Multi-satellite collaborative observation-data transmission combined planning method for agile satellites

Abstract

The invention discloses an agile satellite multi-satellite collaborative observation-data transmission combined planning method, which comprises the following steps: carrying out discrete sampling on a target area, calculating the visibility of each satellite to a sampling point, and obtaining the coverage characteristic of each satellite to the area based on the visibility of the sampling point; based on satellite coverage characteristics, a K-Means clustering algorithm is used to segment the target area into a plurality of sub-areas and distribute the sub-areas to corresponding satellites; boundary adaptive stripe division is carried out on the distributed sub-regions, and the number of stripes is calculated; constructing an observation-data transmission joint planning problem model, solving based on an elite genetic algorithm to obtain a planning result, integrating all satellites, and outputting an optimal scheduling scheme; according to the method, the problem that regional target observation-data transmission combined planning is difficult to solve under multi-satellite cooperation of agile satellites is solved, the coverage of the satellites on the distribution region is ensured, the overlapping rate of observation strips of different satellites is reduced, the overall observation benefit under multi-satellite cooperation can be improved, and the method can be effectively applied to a remote sensing satellite management and control technology.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite data processing and satellite mission planning, and in particular to an agile satellite multi-satellite collaborative observation-data transmission joint planning method. Background Art

[0002] Remote sensing satellites can carry different payloads to image the Earth's surface, and have great application value in environmental surveys, resource exploration, disaster relief and other scenarios. The new generation of Agile Earth Observation Satellite (AEOS) can image targets at different times and angles by adjusting the three-axis attitude, achieving single-turn multi-strip imaging, greatly reducing the mission cycle, and has great development potential. Agile satellites have three maneuvering degrees of freedom: roll, pitch, and yaw, so the visible window time for each target is greatly extended.

[0003] In the field of remote sensing satellite earth observation, regional targets usually refer to targets with a certain area composed of multiple vertices and that cannot be fully covered by a single satellite through a single push-broom imaging. For this type of target, a single satellite needs to repeatedly visit multiple times or use multiple satellites to coordinate observations to achieve full coverage. Agile satellites have an active push-broom imaging mode that can perform attitude maneuvers and earth observations at the same time. This mode greatly improves the observation efficiency and benefits of regional targets.

[0004] At present, the research on agile satellite mission planning for regional targets faces the following problems:

[0005] First, considering the problem of regional repeated observation in the multi-satellite observation process, the existing technology has the problem of regional repeated observation under the coordination of agile satellites, and does not fully consider the coverage characteristics of different satellites in the same area;

[0006] Second, the existing strip segmentation method cannot reflect the ability of agile satellites to push-broom imaging in any direction, and does not consider the dynamic nature of strip segmentation;

[0007] Third, most studies on on-board storage resources are simplified by using the maximum number of observable tasks in a single cycle, which is quite different from the actual mission execution process of satellites.

[0008] Therefore, an agile satellite multi-satellite collaborative observation-data transmission joint planning method is urgently needed to solve the above problems. Summary of the invention

[0009] The purpose of the present invention is to provide an agile satellite multi-satellite collaborative observation-data transmission joint planning method, which can effectively solve the problems existing in the above-mentioned prior art.

[0010] In order to solve the above technical problems, the present invention adopts the following technical solution: an agile satellite multi-satellite collaborative observation-data transmission joint planning method, comprising the following steps:

[0011] S1. Discretely sample the target area, calculate the visibility of each satellite to the sampling point, and obtain the coverage characteristics of each satellite to the area based on the visibility;

[0012] S2, based on the K-Means clustering algorithm, the target area is divided into several sub-areas and assigned to corresponding satellites based on coverage characteristics;

[0013] S3, dividing the allocated sub-areas into boundary-adaptive strips and calculating the number of strips;

[0014] S4. Construct observation-data transmission joint planning problem model:

[0015]

[0016] Where: x ij is a binary decision variable, which means whether the satellite observes strip i in the jth time window. Its value range is x ij ∈{0,1}; Strip i is the i-th strip that the satellite needs to observe; N is the number of time windows for the satellite to target i; T is the total number of strips; T(·) represents the observation time of the strip;

[0017] S5. Solve the observation-data transmission joint planning problem model based on the elite genetic algorithm to obtain the planning results, integrate all satellites and output the optimal scheduling plan;

[0018] Preferably, in step S1, sampling points are generated in the target area at preset intervals, and based on the orbital data and maneuverability of each satellite, its visibility data for all sampling points in the area is calculated, and the satellite's visibility data for all sampling points in the same area is used as the satellite's coverage feature for the area.

[0019] Preferably, the specific calculation process of visibility data is:

[0020] The sampling interval Δr of the target area is preset, starting from the upper left boundary of the target area, sampling is performed inside the entire area at the sampling interval Δr to obtain the longitude and latitude coordinates of all sampling points;

[0021] Input the trajectories of all satellites during the mission cycle and the maximum maneuvering angle of the satellite. Use the maximum maneuvering angle as the half-angle of the satellite's coverable field of view to calculate the visibility data of the satellite to all regional sampling points during the mission cycle. If the sampling point is visible, the visibility data value is 1, otherwise the value is 0.

[0022] Preferably, in step S2, the coverage characteristics of each satellite for the same area are input, and it is determined whether a single satellite can cover the sampling point based on the visibility data of different satellites for the same sampling point:

[0023] If there are sampling points that can be covered by a single satellite, the mean longitude and latitude of the sampling points that can be covered by a single satellite is used as the initial cluster center c of the category to which the satellite k belongs. k Corresponding eigenvalue x ck :

[0024]

[0025] Where: lon and lat represent the longitude and latitude of a point respectively; p represents the sampling point of the single-star observable area of ​​satellite k; n p is the number of observable sampling points of a single satellite, and the remaining satellites select the initial cluster center c according to this formula k ;

[0026] otherwise;

[0027] If there is no single star covered area, start from the sampling point that can be covered by the two stars together, calculate the distance between the two stars and the current initial cluster center, and select the sampling point with the largest distance as the initial cluster center c k ; Continuously increase the number of satellites, calculate the distance between the sampling points that N satellites can jointly cover and the current initial clustering center, and generate the initial clustering center c of the satellite k , until all satellites k covering the area generate initial cluster centers c k ;

[0028] Calculate all sampling points and all cluster centers c k The eigenvalue distance d is used as the initial cluster center c corresponding to the minimum eigenvalue distance k The cluster to which the current sampling point p belongs is taken as the cluster to which the current sampling point p belongs;

[0029]

[0030] Among them, ||||2 represents the 2-norm of the vector; select the c corresponding to the minimum eigenvalue distance k The cluster to which the current point p belongs is taken as the cluster to which the current point p belongs, and the mean of the longitude and latitude of all points in the cluster is taken as the new cluster center feature value of the cluster.

[0031] Calculate the sum of squares of error for the entire data set and the clustering results SSE:

[0032]

[0033] Iterate continuously to generate cluster centers and calculate SSE, repeat the above steps, and when the maximum number of iterations is reached or the cluster center does not change, obtain the final clustering result, obtain the regional sampling points assigned by each satellite to the area according to the category, and calculate the convex hull of all sampling points to generate the sub-area to be observed;

[0034] The above process solves the problem of task allocation for regional targets under multi-satellite collaboration. By solving this problem, the problem of overlapping observation strips between different satellites is avoided. At the same time, the process also takes into account the coverage characteristics of the satellite over the region to ensure the visibility of the satellite to the allocated sub-region. Not only that, the process also reduces the coupling problem of multi-satellite collaboration into multiple single-satellite planning sub-problems, which facilitates the solution of the subsequent planning process.

[0035] Preferably, in step S3, a reference straight line L with a slope k outside the left boundary of the regional target is found, and the number of stripes is calculated based on L as follows:

[0036] The distances between the nearest and farthest points of the target area and L are d min and d max , use the satellite imaging width as the strip segmentation width wid, input the minimum padding width of the area boundary pad, and the minimum strip overlap width wid overlap And the strip division slope k, calculate the strip coverage width D:

[0037] D=d max -d min +2pad;

[0038] Calculate the approximate average width wid of each strip taking into account the overlap ratio men for:

[0039] wid mean =wid-wid overlap / 2;

[0040] The estimated number of stripes is:

[0041]

[0042] in: is the ceiling function.

[0043] Preferably, the stripe division is specifically as follows:

[0044] Initial strip generation: Shift rightward from reference line L Draw the long line segment line1 on the left edge of the strip with slope k, and translate line1 to the right by wid to get line r , calculate the intersection with the area, the length of the farthest intersection is line rlength, draw a rectangular area and record the initial strip boundary point information;

[0045] Among them, the distance between the left edge of the initial strip and the reference straight line L The calculation formula is as follows:

[0046]

[0047] Intermediate strip generation: from the previous strip line r Start, pan left wid overlap Generate a long line segment line1, calculate the intersection point on the left side of the area, and translate wid to the right to get line r , calculate the intersection point on the right side of the area, compare the coordinates of the intersection points on the left and right sides and calculate their respective lengths, take the largest one as the strip length, draw the strip and record the boundary point information;

[0048] Repeat the above process until the long straight line line1 has no intersection with the area, and the strip division is completed.

[0049] According to the above process, each satellite can generate a set of strip tasks to be observed.

[0050] Preferably, in step S4:

[0051] According to the satellite active push-scan imaging rate ω s Calculate strips of different lengths i Required observation time T(Strip i );

[0052]

[0053] Where: Φ i Strip i The geocentric angle between the midpoints of the two ends; ω s is the angular velocity of the satellite orbit; k is the push-broom velocity coefficient, where k = 1.0;

[0054] The observable time window for the target strip is calculated based on the satellite's orbital data and the maximum angle of attitude maneuver during the mission cycle; the observable time window for the target strip is calculated based on the longitude and latitude of the ground station distribution and the minimum elevation angle EA of the data transmission. 0 , calculate the observation windows of all satellites for all targets and the data transmission windows of all ground stations:

[0055]

[0056] Among them: EA 0 is the minimum elevation angle of the ground station; EA is the ground station elevation angle of the satellite at the current position; h is the current orbital height of the satellite; R eis the radius of the earth; α is the angle between the longitude and latitude of the satellite sub-satellite point and the longitude and latitude of the ground station. The calculation formula is as follows:

[0057]

[0058] in: λ represents the latitude and longitude of a point respectively; subscripts k and g represent satellite and ground station.

[0059] The observation-data transmission joint planning problem model is constructed according to the constraint satisfaction model. The model includes: the performance index is the sum of the observation time of the satellite for all observation strips; wherein the constraint satisfaction model includes:

[0060] Constraint 1 is an observation uniqueness constraint, that is, a strip can only be observed once:

[0061]

[0062] Among them: SET T It is the set of satellite strip missions to be observed;

[0063] Constraint 2 is the observation window constraint. The satellite needs to observe the strip within the observable time window of the strip:

[0064]

[0065] Among them: ts ij and ij They represent the start time and end time of the jth time window of satellite k for strip i respectively; t i is the observation start time of satellite k on strip i; d i is the strip observation duration;

[0066] Constraint 3 is the task transfer time constraint. When satellite k continuously observes two strips, it needs to ensure that the difference between the start time of the latter task and the end time of the previous task can meet the satellite attitude maneuvering time:

[0067] t i +d i +tm in ≤t n ;

[0068] Among them: i represents the previous executed task; n represents the next task to be executed; tm in It represents the maneuvering time required to transfer from task i to task n. This time is related to the satellite's maneuvering capability. The specific calculation is described in the subsequent attitude maneuvering modeling section.

[0069] Constraint 4 is the on-board storage constraint. During the planning process, the satellite needs to check whether the storage exceeds the capacity limit:

[0070]

[0071] Among them: mem(t) means the satellite has used fixed storage at the current time t; Indicates the on-board storage required to observe strip i; mem max The maximum available storage capacity.

[0072] Constraint 5 is the data transmission window constraint. The satellite needs to perform the data transmission task within the data transmission window of the ground station:

[0073]

[0074] in: Respectively represent the actual data transmission start and end time of the satellite at the ground station g; ts gj ,te gj They represent the start and end time of the jth data transmission window of the satellite to the ground station g respectively;

[0075] Constraint 6 is the data transmission time constraint. The data of the strip push-scan observation of the satellite must be completely transmitted once within a ground station data transmission window. Due to the limited satellite-to-ground link rate, the total data transmission time corresponding to each strip data transmission needs to be calculated to be less than the data transmission window length:

[0076]

[0077] Where: v t is the link data transmission rate; dt i Trans is the time spent on data transmission of task i; g represents the set of stripe data transmission tasks arranged at the data transmission station g;

[0078] After constructing the above observation-data transmission joint planning model, the solution and constraint checking can be performed in the subsequent steps.

[0079] Preferably, the elite genetic algorithm solution in step S5 includes iterating the following operations until a preset number of iterations is reached, obtaining the maximum sequence of performance indicators, and integrating all satellites to obtain the optimal scheduling solution, and the operations include:

[0080] The current satellite's visibility data (task number, start and end time) for all strips and ground stations are sorted in ascending order by start time. Assuming there are N pieces of data in total, the chromosome length is N. The task encoding method is as follows: observation task encoding set {0,1,2}, data transmission task encoding set {0,3};

[0081] Starting from the first gene of the chromosome, the following operations are performed step by step to obtain all genes: if a gene position i is the first observed band or the previous band in the region, the task gene code is 0, then the gene code is randomly selected with the same probability, otherwise the code is selected by roulette wheel, and the roulette wheel probability is calculated as follows:

[0082]

[0083] Among them: ε is the roulette parameter, its distribution is ε∈[0,1], code(i-1) is the observation method of the previous strip in the same area;

[0084] If the current task is a data transmission task, a roulette wheel selection is performed based on the currently used fixed memory:

[0085]

[0086] Where: P(tr) represents the data transmission probability; w is the interval mapping parameter, which is used to map the original interval [-1,1] to [-w,w]; mem(t) represents the used storage at the current time t; mem max is the maximum usable sequestration amount;

[0087] Perform constraint checks on the generated chromosomes, starting from the first gene that is not 0. For observation tasks, check constraints 1 to 4 one by one. For data transmission tasks, check constraints 5 and 6. If the current gene task meets all constraints, calculate the start and end time of the task, the consumed storage volume, and the corresponding observation posture. If the current gene does not meet the constraints, set the gene code to 0 and the task execution time to -1.

[0088] Chromosome crossover uses the performance index of the parent task planning result as the weight for roulette wheel selection:

[0089]

[0090] Among them: J i represents the individual performance index value, which can be obtained by the above performance index calculation formula; P i represents the probability of individual i being selected; POP represents all individuals in the entire population; after selecting two individuals, a crossover site is randomly generated, and all genes following the crossover site are crossovered;

[0091] Calculate the number of gene mutation points based on chromosome length Select any position in the entire chromosome for mutation, and the mutation method is the same as the above encoding method;

[0092] Perform constraint checks on the generated offspring, calculate the performance indicators of the offspring that meet the constraints, and retain the larger ones.

[0093] Beneficial effect: The present invention provides an agile satellite multi-star collaborative observation-data transmission joint planning method: first, discrete point sampling is performed on the region and the coverage data of the sampling points is calculated, and the region is divided into different satellite task sub-regions using the K-Means clustering task allocation algorithm based on the multi-star coverage characteristics, thereby solving the problem of difficulty in allocating collaborative tasks of multiple agile satellites to targets in the same region; then, a regional adaptive strip division method is used to generate strips to be observed, and satellite observation strips are quickly generated by mathematical calculation, thereby improving the intersection-union ratio of strips and regions; finally, an elite genetic algorithm based on observation-data transmission fusion coding is proposed, and a targeted design is made for the process from population generation to population evolution, thereby solving the problem of difficulty in solving the joint observation-data transmission planning of regional targets under the collaboration of multiple satellites of agile satellites, thereby ensuring the coverage of the allocated area by satellites and reducing the overlap rate of observation strips of different satellites, and improving the overall observation benefit under the collaboration of multiple satellites, thereby effectively being applied to remote sensing satellite management and control technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0095] In the attached picture:

[0096] Figure 1 A flow chart of the planning method of the present invention;

[0097] Figure 2 Schematic diagram for sampling point generation and coverable area calculation;

[0098] Figure 3 This is a simulation diagram of the effect of the K-Means clustering algorithm;

[0099] Figure 4 This is a schematic diagram of the paternal chromosome crossing;

[0100] Figure 5 This is the satellite 5 attitude maneuvering curve and storage usage curve. DETAILED DESCRIPTION

[0101] The embodiments of the present invention are described below in conjunction with the drawings in the embodiments of the present invention. The terms used in the implementation mode of the present invention are only used to explain the specific embodiments of the present invention, and are not intended to limit the present invention. The embodiments of the present application are described below in conjunction with the drawings.

[0102] Example: Reference Figure 1 As shown, an agile satellite multi-satellite collaborative observation-data transmission joint planning method includes the following steps:

[0103] S1. Generate sampling points at preset intervals in the target area, and calculate the visibility data of all sampling points in the area based on the orbital data and maneuverability of each satellite, and use the visibility data of the satellite to all sampling points in the same area as the coverage feature of the satellite to the area;

[0104] The specific calculation process of visibility data is as follows:

[0105] (11) Given a sampling interval Δr of the target area, starting from the upper left boundary of the target area, dot sampling is performed inside the entire area with a sampling interval Δr to obtain the longitude and latitude coordinates of all sampling points. The result is as follows: Figure 2 Shown on the left.

[0106] (12) Input the orbits of all satellites during the mission cycle and the maximum maneuvering angle of the satellite. Take the maximum maneuvering angle as the half angle of the satellite's field of view. Calculate the satellite's visibility to all regional sampling points during the mission cycle. If the sampling point is visible, the visibility data value is 1, otherwise the value is 0. The sampling point visibility result is as follows: Figure 2 Shown on the right.

[0107] (13) The visibility data of all sampling points in the same area are used as the coverage feature of the satellite over the area. The convex hull formed by all sampling points with visibility data value of 1 is the sub-area that the satellite can cover in the area.

[0108] S2. Based on the coverage characteristics of different agile satellites in the same area, the target area is segmented using the K-Means clustering algorithm, and the sub-areas are assigned to the corresponding satellites according to their categories;

[0109] (21) Input the coverage characteristics of each satellite for the same area, and determine whether there is a single satellite that can cover the sampling point based on the visibility data of different satellites for the same sampling point:

[0110] (22) If there are sampling points that can be covered by a single satellite, the mean longitude and latitude of the sampling points that can be covered by a single satellite is used as the initial cluster center c of the category to which the satellite k belongs. k Corresponding eigenvalue x ck :

[0111]

[0112] Where: lon and lat represent the longitude and latitude of a point respectively; p represents the sampling point of the single-star observable area of ​​satellite k; n p is the number of observable sampling points of a single satellite, and the remaining satellites select the initial cluster center c according to this formula k ;

[0113] otherwise;

[0114] (23) If there is no area that can be covered by a single star, start from the sampling point that can be covered by the two stars together, calculate the distance between the two stars and the current initial cluster center, and select the sampling point with the largest distance as the initial cluster center c k ; Continuously increase the number of satellites, calculate the distance between the sampling points that N satellites can jointly cover and the current initial clustering center, and generate the initial clustering center c of the satellite k , until all satellites k covering the area generate initial cluster centers c k ;

[0115] (24) Calculate all sampling points and all cluster centers c k The initial cluster center c corresponding to the minimum eigenvalue distance d k The cluster to which the current sampling point p belongs is taken as the cluster to which the current sampling point p belongs;

[0116]

[0117] Among them: || ||2 represents the 2-norm of the vector; select the c corresponding to the minimum eigenvalue distance k The cluster to which the current point p belongs is taken as the cluster to which the current point p belongs, and the mean of the longitude and latitude of all points in the cluster is taken as the new cluster center feature value of the cluster.

[0118] (25) Calculate the sum of squares of the error of the entire data set and the clustering result SSE:

[0119]

[0120] (26) Continuously iterate to generate cluster centers and calculate SSE, repeat the above steps, and when the maximum number of iterations is reached or the cluster center does not change, obtain the final clustering result, obtain the regional sampling points assigned by each satellite to the area according to the category, calculate the convex hull of all sampling points to generate the sub-area to be observed; Figure 2 As shown;

[0121] The above process solves the problem of task allocation for regional targets under multi-satellite collaboration. By solving this problem, the problem of overlapping observation strips between different satellites is avoided. At the same time, the process also takes into account the coverage characteristics of the satellite over the region to ensure the visibility of the satellite to the allocated sub-region. Not only that, the process also reduces the coupling problem of multi-satellite collaboration into multiple single-satellite planning sub-problems, which facilitates the solution of the subsequent planning process.

[0122] S3, dividing the allocated sub-areas into boundary-adaptive strips and calculating the number of strips;

[0123] Find the reference straight line L with a slope of k outside the left boundary of the regional target. Taking L as the reference, the number of strips is calculated as:

[0124] (31) The distances between the nearest and farthest points of the target area and L are d min and d max , use the satellite imaging width as the strip segmentation width wid, input the minimum padding width of the area boundary pad, and the minimum strip overlap width wid overlap And the strip division slope k, calculate the strip coverage width D:

[0125] D=d max -d min +2pad;

[0126] (32) Calculate the approximate average width wid of each strip considering the overlap rate men for:

[0127] wid mean =wid-wid overlap / 2;

[0128] (33) The estimated result of the number of stripes is:

[0129]

[0130] in: is the ceiling function.

[0131] The strip division is as follows:

[0132] (34) Initial strip generation: Shift rightward from reference line L Draw the long line segment line1 on the left edge of the strip with slope k, and translate line1 to the right by wid to get line r , calculate the intersection with the area, the length of the farthest intersection is line r length, draw a rectangular area and record the initial strip boundary point information;

[0133] Where: The distance between the left edge of the initial strip and the reference straight line L The calculation formula is as follows:

[0134]

[0135] (35) Middle strip generation: from the previous strip line r Start, pan left wid overlap Generate a long line segment line1, calculate the intersection point on the left side of the area, and translate wid to the right to get line r , calculate the intersection point on the right side of the area, compare the coordinates of the intersection points on the left and right sides and calculate their respective lengths, take the largest one as the strip length, draw the strip and record the boundary point information;

[0136] (36) Repeat the above process until the long straight line line1 has no intersection with the area, and the strip division is completed.

[0137] According to the above process, each satellite can generate a set of strip missions to be observed.

[0138] S4. Each satellite calculates the strip observation duration according to the active push-scan imaging rate, and calculates all observation windows and data transmission windows within the mission cycle. Then each agile satellite uses the constraint satisfaction model to model the observation-data transmission joint planning problem:

[0139]

[0140] Among them, x ij is a binary decision variable, which means whether the satellite observes strip i in the jth time window. Its value range is x ij ∈{0,1}; Strip i is the i-th strip that the satellite needs to observe; N is the number of time windows for the satellite to target i; T is the total number of strips; T(·) represents the observation time of the strip;

[0141] (41) According to the satellite active push-scan imaging rate ω s Calculate strips of different lengths i Required observation time T(Strip i );

[0142]

[0143] Where: Φ i Strip i The geocentric angle between the midpoints of the two ends; ω s is the angular velocity of the satellite orbit; k is the push-broom velocity coefficient, where k = 1.0;

[0144] (42) Calculate the observable time window for the target strip based on the satellite's orbital data and the maximum angle of attitude maneuver during the mission cycle; Calculate the observable time window for the target strip based on the longitude and latitude of the ground station distribution and the minimum elevation angle EA of the data transmission. 0 , calculate the observation windows of all satellites for all targets and the data transmission windows of all ground stations:

[0145]

[0146] Among them: EA 0 is the minimum elevation angle of the ground station; EA is the ground station elevation angle of the satellite at the current position; h is the current orbital height of the satellite; R e is the radius of the earth; α is the angle between the longitude and latitude of the satellite sub-satellite point and the longitude and latitude of the ground station. The calculation formula is as follows:

[0147]

[0148] in: λ represents the latitude and longitude of a point respectively; subscripts k and g represent satellite and ground station.

[0149] (43) Based on the constraint satisfaction model, an observation-data transmission joint planning problem model is constructed, the model includes: the performance indicator is the sum of the observation time of the satellite for all observation strips; wherein the constraint satisfaction model includes:

[0150] Constraint 1 is an observation uniqueness constraint, that is, a strip can only be observed once:

[0151]

[0152] Among them: SET T It is the set of satellite strip missions to be observed;

[0153] Constraint 2 is the observation window constraint. The satellite needs to observe the strip within the observable time window of the strip:

[0154]

[0155] Among them: ts ij and ij They represent the start time and end time of the jth time window of satellite k for strip i respectively; t i is the observation start time of satellite k on strip i; d i is the strip observation duration;

[0156] Constraint 3 is the task transfer time constraint. When satellite k continuously observes two strips, it needs to ensure that the difference between the start time of the latter task and the end time of the previous task can meet the satellite attitude maneuvering time:

[0157] t i +d i +tm in ≤t n ;

[0158] Among them: i represents the previous executed task; n represents the next task to be executed; tm in It represents the maneuvering time required to transfer from task i to task n. This time is related to the satellite's maneuvering capability. The specific calculation is described in the subsequent attitude maneuvering modeling section.

[0159] Constraint 4 is the on-board storage constraint. During the planning process, the satellite needs to check whether the storage exceeds the capacity limit:

[0160]

[0161] Among them: mem(t) means the satellite has used fixed storage at the current time t; Indicates the on-board storage required to observe strip i; mem max The maximum available storage capacity.

[0162] Constraint 5 is the data transmission window constraint. The satellite needs to perform the data transmission task within the data transmission window of the ground station:

[0163]

[0164] in: Respectively represent the actual data transmission start and end time of the satellite at the ground station g; ts gj ,te gj They represent the start and end time of the jth data transmission window of the satellite to the ground station g respectively;

[0165] Constraint 6 is the data transmission time constraint. The data of the strip push-scan observation of the satellite must be completely transmitted once within a ground station data transmission window. Due to the limited satellite-to-ground link rate, the total data transmission time corresponding to each strip data transmission needs to be calculated to be less than the data transmission window length:

[0166]

[0167] Where: v t is the link data transmission rate; dt i Trans is the time spent on data transmission of task i; g represents the set of stripe data transmission tasks arranged at the data transmission station g;

[0168] After constructing the above observation-data transmission joint planning model, the solution and constraint checking can be performed in the subsequent steps.

[0169] S5. Solve the observation-data transmission joint planning problem model based on the elite genetic algorithm to obtain the planning results, integrate all satellites and output the optimal scheduling plan;

[0170] The elite genetic algorithm solution includes iterating the following operations until the preset number of iterations is reached, obtaining the maximum sequence of performance indicators, and integrating all satellites to obtain the optimal scheduling solution. The operations include:

[0171] (51) The current satellite's visibility data for all strips and ground stations (task number, start and end time) is sorted in ascending order by start time. Suppose there are N pieces of data in total, then the chromosome length is N. The task coding method is as follows: observation task coding set {0,1,2}, data transmission task coding set {0,3};

[0172] (52) Starting from the first gene of the chromosome, the following operations are performed step by step to obtain all genes: If a gene position i is the first observed band or the previous band in the region, the task gene code is 0, then the gene code is randomly selected with the same probability, otherwise the code is selected by roulette wheel, and the roulette wheel probability is calculated as follows:

[0173]

[0174] Among them: ε is the roulette parameter, its distribution is ε∈[0,1], code(i-1) is the observation method of the previous strip in the same area;

[0175] If the current task is a data transmission task, a roulette wheel selection is performed based on the currently used fixed memory:

[0176]

[0177] Where: P(tr) represents the data transmission probability; w is the interval mapping parameter, which is used to map the original interval [-1,1] to [-w,w]; mem(t) represents the used storage at the current time t; mem max is the maximum usable sequestration amount;

[0178] (53) Perform constraint check on the generated chromosome, starting from the first gene that is not 0. For observation tasks, check constraints 1 to 4 one by one. For data transmission tasks, check constraints 5 and 6. If the current gene task meets all constraints, calculate the start and end time of the task, the consumed storage amount and the corresponding observation posture; if the current gene does not meet any of the constraints, set the gene code to 0 and the task execution time to -1.

[0179] (54) Chromosome crossover uses the performance index of the parent task planning result as the weight to perform roulette selection:

[0180]

[0181] Among them: J i represents the individual performance index value, which can be obtained by the above performance index calculation formula; P i represents the probability of individual i being selected; POP represents all individuals in the entire population; after selecting two individuals, a crossover position is randomly generated, and all genes following the crossover position are crossovered, such as Figure 3 As shown;

[0182] (55) Calculate the number of gene mutation points based on chromosome length Select any position in the entire chromosome for mutation, and the mutation method is the same as the above encoding method;

[0183] (56) Perform constraint checks on the generated offspring, calculate the performance indicators of the offspring that meet the constraints, and retain the larger ones.

[0184] Figure 4 and Figure 5 The results of using the above observation-data transmission joint planning algorithm for 50 randomly generated sub-region strips of one satellite in one day are respectively given. From the results, it can be seen that the attitude angle of the satellite is less than 45° during the entire mission cycle, and the storage usage is less than 100%, which proves the effectiveness of the present invention.

[0185] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. For ordinary technicians in this technical field, after knowing the contents recorded in the present invention, they can make several equivalent changes and substitutions without departing from the principle of the present invention. These equivalent changes and substitutions should also be regarded as belonging to the protection scope of the present invention.

Claims

1. An agile satellite multi-satellite collaborative observation-data transmission joint planning method, characterized in that: The steps include: S1. Discretely sample the target area, calculate the visibility of each satellite to the sampling point, and obtain the coverage characteristics of each satellite to the area based on the visibility; S2, based on the K-Means clustering algorithm, the target area is divided into several sub-areas and assigned to corresponding satellites based on coverage characteristics; S3, dividing the allocated sub-areas into boundary-adaptive strips and calculating the number of strips; S4. Construct observation-data transmission joint planning problem model: Where: x ij is a binary decision variable, which means whether the satellite observes strip i in the jth time window. Its value range is x ij ∈{0,1}; Strip i is the i-th strip that the satellite needs to observe; N is the number of time windows for the satellite to target i; T is the total number of strips; T(·) represents the observation time of the strip; S5. Solve the observation-data transmission joint planning problem model based on the elite genetic algorithm to obtain the planning results, integrate all satellites and output the optimal scheduling plan.

2. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 1, characterized in that: In step S1, sampling points are generated in the target area at preset intervals, and based on the orbital data and maneuverability of each satellite, its visibility data for all sampling points in the area is calculated, and the satellite's visibility data for all sampling points in the same area is used as the satellite's coverage feature for the area.

3. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 2, characterized in that: The specific calculation process of visibility data is as follows: The sampling interval Δr of the target area is preset, starting from the upper left boundary of the target area, sampling is performed inside the entire area at the sampling interval Δr to obtain the longitude and latitude coordinates of all sampling points; The trajectories of all satellites during the mission cycle and the maximum maneuvering angle of the satellite are input, and the visibility data of the satellite to all regional sampling points during the mission cycle is calculated using the maximum maneuvering angle as the half-angle of the satellite's coverable field of view.

4. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 3, characterized in that: In step S2, the coverage characteristics of each satellite for the same area are input, and the visibility data of different satellites for the same sampling point are used to determine whether a single satellite can cover the sampling point: If there are sampling points that can be covered by a single satellite, the mean longitude and latitude of the sampling points that can be covered by a single satellite is used as the initial cluster center c of the category to which the satellite k belongs. k ; otherwise Starting from the sampling point that can be jointly covered by the two satellites, the number of satellites is increased gradually, and the distance between each satellite and the current initial cluster center is gradually calculated. The sampling point with the largest distance is selected as the initial cluster center c. k ; Calculate the distance between the sampling points that N stars can cover together and the current initial cluster center and generate the initial cluster center c of the star k , until all satellites k covering the area generate initial cluster centers c k ; Calculate all sampling points and all cluster centers c k The eigenvalue distance is the initial cluster center c corresponding to the minimum eigenvalue distance k The cluster to which the current sampling point p belongs is taken as the cluster to which the current sampling point p belongs; Calculate the error sum of the entire data set and iterate until the preset condition is reached, generate the cluster center and obtain the clustering result; The regional sampling points assigned by each satellite to the area are obtained according to the category, and the convex hull of all sampling points is calculated to generate the sub-area to be observed.

5. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 1, characterized in that: In step S3, a reference straight line L with a slope of k outside the left boundary of the regional target is found. Taking L as a reference, the number of stripes is calculated as: in: is the ceiling rounding function; D=d max -d min +2pad; Where: D is the strip coverage width; d max and d min are the distances between the nearest and farthest points of the target area and the reference line L respectively; pad is the minimum padding width of the input area boundary; wid mean =wid-wid overlap / 2; wid is the stripe width, wid overlap is the minimum overlap width of the strips.

6. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 5, characterized in that: The strip division is as follows: Shift rightward d from reference line L linel , draw the long line segment line on the left edge of the strip with slope k l , line l Translate wid to the right to get line r , calculate the intersection with the area, the length of the farthest intersection is line r length, draw a rectangular area and record the initial strip boundary point information; From the previous strip line r Start by performing the following operations until the long straight line line1 has no intersection with the area: Move left wid overlap Generate a long line segment line1, calculate the intersection point on the left side of the area, and translate wid to the right to get line r , calculate the intersection point on the right side of the area, compare the coordinates of the intersection points on the left and right sides and calculate their respective lengths, take the largest one as the strip length, draw the strip and record the boundary point information.

7. The method for agile satellite multi-satellite collaborative observation-data transmission joint planning according to claim 6, characterized in that: Distance d linel for:

8. The method for agile satellite multi-satellite coordinated observation-data transmission joint planning according to claim 1, characterized in that: In step S4, the observation time required for strips of different lengths is calculated according to the satellite active push-scan imaging rate; The observable time window for the target strip is calculated based on the satellite's orbital data and the maximum angle of attitude maneuver during the mission cycle; According to the longitude and latitude of the ground stations and the minimum elevation angle of data transmission, the observation windows of all satellites for all targets and the data transmission windows of all ground stations are calculated.

9. The method for agile satellite multi-satellite coordinated observation-data transmission joint planning according to claim 2, characterized in that: In step S4, the observation-data transmission joint planning problem model is constructed based on the constraint satisfaction model, including: Constraint 1: A strip can only be observed once; Constraint 2: The satellite observes the strip within the observable time window of the strip; Constraint 3: The time interval between two consecutive strips observed by the satellite is greater than or equal to the satellite maneuvering time; Constraint 4: Satellites need to check whether storage exceeds capacity limits during planning; Constraint 5: The satellite performs data transmission tasks within the data transmission window of the ground station; Constraint 6: When the satellite is powered on once, the data of the strip push-scan observation must be completely transmitted at one time within the data transmission window of a ground station, and the total data transmission time corresponding to each strip data transmission must be less than the length of the data transmission window.

10. The method for agile satellite multi-satellite coordinated observation-data transmission joint planning according to claim 8, characterized in that: In step S5, the elite genetic algorithm solution includes iterating the following operations until the preset number of iterations is reached, and the maximum sequence of performance indicators is obtained. The optimal scheduling solution can be obtained by integrating all satellites. The operations include: The current satellite's visibility data for all strips and ground stations are sorted in ascending order by start time. Suppose there are N data in total, the chromosome length is N, and the task coding method is as follows: observation task coding set {0,1,2}, data transmission task coding set {0,3}; If a gene position i is the first observed band in the region or the gene code of the previous band is 0, the gene code is randomly selected with the same probability, otherwise the code is selected by roulette; After the chromosome is generated, the corresponding constraint check is performed. For each task whose code is not 0, it is determined whether the task is executable and the task execution start time is planned. If any constraint cannot be met, the gene code is set to 0 and the task execution time is set to -1. The chromosome crossover uses the performance index of the parent task planning result as the weight to perform roulette wheel selection, and crosses all subsequent genes with random crossover positions; Calculate the number of gene mutation points based on chromosome length: in: is the floor rounding function; Choose any position in the chromosome to mutate, and the mutation method is the same as the gene encoding method; Perform constraint checks on the generated offspring, calculate the performance indicators of the offspring that meet the constraints, and retain the larger ones.

Images