithm-related parameters according to the model training information module, and obtain the satellite constellation status monitoring model corresponding to the requirements.

[0024] Furthermore, step 1 specifically includes:

[0025] The algorithm screening module parses the input telemetry data to be processed, determines the input data dimension and the satellite system to which the satellite telemetry data belongs, and combines the expert experience of relevant telemetry data analysis and data application business scenarios. The output results of the algorithm screening module are output to the algorithm library for algorithm matching, and the algorithm recommendation results are output through the back-end. The algorithm recommendation results include recommended algorithms, applicable algorithms, and non-compliant algorithms. Non-compliant algorithms cannot be called.

[0026] Furthermore, each time the user calls the algorithm, the autonomous module records the optimal information in the algorithm screening module analysis process and stores it in the database in Japanese format. By analyzing the log data, excellent scenarios are extracted, and the extracted excellent scenarios are converted into experience records in combination with telemetry data and stored in the database.

[0027] The beneficial effects of the present invention are:

[0028] (1) Through the algorithm model intelligent recommendation module, the system can automatically recommend the optimal algorithm based on the characteristics of telemetry data and business needs, reducing the time spent by scientific researchers on algorithm selection and significantly improving data analysis efficiency.

[0029] (2) Automated recommendations reduce errors caused by improper human selection, ensuring the accuracy and reliability of the analysis process. The system combines expert experience and the underlying technology of the algorithm to select the algorithm that best suits the current business scenario, ensuring the optimization of the analysis results.

[0030] (3) The present invention divides algorithms into recommended algorithms, applicable algorithms, and non-compliant algorithms to help researchers quickly identify the best options and avoid using inappropriate algorithms that affect the quality of analysis.

[0031] (4) Researchers do not need to have an in-depth understanding of the underlying principles of each algorithm. The system automatically recommends the optimal algorithm, which lowers the threshold for learning and using complex algorithms.

[0032] (5) The system continuously accumulates experience and forms a knowledge base during use, providing reference for subsequent similar scenarios and reducing the cost of repeated learning and trial and error. With the development of satellite technology and the emergence of new algorithms, the system can dynamically update the algorithm library and recommendation strategies to keep pace with the times. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A schematic flow chart of a satellite constellation status monitoring model recommendation method based on dimension matching provided by the present invention;

[0034] Figure 2 A diagram showing the working principle of the satellite constellation status monitoring model recommendation method based on dimension matching provided by the present invention;

[0035] Figure 3 This is a structural block diagram of the satellite constellation status monitoring model recommendation system based on dimension matching provided by the present invention. DETAILED DESCRIPTION

[0036] Combine Figure 3 This embodiment is described as follows. Figure 3 As shown in Figure 1, the structure of the satellite constellation status monitoring model recommendation system based on dimension matching includes:

[0037] Algorithmic intelligent recommendation system, algorithm library and database;

[0038] The algorithm library is used to store several algorithms;

[0039] The algorithm intelligent recommendation system is used to obtain qualified algorithms from the algorithm library based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, and match the satellite constellation status monitoring model corresponding to the requirements according to one of the qualified algorithms;

[0040] The database is used to store experience records in the algorithm intelligent recommendation system and algorithm information in the algorithm library.

[0041] The algorithm intelligent recommendation system includes an algorithm screening module, a model training information module and an autonomous learning module; the algorithm screening module obtains the algorithm recommendation results and feeds back to the user based on expert experience, the telemetry code of the telemetry data to be processed, the dimensional information of the telemetry data to be processed and the task scenario of the current telemetry data to be processed; the autonomous learning module is used to record the optimal information in the analysis process of the algorithm screening module, and form an experience record in combination with the telemetry data of the corresponding scenario. The optimal information includes algorithm selection, parameter configuration, telemetry data processing effect and corresponding telemetry data. The model training information module is used to match the algorithm-related parameters with the training information of all models of satellite constellation status monitoring models to obtain a satellite constellation status monitoring model corresponding to user needs.

[0042] The algorithm intelligent recommendation system is connected to a database for two-way communication. The database is used to store experience records and algorithm information in the algorithm library. When the algorithm intelligent recommendation system receives telemetry data to be processed, it calls the experience records in the database to quickly and accurately generate algorithm recommendation results.

[0043] The algorithm library includes several algorithm modules, each of which is used to store an algorithm. Each algorithm module uses a standardized interface design, including an input parameter interface, an output result interface, and a status feedback interface. The standardized interface is used to receive the output parameters of the algorithm screening module and screen the algorithm modules based on the output parameters. By setting up standardized interfaces, seamless interaction between different algorithm modules is ensured. During the encapsulation process, the internal implementation details of the algorithm are hidden, and only the necessary functions are exposed through the interface, thereby improving the stability and security of the system. If a new algorithm needs to be added, it only needs to be encapsulated as a corresponding algorithm module in the same format as the algorithm module and the algorithm information is stored in the database.

[0044] Specific implementation method 2: Combination Figure 1 This embodiment is described as follows. Figure 1 As shown, the steps of the satellite constellation status monitoring model recommendation method based on dimension matching described in this embodiment include:

[0045] S1: The satellite telemetry data to be processed is input into the algorithm intelligent recommendation system. The front-end provides the telemetry code and mission scenario to be analyzed. The back-end parses the telemetry code to obtain relevant information about the telemetry data and obtains basic algorithm information from the database.

[0046] S2: Applicable algorithms are selected from the algorithm library based on the data dimensions. The system further categorizes the algorithms based on expert experience, telemetry codes, and the current task scenario to be analyzed, and returns them to the front-end. After receiving the recommended algorithms, the front-end displays all algorithms and their categories to the user, who can then call them for analysis.

[0047] S3: The backend calls the algorithm based on the data input by the frontend, combined with the basic algorithm information and algorithm library in the database, and returns the algorithm recommendation results to the frontend.

[0048] In this embodiment, the algorithm recommendation results include recommended algorithms, applicable algorithms, and non-compliant algorithms. The algorithms in the recommended algorithm type are the algorithms that the system considers to be the most effective in processing the current business scenario after analysis. The algorithms in the applicable algorithm type are the algorithms that the system considers to be able to process the data in the current business scenario, but with slightly poor performance after analysis. The algorithms in the non-compliant algorithm type are the algorithms that the system determines to be unable to complete the task of analyzing satellite telemetry data after analysis. Forcing the current algorithm to execute may cause the analysis results to be meaningless, affect the analysis results of satellite data by satellite R&D personnel, and cause problems such as reduced task processing efficiency. It is particularly noted that non-compliant algorithms cannot be called by users.

[0049] S4: After the result is returned, the system will determine whether to learn from this recommendation. If so, it will add expert experience and add new algorithm usage scenarios to optimize future recommendation results.

[0050] This embodiment sets up an autonomous update module in the algorithm intelligent recommendation system, so that the system can record the scenarios with excellent analysis results during use, and combine it with the telemetry data of this analysis to accumulate experience. The autonomous update module captures and stores key information in the analysis process in real time, including algorithm selection, parameter configuration, data processing effects, and telemetry data. This information is stored in the form of a structured log to facilitate subsequent query and analysis. Based on the recorded key information such as algorithm selection, parameter configuration, data processing effects, etc., the system extracts scenarios with excellent analysis results by analyzing the recorded log data and converts them into empirical knowledge. This empirical knowledge is stored in the database, including the performance indicators of the algorithm, applicable scenarios, and optimization suggestions. The database adopts a dynamic update mechanism and can be continuously optimized according to new analysis results.

[0051] After completing the algorithm call, the algorithm-related parameters are obtained according to the selected algorithm, the telemetry code and satellite ID of the telemetry data to be processed, and the algorithm-related parameters are input into the model training information module. The algorithm-related parameters are matched with the training information of all types of satellite constellation status monitoring models to obtain the satellite constellation status monitoring model corresponding to the user's needs.

[0052] The entire process significantly improves data analysis efficiency for satellite R&D personnel through intelligent algorithm recommendation and invocation. Automated recommendations reduce errors caused by inappropriate human selection and ensure the accuracy and reliability of the analysis process. The system combines expert experience with underlying algorithm technology to select the most suitable algorithm for the current business scenario, ensuring optimal analysis results and reducing the algorithm learning cost for researchers.

[0053] How it works

[0054] like Figure 2 As shown, the working principle of the satellite constellation status monitoring model recommendation method based on dimension matching proposed in the present invention includes:

[0055] Step 1: The backend parses the JSON string, obtains the specific telemetry parameter name, and determines the telemetry data dimension.

[0056] Step 2: Traverse all algorithm information in the database

[0057] Step 3: Combine algorithm information and telemetry data dimensions to make judgments

[0058] Step 4: Combine the expert experience in the database to match the algorithm.

[0059] dataNames: data name set, dataNames = {dataName1, dataName2,…, dataNamen}.

[0060] D: data dimension, D = |dataNames|

[0061] A: The set of all algorithms, A = {A1, A2, …, Am}.

[0062] Si: The state of the i-th algorithm, which can take values ​​{0, 1, 2}.

[0063] The specific operations for the standardization process in step 3 above are:

[0064] For each algorithm Ai, its state Si is determined by the following rules:

[0065] Dimension matching (Si=0):

[0066] Si=0ifdatasetmini≤D≤datasetmaxi

[0067] The specific operations for standardization in step 4 above are:

[0068] Ai is the detailed information of the algorithm (including name, type, etc.). Si is the state calculated according to the above rules.

[0069] Keyword matching (Si=1):

[0070]

[0071] Other cases (Si=2):

[0072] Si=2otherwise

[0073] The final output is a collection of all algorithms and their states:

[0074] Result={(Ai,Si)∣i=1,2,…,∣A∣}.

[0075] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any technician familiar with the present profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical content disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent replacement and improvement of the above embodiments made according to the technical essence of the present invention, within the spirit and principles of the present invention, without departing from the content of the technical solution of the present invention, shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. Satellite constellation status monitoring model recommendation system based on dimension matching, characterized by: include: Algorithmic intelligent recommendation system, algorithm library and database; The algorithm library is used to store a plurality of algorithms; The algorithm intelligent recommendation system is used to obtain qualified algorithms from the algorithm library based on expert experience, the telemetry code of the telemetry data to be processed, the dimensional information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, and match the satellite constellation status monitoring model corresponding to the requirements according to one of the qualified algorithms; The database is used to store experience records in the algorithm intelligent recommendation system and algorithm information in the algorithm library.

2. The satellite constellation status monitoring model recommendation system based on dimension matching according to claim 1 is characterized in that: The algorithm intelligent recommendation system includes an algorithm screening module, a model training information module and an autonomous learning module; The algorithm screening module obtains the algorithm recommendation results based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, and feeds back to the user; The model training information module is used to match the algorithm-related parameters with the training information of all types of satellite constellation status monitoring models to obtain a satellite constellation status monitoring model corresponding to the requirements; The autonomous learning module is used to record the optimal information in the analysis process of the algorithm screening module, and form an experience record in combination with the telemetry data of the corresponding scenario, wherein the optimal information includes algorithm selection, parameter configuration, telemetry data processing effect and corresponding telemetry data.

3. The satellite constellation status monitoring model recommendation system based on dimension matching according to claim 1 is characterized in that: The algorithm library is communicatively connected to the database and includes a plurality of algorithm modules, each of which is used to store an algorithm, and each of which includes an input parameter interface, an output result interface, and a state feedback interface; The input parameter interface is used to receive the output parameters of the algorithm screening module; The output result interface is used to output the filtered algorithm; The state feedback interface is used to feed back the output parameters of the algorithm screening module and the matching degree of the algorithms in the algorithm library.

4. The satellite constellation status monitoring model recommendation system based on dimension matching according to claim 3 is characterized in that: If you need to add a new algorithm to the algorithm library, you only need to encapsulate the algorithm into a corresponding algorithm module in the same format as the algorithm module and store the algorithm information in the database.

5. The satellite constellation status monitoring model recommendation system based on dimension matching according to claim 1 is characterized in that: The algorithm intelligent recommendation system is bidirectionally connected to a database. When the algorithm intelligent recommendation system receives telemetry data to be processed, the database calls the experience records in the database to quickly and accurately generate algorithm recommendation results.

6. A satellite constellation status monitoring model recommendation method based on dimension matching, applied to a satellite constellation status monitoring model recommendation system based on dimension matching according to any one of claims 1 to 5, characterized in that: include: Step 1: Input the telemetry data to be processed into the algorithm intelligent recommendation system. Based on expert experience, the telemetry code of the telemetry data to be processed, the dimension information of the telemetry data to be processed, and the task scenario of the current telemetry data to be processed, the qualified algorithm is obtained from the algorithm library; Step 2: Select the algorithm corresponding to the requirements from the qualified algorithms, obtain the algorithm-related parameters based on the selected algorithm, the telemetry code and satellite ID of the telemetry data to be processed, match the algorithm-related parameters according to the model training information module, and obtain the satellite constellation status monitoring model corresponding to the requirements.

7. The satellite constellation status monitoring model recommendation method based on dimension matching according to claim 6 is characterized in that: Step 1 specifically includes: The algorithm screening module parses the input telemetry data to be processed, determines the input data dimension and the satellite system to which the satellite telemetry data belongs, and combines the expert experience of relevant telemetry data analysis and data application business scenarios. The output results of the algorithm screening module are output to the algorithm library for algorithm matching, and the algorithm recommendation results are output through the back-end. The algorithm recommendation results include recommended algorithms, applicable algorithms, and non-compliant algorithms. Non-compliant algorithms cannot be called.

8. The satellite constellation status monitoring model recommendation method based on dimension matching according to claim 6 is characterized in that: After each user calls the algorithm, the autonomous module records the optimal information in the algorithm screening module analysis process and stores it in the database in Japanese format. By analyzing the log data, excellent scenarios are extracted, and the extracted excellent scenarios are converted into experience records in combination with telemetry data and stored in the database.

Images