Feedback-driven high earth orbit satellite cooperative observation mission planning iteration method

Abstract

The application discloses a feedback-driven high-orbit satellite cooperative observation task planning iteration method, which comprises the following steps: based on the data of the space target to be observed, the data of the high-orbit satellites participating in the cooperative observation, planning optimization parameters and satellite resource constraints, a structured cooperative observation scheme is generated through a multi-objective optimization algorithm, and is distributed as a ground center retained scheme and a satellite exclusive observation instruction package; based on the observation instruction package and the on-orbit dynamic state, the instruction execution and dynamic adjustment observation action are carried out, the observation results, the cooperative state and the resource consumption data are synchronously and real-timely collected, and the on-orbit execution data set is output; a quantitative index system of the point target detection rate, the multi-satellite cooperation degree and the load resource waste rate is constructed, and the performance bottleneck is identified, and the feedback index and the bottleneck analysis result are generated; based on the PID control thought, a quantitative mapping model of the feedback index and the planning parameter is established, the quantitative mapping of the feedback information and the strategy adjustment is realized, and the planning optimization parameter set is output.

Description

Technical Field

[0001] This invention relates to the field of high-orbit satellite collaborative observation technology, and more specifically to a feedback-driven high-orbit satellite collaborative observation mission planning and iteration method. Background Technology

[0002] High-orbit satellites, with their advantages of wide coverage, strong orbital stability, and unobstructed observation views, can achieve long-term continuous monitoring of targets in medium and low orbits and co-orbits, providing key data support for collision warning, on-orbit services, and space safety assessment.

[0003] However, the execution of high-orbit satellite collaborative observation missions faces three core challenges: highly dynamic target orbits, high difficulty in multi-satellite coordination, and unstable point target detection rate. Traditional high-orbit satellite mission planning often adopts a one-way process: an observation plan is generated based on the initial orbit prediction, and the satellite only transmits back the observation data after executing the plan. This lacks closed-loop optimization of the execution effect and planning strategy, which leads to two major problems: First, the observation plan is disconnected from the actual execution. If a target is not detected due to orbital deviation, the system cannot adjust the subsequent observation priority in time, resulting in missed target detection. Second, the system's capabilities are difficult to continuously improve. Feedback information on key indicators such as point target detection rate and multi-satellite coordination degree is not transformed into planning optimization basis. The planning algorithm always relies on the initial parameters and cannot adapt to the dynamically changing space environment.

[0004] Despite the progress made in domestic and international research on high-orbit satellite collaborative observation and mission planning, existing research still suffers from three major shortcomings in addressing the continuous optimization needs in dynamic space environments: First, the planning process lacks a closed-loop feedback mechanism. The traditional one-way process of offline generation-on-orbit execution-data feedback does not incorporate execution feedback such as point target detection results and multi-satellite collaborative errors into the planning optimization. For example, if a target is not detected in two consecutive observations, the system cannot improve its observation priority through feedback, leading to a persistent risk of missed detection. The planning algorithm always relies on initial orbit prediction and static constraints, and cannot adapt to dynamic changes in the target. Second, optimizing the detection rate of non-focused point targets is crucial. Existing collaborative planning often focuses on maximizing resource utilization and mission efficiency. The goal is to maximize the completion rate, while neglecting core observation indicators. For example, some algorithms improve the task completion rate by compressing the observation time, but reduce the detection rate due to insufficient payload integration time. This logic of emphasizing quantity over quality is out of touch with the needs of precise monitoring in spatial situational awareness and cannot meet the accuracy requirements of collision warning and target recognition. Third, there is a lack of a mapping mechanism between feedback and planning. Even when feedback is introduced in existing research, it is mostly qualitative adjustment. There is a lack of a mapping relationship between quantitative feedback indicators and planning strategy parameters. For example, when the point target detection rate drops from 80% to 65%, the system cannot determine how much priority needs to be increased or how much observation time needs to be added. Fuzzy adjustment leads to poor stability of planning optimization and even over-adjustment, crowding out other target observation resources. Summary of the Invention

[0005] In view of the above problems, the present invention aims to provide a feedback-driven iterative method for planning high-orbit satellite collaborative observation missions to overcome or at least partially solve the above problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A feedback-driven iterative method for planning high-orbit satellite collaborative observation missions includes the following steps: S1. In the planning and generation phase, a structured collaborative observation scheme is generated based on the data of the spatial targets to be observed, the high-orbit satellite data participating in the collaborative observation, the planning optimization parameters, and the satellite resource constraints through a multi-objective optimization algorithm. The scheme is then distributed to the ground center for retention and to satellite-specific observation command packages. S2. During the on-orbit execution phase, based on the satellite's dedicated observation command package and on-orbit dynamic status, commands are executed and observation actions are dynamically adjusted. Observational results data, collaborative status data, and resource consumption data are collected synchronously and in real-time according to timestamps, and an on-orbit execution dataset is output. S3. In the data feedback and evaluation phase, based on the ground center retention scheme and the on-orbit execution dataset, a quantitative indicator system for point target detection rate, multi-satellite coordination and payload resource waste rate is constructed, and performance bottlenecks are identified, generating feedback indicators and bottleneck analysis results. S4. In the strategy correction stage, a quantitative mapping model between feedback indicators and planning parameters is established based on the PID control concept. The feedback indicators and bottleneck analysis results are transformed into optimization parameters for the planning generation stage, realizing the quantitative mapping between feedback information and strategy adjustment, and outputting a set of planning optimization parameters.

[0008] Preferably, in step S1, the space target data to be observed includes the orbit prediction data, target priority, and target size and surface characteristics of the space target to be observed; the high-orbit satellite data participating in the collaborative observation includes the remaining energy, payload working status, field of view coverage and inter-satellite communication bandwidth of the high-orbit satellites participating in the collaborative observation; and the strategy correction parameters include the task priority adjustment coefficient, payload working mode correction amount and multi-satellite synchronization time deviation compensation value from the strategy correction stage. The method for generating a structured collaborative observation scheme is as follows: maximizing the point target detection rate, minimizing resource waste rate, and maximizing collaboration are incorporated into a unified optimization objective. Using the NSGA-II algorithm as the core, decision variables are set as the satellite-target matching relationship, observation time window, and payload operating parameters. Simultaneously, strategy correction parameters are introduced to adjust target weights. Based on satellite resource constraints, the feasibility of candidate schemes generated during the optimization process is verified. Satellite resource constraints include energy constraints, field-of-view constraints, and synchronization constraints. From the candidate schemes that meet the constraints, the scheme with the optimal comprehensive combination of detection rate, resource rate, and collaboration is selected to generate a structured collaborative observation scheme. The scheme is then distributed bidirectionally between the ground center and the satellites. The ground center-retained scheme includes a task allocation table for all satellites, expected performance benchmark values, and constraint boundary conditions. The satellite-specific observation command package includes the target orbit coordinates, observation time window, payload operating parameters, and multi-satellite synchronization command trigger time for each satellite.

[0009] Preferably, in step S2, the satellite's on-orbit dynamic status includes real-time orbital deviation, real-time payload performance, and sudden abnormal signals; The execution of commands and dynamic adjustment of observation actions are as follows: the satellite receives a dedicated observation command packet and parses it into low-level control commands, and then adjusts the satellite's pointing through the attitude control system to ensure that the target enters the payload's field of view; when an emergency occurs, real-time adjustment is triggered; if the satellite's real-time orbit deviation exceeds the threshold, the observation window duration is automatically extended; if the payload signal-to-noise ratio is lower than the threshold, the observation is suspended and the observation interruption status is recorded; if a satellite fails and cannot perform the task, other satellites automatically share its high-priority targets based on preset collaborative backup rules. The on-orbit execution dataset output during the on-orbit execution phase is transmitted back in batches via the satellite-to-ground link. It includes timestamps, satellite IDs, data type tags, complete observation results data, collaborative status data, resource consumption data, and descriptions of sudden anomalies. The transmission strategy prioritizes the transmission of observation data from high-priority targets, delays the transmission of non-critical data, and ensures low packet loss rate through a retransmission mechanism.

[0010] Preferably, in step S3, the input for the data feedback evaluation stage is the on-orbit execution dataset from the on-orbit execution stage, and the ground retention scheme from the planning generation stage, including the expected point target detection rate, expected multi-satellite coordination, expected resource waste rate, as well as the initial orbit prediction data and the initial values ​​of satellite resource constraints. The generated feedback indicators and bottleneck analysis results include: The input raw data is preprocessed by outlier removal, missing value completion, and data alignment. Based on the preprocessed data, the quantitative index system of point target detection rate, multi-satellite coordination degree, and payload resource waste rate is used to calculate and statistically analyze the values ​​of each index by target type and satellite group. Combining index deviation and raw data details, the bottleneck causes are analyzed using the logic of index deviation, data tracing, and cause location. Output a quantitative feedback report, including: a core feedback indicator table, which presents the quantitative results by target type, satellite group, and indicator type, and marks the actual value, expected value, and deviation rate; and a performance bottleneck analysis report, including bottleneck type, scope of impact, key evidence, and preliminary optimization direction.

[0011] Preferably, step S4, the strategy correction stage, includes the following: Based on the bottleneck type, a pre-defined index and parameter mapping rule base is matched. This rule base is a quantitative mapping model of feedback indices and planning parameters built based on engineering experience and PID control logic. Based on each quantitative mapping model, specific optimization parameter values ​​are calculated, while constraint boundaries are introduced to avoid over-adjustment of parameters. The feasibility of the calculated optimization parameters is verified to ensure that the application of the parameters will not cause new constraint conflicts. The set of planning optimization parameters is output, including task priority adjustment coefficients, orbit prediction correction coefficients, payload working parameter correction amounts, and multi-satellite synchronization command adjustment values. The parameter descriptions include the calculation basis, application scope, and validity period of each parameter.

[0012] Preferably, the quantitative indicator system for point target detection rate, multi-satellite coordination degree and payload resource waste rate in step S3 includes the quantitative calculation model for each indicator of point target detection rate, multi-satellite coordination degree and payload resource waste rate, as well as the key factors affecting each indicator. The quantitative calculation model for point target detection rate is as follows:

[0013] Among them, R d Let N be the point target detection rate. s N represents the number of successful detections. p N represents the planned number of observations. i The number of observation interruptions; The key factors affecting the point target detection rate are planning-related factors that are not sudden anomalies, including trajectory prediction errors and load parameter matching degree; The quantitative calculation model for multi-star synergy is as follows:

[0014] in, For multi-satellite coordination, K represents the total number of coordinated observations of the same type of target by the satellite group. The planned synchronization duration for the k-th observation, These are the actual start times of the k-th observation for Satellite 1 and Satellite 2, respectively. Let k be the actual synchronization time difference of the k-th observation. To synchronize the duration of the Pi Champion Project, ; The key factors affecting multi-satellite coordination are planning parameter design issues, including the timing of synchronization command triggering and field-of-view coordination accuracy; The quantitative calculation model for load resource waste rate is as follows:

[0015] in, To reduce the waste rate of payload resources, Ineffective observation energy consumption refers to the total energy consumed by observations that failed without any sudden anomalies. This represents the total energy consumption of the payload observations, i.e., the energy consumed by all observation tasks. The rated power of the m-th load. The invalid observation duration for the m-th load. The total observation duration for the m-th load; The key factors affecting the waste rate of payload resources are the resource allocation strategy during the planning stage, including the redundancy of observation windows and the redundancy of payload parameters.

[0016] Preferably, the quantitative mapping model between feedback indicators and planning parameters based on PID control includes: based on the proportional, integral, and derivative principles of PID control, mapping algorithms are designed for three types of feedback indicators—point target detection rate, multi-satellite coordination, and load resource waste rate—to be mapped to planning parameters. Each mapping algorithm includes deviation calculation, parameter correction, and constraint verification. Mappings are constructed between point target detection rate and task priority adjustment coefficient, multi-satellite coordination and synchronization command lead time, and load resource waste rate and load working time compression coefficient. Lyapunov stability analysis is used to prove theoretical convergence, and multiple iterative examples are used to verify effectiveness.

[0017] The preferred mapping model between point target detection rate and task priority adjustment coefficient is as follows:

[0018] in, Adjust the task priority coefficients for the next iteration. This represents the priority coefficient for the current round. This is the initial value for the priority coefficient. The detection rate deviation in the kth round, The target detection rate is... The actual detection rate in the kth round. This is the proportionality coefficient. The integral coefficient is... and These are the upper and lower bound constraints for the priority coefficient, respectively. When detection rate When this happens, the priority coefficient is restored to its initial value to balance resource allocation; When detection rate At the same time, the proportional term quickly responds to the current deviation, and the integral term eliminates the historical accumulated deviation, jointly generating the priority coefficient for the next round, ensuring that the detection rate gradually converges to the target value.

[0019] The preferred mapping model between multi-satellite coordination degree and synchronization command lead time is as follows:

[0020] in, To allow more time for synchronization instructions in the next iteration. This is the initial value for the synchronization command advance time. To advance the synchronization instructions for the current iteration, The deviation of the coordination degree in the kth round, , The target value for multi-star coordination. The actual multi-star coordination degree in round k. The proportionality coefficient for multi-star coordination deviation. The differential coefficients of the multi-star coordination deviation are... Upper and lower limits are set for the advance time of synchronization instructions.

[0021] Preferably, the mapping model between load resource waste rate and load working time compression coefficient is as follows:

[0022] in, This is the compression factor for the load operating time in the next iteration. The initial value of the compression coefficient is the load operating time. This is the compression factor for the load operating time in the current iteration. The deviation in the load resource waste rate of the first round, , The target value for load resource waste rate, The deviation of the actual load resource waste rate in round k. This is the proportional coefficient for the deviation of the load resource waste rate. This is the integral coefficient for the deviation of the load resource waste rate. The upper and lower limits of the compression coefficient for the load working time are constrained.

[0023] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses a feedback-driven iterative method for planning high-orbit satellite collaborative observation missions, constructing a closed-loop planning architecture of planning generation-on-orbit execution-data feedback evaluation-strategy correction, which significantly improves the point target detection rate, multi-satellite coordination and resource utilization efficiency, while also having the ability to quickly respond to sudden failures, achieve stable convergence and resist environmental interference, providing a feasible engineering solution for planning high-orbit satellite collaborative observation missions; For the first time, on-orbit execution feedback is used as the core input for planning optimization. Through multiple iterations, dynamic adaptation of observation scheme, space environment, and target dynamics is achieved, solving the problem of disconnect between traditional planning and actual execution. A feedback-planning quantitative mapping mechanism guided by point target detection rate is established. For the first time, the point target detection rate is used as the core optimization target. Through a quantitative model, the detection rate deviation and coordination error are transformed into task priority adjustment coefficients and resource allocation corrections, avoiding the instability problem of traditional qualitative adjustment. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0025] Figure 1 A schematic diagram of a feedback-driven high-orbit satellite collaborative observation mission planning iterative method provided by the present invention; Figure 2 The iterative operation flowchart provided for this invention; Figure 3 This is a schematic diagram illustrating the correlation between key influencing factors of the core feedback indicators provided by this invention. Figure 4 A schematic diagram illustrating the iterative process and index changes in the validity verification of the examples provided by this invention; Figure 5 This diagram illustrates a performance comparison between the present invention and traditional frameworks in a typical scenario. Figure 6 A bar chart comparing the core performance indicators of this invention with those of traditional methods in a typical scenario. Figure 7 The fault target detection rate over time is provided by the present invention; Figure 8 The line graph showing the changes of the core performance indicators of this invention with the iteration rounds. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] This invention discloses a feedback-driven iterative method for planning high-orbit satellite cooperative observation missions, such as... Figure 1 and Figure 2 This includes the following steps: S1. In the planning and generation phase, a structured collaborative observation scheme is generated based on the data of the spatial targets to be observed, the high-orbit satellite data participating in the collaborative observation, the planning optimization parameters, and the satellite resource constraints through a multi-objective optimization algorithm. The scheme is then distributed to the ground center for retention and to satellite-specific observation command packages. S2. During the on-orbit execution phase, based on the satellite's dedicated observation command package and on-orbit dynamic status, commands are executed and observation actions are dynamically adjusted. Observational results data, collaborative status data, and resource consumption data are collected synchronously and in real-time according to timestamps, and an on-orbit execution dataset is output. S3. In the data feedback and evaluation phase, based on the ground center retention scheme and the on-orbit execution dataset, a quantitative indicator system for point target detection rate, multi-satellite coordination and payload resource waste rate is constructed, and performance bottlenecks are identified, generating feedback indicators and bottleneck analysis results. S4. In the strategy correction stage, a quantitative mapping model between feedback indicators and planning parameters is established based on the PID control concept. The feedback indicators and bottleneck analysis results are transformed into optimization parameters for the planning generation stage, realizing the quantitative mapping between feedback information and strategy adjustment, and outputting a set of planning optimization parameters.

[0028] To further implement the above technical solution, in step S1, the data of the space target to be observed includes the orbit prediction data of the space target to be observed, the target priority, and the target size and surface characteristics (used to match the payload type). The data of the high-orbit satellites participating in the collaborative observation includes the remaining energy of the high-orbit satellites participating in the collaborative observation, the payload working status (whether the optical / radar payload is available, the current signal-to-noise ratio threshold), the field of view coverage (the field of view coordinates for the next 24 hours calculated based on the satellite orbit), and the inter-satellite communication bandwidth. The strategy correction parameters include the task priority adjustment coefficient, the payload working mode correction amount, and the multi-satellite synchronization time deviation compensation value from the strategy correction stage. The method for generating a structured collaborative observation scheme is as follows: maximizing the point target detection rate, minimizing resource waste rate, and maximizing collaboration are incorporated into a unified optimization objective. Using the NSGA-II algorithm as the core, decision variables are set as the satellite-target matching relationship, observation time window, and payload operating parameters. Simultaneously, strategy correction parameters are introduced to adjust target weights. Based on satellite resource constraints, the feasibility of candidate schemes generated during the optimization process is verified. Satellite resource constraints include energy constraints (energy consumption per observation ≤ 10% of remaining energy), field-of-view constraints (the target remains continuously within the satellite's field of view within the observation window, with a deviation ≤ 0.1°), and synchronization constraints (multi-satellite observation time difference ≤ 1 second, compensating for inter-satellite communication delay). From the candidate schemes that meet these constraints, the scheme with the optimal comprehensive combination of detection rate, resource efficiency, and collaboration is selected to generate a structured collaborative observation scheme. The scheme is then distributed bidirectionally between the ground center and the satellites, including a ground center-retained scheme and satellite-specific observation command packages. The ground center-retained scheme includes the task allocation table for all satellites, expected performance benchmark values, and constraint boundary conditions. The satellite-specific observation command packages include the target orbit coordinates, observation time window, payload operating parameters, and multi-satellite synchronization command trigger time for each satellite.

[0029] To further implement the above technical solution, in step S2, the satellite's on-orbit dynamic status includes real-time orbital deviation, real-time payload performance, and sudden abnormal signals; the satellite's dedicated observation command package is transmitted encrypted via the satellite-to-ground link using the AES-256 encryption algorithm, and the command is verified. The process of executing commands and dynamically adjusting observation actions is as follows: The satellite receives a dedicated observation command packet and parses it into low-level control commands (such as "10:00:00 Start optical payload, exposure time 2s, target coordinates (J2000 coordinate system: X=XXX, Y=XXX, Z=XXX)"). Then, the satellite's pointing is adjusted through the attitude control system to ensure that the target enters the payload's field of view. When an emergency occurs, real-time adjustments are triggered. If the satellite's real-time orbital deviation exceeds the threshold, the observation window duration is automatically extended. If the payload's signal-to-noise ratio is lower than the threshold (optical SNR < 3, radar SNR < 2), observation is paused and the observation interruption status is recorded. If a satellite fails to perform its mission, other satellites automatically share the high-priority target based on preset collaborative backup rules. The on-orbit execution dataset output during the on-orbit execution phase is transmitted back in batches via the satellite-to-ground link. It includes timestamps, satellite IDs, data type tags, complete observation results data, collaborative status data, resource consumption data, and descriptions of sudden anomalies (such as "low payload signal-to-noise ratio at 10:05:00, observation interrupted"). The transmission strategy prioritizes the transmission of observation data of high-priority targets (such as the detection results of faulty satellites), while non-critical data (such as resource consumption details) is transmitted with a delay. A retransmission mechanism is used to ensure low packet loss during transmission.

[0030] In this embodiment, the observation result data includes the point target detection status (success / failure / interruption), the actual coordinates of the target at the detection time (obtained through payload measurement), and the payload output parameters (optical image grayscale value, radar echo signal strength). The collaborative status data includes the multi-satellite synchronization time difference (the deviation between the actual observation time and the planned synchronization time) and the inter-satellite communication delay (the time difference between command transmission and reception). The resource consumption data includes the energy consumption of a single observation, the payload working time, and the storage resource occupation (the amount of observation data stored).

[0031] To further implement the above technical solution, in step S3, the input of the data feedback evaluation stage is the on-orbit execution dataset of the on-orbit execution stage and the ground retention scheme of the planning generation stage, including the expected point target detection rate, expected multi-satellite coordination degree, expected resource waste rate, as well as the initial orbit prediction data and the initial value of satellite resource constraints. The generated feedback indicators and bottleneck analysis results include: The input raw data is preprocessed by outlier removal, missing value completion, and data alignment. Outlier removal involves deleting data that contradicts the detection status and payload parameters, such as successful detection but zero radar echo intensity; missing value completion involves using the average of five adjacent records to fill in the missing data in the coordinated time difference data; data alignment involves grouping by target ID and observation window to align the execution data of multiple satellites with the planning scheme for the same target. Based on the preprocessed data, the quantitative index system of point target detection rate, multi-satellite coordination degree and payload resource waste rate is used to calculate and statistically analyze the values ​​of each index according to target type and satellite group; combined with index deviation and original data details, the bottleneck causes are analyzed using the logic of index deviation, data tracing and cause location. If the point target detection rate is lower than expected, the deviation between the actual coordinates and the predicted coordinates of the source target is considered. If the deviation exceeds 3km, it is determined that the low detection rate is caused by orbit prediction error. If the deviation is ≤1km but the payload SNR is low, it is determined that the low detection rate is caused by insufficient payload performance. If the multi-satellite coordination is lower than expected, the inter-satellite communication delay is considered. If the delay exceeds 0.5 seconds, it is determined that the communication delay causes synchronization failure. If the delay is ≤0.3 seconds but the synchronization command is triggered late, it is determined that the command scheduling deviation causes insufficient coordination. Output a quantitative feedback report, including: a core feedback indicator table, which presents the quantitative results by target type, satellite group, and indicator type, and marks the actual value, expected value, and deviation rate; and a performance bottleneck analysis report, including bottleneck type, scope of impact, key evidence, and preliminary optimization direction.

[0032] To further implement the above technical solution, step S4, the strategy correction stage, specifically includes the following: Based on the bottleneck type, a pre-defined index and parameter mapping rule base is matched. This rule base is a quantitative mapping model of feedback indices and planning parameters built based on engineering experience and PID control logic. Based on each quantitative mapping model, specific optimization parameter values ​​are calculated, while constraint boundaries are introduced to avoid over-adjustment of parameters. The feasibility of the calculated optimization parameters is verified to ensure that the application of the parameters will not cause new constraint conflicts. The set of planning optimization parameters is output, including task priority adjustment coefficients, orbit prediction correction coefficients, payload working parameter correction amounts, and multi-satellite synchronization command adjustment values. The parameter descriptions include the calculation basis, application scope, and validity period of each parameter.

[0033] To further implement the above technical solution, the quantitative indicator system for point target detection rate, multi-satellite coordination degree, and payload resource waste rate in step S3 includes quantitative calculation models for each indicator and key factors affecting each indicator, such as... Figure 3 ; Point target detection rate is a core indicator for measuring whether a planning scheme can accurately capture targets. It is defined as the proportion of successful target detection in effective observations, and the quantitative calculation model is as follows:

[0034] Among them, R d Let N be the point target detection rate. s N represents the number of successful detections. p N represents the planned number of observations. i The number of observation interruptions; The target value for point target detection rate is set to The number of successful detections is the number of records with "detection status = success" collected during the on-orbit execution phase. It must meet the payload parameter thresholds: optical payload SNR≥5 and radar payload SNR≥3. The number of planned observations is the total number of observations of this type of target set during the planning and generation phase. For example, a faulty satellite target is planned to be observed 10 times. The number of observation interruptions is the number of observations terminated during the on-orbit execution phase due to sudden anomalies, such as sudden energy drops or payload failures. These must be excluded by marking them in the anomaly log.

[0035] The key factors affecting the point target detection rate are planning-related factors that are not sudden anomalies, including trajectory prediction errors and load parameter matching degree; Orbit prediction error: If the deviation between the actual target orbit and the predicted orbit is >3km (high-orbit satellite field of view angle) The corresponding spatial deviation will cause the target to leak out of the field of view, at which point N s Reduce, R dDecrease; Load parameter matching: If the optical load exposure time is too short (<1s) resulting in SNR<5, or the radar load power is too low (<80% of rated power) resulting in weak echo signal, the detection success rate will be reduced, which will also lead to R d decline; Multi-satellite coordination efficiency measures the effectiveness of multiple satellites in synchronously observing the same target as planned. It is defined as the ratio of effective synchronous observation time to planned synchronous observation time, calculated by grouping satellites by target type. The quantitative calculation model is as follows:

[0036] in, For multi-star coordination, the target value is set to K represents the total number of collaborative observations by the satellite group of the same type of target. The planned synchronization duration for the k-th observation (set during the planning and generation phase). These are the actual start times of the k-th observation for Satellite 1 and Satellite 2 (collected during the on-orbit execution phase), respectively. Let k be the actual synchronization time difference of the k-th observation. To synchronize the duration of the Pi Champion Project, ; The key factors affecting multi-satellite coordination are planning parameter design issues, including the timing of synchronization command triggering and field-of-view coordination accuracy; Synchronization command triggering time: If the inter-satellite communication delay is not compensated during the planning stage (e.g., satellite 2 needs to trigger the command 0.3 seconds earlier than satellite 1), the actual synchronization time difference will exceed the threshold; Field of view coordination accuracy: If the deviation of the satellite field of view overlap area calculated during the planning stage is >0.1°, it will cause a satellite to leave the field of view in advance, shortening the effective synchronization time; Payload resource waste rate is used to measure whether a planning scheme can make efficient use of satellite resources. It is defined as the proportion of resources consumed by invalid observations to the total resources. Invalid observations refer to observations that fail to detect and do not have any sudden anomalies. Payload resource waste rate is calculated separately for each payload type and satellite (optical and radar payloads have significantly different energy consumption and need to be evaluated separately). Energy consumption is the core metric, and the quantitative calculation model is as follows:

[0037] in, To reduce the waste rate of payload resources, , Ineffective observation energy consumption refers to the total energy consumed by observations that failed without any sudden anomalies. This represents the total energy consumption of the payload observations, i.e., the energy consumed by all observation tasks. The rated power of the m-th load (optical load) radar payload (Refer to actual satellite payload parameters) This is the invalid observation duration for the m-th payload (acquired during the on-orbit execution phase, excluding interruption duration caused by sudden anomalies). The total observation duration for the m-th load; The key factor affecting the waste rate of payload resources is the resource allocation strategy during the planning stage, including the redundancy of observation windows and payload parameters. If the planned observation window is too long (e.g., the target only stays in the field of view for 5 minutes, but a 10-minute observation is planned), it will lead to an increase in ineffective energy consumption. If the exposure time of the optical payload exceeds the requirements (e.g., setting a 5-second exposure when the target brightness is sufficient), or the power of the radar payload is too high (e.g., using the rated power for close-range observation), it will cause energy waste.

[0038] To further implement the above technical solution, the quantitative mapping model between feedback indicators and planning parameters based on the PID control concept includes: based on the proportional, integral, and derivative concepts of PID control, mapping algorithms are designed for three types of feedback indicators—point target detection rate, multi-satellite coordination degree, and load resource waste rate—to be mapped to planning parameters. Each mapping algorithm includes deviation calculation, parameter correction, and constraint verification. Mappings are constructed between point target detection rate and task priority adjustment coefficient, multi-satellite coordination degree and synchronization command lead time, and load resource waste rate and load working time compression coefficient. Lyapunov stability analysis is used to prove theoretical convergence, and effectiveness is verified through multiple iterative examples.

[0039] To further implement the above technical solution, the point target detection rate and task priority adjustment coefficient are mapped using a two-stage proportional and integral (PI) mapping. The proportional stage quickly corrects the current deviation, while the integral stage eliminates the accumulated deviation, avoiding repeated fluctuations in the detection rate. The mapping model is as follows:

[0040] in, Adjust the task priority coefficients for the next iteration. This represents the priority coefficient for the current round. This is the initial value for the priority coefficient. The detection rate deviation in the kth round, The target detection rate is... The actual detection rate in the kth round. This is the proportionality coefficient. Each 1% deviation in detection rate corresponds to a priority coefficient increment of 0.02. The integral coefficient (accumulates multiple rounds of deviation to eliminate persistent error). , and These are the upper and lower limits of the priority coefficient, respectively, to prevent excessively high priority from crowding out other target resources or excessively low priority from causing the target to be missed. When detection rate When this happens, the priority coefficient is restored to its initial value to balance resource allocation; When detection rate At the same time, the proportional term quickly responds to the current deviation, and the integral term eliminates the historical accumulated deviation, jointly generating the priority coefficient for the next round, ensuring that the detection rate gradually converges to the target value.

[0041] To further implement the above technical solution, the multi-satellite coordination degree and synchronization command lead time are mapped using a two-stage proportional and derivative (PD) mechanism. The proportional stage corrects the current synchronization deviation, while the derivative stage predicts the deviation change trend to avoid a sudden drop in coordination degree. The mapping model is as follows:

[0042] in, The advance timing of synchronization commands for the next iteration, i.e., the time required for synchronization commands to be sent in advance during inter-satellite collaborative observation, is used to compensate for communication delays, satellite attitude adjustment time, etc., and to ensure the synchronization of multi-satellite observation actions. This is the initial value for the synchronization command advance time. When the multi-satellite coordination deviation meets the convergence condition, this initial value is restored to balance resources and coordination accuracy. To advance the synchronization instructions for the current iteration, The deviation of the coordination degree in the kth round, , The target value for multi-star coordination. , The actual multi-star coordination degree in round k. This is a proportional coefficient for the multi-star coordination deviation, used to quickly respond to the coordination deviation of the current round, and directly determines the sensitivity of the lead time to the current deviation. This is the differential coefficient of the multi-satellite coordination deviation, used to reflect the rate of change of the deviation, avoiding excessive adjustment of lead time due to deviation fluctuations, and enhancing system stability. To constrain the lead time of synchronization instructions, the upper limit is 1.0 second to avoid resource redundancy caused by excessive lead time, and the lower limit is 0.1 second to ensure basic synchronization accuracy; For every 1% deviation in coordination, the lead time increases by 0.01 seconds. If the deviation increases, the lead time is increased; if the deviation decreases, the correction amount is decreased. like If the level of coordination is met, then ;like After parameter correction, the synchronization command advance time increases, compensating for greater communication delays, and the coordination level is expected to improve in the next round.

[0043] To further implement the above technical solution, the load operating time compression factor β is the compression ratio of the next round of observation time relative to the current round, with an initial value of... (Uncompressed), value range [0.5, 1.0] (to avoid excessive compression leading to a decrease in detection rate). The payload resource waste rate and payload working time compression coefficient are mapped using a two-stage proportional and integral (PI) mechanism. The proportional stage quickly compresses redundant time, while the integral stage prevents the waste rate from rebounding. The mapping model is as follows:

[0044] in, This is the compression factor for the load operating time in the next iteration. The initial value of the compression coefficient is the load operating time. This is the load duration compression factor for the current iteration, used to adjust the effective duration of a single load operation, reducing resource waste caused by invalid observations. Its value range is... , The deviation of the load resource waste rate in round k. , The target value for load resource waste rate, , The deviation of the actual load resource waste rate in round k. This is a proportional coefficient representing the deviation in load resource waste rate, used for rapid response to the current wheel's waste rate deviation. It directly determines the adjustment range of the compression coefficient to the current deviation, and its value is 0.01. This is the integral coefficient for the deviation of the load resource waste rate, used to accumulate historical deviations, eliminate long-standing steady-state errors in the waste rate, and ensure that the waste rate steadily converges to the target value. Its value is 0.003. The upper and lower limits of the compression factor for the payload's working time are constrained. The upper limit of 1.0 ensures that no additional resources are wasted, while the lower limit of 0.5 avoids excessive compression that could lead to insufficient effective observation time.

[0045] In this embodiment, the theoretical convergence is verified using the point target detection rate mapping as an example. Taking the PI mapping between the point target detection rate and the priority coefficient as an example, Lyapunov stability theory is used to prove that: when When, deviation It will gradually converge to 0 (or less than the threshold) with the increase of iteration rounds, specifically: Assumptions: Priority coefficient With detection rate Positive correlation: c is a constant. In engineering, prioritizing the detection rate will increase the observation resources, and the detection rate will inevitably increase. This assumption holds true.

[0046] Scale coefficient in mapping algorithm Integral coefficient In algorithm design, set , The conditions are met.

[0047] Lyapunov function construction: Define Lyapunov functions Obviously And only if and hour, .

[0048] Stability analysis: Calculate the difference of Lyapunov functions between adjacent rounds. ,like Then the mapping mechanism converges:

[0049] Assumption 1, ,and ,Will Substituting, we get:

[0050] Let c = 0.2 (for every 0.1 increase in α, Rd increases by 2%), and substitute it into... , ,but:

[0051] visible along with and Decrease and decrease, substitute Calculation verifiable The mapping mechanism for point target detection rate converges; similarly, it can be proven that the mapping mechanisms for multi-satellite coordination degree and payload resource waste rate also converge.

[0052] Validation of the example: Taking a high-orbit satellite collaborative observation system (4 satellites, 80 targets, including 20 faulty satellite targets) as an example, a five-round iterative process was simulated to verify the improvement effect of the three types of indicators under the mapping mechanism. The initial parameters are as follows: Initial priority coefficient Synchronization command advance time seconds, compression ratio Target value , , ; The iterative process and changes in indicators are as follows Figure 4The point target detection rate increased from 85% in the first round to 90% in the fifth round, the deviation converged from 5% to 0, and the priority coefficient was adjusted to 0.8, verifying that PI mapping can effectively improve the detection rate; the multi-satellite coordination improved from 80% to 85%, the deviation converged from 5% to 0, the synchronization lead time was adjusted to 0.49 seconds, and PD mapping compensated for communication delay, improving the coordination effect; the payload resource waste rate decreased from 20% to 15%, the deviation converged from 5% to 0, and the compression coefficient was adjusted to 0.84, effectively compressing redundant energy consumption.

[0053] Examples show that the mapping mechanism enables the three types of indicators to converge to the target value within 5 iterations, and the parameter adjustment does not exceed the constraint range, verifying the engineering effectiveness of the mechanism.

[0054] In another embodiment, to verify the effectiveness of the method of the present invention in improving the performance of high-orbit satellite collaborative observation and adapting to dynamic space environments, multiple sets of comparative experiments were designed based on a high-orbit satellite collaborative observation simulation platform: First, the hardware and software environment and core module functions of the simulation platform were clarified to ensure that the experimental scenarios closely match engineering practice; second, two typical scenarios, routine monitoring and emergency faults, were set up, and the differences between the two frameworks in three core indicators, namely point target detection rate, multi-satellite coordination degree, and payload resource waste rate, were compared with the traditional non-feedback planning framework; finally, the impact of the number of iterations on the performance of the framework of the present invention was analyzed to verify its stability and convergence.

[0055] To ensure the real-time performance and computational accuracy of the simulation, the hardware and software environment is configured as follows: For the hardware environment, the server uses an Intel Xeon Gold 6330 CPU (24 cores, 48 ​​threads), 128GB DDR4 memory, and 2TB SSD storage, supporting multi-threaded parallel computing (such as synchronous calculations for multi-satellite orbit prediction and collaborative scheduling). The client uses an Intel Core i7-12700H CPU and 32GB memory for experimental parameter configuration and result visualization. For the software environment, the operating system is Ubuntu 22.04LTS (stable support for aerospace simulation tools). The core simulation tools are STK 12.0 (Satellite Tool Kit, used for satellite orbit modeling, field of view calculation, and observation window analysis) and MATLAB R2023a (used for deploying planning algorithms, feedback mapping mechanisms, and data processing). Data interaction uses the MQTT protocol (a lightweight IoT protocol that simulates satellite-to-ground link data transmission, with a latency set to 0.3-0.5 seconds to match the actual communication latency in high orbits).

[0056] The platform comprises four core modules, with each module's functions and parameter settings referencing actual high-orbit satellite engineering: The satellite model module simulates four high-orbit observation satellites (geosynchronous orbit, altitude 35786km, inclination 0°), each equipped with an optical payload and a radar payload. The optical payload has a field of view of 1.5°, angular resolution of 0.8″, maximum detection range of 600km, rated power of 200W, and an SNR threshold of 5. The radar payload has a field of view of 8°, resolution of 6m, maximum detection range of 1200km, rated power of 1000W, and an SNR threshold of 3. Satellite resource constraints include an initial energy of 100kWh (charging rate 200W / h), a maximum single payload operating time of 15 minutes, and an inter-satellite communication bandwidth of 1Mbps. The space target model module simulates 80 targets to be observed, including 60 conventional targets (sunsynchronous orbit, altitude 500-800km). The system simulates two types of interference: solar flares (triggered once every 4 hours, lasting 30 minutes, with an optical payload SNR decrease of 30%-50%) and space debris (100 fragments with a diameter of 0.1-1m, with an optical payload false detection rate of 5%-8%). The framework algorithm deployment module deploys the framework of this invention (fully closed-loop architecture, using the NSGA-II algorithm for planning generation and a PID-based mapping mechanism for strategy correction) and the traditional framework ("offline planning - on-orbit execution" unidirectional process, with the NSGA-II algorithm generating a fixed scheme after it is generated).

[0057] The experiment was divided into three groups to verify performance in different dimensions: Group 1 was a performance comparison experiment in normal scenarios, comparing the average performance (detection rate, coordination, and waste rate) of the two frameworks in normal monitoring scenarios over 24 hours to verify the long-term advantages of the framework of this invention; Group 2 was an emergency scenario response experiment, comparing the changes in indicators of the two frameworks in emergency fault scenarios (abnormal orbit of the faulty target + solar flare) to verify the rapid response capability of the framework of this invention; Group 3 was an experiment on the impact of iteration number, using only the framework of this invention to analyze the relationship between the number of iteration rounds (1-4 rounds) and performance indicators in normal scenarios to verify the convergence and stability of the framework.

[0058] The performance metrics in this embodiment are divided into two categories: core performance metrics and convergence speed metrics. Core performance metrics follow the definitions of the feedback metric system and supplement them with scenario adaptation requirements to ensure consistency with the optimization goals of the fully closed-loop framework. Convergence speed metrics focus on the iterative optimization efficiency of the framework of this invention, and are specifically defined as follows: Core performance metrics include point target detection rate, multi-satellite coordination, and payload resource waste rate. The point target detection rate is subdivided by target type (regular targets, faulty targets). In emergency fault scenarios, it is additionally required that faulty targets reach ≥90% within 2 hours of the fault triggering to verify the framework's rapid support capability for high-priority targets. Multi-satellite coordination is statistically analyzed by satellite group (4 satellites divided into 1-2 groups). Groups 3-4 (average of two groups) are used, requiring an overall coordination rate of ≥85% to ensure the synchronous effectiveness of multi-satellite observations; payload resource waste rate is subdivided and statistically analyzed by payload type (optical payload, radar payload), requiring an overall waste rate of ≤15% to reflect resource utilization efficiency; convergence speed indicators include the number of convergence rounds and the amplitude of index fluctuation, where the number of convergence rounds refers to the number of iteration rounds required for the framework of this invention to reach the threshold of all core performance indicators from the initial state, and the fewer the rounds, the higher the framework response and optimization efficiency; the amplitude of index fluctuation refers to the maximum deviation of the core performance indicators in three consecutive iterations after the framework reaches the convergence state, and a deviation of ≤3% is considered as the framework performance being stable, avoiding over-adjustment or performance rebound.

[0059] Experimental Results and Analysis: Group 1: Comparison of average performance metrics between the two frameworks under standard monitoring scenarios (24 hours, 4 iterations) is as follows: Figure 5 and Figure 6 The framework of this invention significantly outperforms the traditional framework in all three core metrics: Significantly improved detection rate: The overall Rd of this invention reaches 91.2% (meeting the target threshold), while the traditional framework only achieves 72.5%. The difference mainly stems from faulty targets. This invention improves the priority of faulty targets through strategy modification (from 0.5 to 0.8) and increases the observation window, thus increasing the Rd. d2 The detection rate was improved by 34.7%; however, the traditional framework could not adjust the priority, and due to large track errors, the detection rate of faulty targets was only 65.8%.

[0060] Significant optimization in synergy: The R framework in this paper... s The synchronization rate reached 87.8%, while the traditional framework only reached 63.2%. This invention compensates for the synchronization deviation caused by inter-satellite communication delay and environmental interference by correcting the synchronization command advance time (increasing tadv from 0.3 seconds to 0.48 seconds). The traditional framework has fixed synchronization commands, which are prone to coordination failure due to interference (such as satellite attitude fine-tuning during solar flares, with a synchronization time difference of more than 1 second).

[0061] Resource waste is significantly reduced: The R of this invention wThe proportion of invalid observations is only 13.5%, while that of the traditional framework is 28.7%. This invention reduces invalid observations by using a load duration compression coefficient; the traditional framework has a fixed observation window and a high proportion of invalid energy consumption.

[0062] Group 2: Emergency Scenario Response Experiment Results In scenarios involving abnormal orbits of faulty targets and emergency failures caused by solar flares, the changes in the faulty target detection rate over time for both types of frames are as follows: Figure 7 As shown, the rapid response advantage of the framework presented in this paper is significant: Framework of this article: When the fault is triggered in the 8th hour, R d2 The figure is 78%; after one round of iteration (after 1 hour, at the 9th hour), R... d2 Increased to 85%; in the 10th hour (after 2 iterations), R d2 Reaching 92% (meeting emergency mission requirements); during the 9th-10th hour solar flare, R d2 The decrease was only 2% (the decrease in optical payload SNR was compensated by switching radar payloads). Traditional framework: After the fault is triggered in the 8th hour, R d2 The rate continued to decline (due to increased orbital error leading to missed detections), dropping to 58% during the solar flare at the 9th hour, and had not recovered by the 10th hour (without strategy correction), thus failing to meet the requirements of the emergency mission.

[0063] Results analysis: Fast response speed: This invention achieves high-frequency iteration (1 hour / round) within 2 hours, reducing R... d2 The detection rate increased from 78% to 92%, while the traditional framework lacked a feedback mechanism and could not cope with the superposition of sudden failures and interference, resulting in a continuous deterioration of the detection rate. Strong anti-interference capability: When solar flares are triggered, the framework in this paper detects the decrease in the SNR of the optical payload through feedback and automatically increases the observation time of the radar payload (adjusted from 0.9 to 0.95) to avoid a significant drop in the detection rate; the traditional framework has a fixed payload working mode, and there is no alternative solution after the performance of the optical payload deteriorates, resulting in a sharp drop in the detection rate.

[0064] Group 3: The number of iterations affects the experimental results In typical scenarios, the core metrics of this framework change with iteration rounds (1-4 rounds) as follows: Figure 8 As shown, the indicator exhibits a convergence trend of rapid increase followed by stabilization: Point target detection rate: 85.0% in round 1 → 88.2% in round 2 → 91.0% in round 3 → 91.2% in round 4, and stabilized after round 3 (fluctuation ≤ 0.2%). Multi-star synergy: Round 1 80.5% → Round 2 84.3% → Round 3 87.5% → Round 4 87.8%, with fluctuations ≤0.3% after Round 3; Load resource waste rate: 18.2% in round 1 → 15.8% in round 2 → 13.6% in round 3 → 13.5% in round 4, with fluctuations of ≤0.1% after round 3.

[0065] Results analysis: Good convergence: All three types of indicators reached the target threshold within 3 rounds of iteration, and the indicators were basically stable in the 4th round, indicating that the mapping mechanism can effectively promote the indicators to converge toward the target value; High stability: After convergence, the fluctuation range of the index is ≤0.3%, with no over-adjustment or rebound. This is because the integral part in the mapping mechanism eliminates the cumulative deviation, and the proportional part avoids excessive correction of a single deviation, which meets the engineering requirements for stability. Diminishing returns of iteration: The performance improvement is greatest in the first and second rounds, and the improvement is significantly reduced in the third and fourth rounds. This indicates that the framework quickly optimizes the shortcomings in the early stage and enters a period of stable performance in the later stage. There is no need to continue high-frequency iteration, so the iteration cycle can be extended and the consumption of computing resources can be reduced.

[0066] Based on the combined results of the three sets of experiments, this invention demonstrates significant performance advantages, rapid response capabilities, stable convergence, and outstanding anti-interference capabilities in high-orbit satellite collaborative observation missions: In terms of performance, compared to the traditional framework, the overall point target detection rate is improved by 25.8% in normal scenarios, multi-satellite coordination is increased by 38.9%, and payload resource waste rate is reduced by 53.0%, with all core indicators reaching engineering thresholds, thus solving the problem of disconnect between planning and execution in the traditional framework; In terms of response, in emergency scenarios, the fault target detection rate is increased from 78% to 92% within 2 hours, and the payload can be automatically switched to deal with interference, while the detection rate of the traditional framework drops to 58%; In terms of stability, the indicators converge within 3 iterations, with fluctuations ≤0.3% after convergence, without excessive adjustments, and the diminishing returns of iterations support dynamic adjustment of the cycle to balance performance and resources; In terms of anti-interference, the indicators show small fluctuations when facing solar flares and debris interference, while the traditional framework experiences fluctuations of 20%-30%.

[0067] This embodiment establishes a high-orbit satellite collaborative observation simulation platform, designs two scenarios (routine and emergency), and completes three sets of comparative experiments to comprehensively verify the performance of the proposed framework. Experimental results show that the proposed framework, through a closed-loop iteration of "planning-execution-evaluation-correction," significantly improves point target detection rate, multi-satellite coordination, and resource utilization efficiency. It also possesses the ability to quickly respond to sudden faults, achieve stable convergence, and resist environmental interference, providing a feasible engineering solution for high-orbit satellite collaborative observation mission planning. The experimental conclusions also provide data support for the subsequent engineering applications of the framework (such as configuring actual satellite mission parameters and setting iteration cycles).

[0068] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0069] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A feedback-driven iterative method for planning high-orbit satellite collaborative observation missions, characterized in that, Includes the following steps: S1. In the planning and generation phase, a structured collaborative observation scheme is generated based on the data of the spatial targets to be observed, the high-orbit satellite data participating in the collaborative observation, the planning optimization parameters, and the satellite resource constraints through a multi-objective optimization algorithm. The scheme is then distributed to the ground center for retention and to satellite-specific observation command packages. S2. During the on-orbit execution phase, based on the satellite's dedicated observation command package and on-orbit dynamic status, commands are executed and observation actions are dynamically adjusted. Observational results data, collaborative status data, and resource consumption data are collected synchronously and in real-time according to timestamps, and an on-orbit execution dataset is output. S3. In the data feedback and evaluation phase, based on the ground center retention scheme and the on-orbit execution dataset, a quantitative indicator system for point target detection rate, multi-satellite coordination and payload resource waste rate is constructed, and performance bottlenecks are identified, generating feedback indicators and bottleneck analysis results. S4. In the strategy correction stage, a quantitative mapping model between feedback indicators and planning parameters is established based on the PID control concept. The feedback indicators and bottleneck analysis results are transformed into optimization parameters for the planning generation stage, realizing the quantitative mapping between feedback information and strategy adjustment, and outputting a set of planning optimization parameters.

2. The feedback-driven high-orbit satellite collaborative observation mission planning and iterative method according to claim 1, characterized in that, Step S1: The data of the space target to be observed includes the orbit prediction data, target priority, target size and surface characteristics of the space target to be observed; the data of the high-orbit satellites participating in the collaborative observation includes the remaining energy, payload working status, field of view coverage and inter-satellite communication bandwidth of the high-orbit satellites participating in the collaborative observation; the strategy correction parameters include the task priority adjustment coefficient, payload working mode correction amount and multi-satellite synchronization time deviation compensation value from the strategy correction stage. The method for generating a structured collaborative observation scheme is as follows: maximizing the point target detection rate, minimizing the resource waste rate, and maximizing the degree of collaboration are incorporated into a unified optimization objective: taking the NSGA-II algorithm as the core, the decision variables are set as the satellite-target matching relationship, the observation time window, and the payload working parameters, while introducing strategy correction parameters to adjust the target weights; Based on satellite resource constraints, the feasibility of candidate schemes generated during the optimization process is verified. Satellite resource constraints include energy constraints, field of view constraints, and synchronization constraints. From the candidate schemes that meet the constraints, the scheme with the best comprehensive combination of detection rate, resource rate, and coordination is selected to generate a structured collaborative observation scheme. The scheme is then distributed bidirectionally to the ground center and the satellites. The scheme retained by the ground center includes the task allocation table of all satellites, expected performance benchmark values, and constraint boundary conditions. The satellite-specific observation instruction package includes the target orbit coordinates, observation time window, payload operating parameters, and multi-satellite synchronization command trigger time for each satellite.

3. The feedback-driven high-orbit satellite collaborative observation mission planning iterative method according to claim 1, characterized in that, Step S2, the satellite's on-orbit dynamic status includes real-time orbital deviation, real-time payload performance, and sudden abnormal signals; The execution of commands and dynamic adjustment of observation actions are as follows: the satellite receives a dedicated observation command packet and parses it into low-level control commands, and then adjusts the satellite's pointing through the attitude control system to ensure that the target enters the payload's field of view; when an emergency occurs, real-time adjustment is triggered; if the satellite's real-time orbit deviation exceeds the threshold, the observation window duration is automatically extended; if the payload signal-to-noise ratio is lower than the threshold, the observation is suspended and the observation interruption status is recorded; if a satellite fails and cannot perform the task, other satellites automatically share its high-priority targets based on preset collaborative backup rules. The on-orbit execution dataset output during the on-orbit execution phase is transmitted back in batches via the satellite-to-ground link. It includes timestamps, satellite IDs, data type tags, complete observation results data, collaborative status data, resource consumption data, and descriptions of sudden anomalies. The transmission strategy prioritizes the transmission of observation data from high-priority targets, delays the transmission of non-critical data, and ensures low packet loss rate through a retransmission mechanism.

4. The feedback-driven high-orbit satellite collaborative observation mission planning iterative method according to claim 1, characterized in that, Step S3, the input for the data feedback evaluation stage is the on-orbit execution dataset from the on-orbit execution stage, and the ground retention scheme from the planning generation stage, including the expected point target detection rate, expected multi-satellite coordination, expected resource waste rate, as well as the initial orbit prediction data and the initial values ​​of satellite resource constraints; The generated feedback indicators and bottleneck analysis results include: The input raw data is preprocessed by outlier removal, missing value completion, and data alignment. Based on the preprocessed data, the values ​​of each indicator are calculated and statistically analyzed by target type and satellite group through a quantitative index system of point target detection rate, multi-satellite coordination degree and payload resource waste rate; By combining indicator deviations with details of the original data, the bottleneck causes are analyzed using the logic of indicator deviation, data tracing, and cause localization. Output a quantitative feedback report, including: a core feedback indicator table, which presents the quantitative results by target type, satellite group, and indicator type, and marks the actual value, expected value, and deviation rate; and a performance bottleneck analysis report, including bottleneck type, scope of impact, key evidence, and preliminary optimization direction.

5. The feedback-driven high-orbit satellite collaborative observation mission planning iterative method according to claim 1, characterized in that, Step S4, the strategy correction phase, specifically includes the following: Based on the bottleneck type, a pre-defined index and parameter mapping rule base is matched. This rule base is a quantitative mapping model of feedback indices and planning parameters built based on engineering experience and PID control logic. Based on each quantitative mapping model, specific optimization parameter values ​​are calculated, while constraint boundaries are introduced to avoid over-adjustment of parameters. The feasibility of the calculated optimization parameters is verified to ensure that the application of the parameters will not cause new constraint conflicts. The set of planning optimization parameters is output, including task priority adjustment coefficients, orbit prediction correction coefficients, payload working parameter correction amounts, and multi-satellite synchronization command adjustment values. The parameter descriptions include the calculation basis, application scope, and validity period of each parameter.

6. The feedback-driven high-orbit satellite cooperative observation mission planning iterative method according to claim 1, characterized in that, The quantitative indicator system for point target detection rate, multi-satellite coordination degree and payload resource waste rate in step S3 includes the quantitative calculation model for each indicator and the key factors affecting each indicator. The quantitative calculation model for point target detection rate is as follows: Among them, R d Let N be the point target detection rate. s N represents the number of successful detections. p N represents the planned number of observations. i The number of observation interruptions; The key factors affecting the point target detection rate are planning-related factors that are not sudden anomalies, including trajectory prediction errors and load parameter matching degree; The quantitative calculation model for multi-star synergy is as follows: in, For multi-satellite coordination, K represents the total number of coordinated observations of the same type of target by the satellite group. The planned synchronization duration for the k-th observation, These are the actual start times of the k-th observation for Satellite 1 and Satellite 2, respectively. Let k be the actual synchronization time difference of the k-th observation. To synchronize the duration of the Pi Champion Project, ; The key factors affecting multi-satellite coordination are planning parameter design issues, including the timing of synchronization command triggering and field-of-view coordination accuracy; The quantitative calculation model for load resource waste rate is as follows: in, To reduce the waste rate of payload resources, Ineffective observation energy consumption refers to the total energy consumed by observations that failed without any sudden anomalies. This represents the total energy consumption of the payload observations, i.e., the energy consumed by all observation tasks. The rated power of the m-th load. The invalid observation duration for the m-th load. The total observation duration for the m-th load; The key factors affecting the waste rate of payload resources are the resource allocation strategy during the planning stage, including the redundancy of observation windows and the redundancy of payload parameters.

7. The feedback-driven high-orbit satellite cooperative observation mission planning iterative method according to claim 1, characterized in that, The content of establishing a quantitative mapping model between feedback indicators and planning parameters based on the PID control concept includes: based on the proportional, integral, and derivative concepts of PID control, mapping algorithms are designed for three types of feedback indicators—point target detection rate, multi-satellite coordination degree, and load resource waste rate—and their mapping to planning parameters. Each mapping algorithm includes deviation calculation, parameter correction, and constraint verification; mappings are constructed between point target detection rate and task priority adjustment coefficient, multi-satellite coordination degree and synchronization command lead time, and load resource waste rate and load working time compression coefficient; Lyapunov stability analysis is used to prove theoretical convergence, and effectiveness is verified through multiple iterative examples.

8. The feedback-driven high-orbit satellite collaborative observation mission planning iterative method according to claim 7, characterized in that, The mapping model between point target detection rate and task priority adjustment coefficient is as follows: in, Adjust the task priority coefficients for the next iteration. This represents the priority coefficient for the current round. This is the initial value for the priority coefficient. The detection rate deviation in the kth round, The target detection rate is... The actual detection rate in the kth round. This is the proportionality coefficient. The integral coefficient is... and These are the upper and lower bound constraints for the priority coefficient, respectively. When detection rate When this happens, the priority coefficient is restored to its initial value to balance resource allocation; When detection rate At the same time, the proportional term quickly responds to the current deviation, and the integral term eliminates the historical accumulated deviation, jointly generating the priority coefficient for the next round, ensuring that the detection rate gradually converges to the target value.

9. The feedback-driven high-orbit satellite collaborative observation mission planning iterative method according to claim 7, characterized in that, The mapping model between multi-satellite coordination and synchronization command lead time is as follows: in, To allow more time for synchronization instructions in the next iteration. This is the initial value for the synchronization command advance time. To advance the synchronization instructions for the current iteration, The deviation of the coordination degree in the kth round, , The target value for multi-star coordination. The actual multi-star coordination degree in round k. The proportionality coefficient for multi-star coordination deviation. The differential coefficients of the multi-star coordination deviation are... Upper and lower limits are set for the advance time of synchronization instructions.

10. The feedback-driven high-orbit satellite cooperative observation mission planning iterative method according to claim 1, characterized in that, The mapping model between load resource waste rate and load working time compression coefficient is as follows: in, This is the compression factor for the load operating time in the next iteration. The initial value of the compression coefficient is the load operating time. This is the compression factor for the load operating time in the current iteration. The deviation in the load resource waste rate of the first round, , The target value for load resource waste rate, The deviation of the actual load resource waste rate in round k. This is the proportional coefficient for the deviation of the load resource waste rate. This is the integral coefficient for the deviation of the load resource waste rate. The upper and lower limits of the compression coefficient for the load working time are constrained.

Images