Intelligent satellite task cooperative scheduling method and system based on atomic service architecture

Abstract

The application provides a kind of intelligent satellite task cooperative scheduling method and system based on atomic service architecture, which comprises the following steps: splitting satellite tasks into multiple standardized atomic services; Abstracting the hardware devices of heterogeneous satellites into standardized virtual devices; Registering and managing the metadata of atomic services and virtual devices in the service registry center; Real-time sensing of satellite physical resource status; Through the intelligent scheduling engine, according to the metadata and physical resource status, a multi-satellite cooperative scheduling scheme containing the mapping relationship and execution timing of atomic services, virtual devices and physical resources is generated. The application also provides a corresponding system. Through task atomization and resource virtualization, the application improves scheduling flexibility, system scalability and resource utilization.

Description

Technical Field

[0001] This invention relates to the field of satellite scheduling technology, and in particular to an intelligent satellite mission collaborative scheduling method and system based on an atomic service architecture. Background Technology

[0002] Satellite mission planning refers to generating a sequence of executable satellite missions based on user needs, without violating satellite usage constraints, and determining the start and end times of these missions to achieve certain objectives. With the development of satellite technology, the scale of satellite networks is constantly expanding, and mission types are becoming increasingly complex. Complex missions such as large-area target coverage and high-frequency, dense point target observation often require the coordinated operation of multiple satellites. Existing satellite mission scheduling methods mostly adopt a "comprehensive mission allocation" model, directly assigning complete missions to single satellites or fixed satellite groups for execution. This leads to the following technical problems in the actual use of existing technical solutions: the task modules of the current technical solutions have poor reusability, different types of tasks require separate design of scheduling logic, and the scheduling flexibility is insufficient when facing dynamically changing task requirements, making it difficult to adapt to multi-satellite heterogeneous scenarios; in addition, the interfaces of heterogeneous satellite hardware devices are quite different, and there is a lack of standardized adaptation mechanisms, which means that multi-satellite collaboration needs to be customized for specific satellite models, resulting in poor scalability; and the overall use is inconvenient. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the purpose of this invention is to provide an intelligent satellite mission collaborative scheduling method and system based on an atomic service architecture.

[0004] A method for collaborative scheduling of intelligent satellite missions based on an atomic service architecture, provided by the present invention, includes the following steps: S1. Receive and analyze satellite missions; S2. Break down the satellite mission into multiple standardized atomic services with independent functions and clearly defined inputs and outputs; S3. Abstract the hardware of heterogeneous satellites into standardized virtual devices and define the adaptation rules between virtual devices and atomic services; S4. Register, store, and manage the metadata of atomic services, virtual devices, and satellite physical resources. The metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. S5. Generate a multi-satellite collaborative scheduling scheme. The multi-satellite collaborative scheduling scheme is generated based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities. The multi-satellite collaborative scheduling scheme includes the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources. S6. Each satellite, according to the multi-satellite collaborative scheduling scheme, calls physical resources through the virtual device interface to collaboratively execute the corresponding atomic service and provides real-time feedback of execution data; S7. Dynamically adjust the multi-satellite collaborative scheduling scheme based on execution data.

[0005] Furthermore, the atomic services include payload atomic services, attitude control atomic services, communication atomic services, energy atomic services, and data processing atomic services; The atomic service is matched with the virtual device through a standardized interface.

[0006] Furthermore, step S3, which abstracts the hardware of heterogeneous satellites into standardized virtual devices, includes the following steps: S31. Collect hardware equipment parameters of heterogeneous satellites, wherein the hardware equipment includes payload equipment, attitude control equipment, communication equipment, energy equipment and data processing equipment; S32. Standardize the core functions and interfaces of similar hardware devices and generate corresponding virtual device templates; S33. Define the calling interface, resource requirement mapping relationship and fault compatibility mechanism of virtual devices to achieve decoupling of atomic services and hardware devices.

[0007] Furthermore, the state management described in step S4 includes the following steps: S41. Receive registration requests from atomic services and virtual devices, and verify the integrity of metadata; S42. Real-time updates of atomic service availability, virtual device adaptation status, satellite physical resource status, and link communication quality; S43. Respond to the query request from the intelligent scheduling engine and return the correspondence between atomic services, virtual devices and satellite physical resources that meet the adaptation rules. The correspondence is determined based on the adaptation rules between atomic services and virtual devices and the mapping relationship between virtual devices and satellite physical resources.

[0008] Furthermore, step S5 employs an intelligent optimization algorithm to generate a multi-satellite collaborative scheduling scheme; The intelligent optimization algorithm is one or more combinations of deep reinforcement learning algorithm, improved genetic algorithm and particle swarm optimization algorithm. The objective function of the intelligent optimization algorithm includes task completion rate, resource utilization rate, energy consumption cost, execution latency and virtual device adaptation success rate.

[0009] Furthermore, the satellite physical resource status data includes the hardware device operating status, remaining energy, storage capacity, attitude control accuracy, link bandwidth, and virtual device occupancy.

[0010] Furthermore, when the satellites are collaboratively executing their respective atomic services, they achieve real-time interaction between atomic service execution data and virtual device status data through inter-satellite links. An adaptive link switching mechanism is used to dynamically switch between inter-satellite links and satellite-to-ground links to ensure data transmission stability. When the hardware device corresponding to the virtual device fails, the virtual device migration mechanism is triggered to bind a backup hardware device.

[0011] Furthermore, in step S1, the satellite mission is analyzed to obtain the satellite mission's target parameters, mission priority, and constraints.

[0012] According to the present invention, an intelligent satellite mission collaborative scheduling system based on an atomic service architecture includes: Mission parsing module: Receives and parses satellite missions; Atomic service decomposition module: The satellite mission is decomposed into multiple standardized atomic services with independent functions and clearly defined inputs and outputs; Device abstraction module: Abstracts the hardware devices of heterogeneous satellites into standardized virtual devices, and defines the adaptation rules between virtual devices and atomic services; Service Registration Center: Registers, stores, and manages the metadata of atomic services, virtual devices, and satellite physical resources. The metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. Intelligent scheduling engine: Generates multi-satellite collaborative scheduling schemes. These schemes are generated based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities. The multi-satellite collaborative scheduling schemes include the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources. Collaborative execution module: Deployed on each satellite, it calls physical resources through virtual device interfaces to collaboratively execute corresponding atomic services according to the multi-satellite collaborative scheduling scheme, and provides real-time feedback of execution data; The intelligent scheduling engine dynamically adjusts the multi-satellite collaborative scheduling scheme based on the execution data.

[0013] Furthermore, the intelligent scheduling engine includes: The state fusion unit is used to perform noise reduction and normalization processing on the collected satellite hardware device state, virtual device state and link state data; The decision optimization unit is used to load intelligent optimization algorithms, construct a multi-objective optimization model that includes device abstract adaptation constraints, and solve the scheduling scheme. The scheme adaptation unit is used to convert multi-satellite collaborative scheduling schemes into instruction formats that can be recognized by virtual devices, supporting dynamic adjustment of multi-satellite collaborative scheduling schemes and migration adaptation of virtual devices.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Improved scheduling flexibility and task reusability. By breaking down complex tasks into standardized atomic services, services can be flexibly combined according to different task requirements. New task types can reuse or combine existing atomic services without redesigning the underlying scheduling logic, significantly improving the system's flexibility and response speed.

[0015] 2. Enhanced system scalability and heterogeneous compatibility. By abstracting the physical hardware of heterogeneous satellites into standardized virtual devices, the underlying physical differences are shielded. When a new satellite joins the network, only a corresponding virtual device needs to be created for it to be uniformly scheduled, significantly improving the system's scalability and compatibility with heterogeneous resources.

[0016] 3. Improved collaboration efficiency and resource utilization. Through an intelligent scheduling engine, from a global perspective, based on real-time status data and intelligent optimization algorithms, atomic services are dynamically matched to the most suitable virtual devices to find the optimal collaborative execution plan, thereby significantly improving the completion rate, resource utilization, and overall execution efficiency of multi-satellite collaborative tasks. Attached Figure Description

[0017] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 A flowchart illustrating an intelligent satellite mission collaborative scheduling method based on an atomic service architecture, provided as an embodiment of the present invention; Figure 2 A schematic diagram of the structure of an intelligent satellite mission collaborative scheduling system based on an atomic service architecture provided in an embodiment of the present invention; Figure 3 This is a signaling interaction timing diagram for a fault-tolerant scenario in an embodiment of the present invention; Figure 4 This is a schematic diagram of the system implementation environment in an embodiment of the present invention. Detailed Implementation

[0018] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0019] like Figure 1 and Figure 2 As shown, an intelligent satellite mission collaborative scheduling system based on an atomic service architecture includes: The mission parsing module is used to receive satellite mission requirements and parse the target parameters, priorities, and constraints of the satellite mission.

[0020] Atomic Service Decomposition Module: The atomic service decomposition module is used to decompose satellite missions into multiple standardized atomic services with independent functions and clearly defined inputs and outputs.

[0021] Device Abstraction Module: The device abstraction module is used to abstract the hardware devices of heterogeneous satellites into standardized virtual devices and define the adaptation rules between virtual devices and atomic services.

[0022] Service Registry; The Service Registry is used to store atomic service metadata, virtual device metadata, and satellite physical resource data, and provides registration, status update, and matching query functions.

[0023] Intelligent scheduling engine; The intelligent scheduling engine is used to generate and dynamically adjust multi-satellite collaborative scheduling schemes based on satellite physical resource status data, atomic service metadata and virtual device adaptation rules through intelligent optimization algorithms.

[0024] The intelligent scheduling engine includes a state fusion unit, a decision optimization unit, and a scheme adaptation unit. The state fusion unit is used to perform noise reduction and normalization processing on the collected satellite hardware device status, virtual device status, and link status data. The decision optimization unit is used to load intelligent optimization algorithms, construct a multi-objective optimization model that includes device abstract adaptation constraints, and solve the scheduling scheme. The scheme adaptation unit is used to convert the multi-satellite collaborative scheduling scheme into an instruction format that the virtual device can recognize, supporting the dynamic adjustment of the multi-satellite collaborative scheduling scheme and the migration and adaptation of the virtual device.

[0025] The collaborative execution module is deployed on each satellite node. It is used to call satellite physical resources to execute atomic services and provide real-time feedback on execution status data.

[0026] Status awareness module; The status awareness module is used to collect data on the status of satellite hardware devices, virtual devices, links, and execution status.

[0027] A method for collaborative scheduling of intelligent satellite missions based on an atomic service architecture includes the following steps: S1. Receive satellite missions and analyze the mission's target parameters, mission priority, and constraints.

[0028] S2. The satellite mission is broken down into multiple standardized atomic services with independent functions and clearly defined inputs and outputs. Atomic services include payload atomic services, attitude control atomic services, communication atomic services, energy atomic services, and data processing atomic services. Atomic services are matched with virtual devices through standardized interfaces. An atomic service template library is preset during setup. During decomposition, the atomic service decomposition module matches the standard templates in the atomic service template library based on the parsed target parameters. Through functional deconstruction, complex tasks are broken down into smaller executable units.

[0029] S3. The device abstraction module abstracts the hardware devices of heterogeneous satellites into standardized virtual devices, and defines the adaptation rules between virtual devices and atomic services. The device abstraction module abstracts the hardware resources of heterogeneous satellites into standardized virtual devices, including the following steps: S31. Collect hardware parameters of heterogeneous satellites. The hardware includes payload equipment, attitude control equipment, communication equipment, energy equipment, and data processing equipment.

[0030] S32. Standardize the core functions and interfaces of similar hardware devices and generate corresponding virtual device templates.

[0031] S33. Define the calling interface, resource requirement mapping relationship and fault compatibility mechanism of virtual devices to achieve decoupling of atomic services and hardware devices.

[0032] S4. The service registry registers, stores, and manages the metadata of atomic services, virtual devices, and satellite physical resources. This metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. The service registry's state management includes the following steps: S41. Receive registration requests from atomic services and virtual devices, and verify the integrity of metadata.

[0033] S42. Real-time updates of atomic service availability, virtual device adaptation status, satellite physical resource status, and link communication quality.

[0034] S43. Respond to the query request from the intelligent scheduling engine and return the correspondence between atomic services, virtual devices, and satellite physical resources that meet the adaptation rules. The correspondence is determined based on the adaptation rules between atomic services and virtual devices and the mapping relationship between virtual devices and satellite physical resources.

[0035] S5. A multi-satellite collaborative scheduling scheme is generated by the intelligent scheduling engine using intelligent optimization algorithms. The intelligent optimization algorithm is one or more combinations of deep reinforcement learning algorithms, improved genetic algorithms, and particle swarm optimization algorithms. The objective function of the intelligent optimization algorithm includes task completion rate, resource utilization rate, energy consumption cost, execution latency, and virtual device adaptation success rate. The multi-satellite collaborative scheduling scheme is generated by the intelligent scheduling engine based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities. The satellite physical resource status data includes hardware device working status, remaining energy, storage capacity, attitude control accuracy, link bandwidth, and virtual device occupancy. The multi-satellite collaborative scheduling scheme includes the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources.

[0036] S6. Each satellite, according to the multi-satellite collaborative scheduling scheme, calls physical resources through the virtual device interface to collaboratively execute the corresponding atomic services and provides real-time feedback on execution data. When collaboratively executing the corresponding atomic services, each satellite achieves real-time interaction between atomic service execution data and virtual device status data through inter-satellite links. An adaptive link switching mechanism is adopted to dynamically switch between inter-satellite links and satellite-to-ground links to ensure data transmission stability. When the hardware device corresponding to the virtual device fails, the virtual device migration mechanism is triggered to quickly bind a backup hardware device.

[0037] S7. Dynamically adjust the multi-satellite collaborative scheduling scheme based on execution data to optimize subsequent execution.

[0038] The following section provides a more detailed description of the cooperative scheduling method in conjunction with the system: S1. Receive satellite missions from the control center and other sources through the mission parsing module, and parse the mission's target parameters, mission priority, and constraints.

[0039] S2. The parsed satellite mission is split into multiple standardized atomic services with independent functions and clear inputs and outputs through the atomic service splitting module. The atomic services are matched with the virtual device through standardized interfaces. The atomic service template library is preset during setup. During splitting, the atomic service splitting module matches the standard templates in the atomic service template library based on the parsed target parameters. Through functional deconstruction, the complex task is split into smaller executable units.

[0040] S3. The device abstraction module abstracts the hardware devices of heterogeneous satellites into standardized virtual devices, and defines the adaptation rules between virtual devices and atomic services. The device abstraction module abstracts the hardware resources of heterogeneous satellites into standardized virtual devices, including the following steps: S31. Collect hardware parameters of heterogeneous satellites. The hardware includes payload equipment, attitude control equipment, communication equipment, energy equipment, and data processing equipment.

[0041] S32. Standardize the core functions and interfaces of similar hardware devices and generate corresponding virtual device templates.

[0042] S33. Define the calling interface, resource requirement mapping relationship and fault compatibility mechanism of virtual devices to achieve decoupling of atomic services and hardware devices.

[0043] S4. The service registry registers, stores, and manages the metadata of atomic services, virtual devices, and satellite physical resources. This metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. The service registry's state management includes the following steps: S41. Receive registration requests from atomic services and virtual devices, and verify the integrity of metadata.

[0044] S42. Real-time updates of atomic service availability, virtual device adaptation status, satellite physical resource status, and link communication quality.

[0045] S43. Respond to the query request from the intelligent scheduling engine and return the correspondence between atomic services, virtual devices, and satellite physical resources that meet the adaptation rules. The correspondence is determined based on the adaptation rules between atomic services and virtual devices and the mapping relationship between virtual devices and satellite physical resources.

[0046] S5. A multi-satellite collaborative scheduling scheme is generated by the intelligent scheduling engine using an intelligent optimization algorithm. The intelligent optimization algorithm uses a deep reinforcement learning algorithm. The multi-satellite collaborative scheduling scheme is generated by the intelligent scheduling engine based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities.

[0047] The state fusion unit performs noise reduction and normalization processing on the collected satellite hardware status, virtual device status and link status data; the decision optimization unit loads intelligent optimization algorithms, constructs a multi-objective optimization model including device abstract adaptation constraints and solves the scheduling scheme to obtain a multi-satellite collaborative scheduling scheme; the scheme adaptation unit is used to convert the multi-satellite collaborative scheduling scheme into an instruction format that can be recognized by the virtual device.

[0048] S6, the collaborative execution module guides each satellite to collaboratively execute corresponding atomic services by calling physical resources through the virtual device interface according to the multi-satellite collaborative scheduling scheme, and provides real-time feedback of execution data; when each satellite collaboratively executes the corresponding atomic service, it realizes real-time interaction between atomic service execution data and virtual device status data through inter-satellite links, and adopts an adaptive link switching mechanism to dynamically switch between inter-satellite links and satellite-to-ground links to ensure data transmission stability; when the hardware device corresponding to the virtual device fails, the virtual device migration mechanism is triggered to quickly bind a backup hardware device.

[0049] S7. The intelligent scheduling engine dynamically adjusts the multi-satellite collaborative scheduling scheme based on the execution data. The adjustment first compares the execution data with the initial multi-satellite collaborative scheduling scheme to see if there are any differences. If there are differences, the intelligent optimization algorithm is called again to adjust the scheduling scheme. Then, the adjusted scheduling scheme is sent back to the collaborative execution module to guide each satellite to continue to execute according to the new scheduling scheme, forming an integrated closed loop of execution, feedback and adjustment.

[0050] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0051] Example 1 This embodiment provides a basic implementation scheme for an intelligent satellite mission collaborative scheduling method and system based on an atomic service architecture. This scheme achieves flexible and efficient collaborative scheduling of satellite network resources by standardizing and decomposing complex satellite missions and standardizing and abstracting heterogeneous satellite hardware resources.

[0052] Please see Figure 4 This figure illustrates a typical system implementation environment in an embodiment of the present invention. This system environment mainly includes a ground control center and an on-orbit satellite network. The ground control center, as the decision-making core of the entire scheduling system, deploys the core service modules of this embodiment, such as a task parsing module, an atomic service decomposition module, a device abstraction module, a service registration center, and an intelligent scheduling engine as the decision-making core. The satellite network consists of multiple heterogeneous on-orbit satellites, such as satellite A, satellite B, and satellite C. It is understood that these satellites may come from different manufacturers and carry physical hardware with varying performance and interfaces (such as different models of cameras, antennas, thrusters, etc.). The ground control center communicates with each satellite in the satellite network via a space-to-ground link, issuing commands and receiving data. Simultaneously, satellites can exchange data and communicate collaboratively at high speed via inter-satellite links. A technical feature of this invention is that on each satellite, a virtual device layer encapsulates and abstracts the underlying physical hardware, thereby shielding upper-layer applications from hardware differences and providing a unified, standardized calling interface.

[0053] The following will combine Figure 1 and Figure 2 The working process of this embodiment will be explained in detail using a specific task scenario. Figure 1 This illustrates the method flow followed in this embodiment. Figure 2 This demonstrates the system functional module architecture for implementing this method.

[0054] Suppose a ground control center receives a high-priority, complex mission request: "Conduct continuous optical surveys of a country's border (which consists of a polygonal region defined by a series of geographic coordinate points)." This mission must be completed within the next 24 hours, with an imaging resolution better than 1 meter.

[0055] In step S1, the task parsing module deployed at the ground control center receives the task request. This module is used to parse unstructured or semi-structured task descriptions and extract structured task parameters. In this scenario, the task parsing module will parse the following key information: the task type is "area coverage detection," the target area is a polygon composed of a set of latitude and longitude coordinates, the constraints are "resolution < 1 meter" and "completion time < 24 hours," and the task priority is "high." This structured information will be passed to subsequent modules for processing.

[0056] Subsequently, in step S2, the atomic service splitting module receives the parsed task information and, based on a preset rule base, splits this complex detection task into a sequence or directed acyclic graph composed of multiple standardized atomic services. It should be noted that an atomic service is the smallest functional unit constituting a task, possessing independent functionality and clearly defined inputs and outputs. Specifically, in this scenario, the detection mission may be broken down into the following atomic services: 1. Multiple "attitude maneuvering" atomic services, used to adjust the satellite platform's orientation towards the target area. Input parameters may include target attitude quaternions, maximum angular velocity, etc., and the output is the maneuver completion status; 2. Multiple "designated area imaging" atomic services, used to control the optical payload to photograph a designated sub-region. Input parameters may include the coordinates of the target sub-region, required resolution, exposure time, etc., and the output is the generated raw image data; 3. One or more "inter-satellite data transfer" atomic services, used to transfer data between satellites. Input parameters may include the target satellite identifier, data blocks to be transferred, transfer priority, etc., and the output is a successful transfer confirmation; 4. One or more "high-speed data downlink" atomic services, used to transmit the acquired data back to the ground. Input parameters may include the target ground station identifier, data blocks to be transferred, etc., and the output is a successful downlink confirmation. Each atomic service follows a standardized interface definition, such as using an Extensible Markup Language (Extensible Markup Language) or JSON format to describe its functionality, inputs, outputs, and required resources. This standardized breakdown allows different tasks to be built by reusing and combining these atomic services, thereby greatly improving task reusability and scheduling flexibility.

[0057] In step S3, the virtual device abstraction process is performed. This step can be performed before task execution or during system initialization. The satellite network in this embodiment includes two heterogeneous satellites: Satellite A and Satellite B. Satellite A carries a high-performance optical camera with a maximum resolution of 0.8 meters, but its downlink data transmission antenna has a low bandwidth of only 100 Mbps. Satellite B carries a conventional camera with a resolution of 2 meters, but it is equipped with a high-speed phased array antenna, achieving a downlink bandwidth of up to 1 Gbps. The device abstraction module is used to shield these differences in physical hardware. This module defines a standard virtual device template for different types of hardware (such as cameras, antennas, and attitude control systems). For example, the "virtual camera" template can define standard methods such as capture(area, resolution) and set_exposure(time), as well as standard attributes such as max_resolution and swath_width. Accordingly, the device abstraction module will create an instance named "Virtual Camera-A" for the physical camera of satellite A and set its max_resolution attribute to 0.8 meters; simultaneously, it will create an instance named "Virtual Camera-B" for the physical camera of satellite B and set its max_resolution attribute to 2 meters. Similarly, virtual device instances such as "Virtual Data Transmission-A," "Virtual Data Transmission-B," "Virtual Attitude Control-A," and "Virtual Attitude Control-B" will also be created. The core of this step lies in defining the adaptation rules between virtual devices and atomic services. For example, the "Specified Area Imaging" atomic service can be executed by any virtual device of the "Virtual Camera" type.

[0058] In step S4, the metadata of all created atomic services and virtual device instances, as well as the physical resource status of the satellite, are registered with the service registry 40 for unified management. The service registry can be implemented as a dynamic database or directory service to store information on all available "capabilities" (atomic services) and "resources" (virtual devices and their associated physical resources) in the system. For example, it records: "Virtual Camera-A" belongs to satellite A, its compatible physical camera model is "Model-X," its current status is "available," and the real-time physical resource status of its associated satellite A (collected and reported by the status awareness module) is: 85% remaining battery power, 256GB of storage remaining, and the six elements of its current orbit, etc. Simultaneously, it also records the metadata of atomic services such as "imaging a specified area," including function descriptions and standard input / output parameter formats. By updating this status information in real time, the service registry provides the intelligent scheduling engine with a comprehensive resource view required for decision-making.

[0059] As a core component, in step S5, the intelligent scheduling engine generates a scheduling scheme. The intelligent scheduling engine obtains the task requirements (high priority, 1-meter resolution) from the task parsing module and then queries the service registry for resources that meet the task requirements. This query process may include: "finding a 'virtual camera' capable of performing the 'specified area imaging' atomic service and with a resolution attribute better than 1 meter." Based on the metadata it manages, the service registry returns "Virtual Camera-A" as the only virtual device that meets the conditions. However, the bandwidth attribute of "Virtual Data Transmission-A" associated with "Virtual Camera-A" cannot meet the requirement of transmitting a large amount of data within the task time window. At this point, the decision optimization unit of the intelligent scheduling engine will activate an intelligent optimization algorithm (in this embodiment, as an optional implementation, an improved genetic algorithm can be used) to find an optimal solution for multi-satellite collaboration. The objective function of this algorithm is set to comprehensively consider task completion, total energy consumption, and execution time, i.e., seeking the highest cost-effectiveness ratio under the constraints. The algorithm's chromosome can be encoded as a scheduling scheme, i.e., a sequence containing a triple of (atomic service, virtual device, execution timestamp). Through multiple generations of evolution, crossover, and mutation, the algorithm eventually converges to an optimal or suboptimal multi-satellite collaborative scheduling scheme. This scheme clarifies the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources. In this scenario, the generated scheme may be as follows: 1. At time T1: Satellite A's "Virtual Attitude Control-A" performs the "Attitude Maneuver" service, aligning with the first target sub-region; 2. At time T2: Satellite A's "Virtual Camera-A" performs the "Designated Area Imaging" service, acquiring high-resolution images; 3. At time T3: Satellite A, through inter-satellite links, performs the "Inter-Satellite Data Transmission" service, transmitting image data to Satellite B with a suitable orbital position; 4. At time T4: Satellite B's "Virtual Data Transmission-B" performs the "High-Speed ​​Data Downlink" service, efficiently transmitting the received data back to the ground control center; 5. Repeat the above process until all target sub-regions are covered.

[0060] After the plan is generated, in step S6, the system performs collaborative execution and feedback. The intelligent scheduling engine breaks down the above plan and distributes it to satellites A and B respectively via the space-ground link. The collaborative execution module on each satellite is responsible for interpreting and executing the task segment assigned to it. For example, at time T2, the collaborative execution module of satellite A will call the capture method of the local "virtual camera-A". At this time, the virtual device layer will translate this standard call into specific and proprietary control instructions for its underlying "Model-X" physical camera, thereby driving the physical hardware to complete the image capture. During the execution process, the status awareness modules deployed on each satellite will continuously monitor the task execution progress (such as the imaged area), resource consumption (such as the rate of power decline and the increase in storage usage), and equipment health status, and feed this execution data back to the intelligent scheduling engine of the ground control center in real time.

[0061] Finally, in step S7, the system can dynamically adjust based on execution feedback. In this embodiment, if the task executes smoothly, the intelligent scheduling engine will confirm that the task is completed as planned based on the feedback execution data. This real execution data (such as actual energy consumption and maneuver time) will be used to update the resource model in the service registry, for example, to correct the estimated energy consumption model of the "attitude maneuver" service, thereby making subsequent scheduling decisions more accurate. Such a closed-loop process from perception, decision-making, execution to feedback ensures that the system can continuously self-optimize.

[0062] In this embodiment, the system utilizes the high-resolution imaging capability of satellite A and the high-speed data transmission capability of satellite B to collaboratively complete complex tasks that a single satellite cannot efficiently accomplish. This achieves complementary advantages and optimized allocation of heterogeneous resources, fully demonstrating the beneficial effects of the present invention in improving scheduling flexibility and resource utilization.

[0063] Example 2 This embodiment, based on Embodiment 1, focuses on demonstrating the dynamic adjustment capability and fault tolerance mechanism of the present invention in the face of sudden hardware failures. This corresponds to Figure 1 The in-depth application of the dynamic adjustment step S7, and can be combined with Figure 3 The signaling interaction timing shown is used for understanding.

[0064] Following on from Example 1, during the collaborative execution and feedback step S6, it is assumed that at some point, satellite A is performing a "designated area imaging" atomic service. During this execution, a critical temperature sensor reading on its onboard "Model-X" physical camera rises sharply, exceeding a preset safety threshold (e.g., soaring from the normal 45°C to 70°C).

[0065] The status awareness module deployed on satellite A monitors telemetry parameters of various physical hardware devices in real time, including temperature. When it detects this abnormally high temperature data, it immediately determines that the hardware device is in a faulty or high-risk state. Subsequently, the status awareness module generates a fault signal S_Fault, containing the device identifier, fault type (temperature exceeding limit), fault value, and timestamp. This signal is urgently reported to the intelligent scheduling engine of the ground control center via the space-to-ground link. This process... Figure 3 The arrow S_Fault points from "Satellite A (SAT-A)" to "Intelligent Scheduling Engine (ISE)".

[0066] Upon receiving the fault signal S_Fault, the intelligent scheduling engine immediately triggers a dynamic adjustment and fault tolerance process. First, it updates the status information in the service registry, changing the status of "Virtual Camera-A" associated with the faulty physical camera from "available" or "occupied" to "faulty" to prevent subsequent scheduling tasks from being reassigned to this unreliable device.

[0067] Next, the core task of the intelligent scheduling engine is to resume task execution to ensure the continuity of the entire probing mission. To this end, it immediately initiates a resource query Q_Resource to the service registry, such as... Figure 3 The arrow pointing from "Intelligent Scheduling Engine (ISE)" to "Service Registry" is shown in the image. The purpose of this query is to find a backup resource to replace "Virtual Camera-A". The query request can be formalized as: "Find 'Virtual Camera' instances that are currently 'available', capable of performing the 'Specified Area Imaging' atomic service, and whose orbital positions can cover the remaining target area within the next 30 minutes, sorted in descending order of resolution performance."

[0068] Upon receiving a query, the service registry performs a rapid match within its registered resource list. Assume that at this moment, another satellite C in the satellite network (a satellite that was previously performing other low-priority tasks or was in standby mode) has just entered the vicinity of the mission area. Satellite C's device abstraction module had previously created and registered a "virtual camera-C" instance for its onboard medium-resolution camera (e.g., physical resolution of 1.5 meters). The service registry finds that "virtual camera-C" meets the query criteria: its status is "available," it can perform imaging services, and its orbit is reachable. Although its 1.5-meter resolution is inferior to satellite A's 0.8 meters, it is still within the acceptable minor parameters for the mission (e.g., the mission requires better than 2 meters). Therefore, the service registry returns a resource response R_Resource containing complete information about "virtual camera-C" and its parent satellite C to the intelligent scheduling engine, such as... Figure 3 The dashed arrow in the middle represents the data returned from the "Service Registry" to the "Intelligent Scheduling Engine (ISE)".

[0069] Once available alternative resources are acquired, the intelligent scheduling engine's decision optimization unit immediately performs a task replanning. It remaps all unexecuted "designated area imaging" atomic services allocated to "virtual camera-A" in the original scheduling scheme to "virtual camera-C". Simultaneously, it generates new "attitude maneuver" atomic service instructions for satellite C, precisely targeting the new target sub-region. Understandably, this replanning process also considers switching costs, such as the additional energy consumption and time required for satellite C to perform attitude maneuvers. Ultimately, a dynamically adjusted, entirely new multi-satellite collaborative scheduling scheme is generated.

[0070] Subsequently, the intelligent scheduling engine's scheme adaptation unit converts the portion of the new scheme related to satellite C into a specific executable instruction, namely the new instruction C_New. This instruction is sent to satellite C via the space-to-ground link, such as... Figure 3 The arrow pointing from "Intelligent Scheduling Engine (ISE)" to "Satellite C (SAT-C)" is shown in the image.

[0071] Upon receiving the new instruction C_New, the collaborative execution module on satellite C immediately begins execution. It invokes the local "Virtual Attitude Control-C" and "Virtual Camera-C" interfaces to complete attitude adjustment and imaging tasks. After the task begins execution, its state awareness module sends an execution feedback signal F_Exec to the intelligent scheduling engine to confirm that the task has been successfully executed. Figure 3 The dashed arrow in the middle indicates the return from "Satellite C (SAT-C)" to "Intelligent Scheduling Engine (ISE)".

[0072] In this embodiment, the process of dynamically and automatically migrating atomic services from a faulty virtual device to another healthy virtual device of the same type is called the virtual device migration mechanism. Through this mechanism, the entire complex collaborative detection task can continue to be completed with only a small loss of time delay and data quality (resolution reduced from 0.8 meters to 1.5 meters) caused by fault switching. This fully demonstrates that the "perception-feedback-adjustment" closed-loop control system constructed by the present invention can guarantee the continuity and success rate of the entire collaborative task in the face of a single point of hardware failure, exhibiting high robustness and fault tolerance.

[0073] Furthermore, as a preferred implementation, the system can also utilize an adaptive link switching mechanism to ensure data transmission stability during collaborative execution. For example, in the scenario of Example 1, if the inter-satellite link suffers quality degradation due to space environment interference while satellite A is transmitting data to satellite B, the status awareness module will report the link status. During dynamic adjustments, the intelligent scheduling engine may decide to suspend inter-satellite transmission by satellite A and wait until it flies past a friendly ground station before directly transmitting the data via a higher-quality space-to-ground link. This flexible switching between different types of links further enhances the system's reliability.

[0074] Example 3 This embodiment will delve into the advanced decision-making capabilities of the intelligent scheduling engine in handling more complex, multi-task competition scenarios, particularly its specific implementation when employing deep reinforcement learning algorithms. In this embodiment, the internal structure of the intelligent scheduling engine is as follows: Figure 2 As shown, it may include a state fusion unit, a decision optimization unit, and a scheme adaptation unit.

[0075] Imagine a more complex business scenario: Within the same time window, the system needs to handle three concurrent tasks with different priorities: Task T1: Highest priority. Emergency single-point target observation of a sudden event (such as a volcanic eruption), requiring at least one high-quality image to be acquired within 15 minutes. Task T2: Medium priority. Capturing meteorological cloud images of a large sea area to support flight route planning, requiring coverage of more than 70% of the area within 6 hours. Task T3: Low priority. Utilizing idle time to transmit routine remote sensing data stored on multiple satellites.

[0076] At this point, there are 5 heterogeneous satellites available in the satellite network, but their resources such as energy, computing power, storage space, available payload, and communication link bandwidth are all limited, and there is obvious resource competition and conflict between atomic services performing different tasks.

[0077] In this scenario, the intelligent scheduling engine operates as follows: The state fusion unit first acts as an information processing center, periodically (e.g., every 10 seconds) collecting massive amounts of heterogeneous state information from the service registry and the state awareness modules of all satellites. This information includes, but is not limited to: 1. Real-time orbital parameters and predicted orbits for all satellites; 2. Physical resource status of each satellite: remaining power (unit: watt-hours), available storage capacity (unit: gigabytes), CPU load, and health status of each physical hardware component (such as temperature and voltage); 3. Status of all virtual devices: such as whether "Virtual Camera-A" is currently occupied and its cooling status; 4. Quality of all communication links: including real-time bandwidth, signal-to-noise ratio, and latency of space-to-ground and inter-satellite links; 5. Queues of atomic services to be executed for all current tasks: For example, task T1 has one “urgent imaging” atomic service to be assigned, T2 has dozens of “area scan” atomic services to be assigned, and T3 has hundreds of “data return” atomic services to be assigned.

[0078] The state fusion unit preprocesses this raw data, including data cleaning, noise reduction, normalization (e.g., normalizing all power values ​​to the [0, 1] interval), and feature engineering (e.g., calculating the visibility window to the ground over a future period based on orbital features). Finally, it integrates all the information into a fixed-length, high-dimensional state vector S. This state vector S provides a comprehensive and quantitative description of the current satellite network resources and mission status and will serve as input to the deep reinforcement learning model.

[0079] The processed state information is transmitted to the decision optimization unit, which serves as the decision-making core of the intelligent scheduling engine. In this embodiment, this unit loads a deep reinforcement learning model pre-trained through extensive simulations, such as a proximal policy optimization model based on the Actor-Critic framework. The model operates as follows: State space: the set of high-dimensional state vectors S generated by the state fusion unit.

[0080] Action Space: Defines all possible decisions the scheduling engine can make at each decision moment. This is a huge, combinatorial, discrete action space. An "action" A can be defined as "assigning an atomic service to be executed to a currently available and compatible virtual device for execution".

[0081] Reward Function: The design of the reward function is crucial, as it guides the reinforcement learning model to learn the desired scheduling behavior. In this embodiment, the reward function R is designed as a sophisticated multi-objective weighted function to evaluate the selected action at each decision time t. Advantages and disadvantages:

[0082] in: – They represent the actions to be performed. Then, the incremental completion rate of tasks T1, T2, and T3. For example, completing an emergency imaging task... It will get a large positive value.

[0083] – It is the weighting coefficient for task priority, and (For example, it can be set to) This ensures that the model will prioritize completing high-priority tasks.

[0084] – It is the normalized value of the total resource cost consumed in performing the action, such as the weighted sum of power consumption and storage consumption. It is the weight of cost penalty.

[0085] – It is a huge negative penalty term. If the deadline for task T1 is approaching and the model has not yet completed it, this penalty term will be triggered, thus guiding the model to avoid timeouts for high-priority tasks at all costs.

[0086] During the decision-making process, the decision optimization unit inputs the current state vector S into the policy network (Actor) of the near-end policy optimization model. This network outputs a probability distribution for all possible actions A in the current state. This means the model evaluates the long-term value of allocating each pending atomic service to each available virtual device. For example, the model might calculate that while immediately having satellite E perform the T2 scan task would yield some short-term rewards, having satellite E first perform an attitude maneuver to prepare for the T1 emergency imaging task, which can only begin 10 minutes later, would be a better choice in terms of long-term cumulative rewards. The engine samples or selects the action with the highest probability based on this probability distribution as the decision for the current moment.

[0087] The abstract decisions output by the decision optimization unit are then transformed into specific, executable scheduling instructions by the scheme adaptation unit. These instructions are subsequently sent to the corresponding satellites for execution.

[0088] Through this continuous cycle of "state awareness-decision-execution-reward feedback," the intelligent scheduling engine based on deep reinforcement learning can autonomously learn and execute a scheduling strategy that maximizes long-term overall benefits in a dynamic, highly competitive, and complex environment. It can not only handle task priorities but also predict future resource availability and make forward-looking decisions. Its comprehensive benefits in terms of task completion rate, resource utilization, and system response speed far surpass scheduling methods based on fixed rules or traditional heuristic optimization algorithms.

[0089] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features of the present invention can be arbitrarily combined with each other.

Claims

1. A method for collaborative scheduling of intelligent satellite missions based on an atomic service architecture, characterized in that, Including the following steps: S1. Receive and analyze satellite missions; S2. Break down the satellite mission into multiple standardized atomic services with independent functions and clearly defined inputs and outputs; S3. Abstract the hardware of heterogeneous satellites into standardized virtual devices and define the adaptation rules between virtual devices and atomic services; S4. Register, store, and manage the metadata of atomic services, virtual devices, and satellite physical resources. The metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. S5. Generate a multi-satellite collaborative scheduling scheme. The multi-satellite collaborative scheduling scheme is generated based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities. The multi-satellite collaborative scheduling scheme includes the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources. S6. Each satellite, according to the multi-satellite collaborative scheduling scheme, calls physical resources through the virtual device interface to collaboratively execute the corresponding atomic service and provides real-time feedback of execution data; S7. Dynamically adjust the multi-satellite collaborative scheduling scheme based on execution data.

2. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, The atomic services include payload atomic services, attitude control atomic services, communication atomic services, energy atomic services, and data processing atomic services; The atomic service is matched with the virtual device through a standardized interface.

3. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, Step S3, which abstracts the hardware of heterogeneous satellites into standardized virtual devices, includes the following steps: S31. Collect hardware equipment parameters of heterogeneous satellites, wherein the hardware equipment includes payload equipment, attitude control equipment, communication equipment, energy equipment and data processing equipment; S32. Standardize the core functions and interfaces of similar hardware devices and generate corresponding virtual device templates; S33. Define the calling interface, resource requirement mapping relationship and fault compatibility mechanism of virtual devices to achieve decoupling of atomic services and hardware devices.

4. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 3, characterized in that, The state management described in step S4 includes the following steps: S41. Receive registration requests from atomic services and virtual devices, and verify the integrity of metadata; S42. Real-time updates of atomic service availability, virtual device adaptation status, satellite physical resource status, and link communication quality; S43. Respond to the query request from the intelligent scheduling engine and return the correspondence between atomic services, virtual devices and satellite physical resources that meet the adaptation rules. The correspondence is determined based on the adaptation rules between atomic services and virtual devices and the mapping relationship between virtual devices and satellite physical resources.

5. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, Step S5 uses an intelligent optimization algorithm to generate a multi-satellite collaborative scheduling scheme. The intelligent optimization algorithm is one or more combinations of deep reinforcement learning algorithm, improved genetic algorithm and particle swarm optimization algorithm. The objective function of the intelligent optimization algorithm includes task completion rate, resource utilization rate, energy consumption cost, execution latency and virtual device adaptation success rate.

6. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, The satellite physical resource status data includes the working status of hardware devices, remaining energy, storage capacity, attitude control accuracy, link bandwidth, and virtual device occupancy.

7. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, When each satellite is collaboratively executing its corresponding atomic service, it achieves real-time interaction between atomic service execution data and virtual device status data through inter-satellite links. An adaptive link switching mechanism is used to dynamically switch between inter-satellite links and satellite-to-ground links to ensure data transmission stability. When the hardware device corresponding to the virtual device fails, the virtual device migration mechanism is triggered to bind a backup hardware device.

8. The intelligent satellite mission collaborative scheduling method based on atomic service architecture according to claim 1, characterized in that, In step S1, the satellite mission is analyzed to obtain the satellite mission's target parameters, mission priority, and constraints.

9. A smart satellite mission collaborative scheduling system based on an atomic service architecture, characterized in that, include: Mission parsing module: Receives and parses satellite missions; Atomic service decomposition module: The satellite mission is decomposed into multiple standardized atomic services with independent functions and clearly defined inputs and outputs; Device abstraction module: Abstracts the hardware devices of heterogeneous satellites into standardized virtual devices, and defines the adaptation rules between virtual devices and atomic services; Service Registration Center: Registers, stores, and manages the metadata of atomic services, virtual devices, and satellite physical resources. The metadata includes service function descriptions, virtual device interface parameters, physical resource attributes, and execution constraints. Intelligent scheduling engine: Generates multi-satellite collaborative scheduling schemes. These schemes are generated based on satellite physical resource status data, atomic service metadata, virtual device adaptation rules, and task priorities. The multi-satellite collaborative scheduling schemes include the mapping relationship and execution sequence of atomic services, virtual devices, and satellite physical resources. Collaborative execution module: Deployed on each satellite, it calls physical resources through virtual device interfaces to collaboratively execute corresponding atomic services according to the multi-satellite collaborative scheduling scheme, and provides real-time feedback of execution data; The intelligent scheduling engine dynamically adjusts the multi-satellite collaborative scheduling scheme based on the execution data.

10. The intelligent satellite mission collaborative scheduling system based on atomic service architecture according to claim 9, characterized in that, The intelligent scheduling engine includes: The state fusion unit is used to perform noise reduction and normalization processing on the collected satellite hardware device state, virtual device state and link state data; The decision optimization unit is used to load intelligent optimization algorithms, construct a multi-objective optimization model that includes device abstract adaptation constraints, and solve the scheduling scheme. The scheme adaptation unit is used to convert multi-satellite collaborative scheduling schemes into instruction formats that can be recognized by virtual devices, supporting dynamic adjustment of multi-satellite collaborative scheduling schemes and migration adaptation of virtual devices.

Images