Low-orbit large-scale remote sensing constellation task scheduling method oriented to ground multi-attention target continuous observation application

Abstract

The invention discloses a low-orbit large-scale remote sensing constellation task scheduling method for ground multi-attention target continuous observation application, which comprises the steps of S1, task allocation model construction: establishing a ground multi-attention target continuous observation task allocation model of a low-orbit large-scale remote sensing satellite constellation, the task distribution model describes the distribution process of observation tasks, stipulates conditions and target functions met by the optimal observation scheme, and comprehensively considers the maximum observation gap window constraint and various strong constraints, so that the overall task benefit value of the optimal observation scheme is maximum; and S2, solving the task allocation model: solving the task allocation model by using a solving algorithm formed by adding constraint check and a correction operator on the basis of a genetic algorithm to obtain an optimal observation scheme. The method has the advantages that the maximum observation gap is used as an index for measuring the continuity of the observation task according to the demand characteristics of the continuous observation task, the method has the advantage of considering both the observation time and the observation continuity, and the accuracy and efficiency of task solving are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of constellation mission planning, and in particular to a low-orbit large-scale remote sensing constellation mission scheduling method for continuous observation of multiple ground targets. Background Art

[0002] At present, the earth observation satellite system has shown a trend of satellite miniaturization and constellation clustering. With the increase in the number of constellation satellites and the increase in the types of imaging payloads, the constellation's coverage frequency of ground targets has increased, the revisit cycle has shortened, and the ability to complete earth observation tasks has been continuously enhanced. At the same time, with the continuous expansion of the application field of remote sensing technology, users' needs have become more diversified. For example, in disaster disposal and environmental monitoring, the need for real-time monitoring of forest fires has been proposed to reduce casualties and improve work efficiency. In order to achieve such needs, it is necessary to improve the ability of earth observation satellite constellations to continuously observe ground targets of interest for a period of time. However, on the one hand, in the field of imaging satellite task allocation, there are few studies on the allocation of tasks for continuous observation of ground targets by constellations, and no clear indicators for measuring the continuity of constellations in performing tasks for observing ground targets are given in the research; on the other hand, the continuous expansion of the constellation scale has led to an increasing amount of calculations for solving task allocation and low calculation efficiency. In view of this situation, it is urgent to study a task allocation model and solution method that uses multiple low-orbit remote sensing satellites to relay work to achieve continuous observation of multiple key targets, so as to improve the cost-effectiveness of low-orbit earth observation constellations in performing tasks for continuous observation of multiple key targets and improve continuity.

[0003] In summary, there is an urgent need to propose a task allocation model for the continuous observation of multiple ground targets by a remote sensing satellite constellation and a task solving method based on an intelligent optimization algorithm. Summary of the invention

[0004] The object of the present invention is to provide a low-orbit large-scale remote sensing constellation task scheduling method for continuous observation of multiple ground targets, thereby solving the above-mentioned problems existing in the prior art.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A low-orbit large-scale remote sensing constellation task scheduling method for continuous observation of multiple ground-based targets includes the following steps:

[0007] S1. Task allocation model construction:

[0008] A task allocation model for continuous observation of multiple ground targets by a large-scale remote sensing satellite constellation in low orbit is established. The task allocation model describes the observation task allocation process, specifies the conditions and objective functions that the optimal observation scheme must satisfy, and comprehensively considers the maximum observation gap window constraint and multiple strong constraints to maximize the overall task benefit value of the optimal observation scheme.

[0009] S2. Task allocation model solution:

[0010] The task allocation model is solved by adding constraint checking and correction operators to the genetic algorithm to obtain the optimal observation plan.

[0011] Preferably, the input of the task allocation model includes:

[0012] A satellite resource set, wherein the satellite resource set includes the number of six orbits, the type of imaging payload, the optimal resolution of the imaging payload, and the cost of the satellite performing an observation mission;

[0013] A task information set, wherein the task information set includes the geographical location of the observed target, the observation time period required by the task, the task priority, the first observation interval required by the task, and the minimum resolution required by the task;

[0014] An available observation data set includes a satellite to which the observation data belongs, an observed target to which the observation data belongs, and an observation time window.

[0015] Preferably, step S1 specifically includes the following contents:

[0016] S11, calculate the actual time coverage rate of each task and the cost consumed by the satellite to perform the task, normalize the time coverage rate and the consumption cost of each task and calculate the difference, then use the task priority to weight the corresponding task, and sum the weights of all tasks to obtain the overall task benefit value;

[0017] S12. Based on the overall task benefit value, a task allocation model is established with the goal of maximizing the overall task benefit value and with the maximum observation gap window constraint and multiple strong constraints as restriction conditions.

[0018] Preferably, the multiple strong constraints include task priority constraints, observation time constraints, lighting condition constraints, meteorological condition constraints, image type constraints, and resolution constraints;

[0019] The observation time constraint is to consider the observation time required by the mission; the lighting condition constraint is that the optical imaging payload must be used when the lighting conditions meet the imaging requirements; the meteorological condition constraint is that the optical imaging payload can only obtain valid images on sunny days; the image type constraint is that the image type must be consistent with the image type required by the mission; the resolution constraint is that the image resolution must be less than or equal to the minimum resolution required by the mission.

[0020] Preferably, a task allocation model is mathematically described to arrange n observation tasks for m different imaging payloads, thereby improving the overall task efficiency while meeting the user requirements of each task;

[0021] If tasks i and j occupy the same imaging load k during execution, and task i is executed before task j, then after task i is completed, task j can be executed immediately based on basic simplifications and assumptions;

[0022] The basic simplifications and assumptions include simplifying the observation target to a point target; assuming that the satellite cannot be constrained by energy; assuming that the command injection and data return during the mission are not constrained; assuming that the attitude conversion of the optical earth observation satellite is an instantaneous process;

[0023] When different tasks are observed by the same imaging payload in the same period of time, if they are observed by the SAR imaging payload at the same time, these tasks can be executed simultaneously; if they are observed by the visible light imaging payload at the same time, if the distance between the observed targets does not exceed the width of the visible light imaging payload, these tasks can be executed simultaneously. If the observed targets are relatively dispersed and the distance exceeds the width of the visible light imaging payload, the task with higher priority will be executed first. If the priorities are the same, the task with smaller coverage of the current task will be executed.

[0024] Among them, the imaging mode of the SAR imaging payload is scanning imaging; the imaging mode of the visible light imaging payload is frame imaging, taking into account the lateral swing in the rolling direction.

[0025] Preferably, the conditions that the optimal observation plan should meet include: each task can only be executed within its time window; each task can occupy any imaging payload in the imaging payload set that meets its requirements, and can be executed by multiple imaging payloads at the same time; each imaging payload can only execute one observation task at any time; the execution time of each observation task can be less than its required execution time, but must meet the maximum observation interval requirement of the task and maximize the overall task benefit value.

[0026] Preferably, step S2 specifically includes the following contents:

[0027] S21. The user proposes observation task requirements, including the geographical coordinates of the observation target, observation time period, maximum observation interval, task priority, image type requirements, and minimum resolution requirements;

[0028] S22. Acquire remote sensing satellite resource data, including the number of satellites in the remote sensing satellite constellation, the number of six orbits of the satellite, the type of satellite imaging payload, the resolution of the satellite imaging payload, and the maximum roll angle of the optical satellite in the rolling direction;

[0029] S23, without considering any constraints, calculating the observation data of all satellites on all mission targets; each piece of observation data includes imaging payload, observation target, imaging start time and imaging end time;

[0030] S24, preprocessing all observation data to filter out observation data that meets the maximum gap window of the observation task;

[0031] S25, for the observation data that meets the maximum gap window of the observation task, a genetic algorithm with constraint checking and correction operators is used to solve and obtain the optimal observation plan;

[0032] S26. Output an optimal observation plan report, which includes satellite name, target name, observation start time, observation end time and usage status.

[0033] Preferably, step S24 specifically includes the following contents:

[0034] S241, receiving observation task requirements proposed by users;

[0035] S242, judging the observation task requirement. When the image type requirement of the task is optical image and the observed meteorological conditions are not suitable for observation using optical imaging satellites, the task requirement cannot be met, a relevant prompt is given, and the task requirement does not participate in subsequent calculations;

[0036] S243, reading all observation data that meet the required observation task requirements;

[0037] S244, selecting an observable time window that meets the task requirements according to the observation time constraint, the illumination condition constraint, the meteorological condition constraint, the image type constraint, and the resolution constraint;

[0038] S245. Calculate the maximum observation gap window for each task when all observation time windows that meet the task requirements are used. If the maximum observation gap window is larger than the maximum observation gap window required by the task, the task requirements cannot be met under the existing conditions. If the maximum observation gap window is less than or equal to the maximum observation time window required by the task, the task requirements can be met under the existing conditions and further optimization calculations can be performed.

[0039] Preferably, step S25 specifically includes the following contents:

[0040] S251, input the number of evolutions and population size of the genetic algorithm;

[0041] S252, performing population initialization;

[0042] S253, performing constraint check on the initial population, and correcting the individuals whose maximum observation gap window does not meet the task requirements;

[0043] S254, performing iterative calculations, each of which includes calculating the fitness, selection, crossover, mutation, recombination, constraint checking and correction of each individual in the population;

[0044] S255. From the last generation of population, select the individual with the largest fitness as the result of genetic algorithm calculation;

[0045] S256. According to the result of the genetic algorithm, the optimal individual genotype is decoded to form an optimal observation plan; the observation plan includes the satellites used for each mission and the corresponding observation time window.

[0046] Preferably, step S253 is specifically as follows: first, a population is received, and then the genes of each individual are read one at a time, the actual maximum observation gap window of each observation task in the individual is calculated, and whether the actual maximum observation gap window of each observation task meets the requirements is determined. If so, the individual is saved until all individuals in the population are traversed; if not, one of the genotypes of the individual is changed from 0 to 1, and the check is recalculated. After the task requirements are met, the individual is saved until all individuals in the population are traversed.

[0047] The beneficial effects of the present invention are: 1. In view of the characteristics of continuous observation task requirements, the present invention uses the maximum observation gap as an indicator to measure the continuity of observation tasks. Compared with the traditional task time coverage index, it has the advantages of taking into account both observation time and observation continuity, and improves the accuracy and efficiency of task solving. 2. The present invention is aimed at the continuous observation task allocation model for multiple targets of interest for low-orbit constellations, mainly based on the maximum gap window constraint, and integrates a variety of strong constraints, such as task priority, weather conditions, observation time period, minimum resolution and image type requirements, so that the task allocation model is closer to the actual engineering situation. 3. The present invention uses genetic algorithms to solve the problem of low solution efficiency for the allocation of continuous observation tasks under the constructed strong constraints; by constructing constraint detection and correction operators, the problem of the population evolving in the direction of increasing the maximum observation gap is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the side swing of the optical imaging load in the rolling direction according to an embodiment of the present invention;

[0049] Figure 2 is a schematic diagram of an optical imaging payload satellite observing multiple targets simultaneously in an embodiment of the present invention;

[0050] Figure 3 is a flow chart of a task scheduling method in an embodiment of the present invention;

[0051] Figure 4 is a flow chart of observation time window data preprocessing in an embodiment of the present invention;

[0052] Figure 5 is a flow chart of genetic algorithm optimization solution in an embodiment of the present invention;

[0053] Figure 6 Schematic diagram of two evolutionary directions of populations in an embodiment of the present invention;

[0054] Figure 7 is a flow chart of constraint checking and correction in an embodiment of the present invention;

[0055] Figure 8 is a schematic diagram of a crossover operator in an embodiment of the present invention;

[0056] Fig. 9 is a schematic diagram of a mutation operator in an embodiment of the present invention;

[0057] Fig.10 is a schematic diagram of a reorganization operator in an embodiment of the present invention;

[0058] Fig.11 is a schematic diagram of a time window within the time required for selecting a task in an embodiment of the present invention;

[0059] Fig.12 is a schematic diagram of the maximum observation gap window in an embodiment of the present invention;

[0060] Fig.13 is a schematic diagram of an earth observation constellation used in an embodiment of the present invention;

[0061] Fig.14 Schematic diagram of the convergence curve of the genetic algorithm in the embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation methods described herein are only used to explain the present invention and are not used to limit the present invention.

[0063] At present, the satellite remote sensing industry is developing rapidly, and it can provide users with high-precision products in a very short response time. With the development of technologies such as artificial intelligence and big data, the application field of remote sensing technology is constantly expanding. With the development of remote sensing technology, users' needs are more diversified, such as larger data volume, higher precision, and larger coverage. Among them, the task demand for continuous observation of multiple ground targets is particularly urgent. However, there is currently little research on how to meet the task demand for continuous observation of multiple ground targets, and there is a lack of research on how to set indicators to measure the continuity of constellation observation of ground targets.

[0064] Therefore, the present invention proposes a low-orbit large-scale remote sensing satellite constellation task allocation model for continuous observation of multiple ground targets and a solution algorithm. The model describes the observation task allocation process, stipulates the conditions and objective functions satisfied by the optimal observation plan, and comprehensively considers the maximum observation gap window constraint and other strong constraints to maximize the overall task benefit value of the optimal observation plan. The solution algorithm adds constraint checking and correction operators on the basis of the genetic algorithm to improve the accuracy and efficiency of the solution, and provide a reference for the allocation of earth observation tasks under future large-scale star clusters. The method of the present invention mainly includes two aspects:

[0065] 1. Task allocation model construction:

[0066] A task allocation model for continuous observation of multiple ground targets by a low-orbit large-scale remote sensing satellite constellation is established. The task allocation model describes the observation task allocation process, stipulates the conditions and objective function satisfied by the optimal observation plan, and comprehensively considers the maximum observation gap window constraint and multiple strong constraints to maximize the overall task benefit value of the optimal observation plan.

[0067] In this embodiment, a model for allocating continuous observation tasks of multiple ground targets by a low-orbit large-scale remote sensing satellite constellation is established, in which the maximum observation gap window of the observation task is proposed as an indicator for measuring the continuity of the observation task. The main contents include:

[0068] 1.1. Make simplifications and assumptions. Simplify the observation target to a point target. Assume that the satellite is not constrained by energy, that the command injection and data return are not constrained during the mission, and that the attitude conversion of the optical earth observation satellite is an instantaneous process.

[0069] 1.2. Model input, including satellite resource set, mission information set, and available observation data set.

[0070] (1) The satellite resource set includes: the number of satellite orbits, the type of imaging payload, the optimal resolution of the imaging payload, and the cost of the satellite performing an observation mission. The satellite resource set is denoted as Sat: Sat = {sat k |k∈M={1,...,m}},sat k represents the kth satellite, and m represents the number of satellites. Its related parameters include: k ={a k ,i k ,e k ,Ω k ,ω k ,τ k}Six satellite orbit numbers, img_s k ∈{O,S}, where O represents the visible light imaging payload satellite and S represents the SAR imaging payload satellite; res_sk Represents the optimal resolution of the imaging payload; sat_cost k It represents the cost of a satellite performing an observation mission. In the present invention, this cost is considered as a constant value.

[0071] (2) The task information set includes: the geographical location of the observed target, the observation time period required by the task, the task priority, the first observation interval required by the task, and the minimum resolution required by the task. The task information set is denoted as, Task: Task = {task j |j∈J={1,...,n}},task j represents the jth task, and n represents the number of tasks. Its related parameters include: j =[qs j ,qe j ] represents the observation time period required by the task, qs j Indicates the start time, qe j Indicates the end time; pri j Represents the task priority, which is used to distinguish the importance of different tasks. j Represents the minimum image resolution required by the task, img_t j ∈{O,S,OS} represents the image type required by the task, where O represents visible light images, S represents SAR images, OS represents no image type requirement, and atmospheric j ={0,1} represents the local meteorological conditions of the observed target, where 0 represents cloudy weather, 1 represents clear weather, and δ j Represents the maximum observation gap time required by the task.

[0072] (3) Available observation data sets include: the satellite to which the observation data belongs, the observed object to which the observation data belongs, and the observation time window. The observation data set is denoted as TW: TW = {tw i,j |i∈{1,...,K all},j∈J},tw i,j represents the i-th observation data, and the observation data belongs to the j-th task, K all represents the number of observations. Its related parameters include: sat k,j Indicates the satellite number to which the observation data belongs, task j Represents the task number to which the observation data belongs, p i,j Represents the observation time window of the observation data, where p i,j =[ps i,j ,pe i,j ] represents the time window of observation, ps i,j Indicates the start time, pei,j Indicates the end time. i :x i ∈{0,1} indicates the usage of the window, 1 indicates that the window is in use, and 0 indicates that the window is not in use.

[0073] 1.3. Explain the satellite earth observation model used. Figure 1 As shown, the present invention takes into account SAR imaging payload satellites and visible light imaging payload satellites. The SAR imaging payload satellite uses scanning imaging, while the visible light imaging payload satellite uses frame imaging, taking into account the side swing in the rolling direction.

[0074] See attached Fig.13 , is a schematic diagram of a large-scale low-orbit constellation used in this embodiment. The constellation includes 2308 satellites, including 1154 optical imaging payload satellites and 1154 SAR imaging payload satellites.

[0075] 1.4. Mathematically describe the task allocation model: arrange n observation tasks (task set Task) for m different imaging payloads (satellite set Sat) to improve the overall task efficiency while meeting the user needs of each task.

[0076] In this embodiment, the specific requirements of the four observation tasks are shown in Table 1.

[0077] Table 1 Observation mission requirements

[0078]

[0079] 1.5. If tasks i and j occupy the same imaging payload k during execution, and task i is executed before task j, then after task i is completed, task j can be executed immediately based on the basic assumption that the satellite attitude transition time is not considered.

[0080] Due to the limitations of the optimal resolution of the imaging payload, the type of imaging payload, and the imaging time, some tasks may not be able to be performed.

[0081] When different tasks are observed by the same imaging payload at the same time, if they are observed by the SAR imaging payload at the same time, these tasks can be executed simultaneously. If they are observed by the visible light imaging payload at the same time, if the distance between the observed targets does not exceed the visible light imaging payload width (such as Figure 2 Left), these tasks can be performed simultaneously; if the observed targets are relatively dispersed and the distance exceeds the width of the visible light imaging payload (such as Figure 2 Right), at this time, the task with higher priority is executed first. If the priorities are the same, the task with smaller coverage of the current task is executed.

[0082] 1.6. The conditions that an optimal task allocation plan (observation plan) should meet are as follows: Each task can only be executed within its time window; Each task can occupy any imaging payload in the set of imaging payloads that meet its requirements, and can be executed by multiple imaging payloads at the same time; Each imaging payload can only execute one observation task at any time; The execution time of each task can be less than its required execution time, but it must meet the maximum observation gap requirement of the task and maximize the overall task benefit.

[0083] 1.7. Calculation process of overall mission benefit value:

[0084] First, calculate the actual time coverage of a task;

[0085] qrt j =(∪ j p i,j x i )∩qt j

[0086] Then calculate the cost consumed by the satellite to perform the mission;

[0087] cost j =∑ j x i sat_cost k,j , j∈J

[0088] Normalize the time coverage and cost and calculate the difference; then use the task priority to weight them;

[0089]

[0090] Finally, sum the weights of all tasks;

[0091]

[0092] 1.8. Based on the above symbolic definitions and mathematical descriptions, establish a task allocation model;

[0093]

[0094]

[0095] or img_s k,j =img_t j

[0096] or img_s k,j =S

[0097] 2. Task allocation model solution:

[0098] The task allocation model is solved by adding constraint checking and correction operators to the genetic algorithm to obtain the optimal observation plan.

[0099] In this embodiment, based on the genetic algorithm, a constraint check and correction operator are added to solve the problem that the population may evolve in the wrong direction, and a task solving method is designed, such as Figure 3 As shown, it specifically includes the following contents:

[0100] 2.1. The user proposes observation task requirements, including the geographic coordinates of the observation target, observation time period, maximum observation interval, task priority, image type requirements, and minimum resolution requirements.

[0101] 2.2. Obtain remote sensing satellite resource data, including the number of satellites in the remote sensing satellite constellation, the number of six orbits of the satellite, the type of satellite imaging payload, the resolution of the satellite imaging payload, and the maximum roll angle of the optical satellite in the rolling direction.

[0102] 2.3. Without considering any constraints, calculate the observation data of all satellites on all mission targets; each observation data includes imaging payload, observation target, imaging start time and imaging end time.

[0103] 2.4. Preprocess all observation data to filter out observation data that meets the maximum gap window of the observation task.

[0104] like Figure 4 As shown, the specific execution process of this step is:

[0105] (1) Receive observation task requirements proposed by users.

[0106] (2) Determine the observation task requirements. When the image type requirement of a task is an optical image and the observed meteorological conditions are not suitable for observation using an optical imaging satellite, the task requirement cannot be met, a relevant prompt is given, and the task requirement does not participate in subsequent calculations.

[0107] (3) Read all observation data that meets the required observation task requirements.

[0108] (4) Filter the observable time window that meets the mission requirements based on the observation time constraint, lighting condition constraint, meteorological condition constraint, image type constraint, and resolution constraint. The observation time constraint refers to the observation time required by the mission; the lighting condition constraint refers to the optical imaging payload must be used when the lighting conditions meet the imaging requirements; the meteorological condition constraint refers to the optical imaging payload only obtaining valid images on sunny days; the image type constraint refers to the image type that must be consistent with the image type required by the mission; and the resolution constraint refers to the image resolution that must be less than or equal to the minimum resolution required by the mission.

[0109] For the interception process of the observation window, see Fig.11 For the observation window that is completely within the task time window, the entire window is selected; for the observation window that is partially within the task time window, the intersection with the task time window is taken; for the observation window that is completely outside the task time window, the entire window is not selected.

[0110] Maximum observation gap window see Fig.12 , that is, take the union of all observation windows of the task and then take the complement of the task time window. The largest window in the set is the maximum observation gap window.

[0111] In this embodiment, after data preprocessing, the available observation data of each task can be found in Table 2.

[0112] Table 2 Available observation data after data preprocessing

[0113]

[0114]

[0115]

[0116] (5) Calculate the maximum observation gap window of each task when all observation time windows are used. If the maximum observation gap window is larger than the maximum observation gap window required by the task, the task requirements cannot be met under the existing conditions. If the maximum observation gap window is less than or equal to the maximum observation time window required by the task, the task requirements can be met under the existing conditions and further optimization calculations can be performed.

[0117] 2.5. For the observation data that meets the maximum gap window of the observation task, a genetic algorithm with constraint checking and correction operators is used to solve and obtain the optimal observation plan.

[0118] like Figure 5 As shown, the specific execution process of this step is:

[0119] (1) Input the number of evolutions and population size of the genetic algorithm;

[0120] (2) Initialize the population;

[0121] (3) Check the constraints of the initial population and correct the individuals whose maximum observation gap window does not meet the task requirements;

[0122] (4) Perform iterative calculations. Each iterative calculation includes calculating the fitness, selection, crossover, mutation, recombination, constraint checking and correction of each individual in the population. Figure 8The crossover process used by the genetic algorithm is shown, and the operator used is the single-point crossover operator. Fig. 9 It shows the mutation process used by the genetic algorithm, and the operator used is the single-point mutation operator. Fig.10 It shows the recombination process used by the genetic algorithm. The number of offspring populations may decrease. At this time, individuals with large fitness in the parent generation should be selected and added to the offspring population to keep the number of individuals in each generation stable.

[0123] (5) From the last generation of population, select the individual with the largest fitness as the result of genetic algorithm calculation;

[0124] (6) Based on the results of the genetic algorithm, the optimal individual genotype is decoded to form an optimal observation plan; the observation plan includes the satellites used for each mission and the corresponding observation time window.

[0125] In this embodiment, Figure 6 As shown in the figure, during evolutionary computation, the population will develop in two directions: one is that the number of observation gap windows increases, the maximum observation gap window decreases, and the time coverage and observation task continuity are improved; the other is that the number of observation gap windows decreases, but the maximum observation gap window increases. Although the time coverage is improved, the continuity of the observation task is reduced, resulting in the task plan not meeting the maximum observation gap window constraint and becoming an infeasible solution. In order to suppress the second development direction, a constraint detection and correction operator is proposed to perform constraint checks on the population after crossover, mutation and other operations, and to correct the infeasible solutions found.

[0126] like Figure 7 As shown in , the specific correction process is to first receive the population, then read the genes of each individual at a time, calculate the actual maximum observation gap window of each observation task in the individual, and determine whether the actual maximum observation gap window of each observation task meets the requirements. If so, save the individual until all individuals in the population are traversed; if not, change one of the genotypes of the individual from 0 to 1, recalculate and check, and save the individual after meeting the task requirements until all individuals in the population are traversed. The convergence curve of the genetic algorithm solution can be found in Fig.14 .

[0127] 2.6. Output the optimal observation plan report, which includes satellite name, target name, observation start time, observation end time and usage status.

[0128] In this embodiment, after solving, the observation scheme is obtained, see Table 3.

[0129] Table 3 Observation task allocation scheme obtained by solution

[0130]

[0131]

[0132]

[0133] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0134] The present invention provides a low-orbit large-scale remote sensing constellation task scheduling method for continuous observation of multiple ground targets. The present invention uses the maximum observation gap as an indicator to measure the continuity of the observation task in view of the characteristics of the continuous observation task requirements. Compared with the traditional task time coverage index, it has the advantages of taking into account both the observation time and the observation continuity, and improves the accuracy and efficiency of task solving. The present invention is aimed at the low-orbit constellation continuous observation task allocation model for multiple targets, mainly based on the maximum gap window constraint, and integrates a variety of strong constraints, such as task priority, weather conditions, observation time period, minimum resolution and image type requirements, so that the task allocation model is closer to the actual engineering situation. The present invention uses a genetic algorithm to solve the problem of low solution efficiency for the continuous observation task allocation under the constructed strong constraint conditions; by constructing a constraint detection and correction operator, the problem of the population evolving in the direction of increasing the maximum observation gap is solved.

[0135] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A low-orbit large-scale remote sensing constellation task scheduling method for continuous observation of multiple ground-based targets, characterized by: The following steps are included: S1. Task allocation model construction: A task allocation model for continuous observation of multiple ground targets by a large-scale remote sensing satellite constellation in low orbit is established. The task allocation model describes the observation task allocation process, specifies the conditions and objective functions that the optimal observation scheme must satisfy, and comprehensively considers the maximum observation gap window constraint and multiple strong constraints to maximize the overall task benefit value of the optimal observation scheme. S2. Task allocation model solution: The task allocation model is solved by adding constraint checking and correction operators to the genetic algorithm to obtain the optimal observation plan.

2. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 1, characterized in that: The input of the task allocation model includes, A satellite resource set, wherein the satellite resource set includes the number of six orbits, the type of imaging payload, the optimal resolution of the imaging payload, and the cost of the satellite performing an observation mission; A task information set, wherein the task information set includes the geographical location of the observed target, the observation time period required by the task, the task priority, the first observation interval required by the task, and the minimum resolution required by the task; An available observation data set includes a satellite to which the observation data belongs, an observed target to which the observation data belongs, and an observation time window.

3. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 1, characterized in that: Step S1 specifically includes the following contents: S11, calculate the actual time coverage rate of each task and the cost consumed by the satellite to perform the task, normalize the time coverage rate and the consumption cost of each task and calculate the difference, then use the task priority to weight the corresponding task, and sum the weights of all tasks to obtain the overall task benefit value; S12. Based on the overall task benefit value, a task allocation model is established with the goal of maximizing the overall task benefit value and with the maximum observation gap window constraint and multiple strong constraints as restriction conditions.

4. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 3 is characterized in that: A variety of strong constraints include task priority constraints, observation time constraints, lighting condition constraints, meteorological condition constraints, image type constraints, and resolution constraints; The observation time constraint is to consider the observation time required by the mission; the lighting condition constraint is that the optical imaging payload must be used when the lighting conditions meet the imaging requirements; the meteorological condition constraint is that the optical imaging payload can only obtain valid images on sunny days; the image type constraint is that the image type must be consistent with the image type required by the mission; the resolution constraint is that the image resolution must be less than or equal to the minimum resolution required by the mission.

5. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 1, characterized in that: A mathematical description of the task allocation model is provided to arrange n observation tasks for m different imaging payloads, so as to improve the overall mission efficiency while meeting the user requirements of each task. If tasks i and j occupy the same imaging load k during execution, and task i is executed before task j, then after task i is completed, task j can be executed immediately based on basic simplifications and assumptions; The basic simplifications and assumptions include simplifying the observation target to a point target; assuming that the satellite cannot be constrained by energy; assuming that the command injection and data return during the mission are not constrained; It is assumed that the attitude transformation of optical earth observation satellite is an instantaneous process; When different tasks are observed by the same imaging payload in the same period of time, if they are observed by the SAR imaging payload at the same time, these tasks can be executed simultaneously; if they are observed by the visible light imaging payload at the same time, if the distance between the observed targets does not exceed the width of the visible light imaging payload, these tasks can be executed simultaneously. If the observed targets are relatively dispersed and the distance exceeds the width of the visible light imaging payload, the task with higher priority will be executed first. If the priorities are the same, the task with smaller coverage of the current task will be executed. Among them, the imaging mode of the SAR imaging payload is scanning imaging; the imaging mode of the visible light imaging payload is frame imaging, taking into account the lateral swing in the rolling direction.

6. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 1, characterized in that: The conditions that the optimal observation plan should meet include: each task can only be executed within its time window; each task can occupy any imaging payload in the imaging payload set that meets its requirements, and can be executed by multiple imaging payloads at the same time; each imaging payload can only execute one observation task at any time; the execution time of each observation task can be less than its required execution time, but must meet the maximum observation interval requirement of the task and maximize the overall task benefit value.

7. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 1, characterized in that: Step S2 specifically includes the following contents: S21. The user proposes observation task requirements, including the geographical coordinates of the observation target, observation time period, maximum observation interval, task priority, image type requirements, and minimum resolution requirements; S22. Acquire remote sensing satellite resource data, including the number of satellites in the remote sensing satellite constellation, the number of six orbits of the satellite, the type of satellite imaging payload, the resolution of the satellite imaging payload, and the maximum roll angle of the optical satellite in the rolling direction; S23, without considering any constraints, calculating the observation data of all satellites on all mission targets; each piece of observation data includes imaging payload, observation target, imaging start time and imaging end time; S24, preprocessing all observation data to filter out observation data that meets the maximum gap window of the observation task; S25, for the observation data that meets the maximum gap window of the observation task, a genetic algorithm with constraint checking and correction operators is used to solve and obtain the optimal observation plan; S26. Output an optimal observation plan report, which includes satellite name, target name, observation start time, observation end time and usage status.

8. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 7, characterized in that: Step S24 specifically includes the following contents: S241, receiving observation task requirements proposed by users; S242, judging the observation task requirement. When the image type requirement of the task is optical image and the observed meteorological conditions are not suitable for observation using optical imaging satellites, the task requirement cannot be met, a relevant prompt is given, and the task requirement does not participate in subsequent calculations; S243, reading all observation data that meet the required observation task requirements; S244, selecting an observable time window that meets the task requirements according to the observation time constraint, the illumination condition constraint, the meteorological condition constraint, the image type constraint, and the resolution constraint; S245, calculating the maximum observation gap window of each task when all observation time windows that meet the task requirements are used. If the maximum observation gap window is larger than the maximum observation gap window required by the task, the task requirements cannot be met under the existing conditions; If the maximum observation gap window is less than or equal to the maximum observation time window required by the task, then under the existing conditions, the task requirements can be met and further optimization calculations can be performed.

9. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 7, characterized in that: Step S25 specifically includes the following contents: S251, input the number of evolutions and population size of the genetic algorithm; S252, performing population initialization; S253, performing constraint check on the initial population, and correcting the individuals whose maximum observation gap window does not meet the task requirements; S254, performing iterative calculations, each of which includes calculating the fitness, selection, crossover, mutation, recombination, constraint checking and correction of each individual in the population; S255. From the last generation of population, select the individual with the largest fitness as the result of genetic algorithm calculation; S256. According to the result of the genetic algorithm, the optimal individual genotype is decoded to form an optimal observation plan; the observation plan includes the satellites used for each mission and the corresponding observation time window.

10. The method for scheduling low-orbit large-scale remote sensing constellation tasks for continuous observation of multiple ground-based targets according to claim 9, characterized in that: Specifically, step S253 first receives the population, then reads the genes of each individual at a time, calculates the actual maximum observation gap window of each observation task in the individual, and determines whether the actual maximum observation gap window of each observation task meets the requirements. If so, the individual is saved until all individuals in the population are traversed; if not, one of the genotypes of the individual is changed from 0 to 1, and the calculation and inspection are recalculated. After the task requirements are met, the individual is saved until all individuals in the population are traversed.

Images