Method, system and equipment for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system

Abstract

The invention relates to a method, system and device for scheduling tasks of low-orbit constellations for dynamic targets in space under a fully distributed satellite system, belonging to the field of on-orbit mission planning technology for spacecraft, and solving the problems of high technical cost, low flexibility and low stability of low-orbit constellation task scheduling. The method of the invention comprises: using an autonomous distributed task scheduling method based on CCGA to solve the multi-satellite scheduling problem, decomposing the multi-satellite scheduling problem into a single-satellite scheduling problem, and solving the optimal scheduling solution for each satellite separately to form an overall optimal scheduling solution. The scheduling process consists of local optimization and global negotiation. Each satellite independently explores the scheduling problem of the satellite for local optimization. At the same time, a ring topology is used to form an inter-satellite communication network for global negotiation within the constellation. The observation effect of the scheduling solution is comprehensively evaluated by the weighted sum of the satellite contribution, the number of switching times and the satellite load. The task benefit evaluation index is introduced to assist in the optimization of the single-satellite scheduling solution. The invention is suitable for low-orbit constellation task scheduling.

Description

Technical Field

[0001] The present application relates to the technical field of on-orbit mission planning for spacecraft, and in particular to the task scheduling of dynamic space targets by low-orbit constellations in a fully distributed satellite system. Background Art

[0002] Low Earth Orbit (LEO) constellations provide status information on dynamic space targets. With their wide coverage and strong detection capabilities, LEO constellations can effectively monitor and track dynamic space targets. As the number of satellites in orbit continues to increase, constellation systems become increasingly complex. The high-speed motion of space targets also presents new challenges for constellation scheduling. Within a limited timeframe and specific constraints, the rational allocation and optimal scheduling of satellite resources are crucial for achieving precise positioning and tracking of space targets.

[0003] The key to mission scheduling lies in optimizing the complex allocation relationship between targets and resources, resolving target conflicts and resource competition. Mission information for low-orbit constellations is either recorded by ground stations or generated onboard. Because space targets are dynamic and real-time, and cover a vast geographical area during flight, a single satellite has a limited field of view. To continuously observe a target over time, multiple satellites are required to relay detection.

[0004] Traditional satellite mission scheduling relies on centralized ground-based management, with ground personnel coordinating and planning based on satellite status information and mission requirements. This approach requires ample communication time between the satellite and the ground, incurring high communication costs and consuming significant resources and time. With the continuous advancement of space technology, the capabilities of onboard processors are constantly increasing, demonstrating a trend towards onboard autonomy. Offloading task scheduling algorithms to onboard computers can improve emergency response capabilities and reduce satellite operations' reliance on the ground.

[0005] To effectively implement autonomous scheduling of onboard missions, it is necessary to establish a rapid response mechanism based on limited resource capabilities and design an algorithm for autonomous, collaborative multi-satellite task scheduling. Currently, a hierarchical structure is often employed, dividing the satellites in a constellation into two types, master and slave, according to their functions. The master satellite, as the core of the collaborative planning process, first analyzes the mission's urgency and the visibility of the observation target, decomposes the mission, and then facilitates inter-satellite information exchange to ensure that each slave satellite receives mission information in a timely and accurate manner. Slave satellites receive mission information from the master satellite and independently plan their own plans based on their own resources and environmental conditions. The master satellite then summarizes and evaluates these plans, optimizing the scheduling plan through continuous information exchange.

[0006] This approach places high demands on the master satellite's capabilities, requiring each satellite to communicate with it, resulting in excessive communication overhead and low system stability. A failure in the master satellite can lead to failure of the entire system. Currently, a fully distributed architecture without a central node is becoming the future trend in the development of multi-satellite collaborative systems. This architecture not only ensures the successful execution of onboard mission scheduling but also plays a vital role in improving the overall stability of the system. Summary of the Invention

[0007] The purpose of the present invention is to solve the problems of high cost, low flexibility and low stability of existing low-orbit constellation task scheduling technology, and to provide a task scheduling method, system and equipment for low-orbit constellations to dynamic targets in space in a fully distributed satellite system.

[0008] The present invention is achieved through the following technical solutions. On one hand, the present invention provides a task scheduling method for a low-orbit constellation to a dynamic target in space in a fully distributed satellite system. Resource allocation within the constellation is determined by negotiation among multiple satellites, and each satellite only plans its own scheduling plan. The method includes:

[0009] The satellite that receives the mission information is determined to be the mission initiator. Based on the ephemeris and target status information, the visibility conditions of all satellites and targets in the constellation are determined to obtain the visible time window.

[0010] According to the visible time window, the mission cycle is decomposed into a series of subtasks to form the mission set information; each subtask represents the observation of a target within the time period and determines the corresponding visible satellite resources;

[0011] The mission initiating satellite broadcasts the mission set information to the remaining satellites in the constellation that are visible to the target. Each satellite organizes the subtasks that it can perform into its own mission set.

[0012] According to the local satellite task set, each satellite selects subtasks with non-overlapping time windows to form the local satellite scheduling plan, and during the task initialization process, generates a series of scheduling plans to form the local satellite scheduling plan set;

[0013] By integrating the scheduling plans of all satellites, satellites that perform the same subtask are selected and rejected. The expected weight of these satellites performing the subtask is calculated based on the subtask priority and duration. The two satellites that are most suitable for performing the subtask are determined, and an overall constellation scheduling plan for the mission period is obtained.

[0014] The overall constellation scheduling plan is optimized through local optimization and global negotiation within the constellation, specifically including:

[0015] The co-evolutionary genetic algorithm is used to locally optimize the single-satellite scheduling scheme, specifically including:

[0016] Subpopulations representing different satellite solutions are distributed on the corresponding satellites, including a set of local satellite scheduling solutions for the corresponding satellites. The initial solution of the local satellite scheduling solution set is randomly generated according to the time window and is continuously updated using the iteration and optimization of the algorithm;

[0017] Evaluate each plan within the subpopulation and calculate the fitness value of the overall constellation scheduling plan obtained by integrating the plan with other satellite plans. The fitness value is obtained based on the mission gain MP and the observation effect EP. The mission gain MP is used to measure the number of successfully scheduled subtasks, and the observation effect EP is used to measure the comprehensive effect of the scheduling plan on target observation accuracy, sensor efficiency, and satellite payload.

[0018] According to the fitness value, the individual with the highest fitness value in each satellite subpopulation is used as the latest scheduling plan for the satellite;

[0019] Global negotiation and communication are carried out within the constellation. Each satellite continuously sends its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites.

[0020] When all satellite plans are no longer updated, the plan optimization phase ends and a multi-satellite mission scheduling plan is obtained; each satellite performs its mission according to its own satellite scheduling plan and observes the target.

[0021] Furthermore, a sun-synchronous near-circular orbit is adopted, and the constellation is evenly deployed at the same orbital altitude in a Walker-δ configuration.

[0022] Furthermore, the calculation formula of the fitness value is:

[0023] maxf=v1·MP+v2·EP

[0024] Wherein, v1 and v2 represent weight coefficients, v1+v2=1.

[0025] Furthermore, the calculation formula for task income MP is:

[0026]

[0027] Among them, STNum_do represents the number of successfully executed subtasks, that is, within the subtask time window, there are two satellites observing the target at the same time, and STNum_all represents the total number of subtasks.

[0028] Furthermore, the calculation formula of the observation effect EP is:

[0029] EP=ω1·sumDegree+ω2·SwitchNum+ω3·sumLoad

[0030] ω1+ω2+ω3=1

[0031] Among them, sumDegree represents the satellite contribution, SwitchNum represents the number of switches, and sumLoad represents the satellite load.

[0032] Furthermore, the global negotiation communication link adopts a ring topology structure and an inter-satellite negotiation strategy based on receiving, updating and sending.

[0033] Furthermore, the two satellites most suitable for performing the subtask are determined as follows:

[0034] When a satellite has multiple consecutive subtasks corresponding to each target, the subtasks of different targets are defined as conflicting subtasks. The expected weight of each conflicting subtask is calculated based on the priority and duration of the subtask, which represents the benefit of scheduling other tasks in the time window.

[0035] When multiple satellites can perform the same subtask, the satellite with the lowest expected weight is selected to perform the subtask.

[0036] In a second aspect, the present invention provides a task scheduling system for low-orbit constellations to dynamic space targets in a fully distributed satellite system, the system comprising:

[0037] The mission initiation module is used to determine the satellite that receives the mission information as the mission initiator. Based on the ephemeris and target status information, it determines the visibility conditions of all satellites and targets in the constellation to obtain the visible time window. Based on the visible time window, the mission cycle is decomposed into a series of subtasks to form the task set information. Each subtask represents the observation of a target within the time period and determines the corresponding visible satellite resources.

[0038] The mission broadcast module is used by the mission initiating satellite to broadcast the mission set information to the remaining satellites in the constellation that are visible to the target. Each satellite organizes the subtasks that it can perform into its own mission set;

[0039] The local satellite scheduling plan generation module is used to select subtasks with non-overlapping time windows for each satellite according to the local satellite task set to form the local satellite scheduling plan, and generate a series of scheduling plans to form the local satellite scheduling plan set during the task initialization process;

[0040] The solution integration and execution determination module is used to integrate the scheduling plans of all satellites, select and reject satellites that perform the same subtask, calculate the weighted expectation of these satellites performing the subtask based on the subtask priority and duration, determine the two satellites that are most suitable for performing the subtask, and obtain an overall constellation scheduling plan within the mission cycle;

[0041] The solution optimization module is used to optimize the overall constellation scheduling plan through local optimization and global negotiation within the constellation. Specifically, it includes:

[0042] The co-evolutionary genetic algorithm is used to locally optimize the single-satellite scheduling scheme, specifically including:

[0043] Subpopulations representing different satellite solutions are distributed on the corresponding satellites, including a set of local satellite scheduling solutions for the corresponding satellites. The initial solution of the local satellite scheduling solution set is randomly generated according to the time window and is continuously updated using the iteration and optimization of the algorithm;

[0044] Evaluate each plan within the subpopulation and calculate the fitness value of the overall constellation scheduling plan obtained by integrating the plan with other satellite plans. The fitness value is obtained based on the mission gain MP and the observation effect EP. The mission gain MP is used to measure the number of successfully scheduled subtasks, and the observation effect EP is used to measure the comprehensive effect of the scheduling plan on target observation accuracy, sensor efficiency, and satellite payload.

[0045] According to the fitness value, the individual with the highest fitness value in each satellite subpopulation is used as the latest scheduling plan for the satellite;

[0046] The global negotiation module is used for global negotiation and communication within the constellation. Each satellite continuously sends its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites.

[0047] The multi-satellite task scheduling scheme acquisition module is used to obtain the multi-satellite task scheduling scheme when all satellite schemes are no longer updated and the scheme optimization phase ends; each satellite performs its mission according to its own scheduling scheme and observes the target.

[0048] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the steps of a task scheduling method for low-orbit constellations to dynamic space targets in a fully distributed satellite system as described above are executed.

[0049] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein a plurality of computer instructions are stored in the computer-readable storage medium, and the plurality of computer instructions are used to enable a computer to execute a task scheduling method for low-orbit constellations to dynamic targets in space under a fully distributed satellite system as described above.

[0050] Beneficial effects of the present invention:

[0051] The present invention provides an optimization method for solving multi-satellite multi-objective scheduling problems in a distributed satellite system based on a co-evolutionary genetic algorithm. The method adopts a non-hierarchical fully distributed structure, so that each satellite has the same decision-making ability, and ensures the consistency of information cognition between nodes through inter-satellite negotiation. By mobilizing all nodes to cooperate with each other, the overall planning of the entire system is carried out. This structure has higher flexibility and stability, is more reliable in actual operation, and will not cause the problem of paralysis of the entire system due to the failure of a satellite, thereby greatly improving the flexibility, stability and scalability of the system.

[0052] For low-orbit constellations observing dynamic targets in space, this paper optimizes the task scheduling problem using a fully distributed approach with decentralized nodes, effectively enabling autonomous onboard scheduling for rapid response missions. The basic concept is to transform the multi-satellite, multi-target task scheduling problem into a series of single-satellite scheduling problems. Each satellite independently explores its own scheduling solution and then integrates it into an overall scheduling solution. An optimization model is established under constraints, using the number of completed subtasks, satellite contribution, number of handoffs, and satellite payload as evaluation indicators. A collaborative evolutionary genetic algorithm (CCGA) is used to solve this optimization problem and obtain the optimal scheduling solution.

[0053] This invention addresses the problem of centralized methods, which rely heavily on central nodes and suffer from poor system stability. It employs an autonomous distributed task scheduling method based on CCGA to solve the multi-satellite scheduling problem. It decomposes the multi-satellite scheduling problem into a series of single-satellite scheduling problems, solving the optimal scheduling solution for each satellite separately to form an overall optimal scheduling solution.

[0054] The scheduling process consists of local optimization and global negotiation. Each satellite independently explores its own scheduling sub-problem for local optimization. Simultaneously, a ring-shaped intersatellite communication network is employed to facilitate global negotiation within the constellation, maintaining consistent information awareness. The observed effectiveness of the scheduling solution is comprehensively evaluated using a weighted sum of satellite contribution, number of handoffs, and satellite payload. Furthermore, because individual satellites have limited global information awareness, which can lead to unsuccessful scheduling of individual subtasks, the task benefit metric is introduced to assist in optimizing the scheduling solution for each satellite. Through continuous iteration, the optimal solution to the optimization problem is ultimately achieved. This algorithm leverages the characteristics of a fully distributed satellite system to perform distributed optimization of the scheduling problem, improving system stability and effectively reducing the number of satellite handoffs.

[0055] The present invention is applicable to the task scheduling of low-orbit constellations to dynamic space targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0057] Figure 1 This is a flowchart for multi-satellite autonomous distributed task scheduling;

[0058] Figure 2 This is the flow chart of intersatellite negotiation;

[0059] Figure 3 is the fitness value change corresponding to the centralized method;

[0060] Figure 4 is the change of satellite contribution corresponding to the centralized method;

[0061] Figure 5 The change in the number of satellite switching times corresponding to the centralized method;

[0062] Figure 6 is the satellite load change corresponding to the centralized method;

[0063] Figure 7 The fitness value change corresponding to the task scheduling method of the low-orbit constellation for space dynamic targets in a fully distributed satellite system according to the present invention;

[0064] Figure 8 The task benefit change corresponding to the task scheduling method of the present invention for low-orbit constellations to space dynamic targets in a fully distributed satellite system;

[0065] Figure 9 The observation effect changes corresponding to the task scheduling method of the low-orbit constellation for space dynamic targets in a fully distributed satellite system applicable to the present invention;

[0066] Figure 10 The satellite contribution changes corresponding to the task scheduling method of the low-orbit constellation for space dynamic targets in a fully distributed satellite system applicable to the present invention;

[0067] Figure 11 The number of switching changes corresponding to the task scheduling method of the low-orbit constellation for space dynamic targets in a fully distributed satellite system applicable to the present invention;

[0068] Figure 12 Satellite load changes corresponding to the task scheduling method of the present invention for low-orbit constellations to space dynamic targets in a fully distributed satellite system;

[0069] Figure 13 Schedule Gantt charts for tasks. DETAILED DESCRIPTION

[0070] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but are not to be construed as limiting the present invention.

[0071] Embodiment 1: A method for scheduling tasks for dynamic space targets in a low-orbit constellation in a fully distributed satellite system. Resource allocation within the constellation is determined by negotiation among multiple satellites, and each satellite only plans its own scheduling plan. The method includes:

[0072] The satellite that receives the mission information is determined to be the mission initiator. Based on the ephemeris and target status information, the visibility conditions of all satellites and targets in the constellation are determined to obtain the visible time window.

[0073] According to the visible time window, the mission cycle is decomposed into a series of subtasks to form the mission set information; each subtask represents the observation of a target within the time period and determines the corresponding visible satellite resources;

[0074] The mission initiating satellite broadcasts the mission set information to the remaining satellites in the constellation that are visible to the target. Each satellite organizes the subtasks that it can perform into its own mission set; the local mission set is the mission set of each satellite.

[0075] According to the local satellite task set, each satellite selects subtasks with non-overlapping time windows to form the local satellite scheduling plan, and during the task initialization process, generates a series of scheduling plans to form the local satellite scheduling plan set;

[0076] By integrating the scheduling plans of all satellites, satellites that perform the same subtask are selected and rejected. The expected weight of these satellites performing the subtask is calculated based on the subtask priority and duration. The two satellites that are most suitable for performing the subtask are determined, and an overall constellation scheduling plan for the mission period is obtained.

[0077] The overall constellation scheduling plan is optimized through local optimization and global negotiation within the constellation, specifically including:

[0078] The co-evolutionary genetic algorithm is used to locally optimize the single-satellite scheduling scheme, specifically including:

[0079] Subpopulations representing different star plans are distributed on the corresponding stars, including the set of local scheduling plans for the corresponding stars. This set is randomly generated according to the time window at the beginning of the algorithm, that is, the initial scheduling plan set (i.e., the initial solution), but it is continuously updated with the iteration and optimization of the algorithm.

[0080] Evaluate each plan within the subpopulation and calculate the fitness value of the overall constellation scheduling plan obtained by integrating the plan with other satellite plans. The fitness value is obtained based on the mission gain MP and the observation effect EP. The mission gain MP is used to measure the number of successfully scheduled subtasks, and the observation effect EP is used to measure the comprehensive effect of the scheduling plan on target observation accuracy, sensor efficiency, and satellite payload.

[0081] According to the fitness value, the individual with the highest fitness value in each satellite subpopulation is used as the latest scheduling plan for the satellite;

[0082] In this implementation, before evaluating fitness values, satellites within the constellation must communicate with each other to obtain the optimal scheduling solutions of other satellites (the initial communication process obtains the initial scheduling solutions of other satellites), forming the current knowledge. A higher fitness value indicates a better overall solution, combined with the solutions of other satellites. By comparing the fitness values ​​of each solution within the corresponding subpopulation, the optimal scheduling solution for each satellite under the current knowledge can be obtained.

[0083] Global negotiation and communication are carried out within the constellation. Each satellite continuously sends its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites.

[0084] When all satellite plans are no longer updated, the plan optimization phase ends and a multi-satellite mission scheduling plan is obtained; each satellite performs its mission according to its own satellite scheduling plan and observes the target.

[0085] This implementation adopts a divide-and-conquer strategy to reduce the complexity of global optimization. It addresses the centralized approach's strong dependence on central nodes and poor system stability by employing an autonomous distributed task scheduling method based on CCGA. The multi-satellite scheduling problem is decomposed into a series of individual satellite scheduling problems, with the optimal scheduling solution for each satellite being solved separately to form the overall optimal scheduling solution. The scheduling process consists of local optimization and global negotiation. Each satellite independently explores its own scheduling sub-problem for local optimization. The observed performance of the scheduling solution is comprehensively evaluated using a weighted sum of satellite contribution, number of handoffs, and satellite load. Furthermore, because individual satellites have limited global information, which can lead to unsuccessful scheduling of individual subtasks, task yield is introduced as an evaluation metric to assist in optimizing the individual satellite scheduling solution. Through continuous iteration, the optimal solution to the optimization problem is ultimately achieved. This algorithm leverages the characteristics of a fully distributed satellite system to perform distributed optimization of the scheduling problem, improving system stability and effectively reducing the number of satellite handoffs.

[0086] Implementation method 2: This implementation method further defines the task scheduling method for a low-orbit constellation to a dynamic space target in a fully distributed satellite system as described above. In this implementation method, the tasks of the low-orbit constellation to a dynamic space target are further defined, specifically including:

[0087] The constellation adopts a sun-synchronous near-circular orbit and is evenly deployed at the same orbital altitude in a Walker_δ configuration to achieve successive observation of multiple ballistic missile targets in mid-flight.

[0088] Implementation method 3: This implementation method further limits the task scheduling method for low-orbit constellations of dynamic space targets in a fully distributed satellite system as described above. In this implementation method, the fitness value is further limited, specifically including:

[0089] The calculation formula of the fitness value is:

[0090] maxf=v1·MP+v2·EP

[0091] Wherein, v1 and v2 represent weight coefficients, v1+v2=1.

[0092] This embodiment provides a method for calculating the fitness value, which can effectively evaluate the fitness of the overall constellation scheduling solution obtained by integrating the solution with other satellite solutions.

[0093] Implementation method 4: This implementation method further defines the task scheduling method for dynamic space targets in a low-orbit constellation under a fully distributed satellite system as described above. In this implementation method, the task profit MP is further defined, specifically including:

[0094] The calculation formula for mission income MP is:

[0095]

[0096] Among them, STNum_do represents the number of successfully executed subtasks, that is, within the subtask time window, there are two satellites observing the target at the same time, and STNum_all represents the total number of subtasks.

[0097] This embodiment provides a method for calculating task profit MP, which is used to improve the accuracy and rationality of fitness values.

[0098] Implementation method 5: This implementation method further defines the task scheduling method for low-orbit constellations of dynamic space targets in a fully distributed satellite system as described above. In this implementation method, the observation effect EP is further defined, specifically including:

[0099] The calculation formula of observation effect EP is:

[0100] EP=ω1·sumDegree+ω2·SwitchNum+ω3·sumLoad

[0101] ω1+ω2+ω3=1

[0102] Among them, sumDegree represents the satellite contribution, SwitchNum represents the number of switches, and sumLoad represents the satellite load.

[0103] This embodiment provides a calculation method for the observation effect EP, which is used to improve the accuracy and rationality of the fitness value.

[0104] Implementation method 6: This implementation method further limits the task scheduling method for low-orbit constellations of dynamic space targets in a fully distributed satellite system as described above. In this implementation method, the global negotiation communication is further limited, specifically including:

[0105] The global negotiation communication link adopts a ring topology structure and an inter-satellite negotiation strategy based on receiving, updating and sending.

[0106] This embodiment adopts a ring topology structure to form an inter-satellite communication network for global negotiation within the constellation to maintain consistency in information cognition.

[0107] Embodiment 7: This embodiment further defines the above-described method for scheduling tasks for dynamic targets in a low-orbit constellation in a fully distributed satellite system. In this embodiment, the method for determining the two satellites most suitable for executing the subtask is further defined, specifically including:

[0108] The two satellites most suitable for performing the subtask are determined as follows:

[0109] When a satellite has multiple consecutive subtasks corresponding to each target, the subtasks of different targets are defined as conflicting subtasks. The expected weight of each conflicting subtask is calculated based on the priority and duration of the subtask, which represents the benefit of scheduling other tasks in the time window.

[0110] When multiple satellites can perform the same subtask, the satellite with the lowest expected weight is selected to perform the subtask.

[0111] In this implementation, satellite Si may have a continuous subtask corresponding to target Tj. Similarly, there may be a series of subtasks corresponding to other targets. Subtasks for different targets may conflict in time. Therefore, satellite Si can only select one of these conflicting subtasks for execution. Satellite Si calculates a weighted expectation for each conflicting subtask based on its priority and duration. This represents the benefit of scheduling other tasks within that time window. Since a subtask can be executed by different satellites, when multiple satellites are willing to execute the subtask (i.e., all of these satellites have the subtask in their individual satellite scheduling plans), they can select the subtask based on the weighted expectation. The lower the weighted expectation, the more appropriate it is.

[0112] Embodiment 8: This embodiment is an example of a task scheduling method for a low-orbit constellation to a dynamic space target in a fully distributed satellite system as described above, specifically including:

[0113] 1) The present invention targets mission scenarios in which the constellation adopts a sun-synchronous near-circular orbit and is evenly deployed at the same orbital altitude in a Walker-δ configuration to achieve successive observation of multiple ballistic missile targets in mid-flight.

[0114] 2) The satellite that receives the mission information serves as the mission initiator. Combining the ephemeris and target status information, it determines the visibility conditions for all satellites and targets in the constellation, determining the visible time window. Based on the visible time window, the mission cycle is decomposed into a series of subtasks, forming a task set. Each subtask represents the observation of a target within that time period and identifies the corresponding visible satellite resources. The mission initiator then broadcasts the task set information to the remaining satellites in the constellation that are visible to the target (the set of working satellites). Each satellite organizes the executable subtasks into its own task set.

[0115] 3) Based on the constellation's mission set, each satellite selects subtasks with non-overlapping time windows to form its initial scheduling plan. By integrating the scheduling plans of all satellites, selecting and rejecting satellites willing to perform the same subtask, and retaining the two most suitable satellites for that subtask, an overall constellation scheduling plan for the mission cycle is obtained.

[0116] 4) The overall constellation scheduling plan is optimized through local optimization and global negotiation within the constellation. The task scheduling process is as follows: Figure 1 The number on the subtask represents the sequence of the observation target and its subtask, such as 103 is the third subtask of the first observation target, S1, S2, ..., S n are the satellite numbers respectively.

[0117] 5) The multi-satellite scheduling problem is transformed into a series of single-satellite scheduling problems, and CCGA is used to locally optimize the single-satellite scheduling solutions. Each subpopulation corresponds to a set of single-satellite scheduling solutions for a satellite. Each subpopulation evolves relatively independently to explore its own scheduling problem. Each individual solution within a subpopulation is evaluated, and the fitness of the overall constellation scheduling solution, obtained by integrating it with the solutions of other satellites, is calculated. This fitness value is composed of the mission benefit (MP) and the observation effect (EP).

[0118] maxf=v1·MP+v2·EP

[0119] Mission Profit (MP) is used to measure the number of successfully scheduled subtasks. STNum_do represents the number of successfully executed subtasks, meaning that two satellites simultaneously observed the target within the subtask's time window. STNum_all represents the total number of subtasks.

[0120]

[0121] The observation effect EP is used to comprehensively measure the overall effect of the scheduling scheme. It is the weighted sum of the satellite contribution sumDegree, the number of switches SwitchNum and the satellite load sumLoad.

[0122] EP=ω1·sumDegree+ω2·SwitchNum+ω3·sumLoad

[0123] ω1+ω2+ω3=1

[0124] 6) In the solution optimization phase, in order to assist in the fitness value evaluation process, global negotiation and communication are required within the constellation. The specific process is as follows: Figure 2 As shown in the figure, the communication link adopts a ring topology and an inter-satellite negotiation strategy based on receiving, updating, and sending. Each satellite continuously transmits its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites. When all satellite plans are no longer updated, the plan optimization phase ends, resulting in a task scheduling plan unanimously agreed upon by all satellites. Each satellite executes its mission and observes the target according to its own scheduling plan.

[0125] Finally, the technical effects of the present invention are given in conjunction with the accompanying drawings:

[0126] The mission's initial UTCG time is 2024-07-11 05:10:00. The Walker-δ constellation configuration is 24 / 3 / 2, and the six orbit numbers of the first satellite are shown in Table 1.

[0127] Table 1 Six numbers of the first satellite orbit

[0128]

[0129] Three targets to be observed are generated using STK software. The specific parameters are shown in Table 2.

[0130] Table 2 Target missile take-off and landing point parameters

[0131] Starting point latitude and longitude (°) Latitude and longitude of landing point (°) Goal 1 (42,124) (47,-122) Goal 2 (43,122) (37,-118) Goal 3 (45,121) (44,-119)

[0132] The centralized method is used for optimization, and the changes in fitness value and each optimization index are as follows: Figures 3 to 6 As shown. The optimization scheme of this application is used for simulation and solution. All satellites communicate with each other once as one round. The maximum number of rounds is g = 100. The number of generations for each satellite subpopulation to be updated through genetic operations in each round is Gen = 10. The weights of the objective function are: ω1 = 0.4, ω2 = 0.2, ω3 = 0.4, v1 = 0.7, v2 = 0.3. The optimization process is shown in Figures 7 to 12 The Gantt chart of the optimal scheduling solution obtained after 100 rounds of negotiation is shown in Figure 13 As shown in the figure, each square represents a subtask, and the row it occupies corresponds to the satellite executing that subtask. For example, 301 represents the first subtask for the third observation target, observed by the 11th and 12th satellites. Simulation results show that this distributed optimization method reaches convergence by the 60th iteration, with all tasks successfully executed. Compared to centralized optimization methods, it effectively reduces the number of sensor switching times.

Claims

1. A task scheduling method for low-orbit constellations to dynamic space targets in a fully distributed satellite system, characterized in that: Resource allocation within the constellation is determined by negotiation among multiple satellites, with each satellite only planning its own scheduling plan. The method includes: The satellite that receives the mission information is determined to be the mission initiator. Based on the ephemeris and target status information, the visibility conditions of all satellites and targets in the constellation are determined to obtain the visible time window. According to the visible time window, the mission cycle is decomposed into a series of subtasks to form the mission set information; each subtask represents the observation of a target within the time period and determines the corresponding visible satellite resources; The mission initiating satellite broadcasts the mission set information to the remaining satellites in the constellation that are visible to the target. Each satellite organizes the subtasks that it can perform into its own mission set. According to the local satellite task set, each satellite selects subtasks with non-overlapping time windows to form the local satellite scheduling plan, and during the task initialization process, generates a series of scheduling plans to form the local satellite scheduling plan set; By integrating the scheduling plans of all satellites, satellites that perform the same subtask are selected and rejected. The expected weight of these satellites performing the subtask is calculated based on the subtask priority and duration. The two satellites that are most suitable for performing the subtask are determined, and an overall constellation scheduling plan for the mission period is obtained. The overall constellation scheduling plan is optimized through local optimization and global negotiation within the constellation, specifically including: The co-evolutionary genetic algorithm is used to locally optimize the single-satellite scheduling scheme, specifically including: Subpopulations representing different satellite solutions are distributed on the corresponding satellites, including a set of local satellite scheduling solutions for the corresponding satellites. The initial solution of the local satellite scheduling solution set is randomly generated according to the time window and is continuously updated using the iteration and optimization of the algorithm; Evaluate each plan within the subpopulation and calculate the fitness value of the overall constellation scheduling plan obtained by integrating the plan with other satellite plans. The fitness value is obtained based on the mission gain MP and the observation effect EP. The mission gain MP is used to measure the number of successfully scheduled subtasks, and the observation effect EP is used to measure the comprehensive effect of the scheduling plan on target observation accuracy, sensor efficiency, and satellite payload. According to the fitness value, the individual with the highest fitness value in each satellite subpopulation is used as the latest scheduling plan for the satellite; Global negotiation and communication are carried out within the constellation. Each satellite continuously sends its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites. When all satellite plans are no longer updated, the plan optimization phase ends and a multi-satellite mission scheduling plan is obtained; each satellite performs its mission according to its own satellite scheduling plan and observes the target.

2. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 1, characterized in that: Using a sun-synchronous near-circular orbit, the constellation is evenly deployed at the same orbital altitude in a Walker-δ configuration.

3. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 1, characterized in that: The calculation formula of the fitness value is: maxf=v1·MP+v2·EP Wherein, v1 and v2 represent weight coefficients, v1+v2=1.

4. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 3, characterized in that: The calculation formula for mission income MP is: Among them, STNum_do represents the number of successfully executed subtasks, that is, within the subtask time window, there are two satellites observing the target at the same time, and STNum_all represents the total number of subtasks.

5. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 4, characterized in that: The calculation formula of observation effect EP is: EP=ω1·sumDegree+ω2·SwitchNum+ω3·sumLoad ω1+ω2+ω3=1 Among them, sumDegree represents the satellite contribution, SwitchNum represents the number of switches, and sumLoad represents the satellite load.

6. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 1, characterized in that: The global negotiation communication link adopts a ring topology structure and an inter-satellite negotiation strategy based on receiving, updating and sending.

7. The method for scheduling tasks of low-orbit constellations for dynamic space targets in a fully distributed satellite system according to claim 1, characterized in that: The two satellites most suitable for performing the subtask are determined as follows: When a satellite has multiple consecutive subtasks corresponding to each target, the subtasks of different targets are defined as conflicting subtasks. The expected weight of each conflicting subtask is calculated based on the priority and duration of the subtask, which represents the benefit of scheduling other tasks in the time window. When multiple satellites can perform the same subtask, the satellite with the lowest expected weight is selected to perform the subtask.

8. A task scheduling system for low-orbit constellations to dynamic space targets in a fully distributed satellite system, characterized by: The system comprises: The mission initiation module is used to determine the satellite that receives the mission information as the mission initiator. Based on the ephemeris and target status information, it determines the visibility conditions of all satellites and targets in the constellation to obtain the visible time window. Based on the visible time window, the mission cycle is decomposed into a series of subtasks to form the task set information. Each subtask represents the observation of a target within the time period and determines the corresponding visible satellite resources. The mission broadcast module is used by the mission initiating satellite to broadcast the mission set information to the remaining satellites in the constellation that are visible to the target. Each satellite organizes the subtasks that it can perform into its own mission set; The local satellite scheduling plan generation module is used to select subtasks with non-overlapping time windows for each satellite according to the local satellite task set to form the local satellite scheduling plan, and generate a series of scheduling plans to form the local satellite scheduling plan set during the task initialization process; The solution integration and execution determination module is used to integrate the scheduling plans of all satellites, select and reject satellites that perform the same subtask, calculate the weighted expectation of these satellites performing the subtask based on the subtask priority and duration, determine the two satellites that are most suitable for performing the subtask, and obtain an overall constellation scheduling plan within the mission cycle; The solution optimization module is used to optimize the overall constellation scheduling plan through local optimization and global negotiation within the constellation. Specifically, it includes: The co-evolutionary genetic algorithm is used to locally optimize the single-satellite scheduling scheme, specifically including: Subpopulations representing different satellite solutions are distributed on the corresponding satellites, including a set of local satellite scheduling solutions for the corresponding satellites. The initial solution of the local satellite scheduling solution set is randomly generated according to the time window and is continuously updated using the iteration and optimization of the algorithm; Evaluate each plan within the subpopulation and calculate the fitness value of the overall constellation scheduling plan obtained by integrating the plan with other satellite plans. The fitness value is obtained based on the mission gain MP and the observation effect EP. The mission gain MP is used to measure the number of successfully scheduled subtasks, and the observation effect EP is used to measure the comprehensive effect of the scheduling plan on target observation accuracy, sensor efficiency, and satellite payload. According to the fitness value, the individual with the highest fitness value in each satellite subpopulation is used as the latest scheduling plan for the satellite; The global negotiation module is used for global negotiation and communication within the constellation. Each satellite continuously sends its latest plan to its neighboring satellites and optimizes and updates its own plan based on the latest plans received from other satellites. The multi-satellite task scheduling scheme acquisition module is used to obtain the multi-satellite task scheduling scheme when all satellite schemes are no longer updated and the scheme optimization phase ends; each satellite performs its mission according to its own scheduling scheme and observes the target.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor runs the computer program stored in the memory, the steps of the method according to any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a plurality of computer instructions, and the plurality of computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images