Knowledge graph-based remote control instruction construction method and management system for low-orbit flyer cluster

Abstract

The invention provides a knowledge graph-based remote control instruction construction method and management system for a low-orbit flyer cluster, designs an instruction metadata model based on knowledge ontology, and realizes classification, treatment and semantic association of instruction elements by constructing a'concept layer-strength layer-instruction layer 'multi-level instruction generation mechanism. A business-driven instruction collaboration network concept is provided, and a multi-satellite, multi-system and multi-task collaboration instruction cluster generation mechanism oriented to a complex task scene is supported. Trial is carried out on the realized low earth orbit satellite remote control instruction generation and management system, and the feasibility of the method and the system is verified. According to the method, the flexibility and expandability of an instruction system are improved, support is provided for efficient configuration, generalization multiplexing, new instruction discovery, safety and reliability of future complex instructions, and deep exploration can be continued in future research work. The method is not only suitable for the field of low-orbit satellites, but also has reference significance for other satellite system networks.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph instruction set technology, and in particular to a remote control instruction generation method and management platform for low-orbit flying object clusters. Background Art

[0002] Low-Earth orbit (LEO), also known as near-Earth orbit, refers to an orbit in which a spacecraft is at a relatively low altitude from the ground. Generally, a near-circular orbit with an altitude of less than 2,000 kilometers is called a LEO. Since LEO objects are relatively close to the ground, most LEO objects for earth observation, geodesy, space stations, and some new communication object cluster systems use LEO. Among them, LEO constellations are more suitable for building large-scale LEO networking and are the inevitable choice for the LEO Internet. The advantages of LEO communication systems are short transmission delay, good stability, low link loss, and rich application scenarios. Multi-satellite networking can achieve global coverage. The characteristics of high-orbit and low-orbit flying objects are that there are fewer low-orbit flying objects, but the coverage area of ​​a single low-orbit flying object is relatively wide. If a single low-orbit flying object is damaged, it may affect the normal operation of the entire low-orbit flying object communication system; while there are a large number of low-orbit flying objects in low-orbit communication, presenting a mesh structure. Even if there are problems with individual low-orbit flying objects, the entire network can still continue to provide reliable and continuous communication services, which is in line with the development trend of low-orbit flying object Internet. With the advancement of low-orbit flying object miniaturization technology, the development of information technology and the reduction of low-orbit flying object launch costs, low-orbit flying object cluster systems have gradually attracted attention. High-tech companies such as Amazon, Google, Facebook, and SpaceX have invested in the field of low-orbit and low-orbit flying object communications and proposed more than ten low-orbit and low-orbit flying object communication system solutions such as OneWeb and Starlink, with the goal of achieving global Internet coverage. my country's low-orbit and low-orbit flying object constellation plan has been launched, and the scale is expected to increase. The domestic low-orbit flying object Internet system plan has begun to take shape. After overall coordination and standardization, my country's low-orbit broadband communication low-orbit flying objects will begin an intensive launch period from 2023 to 2025. By 2027, the total number of my country's low-orbit low-orbit flying objects is expected to reach 3,960.

[0003] With the increase in the number of LEO and LEO flying objects and the expansion of their application fields, the generation and management of remote control commands for flying object clusters face challenges. Due to the different types of LEO and LEO flying objects and different initial designs, the functional modules of each subsystem of LEO flying objects are complex and changeable, and the parameters of each command are very different, which usually requires dynamic configuration of multiple satellites, multiple commands, and multiple parameters. Furthermore, the remote control commands of LEO and LEO flying objects often have problems such as large number, scattered and changeable parameters, difficulty in structuring and engineering, difficulty in reuse, and difficulty in expansion. Traditional remote control of LEO flying objects is usually implemented by directly solidifying the remote control logic into the program. This method has the disadvantages of a long R&D cycle, cumbersome modification process, and high difficulty in adding new ones in orbit. It is also limited to the command and parameter encapsulation of a certain type of LEO flying object and lacks versatility. In terms of low-orbit flight object command generation technology, the low-orbit flight object telemetry data processing method based on Extensible Telemetry Protocol (XTCE) is one of them. XTCE is an extensible telemetry transmission protocol using source and receiver representation (SRT). Its core idea is to decompose telemetry data into equivalent data units, so that unified management and integration of data can be achieved. The remote control metadata model based on the improved XTCE model reconstructs the space mission telemetry and remote control information description model and introduces the command set element, so that it can support more general commands and has higher scalability. Domestic low-orbit flight object mission planning and command management are mostly aimed at remote sensing imaging tasks. For example, the Harbin Institute of Technology team introduced multiple types of synergistic relationships between subtasks in multi-satellite collaborative mission planning, namely classical logical relationships, temporal logical relationships, and resource demand constraint relationships to improve risk control. NASA Goddard Space Flight Center in the United States released the Core Flight System in 2015. C. Mitchell et al. designed a modular command and data processing architecture for the KySat-2 CubeSat microsystem, which can call required resources according to mission planning. Stephen L. Dorton and others proposed a "user-centric" approach to designing the human-machine command interaction mode of space systems to improve usability and realize scene perception functions. The Tokyo University team proposed a "Command-centric architecture (C2A)" (command center architecture), which ensures that the loss of storage memory will not occur when switching between aircraft functions and modes, and improves the flexibility and safety of the command center's reconfiguration by planning and describing all the command behaviors of the aircraft.

[0004] In summary, there is little research on the remote control command generation technology for low-orbit flying object clusters. The main technical bottlenecks include:

[0005] 1) It is limited to the command and parameter encapsulation of a certain type of low-orbit flying objects, and lacks the universal implementation of the entire process; 2) It lacks unified metadata management, and does not implement the modification and expansion functions of metadata; 3) It lacks algorithm support and is not suitable for certain specific complex instructions; 4) It lacks instruction cluster solutions and is not suitable for the encoding and construction of a large number of instructions in actual multi-satellite scenarios; 5) It does not implement interaction with other low-orbit flying object functional modules or systems.

[0006] According to the characteristics of the instruction sets of different types of low-orbit and low-orbit flying objects, combined with the actual business needs of sending instructions, this application designs a remote control instruction generation method and management system for low-orbit and low-orbit flying objects, supporting remote control instruction cluster generation, multi-satellite expansion and multi-channel measurement and control. This method is based on the remote control instructions of low-orbit and low-orbit flying objects, and uses big data, data governance, knowledge ontology and other related technologies. On the basis of analyzing the remote control instruction rules of low-orbit and low-orbit flying objects, through the extraction and governance of its metadata, it realizes the flexible encapsulation and call of single instructions, instruction sets, and instruction sequences. It can realize online modification and expansion through the entry and editing of metadata and low-orbit flying object instruction data. At the same time, it integrates the logic judgment, algorithm library, verification encryption, operation log and abnormal alarm functions of low-orbit flying object instructions, supports multi-satellite expansion and multi-channel measurement and control, and can be connected and collaborated with other low-orbit flying object subsystems, such as telemetry, network management, simulation, and visualization systems, through interfaces, messages, etc. The method and system proposed in this application realize dynamic integration in multi-satellite expansion, command cluster editing and parameter entry, solving the problem of lack of flexibility and scalability of current remote control command systems; at the same time, full consideration is given to the adaptation and collaboration with other flying object cluster systems; in terms of execution efficiency, security and availability, there are also significant improvements compared with existing similar systems. Summary of the invention

[0007] The present invention proposes a remote control command construction method and management system for a low-orbit flying object cluster based on a knowledge graph, which solves the problem of the current remote control command system lacking flexibility and scalability in the prior art; at the same time, it fully considers the adaptation and collaboration with other flying object cluster systems. The technical solution of the present invention is implemented as follows: The remote control command construction method for a low-orbit flying object cluster based on a knowledge graph includes the following steps:

[0008] Step S1: Construct a remote control command data model. Based on the concept of ontology, perform semantic mapping between different metadata standards to describe the complex association inertia between command objects and different information systems, thereby constructing an organization, management and retrieval model for metadata information from the server to the command end;

[0009] Step S2: Based on the remote control instruction data model constructed in step S1, a single instruction is defined to target a specific action or task, thereby ensuring a small granularity of action execution; an instruction cluster is generated based on an instruction collaboration network and driven by business needs;

[0010] Step S3: Based on single instructions, instruction rules and instruction clusters, an instruction execution system is constructed to extract and manage their metadata, and to realize flexible encapsulation and calling of single instructions, instruction clusters and instruction sequences. Online modification and expansion are realized through the entry and editing of metadata and low-orbit flight object instruction data. At the same time, the logic judgment, algorithm library, verification encryption, operation log and abnormal alarm functions of low-orbit flight object instructions are integrated, and the expansion and multi-channel measurement and control of low-orbit flight object clusters are supported. It can also be connected and collaborated with other low-orbit flight object cluster subsystems, including telemetry, network management, simulation, and visualization systems. The connection and collaboration are realized through interfaces and messages.

[0011] As a preferred technical solution, the instruction execution system is an abstract complex network, the instructions are nodes on the abstract network, and the mutual cooperation relationship between the instructions in the business system is the edge between the abstract network nodes, that is, the business instruction cluster is defined as an instruction cooperation network constructed by a business-oriented instruction set through a relational network, denoted as G=(C,R), where C represents the set of instructions as network nodes, and R represents the set of relationships between instructions, including timing relationships, device concurrency relationships, task sequence relationships, and delayed execution relationships.

[0012] As a preferred technical solution, a single instruction includes real-time instructions, planning instructions, direct instructions and indirect instructions; the instruction core metadata includes three elements, namely, instruction content elements, instruction object elements and instruction management elements; wherein the instruction content element reflects the main content of the object or task served by the instruction, including the PCAT task type, the specific remote control operation in the remote control task package, and the form and content of the instruction parameters; the instruction object element reflects the equipment and system objects operated by the instruction, including the type of low-orbit flying object, subsystem, execution unit PID or data object; the instruction management element is information related to instruction use control, authority management, dynamic customization, confidentiality protection, and long-term preservation mechanism and process.

[0013] As a preferred technical solution, generating an instruction cluster based on an instruction collaboration network in step S2 includes: defining a set of multiple instructions as a remote control instruction cluster, defining a single instruction numbered i on a low-orbit flying object S as C_i(S, D, T), wherein D represents the instruction's dependency on the system on the low-orbit flying object S, and T represents the instruction's dependency on the task package on the low-orbit flying object. When faced with new service requirements, the representation method of instruction = collaborative network supports the discovery of business instruction clusters using machine learning or artificial intelligence methods, and provides feasibility for the generation, verification and management of possible massive complex instruction tasks. The collaborative relationship between consecutive instructions in an instruction operation sequence is captured using a sequential multi-hop networking strategy, and a higher-quality instruction collaboration relationship network is reconstructed. Then, path planning, community discovery, and link prediction are used to discover business instruction clusters, thereby reducing the running time of the instruction generation solution and reducing the impact of sequence noise on the quality of subsequent business instruction cluster discovery.

[0014] As a preferred technical solution, the generation of the instruction cluster based on business demand drive in step S2 includes the expression form of business service and instruction collaborative network, and the semantic mapping can be expressed as:

[0015] O T (R a , R ts , R p )→G(C,R)

[0016] Semantic mapping is the mapping from business to instruction data types and features, where R a , R ts , R p The attribute constraints, spatiotemporal constraints, and preference constraints between subtasks are expressed separately, including the following steps:

[0017] 5) Decompose and allocate the actual mission objectives of each low-orbiting object and each system by clarifying the principles of task allocation and decomposition;

[0018] 6) According to the decomposed single task planning goal, single task planning is carried out, and then the semantic constraints and conceptual relationships of the ontology model are used to associate the corresponding task requirements with the instruction goals to establish a single instruction generation model;

[0019] 7) Based on the semantic constraints between tasks and the association framework of the low-orbit flying object concept layer, multiple single instructions are constructed into an instruction coordination network;

[0020] 8) Perform instruction verification, instruction distribution, and enter and store valid instructions.

[0021] A remote control command generation and management system for a low-orbit flying object cluster uses the above-mentioned remote control command generation method for a low-orbit flying object cluster, and includes the following modules:

[0022] An extraction module is used to obtain the factory instruction set of each low-orbit flying object subsystem corresponding to each low-orbit flying object, perform dimension combing on the factory instruction set, and extract the metadata model;

[0023] An input module, used for inputting each remote control instruction in the factory instruction set into a database for storage according to the metadata model;

[0024] A configuration module, for performing parameter configuration on a remote control instruction content when selecting a remote control instruction content from the database, and generating a remote control instruction;

[0025] An encoding module, used for encoding the remote control command to generate a command code corresponding to the remote control command when the remote control command passes the verification;

[0026] A sending module, used for sending the command code to the corresponding low-orbit flying object, so that the low-orbit flying object performs a corresponding operation according to the command code;

[0027] An assembling module, used to assemble the plurality of remote control instructions into a remote control instruction set according to instruction assembly sequence and sending time interval according to use requirements;

[0028] A storage module, used for entering the commonly used remote control command set into a database for storage when the remote control command set is a commonly used remote control command set;

[0029] The modification module is used to modify the remote control instruction when the remote control instruction fails to pass the verification until the remote control instruction passes the verification.

[0030] Compared with the prior art, this solution has the following beneficial effects:

[0031] 1) Aiming at the bottlenecks of scattered, difficult to reuse and difficult to expand command elements in low-orbit satellite networks, a command metadata model based on knowledge ontology is designed. By constructing a multi-level command generation mechanism of "concept layer-capability layer-command layer", the command elements are classified, governed and semantically associated. The concept of business-driven "command coordination network" is proposed to support the command cluster generation mechanism for multi-satellite, multi-system and multi-task coordination in complex mission scenarios;

[0032] 2) A remote control command generation and management system for low-orbit flying object clusters was provided, and the low-orbit satellite remote control command generation and management system was implemented and tested, verifying the feasibility of the method and system. The method improves the flexibility and scalability of the command system, and provides support for efficient configuration, generalization and reuse of complex commands, discovery of new commands, and safety and reliability in the future;

[0033] 3) The knowledge graph-based remote control command construction method for low-orbit aircraft clusters in this application is not only applicable to the field of low-orbit aircraft, but also has reference significance for other low-orbit aircraft cluster system networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 This is a classification diagram of remote control instructions of the present invention;

[0036] Figure 2 This is a schematic diagram showing an example of the relationship between instruction metadata and instruction body of the present invention;

[0037] Figure 3 This is a schematic diagram of the core metadata of the low-orbit satellite network remote control command of the present invention;

[0038] Figure 4 The instruction metadata association model and instruction generation mechanism diagram of the present invention;

[0039] Figure 5 A flow chart of the instruction cluster generation method driven by business requirements of the present invention;

[0040] Figure 6 Construct an average time comparison graph for the instruction cluster of the present invention. DETAILED DESCRIPTION

[0041] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Based on the concept of ontology, this application provides a method for constructing remote control instructions for low-orbit flying object clusters based on knowledge graphs. By constructing a multi-level instruction generation mechanism of "concept layer-capability layer-instruction layer", the instruction elements are classified, governed and semantically associated, and the concept of business-driven "instruction collaboration network" is proposed to support the instruction cluster generation mechanism for multi-satellite, multi-system and multi-task collaboration in complex mission scenarios. The problem of lack of flexibility and scalability of the current remote control instruction system in the prior art is solved; at the same time, the adaptation and collaboration with other flying object cluster systems are fully considered. This embodiment uses low-orbit satellites as an example for explanation.

[0043] The specific steps include:

[0044] 1. Construction of instruction metadata model based on ontology

[0045] As a standardized description of domain concepts and the relationships between concepts, knowledge ontology is standardized, clear, formalized, and shared. The goal of knowledge ontology is to capture knowledge in related fields, provide a common understanding of knowledge in the field, determine commonly recognized vocabulary in the field, and give clear definitions of mutual relationships from different levels of formal models. By using ontology to perform semantic mapping between different metadata standards, the complex relationship inertia between instruction objects and different information systems can be described, thereby providing models and methods for the organization, management, and retrieval of metadata information from the server to the instruction end. See Figure 1 .

[0046] Single instructions can be divided into real-time instructions and planning instructions according to the execution method, and can be divided into direct instructions and indirect instructions according to the parameter method. Figure 2 and Figure 3 , indirect instructions are subdivided according to the number of parameters. The core metadata of low-orbit satellite network remote control instructions can be mainly divided into three elements (this application only lists some examples), namely instruction content elements, instruction object elements and instruction management elements. Among them, 1) the instruction content element reflects the main content of the object or task served by the instruction, mainly including the PCAT task type, the specific remote control operation in the remote control task package, and the form and content of the instruction parameters; 2) the instruction object element reflects the equipment and system objects of the instruction operation, including satellite types, subsystems, execution unit PID or data objects, etc.; 3) the instruction management object is related to the mechanisms and processes such as instruction use control, authority management, dynamic customization, confidentiality protection, and long-term preservation.

[0047] 2. Instruction metadata association model

[0048] The instruction metadata association model describes the system business instruction objects and the relationships between instruction objects. The relationship between instruction objects is the main object of user operation in this system. Users can create, adjust and delete the relationship between instructions. CUBRC, Inc. has built a space domain ontology, which mainly includes concepts such as outer space ontology, event ontology, aircraft ontology, and aircraft mission ontology. Figure 4 ,This application is aimed at the instruction generation mechanism based on ,remote control instruction ontology, and implements the metadata association model through ,the management of knowledge concepts. The model is mainly divided into the ,concept layer, the strength layer and the instruction generation layer.

[0049] 1) The concept layer defines the main concepts involved in command generation, the relationship between concepts, and the attribute information of concepts. It mainly includes concepts such as command, satellite, administrator, resource, system, task, service, user, and external ontology such as time and geospace. The concepts are connected by relationships to form mutual associations or semantic constraints.

[0050] 2) The instance layer uses the concept layer as a framework to construct specific instances of instruction metadata elements, build relationships such as association, dependency, subordination, and constraints between elements of instruction metadata, and form specific constraints between metadata. For example, the low-orbit satellite No. 1 is an instance of the satellite concept, which is equipped with an attitude and orbit control subsystem (i.e., it forms equipment constraints on the subsystem). The attitude and orbit control subsystem further forms equipment constraints on the reaction wheel of the execution unit (i.e., the reaction wheel belongs to this subsystem), and forms functional constraints on the remote control task (such as PCAT for satellite services), and finally determines the parameter content and form.

[0051] 3) The command generation layer forms a single remote control command based on the code number of the remote control command elements according to the coding method set by the rules or protocol, and performs verification. Since the command is generated under the guidance of the concept layer and the instance layer, the functionality and reliability of the command are also guaranteed.

[0052] 3.1 Instruction cluster representation method based on instruction collaboration network

[0053] The original intention of designing a single instruction is to only target specific actions or specific tasks, so as to ensure the small granularity of action execution. However, in actual satellite business scenarios, there are often a large number of complex service requirements, involving the coordination between multiple satellites, multiple systems, and multiple tasks, so as to generate multiple instructions to jointly complete complex tasks. The text defines a multi-instruction set as a "remote control instruction cluster". Define a single instruction numbered i for a low-orbit flying object S as C_i(S,D,T), where D represents the dependency of the instruction on the system on the low-orbit flying object S, and T represents the dependency of the instruction on the task package on the low-orbit flying object. The traditional method uses the instruction operation sequence as the management strategy for multiple instructions, that is, the commonly used instruction operation sequence is simply stored and called when the same service requirements are encountered. However, the instruction sequence method has two shortcomings: 1) The instruction collaboration relationship is not clearly expressed, and the instruction sequence only reflects a simple sequential execution relationship, and cannot form constraints on complex dependencies; 2) The generation method of service requirements is not flexible. The instruction sequence is only applicable to instruction calls under the same requirements, and there is no recommendation mechanism for the instruction generation method for combined tasks or new tasks. Since the business functions of the business systems deployed on the satellite system do not overlap, the instructions within the business systems collaborate with each other to complete business functions, forming an organic whole for external services.

[0054] The instruction execution system is an abstract complex network, the instructions are nodes on the abstract network, and the mutual cooperation relationship between instructions in the business system is the edge between abstract network nodes, that is, the business instruction cluster is defined as an instruction cooperation network constructed by a business-oriented instruction set through a relationship network, denoted as G = (C, R), where C represents the set of instructions as network nodes, and R represents the set of relationships between instructions, including timing relationships, device concurrency relationships, task sequence relationships, and delayed execution relationships. When facing new service requirements, the representation method of instruction = cooperative network supports the discovery of business instruction clusters using machine learning or artificial intelligence methods, and provides feasibility for the generation, verification and management of possible massive complex instruction tasks. The sequential multi-hop networking strategy is used to capture the cooperation relationship between continuous instructions in the instruction operation sequence, reconstruct a higher quality instruction cooperation relationship network, and then use path planning, community discovery, and link prediction to discover business instruction clusters, so as to reduce the running time of the instruction generation solution and reduce the impact of sequence noise on the quality of subsequent business instruction cluster discovery.

[0055] 3.2 Business Demand-Driven Instruction Cluster Generation Method

[0056] Generating instruction clusters based on business demand drivers includes business service and instruction collaborative network expressions. The semantic mapping can be expressed as:

[0057] O T , (R a , R ts , R p)→G(C,R)

[0058] Semantic mapping is the mapping from business to instruction data types and features, where R a , R ts , R p Respectively express the attribute constraints, spatiotemporal constraints, and preference constraints between subtasks, see Figure 5 , specifically including the following steps:

[0059] 1) By clarifying the principles of task allocation and decomposition, decompose and allocate the actual mission objectives of each low-orbiting object and each system;

[0060] 2) According to the decomposed single-task planning objectives, single-task planning is performed, and then the semantic constraints and conceptual relationships of the ontology model are used to associate the corresponding task requirements with the instruction objectives to establish a single instruction generation model;

[0061] 3) Based on the semantic constraints between tasks and the association framework of the low-orbit flying object conceptual layer, multiple single instructions are constructed into an instruction coordination network; instruction verification and instruction distribution are performed, and valid instructions are entered and stored.

[0062] 4. Architecture design of low-orbit satellite remote control command generation system

[0063] The low-orbit satellite remote control command generation system designed in the present application realizes the extraction and management of its metadata on the basis of fully analyzing the low-orbit satellite remote control command rules, and realizes the flexible encapsulation and calling of single instructions, instruction sets, and instruction sequences. It can realize online modification and expansion through the entry and editing of metadata and satellite command data, and integrates the logic judgment, algorithm library, verification encryption, operation log and abnormal alarm functions of satellite commands, supports multi-satellite expansion and multi-channel measurement and control, and can connect and collaborate with other satellite subsystems, such as telemetry, network management, simulation, and visualization systems, through interfaces, messages, etc.

[0064] The logical framework of the low-orbit satellite remote control command generation and management system is as follows: Figure 6 As shown, specifically including:

[0065] Analyze the low-orbit satellite constellation, including the corresponding satellite subsystems of experimental satellites, communication satellites, remote sensing satellites, detection satellites, etc., as well as the different instruction sets of each subsystem, and comprehensively parse and extract their metadata; build the underlying architecture of big data and data governance to realize the structured and procedural processing of metadata; according to the metadata model, use the front-end entry or one-click file import method to enter each instruction into the database according to the dimensions set by the metadata rules, such as instruction code, virtual channel identification word, subsystem, remote control package name, parameter id, parameter form, parameter description, and sending delay; according to the metadata model, use the front-end entry or one-click file import method to enter the parameters of each instruction according to the dimensions set by the metadata rules, such as the number of parameters, form, type, base, value, and checksum The code, default value, algorithm, etc. are entered into the database; for the sending and execution of a single instruction, the user side flexibly edits and configures the instruction and its various parameter items through the system front-end interface, combined with the instruction and parameter data settings in the database. After submission, the system background automatically completes the packaging of this single instruction, as well as subsequent calculations, range logic judgments, verification, scrambling, instruction code generation and other operations, and sends it to the corresponding satellite through the ground antenna; the system has developed an instruction cluster (instruction set and instruction sequence) generation method based on multiple instructions, and realized the combined configuration and sending of multiple instructions. The specific method is to realize the flexible configuration of parameter analysis, assembly order, sending interval, logic verification, etc. of each instruction in the instruction cluster on the basis of a single instruction, and finally the system background automatically completes the packaging, verification and satellite launch of the instruction cluster. The difference between an instruction set and an instruction sequence is that an instruction sequence is a commonly used instruction set, which can be serialized and stored after configuration to facilitate adjustment and reuse next time, while an instruction set is not stored and has no reuse function. The system supports exporting and importing instruction clusters in the form of files, and supports sending instruction clusters to satellites through manual calls or scheduled scheduling. At the same time, based on big data technology, auxiliary functions such as instruction cluster (instruction sets and instruction sequences) historical data query, log management, and message systems have been developed. The system instruction sending module docks and calls the antenna and other equipment of the ground gateway system, and supports the underlying implementation and expansion based on different measurement and control channels. At the same time, the system can interact with other satellite systems by calling relevant interfaces or docking through the message system.

[0066] 5.1 Implementation and Functional Verification of Low-Earth Satellite Remote Control Command Management System

[0067] A remote control command generation and management system for a low-orbit flying object cluster uses the above-mentioned remote control command generation method for a low-orbit flying object cluster, and includes the following modules:

[0068] An extraction module is used to obtain the factory instruction set of each low-orbit flying object subsystem corresponding to each low-orbit flying object, perform dimension combing on the factory instruction set, and extract the metadata model;

[0069] An input module, used for inputting each remote control instruction in the factory instruction set into a database for storage according to the metadata model;

[0070] A configuration module, for performing parameter configuration on a remote control instruction content when selecting a remote control instruction content from the database, and generating a remote control instruction;

[0071] An encoding module, used for encoding the remote control command to generate a command code corresponding to the remote control command when the remote control command passes the verification;

[0072] A sending module, used for sending the command code to the corresponding low-orbit flying object, so that the low-orbit flying object performs a corresponding operation according to the command code;

[0073] An assembling module, used to assemble the plurality of remote control instructions into a remote control instruction set according to instruction assembly sequence and sending time interval according to use requirements;

[0074] A storage module, used for entering the commonly used remote control command set into a database for storage when the remote control command set is a commonly used remote control command set;

[0075] The modification module is used to modify the remote control instruction when the remote control instruction fails to pass the verification until the remote control instruction passes the verification.

[0076] 5.2 Instructions for building performance tests

[0077] This paper tests a. the method of building and integrating single instructions into instruction clusters and b. the instruction cluster construction method based on the metadata model in this paper. This paper counts three types of instruction clusters: operation and control management tasks, digital payload tasks, and laser payload tasks. The operation and control management tasks contain an average of 20.5 individual instructions, the digital payload tasks contain an average of 16.3 individual instructions, and the laser payload tasks contain an average of 52 individual instructions.

[0078] Depend on Figure 6 It can be seen that since the single instruction integration method usually requires configuration and review of each single instruction, the construction time is positively correlated with the number of task instructions. However, since the method in this paper adopts an ontology-based metadata model, it automatically completes the numerical configuration, instruction verification, instruction combination and other tasks during the configuration process, greatly shortening the instruction cluster construction time for complex tasks. At the same time, since this method realizes the storage management of instruction clusters and provides the reusability of instructions for the same or similar tasks, the instruction cluster construction time can be further reduced to the ms level, thereby improving the instruction management efficiency of the operation control system.

[0079] In general, the remote control command generation and management technology for low-orbit satellite networks proposed in this paper has the following characteristics and advantages:

[0080] Configuration flexibility: Through unified metadata parsing and management of various instructions from different satellite subsystems, the more complex parts involved in the instruction parameters such as large file escape segmentation, verification, encryption, calculation, as well as the instruction execution order and delay settings in the instruction cluster are encapsulated at the metadata level, and dynamic online expansion is supported. It can realize dynamic encapsulation and flexible calling of instruction clusters for specific business application scenarios.

[0081] Command reusability: Dynamic integration is achieved in command cluster editing and parameter entry, supporting multi-satellite expansion, solving the problem of lack of flexibility and scalability of the current remote control command system, while fully considering the adaptation and collaboration with other satellite systems, and realizing the full-process universal application of satellite remote control commands. Through integrated interfaces and message notifications, data channels and interactive operations with other satellite functional modules or subsystems are opened up, with good usability, scalability and generalization, supporting the expansion of data volume, business areas and application scenarios.

[0082] New instruction discovery: By designing an ontology-based instruction element framework, we support the subsequent algorithm-based automatic discovery solution for business instruction clusters. When faced with new service requirements, the representation method of the instruction collaboration network supports the discovery of business instruction clusters using machine learning or artificial intelligence methods, and provides feasibility for the generation, verification and management of massive and complex instruction tasks in the future.

[0083] Safety and reliability: By using the instruction cluster construction method of satellite knowledge ontology, the system natively provides semantic associations and constraints between instruction elements and between instructions, so that instruction generation has self-checking and self-protection mechanisms. In the face of system failure or external attack, it provides better functional reliability, information security and recoverability for satellite networks. The system also provides flexible algorithm, verification, encryption and other functional modules for aspect development, supports dynamic expansion of each module, and integrates detailed logs and abnormal warning modules, so that each operation can be documented and each abnormality can be warned in real time.

[0084] Aiming at the bottlenecks of scattered, difficult to reuse and difficult to expand command elements in low-orbit satellite networks, this paper designs a command metadata model based on knowledge ontology. By constructing a multi-level command generation mechanism of "concept layer-capability layer-command layer", the command elements are classified, governed and semantically associated. The concept of business-driven "command collaboration network" is proposed to support the command cluster generation mechanism for multi-satellite, multi-system and multi-task collaboration in complex mission scenarios. This paper has carried out a trial in the realized low-orbit satellite remote control command generation and management system, and verified the feasibility of the method and system. The method improves the flexibility and scalability of the command system, and provides support for the efficient configuration, generalization reuse, new command discovery and safety and reliability of complex commands in the future. It can be further explored in future research work. The method in this paper is not only applicable to the field of low-orbit satellites, but also has reference significance for other satellite system networks.

[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing remote control instructions for a low-orbit flying object cluster based on a knowledge graph, characterized in that: The following steps are involved: Step S1: Construct a remote control command data model. Based on the concept of ontology, perform semantic mapping between different metadata standards to describe the complex association inertia between command objects and different information systems, thereby constructing an organization, management and retrieval model for metadata information from the server to the command end; Step S2: Based on the remote control instruction data model constructed in step S1, a single instruction is defined to target a specific action or task, thereby ensuring a small granularity of action execution; an instruction cluster is generated based on an instruction collaboration network and driven by business needs; Step S3: Based on single instructions, instruction rules and instruction clusters, an instruction execution system is constructed to extract and manage their metadata, and to realize flexible encapsulation and calling of single instructions, instruction clusters and instruction sequences. Online modification and expansion are realized through the entry and editing of metadata and low-orbit flight object instruction data. At the same time, the logic judgment, algorithm library, verification encryption, operation log and abnormal alarm functions of low-orbit flight object instructions are integrated, and the expansion and multi-channel measurement and control of low-orbit flight object clusters are supported. It can also be connected and collaborated with other low-orbit flight object cluster subsystems, including telemetry, network management, simulation, and visualization systems. The connection and collaboration are realized through interfaces and messages.

2. The method for constructing remote control instructions for a low-orbit flying object cluster based on a knowledge graph as claimed in claim 1, characterized in that: The instruction execution system is an abstract complex network, instructions are nodes on the abstract network, and the mutual cooperation relationship between instructions in the business system is the edge between abstract network nodes, that is, the business instruction cluster is defined as an instruction cooperation network constructed by a business-oriented instruction set through a relational network, denoted as G=(C,R), where C represents the set of instructions as network nodes, and R represents the set of relationships between instructions, including timing relationships, device concurrency relationships, task sequence relationships, and delayed execution relationships.

3. The method for constructing remote control instructions for a low-orbit flying object cluster based on a knowledge graph as claimed in claim 1, characterized in that: The single instructions include real-time instructions, planning instructions, direct instructions and indirect instructions; the instruction core metadata includes three elements, namely, instruction content elements, instruction object elements and instruction management elements; wherein, the instruction content element reflects the main content of the object or task served by the instruction, including the PCAT task type, the specific remote control operation in the remote control task package, and the form and content of the instruction parameters; the instruction object element reflects the equipment and system objects of the instruction operation, including the type of low-orbit flying object, subsystem, execution unit PID or data object; the instruction management element is information related to instruction use control, authority management, dynamic customization, confidentiality protection, and long-term preservation mechanism and process.

4. The method for constructing remote control instructions for a low-orbit flying object cluster based on a knowledge graph as claimed in claim 2, characterized in that: The step S2 of generating an instruction cluster based on the instruction collaboration network includes: defining a set of multiple instructions as a remote control instruction cluster, defining a single instruction numbered i for a low-orbit flying object S as C_i(S,D,T), wherein D represents the instruction's dependency on the system on the low-orbit flying object S, and T represents the instruction's dependency on the task package on the low-orbit flying object. When faced with new service requirements, the representation method of the instruction collaboration network supports the discovery of business instruction clusters using machine learning or artificial intelligence methods, and provides feasibility for the generation, verification and management of possible massive complex instruction tasks. The sequential multi-hop networking strategy is used to capture the collaborative relationship between a section of continuous instructions in the instruction operation sequence, and a higher-quality instruction collaboration relationship network is reconstructed. Then, path planning, community discovery, and link prediction are used to discover business instruction clusters, thereby reducing the running time of the instruction generation solution and reducing the impact of sequence noise on the quality of subsequent business instruction cluster discovery.

5. The method for constructing remote control instructions for a low-orbit flying object cluster based on a knowledge graph as claimed in claim 1, characterized in that: The generation of the instruction cluster based on the business demand drive in step S2 includes the business service and instruction collaborative network expression form, and the semantic mapping can be expressed as: O T (R a ,R ts ,R p )→G(C,R) Semantic mapping is the mapping from business to instruction data types and features, where R a , R ts , R p The attribute constraints, spatiotemporal constraints, and preference constraints between subtasks are expressed separately, including the following steps: 1) By clarifying the principles of task allocation and decomposition, decompose and allocate the actual mission objectives of each low-orbiting object and each system; 2) According to the decomposed single-task planning objectives, single-task planning is performed, and then the semantic constraints and conceptual relationships of the ontology model are used to associate the corresponding task requirements with the instruction objectives to establish a single instruction generation model; 3) Based on the semantic constraints between tasks and the association framework of the low-orbit flying object concept layer, multiple single instructions are constructed into an instruction coordination network; 4) Perform instruction verification, instruction distribution, and enter and store valid instructions.

6. A remote control command generation and management system for a low-orbit flying object cluster, characterized in that: The remote control command generation method for a low-orbit flying object cluster as described in any one of claims 1 to 5 above is used, comprising the following modules: An extraction module is used to obtain the factory instruction set of each low-orbit flying object subsystem corresponding to each low-orbit flying object, perform dimension combing on the factory instruction set, and extract the metadata model; An input module, used for inputting each remote control instruction in the factory instruction set into a database for storage according to the metadata model; A configuration module, for performing parameter configuration on a remote control instruction content when selecting a remote control instruction content from the database to generate a remote control instruction; An encoding module, used for encoding the remote control command to generate a command code corresponding to the remote control command when the remote control command passes the verification; A sending module, used for sending the command code to the corresponding low-orbit flying object, so that the low-orbit flying object performs a corresponding operation according to the command code; An assembling module, used for assembling a plurality of remote control instructions into a remote control instruction set according to instruction assembly sequence and sending time interval according to use requirements; A storage module, used for entering the commonly used remote control command set into a database for storage when the remote control command set is a commonly used remote control command set; The modification module is used to modify the remote control instruction when the remote control instruction fails to pass the verification until the remote control instruction passes the verification.

Images