Satellite autonomous task planning system based on weather forecast

Abstract

The invention relates to the technical field of communication data processing, in particular to a satellite autonomous task planning system based on weather forecast, which comprises a meteorological data receiving module, an on-satellite weather forecast processing module, an autonomous task decision module and an instruction execution verification module. And the meteorological data receiving module obtains the global gridding weather forecast product through the satellite-ground link, and generates standardized format data after analysis. And the on-satellite weather prediction processing module fuses meteorological data and orbit parameters, calculates an effective imaging probability according to an observation condition evaluation model, and outputs an executable degree quantization map. And the autonomous task decision module generates a satellite control instruction sequence by adopting a dynamic rolling optimization strategy. And the instruction execution verification module executes the instruction, records imaging quality data and generates a feedback log. The system eliminates ground instruction transmission delay through on-satellite closed-loop decision, avoids cloud interference by using multi-source meteorological data fusion and a real-time feedback mechanism, and improves disaster emergency response efficiency and data acquisition effectiveness.

Description

Technical Field

[0001] This invention relates to the field of communication data processing technology, and in particular to a satellite autonomous mission planning system based on weather forecasting. Background Technology

[0002] Satellite mission planning is the process of systematically arranging satellite operation missions. Based on the principles of orbital dynamics and the coverage characteristics of ground stations, it analyzes mission requirements and resource constraints, and rationally allocates observation periods, communication windows, and energy budgets through optimization algorithms. Since satellite resources are limited and there is competition between missions, the planning system must dynamically schedule missions according to mission priorities and conflict resolution mechanisms to maintain operational efficiency and mission continuity.

[0003] Existing satellite mission planning technologies suffer from the following technical pain points in practical applications: Firstly, the system's heavy reliance on manual command transmission from the ground control center introduces communication delays and human processing time, resulting in slow mission response speeds and an inability to meet real-time requirements. Secondly, the influence of variable atmospheric conditions such as clouds and haze often obstructs data acquisition by optical sensors, rendering many imaging missions ineffective and causing significant resource waste. For example, in disaster emergency response scenarios, satellites need to quickly acquire images of disaster areas, but command transmission can take several hours, and sudden weather changes can lead to low image quality or complete image loss, severely impacting decision-making efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a satellite autonomous mission planning system based on weather forecasts. This invention solves the technical problems caused by communication delays introduced by manual command transmission from the ground control center and interference from variable atmospheric conditions with optical sensors, resulting in slow satellite mission response, invalid observation data, and low resource utilization, especially affecting decision-making efficiency in disaster emergency scenarios.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows:

[0006] This invention provides a satellite autonomous mission planning system based on weather forecasting, comprising:

[0007] The meteorological data receiving module is used to acquire global gridded weather forecast products provided by ground meteorological data services through the satellite-to-ground data link, parse and decode the weather forecast products, extract meteorological forecast information in the satellite's transit area, and convert the meteorological forecast information into standardized format data for use by the on-board processing unit.

[0008] The meteorological data receiving module is connected to the on-board weather forecasting and processing module and is used to provide the standardized format data to the on-board weather forecasting and processing module.

[0009] The on-board weather forecasting and processing module is used to receive the standardized format data, fuse the standardized format data with satellite orbit forecast data and payload parameters, calculate the effective imaging probability within the satellite's Earth observation window based on the observation condition evaluation model, and generate an executable quantifiable map of each observation target within a future orbital cycle.

[0010] The on-board weather forecasting and processing module is connected to the autonomous mission decision-making module and is used to provide the executable quantification map to the autonomous mission decision-making module.

[0011] The autonomous mission decision-making module is used to receive the executable quantification map, combine it with the pre-injected observation mission priority list and the current resource status of the satellite, adopt a dynamic rolling optimization strategy to make planning decisions, and output the satellite attitude control command sequence and payload working parameter set.

[0012] The autonomous mission decision-making module is connected to the instruction execution verification module and is used to provide the instruction execution verification module with the satellite attitude control instruction sequence and payload working parameter set.

[0013] The instruction execution verification module is used to receive the satellite attitude control instruction sequence and payload working parameter set, send the satellite attitude control instruction sequence and payload working parameter set to the satellite platform attitude control system and payload control system to complete the satellite's Earth orientation, attitude adjustment and imaging, record the meteorological conditions and imaging quality assessment results at the actual imaging time, and generate a mission execution feedback log.

[0014] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the meteorological data receiving module acquiring global gridded weather forecast products provided by ground meteorological data services through a satellite-to-ground data link;

[0015] The meteorological data receiving module also acquires real-time cloud images and water vapor data from other meteorological satellites via inter-satellite links or direct reception.

[0016] The meteorological data receiving module also connects to the satellite's own remote sensing historical data to build a historical weather pattern library;

[0017] The meteorological data receiving module integrates the global gridded weather forecast product, the real-time cloud image and water vapor data, and the data in the historical weather model library using data assimilation technology. With the ground forecast as the background field, it uses the real-time received observation data for correction to generate an on-board analysis field product.

[0018] The meteorological data receiving module provides the on-board analysis field products to the on-board weather forecasting and processing module.

[0019] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the on-board weather prediction processing module uses a lightweight convolutional neural network or a graph neural network as an observation condition evaluation model;

[0020] The input to the observation condition assessment model includes multi-band, multi-element data cubes from the on-board analysis field products of the meteorological data receiving module, as well as observation geometric information after the fusion of satellite orbit forecast data and payload parameters;

[0021] The output of the observation condition assessment model includes an imaging probability map, an uncertainty map, and a suggestion for the optimal observation band.

[0022] The on-board weather prediction processing module integrates the imaging probability map, uncertainty map, and optimal observation band suggestion to generate the executable quantization map.

[0023] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the instruction execution verification module uses the recorded meteorological conditions at the actual imaging time as actual weather data;

[0024] The instruction execution verification module uses the recorded actual imaging quality assessment results as the actual imaging quality data.

[0025] The instruction execution verification module calculates the deviation between actual weather data and predicted weather data;

[0026] The instruction execution verification module provides deviation data and actual imaging quality data to the on-board weather prediction and processing module.

[0027] The on-board weather prediction processing module performs online incremental learning or federated learning to fine-tune the model weights of the neural network model based on deviation data and actual imaging quality data.

[0028] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the autonomous mission decision-making module constructs a lightweight digital twin of the satellite platform and the mission on-board.

[0029] The digital twin synchronizes the satellite's fuel status, power status, storage status, and payload health status in real time.

[0030] The autonomous mission decision-making module receives an executable quantified map from the on-board weather prediction and processing module;

[0031] The autonomous task decision-making module inputs the executable quantification graph into the digital twin;

[0032] The autonomous task decision-making module performs Monte Carlo tree search or evolutionary algorithm simulation within the digital twin;

[0033] The autonomous mission decision-making module outputs a sequence of satellite attitude control commands and a set of payload operating parameters based on the simulation results.

[0034] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the multi-objective optimization decision-making simultaneously considers multiple optimization objectives, including maximizing long-term scientific returns, minimizing emergency mission response time, optimizing energy and fuel consumption, and avoiding high uncertainty areas;

[0035] The autonomous task decision-making module performs collaborative optimization of at least two optimization objectives during digital twin simulation.

[0036] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the autonomous mission decision module exchanges observation intentions with other satellites through inter-satellite links and performs distributed collaborative mission planning in a digital twin;

[0037] For all instruction sequences generated by autonomous decision-making, a hash digest is generated using lightweight blockchain technology and stored in an append-only log.

[0038] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: the instruction execution verification module, while imaging, uses the payload's fast preview mode to perform real-time on-orbit image quality assessment and generate actual imaging utility values.

[0039] The actual imaging utility value is compared with the expected value of the prediction module to generate deviation data;

[0040] The deviation data, together with the actual meteorological conditions, is fed back as a feedback data stream to the on-board weather forecasting and processing module in real time.

[0041] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention also includes: when the imaging quality is lower than a preset threshold, the instruction execution verification module immediately interrupts the current mission and sends an interruption signal and a resource release signal to the autonomous mission decision module;

[0042] The autonomous task decision-making module invokes the digital twin to replan and utilize the saved time and resources to execute backup tasks.

[0043] Furthermore, the satellite autonomous mission planning system based on weather forecasting described in this invention is characterized by further comprising: the mission execution feedback log being transmitted back to the ground system via a satellite-to-ground link, used for post-event analysis of the accuracy of weather forecasting and the effectiveness of autonomous decision-making, and used to support continuous optimization of the model.

[0044] Beneficial effects of this invention:

[0045] This invention eliminates the need for manual ground-based command transmission through an onboard autonomous decision-making system. The meteorological data receiving module directly processes multi-source meteorological data to generate standardized format data, while the onboard weather forecasting and processing module generates executable quantifiable maps in orbit. This enables the satellite to autonomously initiate observation tasks based on a pre-set task priority list in disaster emergency scenarios, solving the response lag problem caused by the traditional command transmission delay of several hours. It employs multi-source data assimilation and online learning mechanisms to improve meteorological forecast accuracy. The onboard weather forecasting and processing module integrates ground forecasts, real-time cloud images, and historical weather model libraries to generate analysis field products. The observation condition assessment model outputs imaging probability and uncertainty maps, guiding the autonomous task decision-making module to dynamically avoid high-uncertainty areas and optimize imaging band selection, significantly reducing the invalid shooting rate caused by cloud interference. A closed-loop feedback optimization link is constructed. The command execution verification module generates real-time actual imaging quality assessment results and meteorological deviation data, which are fed back to the prediction module to trigger incremental model learning. Simultaneously, a digital twin synchronizes satellite resource status to achieve multi-objective collaborative optimization. When imaging quality is detected to be below a threshold, the task is immediately interrupted and a backup observation is replanned, improving the satellite's resource utilization and data acquisition effectiveness under complex meteorological conditions. Attached Figure Description

[0046] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0047] Figure 1 The system architecture diagram of the satellite autonomous mission planning system based on weather forecasting provided in the embodiments of the present invention is shown. Detailed Implementation

[0048] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. 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. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0049] Please see Figure 1 The present invention provides a satellite autonomous mission planning system based on weather forecasting, comprising:

[0050] The meteorological data receiving module is used to acquire global gridded weather forecast products provided by ground meteorological data services through the satellite-to-ground data link, parse and decode the weather forecast products, extract meteorological forecast information in the satellite's transit area, and convert the meteorological forecast information into standardized format data for use by the on-board processing unit.

[0051] The meteorological data receiving module is connected to the on-board weather forecasting and processing module and is used to provide the standardized format data to the on-board weather forecasting and processing module.

[0052] The on-board weather forecasting and processing module is used to receive the standardized format data, fuse the standardized format data with satellite orbit forecast data and payload parameters, calculate the effective imaging probability within the satellite's Earth observation window based on the observation condition evaluation model, and generate an executable quantifiable map of each observation target within a future orbital cycle.

[0053] The on-board weather forecasting and processing module is connected to the autonomous mission decision-making module and is used to provide the executable quantification map to the autonomous mission decision-making module.

[0054] The autonomous mission decision-making module is used to receive the executable quantification map, combine it with the pre-injected observation mission priority list and the current resource status of the satellite, adopt a dynamic rolling optimization strategy to make planning decisions, and output the satellite attitude control command sequence and payload working parameter set.

[0055] The autonomous mission decision-making module is connected to the instruction execution verification module and is used to provide the instruction execution verification module with the satellite attitude control instruction sequence and payload working parameter set.

[0056] The instruction execution verification module is used to receive the satellite attitude control instruction sequence and payload working parameter set, send the satellite attitude control instruction sequence and payload working parameter set to the satellite platform attitude control system and payload control system to complete the satellite's Earth orientation, attitude adjustment and imaging, record the meteorological conditions and imaging quality assessment results at the actual imaging time, and generate a mission execution feedback log.

[0057] The following are supplementary explanations of the technical content of the specification, strictly adhering to patent drafting standards and technical consistency requirements:

[0058] The meteorological data receiving module receives global gridded weather forecast products transmitted from ground-based meteorological data services via a satellite-to-ground data link. The module performs parsing and decoding operations on these weather forecast products, extracting weather forecast information for the areas the satellite is about to pass over. Finally, the module converts the weather forecast information into standardized data format that conforms to the onboard processing unit's calling specifications.

[0059] The meteorological data receiving module establishes a communication link with the onboard weather forecasting and processing module via a physical connection or data interface. The meteorological data receiving module transmits standardized format data to the onboard weather forecasting and processing module, thus enabling onboard transmission of meteorological information.

[0060] The onboard weather forecasting and processing module receives standardized format data output from the meteorological data receiving module. It then fuses this standardized data with satellite orbit forecast data and payload parameters. The module invokes a pre-built observation condition assessment model and calculates the effective imaging probability of the satellite's Earth observation window based on the fused dataset. Finally, the module generates an executable quantifiable map of each observed target for the entire future orbital period, mapping meteorological suitability values ​​to spatial coordinates.

[0061] The onboard weather forecasting and processing module establishes a transmission channel with the autonomous mission decision-making module via a data bus. The onboard weather forecasting and processing module transmits executable quantifiable maps to the autonomous mission decision-making module, providing meteorological assessment basis for mission planning.

[0062] The autonomous mission decision-making module receives executable quantifiable maps from the onboard weather forecasting and processing module. It then accesses a pre-configured list of observation mission priorities in the satellite's storage unit, while simultaneously acquiring real-time information on the satellite's current energy status, storage space, and payload status. The module employs a dynamic rolling optimization strategy for decision-making, iteratively updating the mission sequence within continuous time windows. Finally, the module outputs a sequence of satellite attitude control commands and a matching set of payload operating parameters. The command sequence includes three-dimensional spatial pointing commands and timestamps.

[0063] The autonomous mission decision-making module connects to the command execution verification module via a command channel. The autonomous mission decision-making module sends satellite attitude control command sequences and payload operating parameter sets to the command execution verification module, thereby distributing control commands.

[0064] The command execution verification module receives the satellite attitude control command sequence and payload operating parameter set transmitted by the autonomous mission decision module. The module then sends the command sequence and parameter set to the satellite platform attitude control system and the payload control system's actuators. The satellite platform attitude control system performs Earth orientation and attitude adjustment according to the command sequence, while the payload control system configures imaging parameters and executes imaging based on the parameter set. During imaging, the command execution verification module simultaneously records actual meteorological parameters and imaging quality assessment results. Finally, the module integrates the recorded data to generate a mission execution feedback log, which includes timestamps, location coordinates, and quality assessment indicators.

[0065] The data flow between modules is as follows: meteorological data is received, processed, and converted into a standardized format, then input into the forecasting module to generate quantitative maps; the maps are input into the decision-making module, which combines task priority and resource status to output control commands; the commands are input into the execution module to drive satellite actions and generate feedback logs. The entire process forms a closed-loop technology chain from data acquisition to task execution and effect verification.

[0066] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the meteorological data receiving module acquiring global gridded weather forecast products provided by ground meteorological data services through a satellite-to-ground data link;

[0067] The meteorological data receiving module also acquires real-time cloud images and water vapor data from other meteorological satellites via inter-satellite links or direct reception.

[0068] The meteorological data receiving module also connects to the satellite's own remote sensing historical data to build a historical weather pattern library;

[0069] The meteorological data receiving module integrates the global gridded weather forecast product, the real-time cloud image and water vapor data, and the data in the historical weather model library using data assimilation technology. With the ground forecast as the background field, it uses the real-time received observation data for correction to generate an on-board analysis field product.

[0070] The meteorological data receiving module provides the on-board analysis field products to the on-board weather forecasting and processing module.

[0071] This invention's meteorological data receiving module receives global gridded weather forecast products transmitted from ground-based meteorological data services via a satellite-to-ground data link. The module also receives real-time cloud images and water vapor data broadcast by other on-orbit meteorological satellites via inter-satellite links. Furthermore, the module utilizes historical remote sensing data stored locally on the satellite to construct a historical weather model database. The module employs a data assimilation algorithm to process three types of data sources: global gridded weather forecast products as the background field, real-time cloud images and water vapor data as the observation field, and the historical weather model database as the statistical correction benchmark. The module performs multi-source data fusion calculations to generate an on-board analysis field product with enhanced spatiotemporal resolution. Finally, the module transmits the on-board analysis field product to the on-board weather prediction processing module via the on-board data bus.

[0072] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the on-board weather prediction processing module uses a lightweight convolutional neural network or a graph neural network as an observation condition evaluation model.

[0073] The input to the observation condition assessment model includes multi-band, multi-element data cubes from the on-board analysis field products of the meteorological data receiving module, as well as observation geometric information after the fusion of satellite orbit forecast data and payload parameters;

[0074] The output of the observation condition assessment model includes an imaging probability map, an uncertainty map, and a suggestion for the optimal observation band.

[0075] The on-board weather prediction processing module integrates the imaging probability map, uncertainty map, and optimal observation band suggestion to generate the executable quantization map.

[0076] This invention's onboard weather prediction processing module uses a lightweight convolutional neural network as the computational core of the observation condition assessment model. The observation condition assessment model receives onboard analysis field products from the meteorological data receiving module, extracting multi-band spectral features and multi-element meteorological parameters to form a data cube. Simultaneously, the model receives orbital parameters from the satellite orbit prediction system and sensor parameters from the payload control system, fusing them to generate an observation geometric information matrix. The model outputs three sets of data products: an imaging probability map with a grid-marked success probability value for each observation area; an uncertainty map quantifying the confidence level of the prediction results; and an optimal observation band suggestion providing spectral band selection schemes under different meteorological conditions. The onboard weather prediction processing module integrates these three sets of output data into a quantized map with a unified coordinate system, serving as an executable quantized map.

[0077] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the instruction execution verification module uses the recorded meteorological conditions at the actual imaging time as actual weather data;

[0078] The instruction execution verification module uses the recorded actual imaging quality assessment results as the actual imaging quality data.

[0079] The instruction execution verification module calculates the deviation between actual weather data and predicted weather data;

[0080] The instruction execution verification module provides deviation data and actual imaging quality data to the on-board weather prediction and processing module.

[0081] The on-board weather prediction processing module performs online incremental learning or federated learning to fine-tune the model weights of the neural network model based on deviation data and actual imaging quality data.

[0082] The instruction execution verification module of this invention records actual meteorological parameters to form actual weather data during satellite imaging. The module performs a rapid quality assessment on the imaging data to generate actual imaging quality data. It then calls upon predicted weather data provided by the onboard weather prediction processing module to calculate the numerical deviation from the actual weather data. The module transmits the deviation data and actual imaging quality data to the onboard weather prediction processing module via the onboard communication link. Upon receiving the feedback data, the onboard weather prediction processing module initiates a model update procedure: adjusting the neural network weight parameters using online incremental learning; or aggregating and updating parameters with other satellite models under a federated learning framework. This model fine-tuning process allows the prediction model to continuously adapt to actual meteorological changes.

[0083] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the autonomous mission decision-making module constructing a lightweight digital twin of the satellite platform and the mission on-board.

[0084] The digital twin synchronizes the satellite's fuel status, power status, storage status, and payload health status in real time.

[0085] The autonomous mission decision-making module receives an executable quantified map from the on-board weather prediction and processing module;

[0086] The autonomous task decision-making module inputs the executable quantification graph into the digital twin;

[0087] The autonomous task decision-making module performs Monte Carlo tree search or evolutionary algorithm simulation within the digital twin;

[0088] The autonomous mission decision-making module outputs a sequence of satellite attitude control commands and a set of payload operating parameters based on the simulation results.

[0089] This invention's autonomous mission decision-making module deploys a lightweight digital twin on the onboard computing platform. The digital twin acquires real-time information via the satellite bus, including propellant remaining quantity (representing fuel status), battery state of charge (representing battery capacity), remaining solid-state memory capacity (representing storage status), and payload sensor health status codes (representing payload health status). The autonomous mission decision-making module receives executable quantifiable maps transmitted from the onboard weather forecasting and processing module. It then inputs these executable quantifiable maps into the digital twin environment. Within the digital twin, a Monte Carlo tree search algorithm is executed to simulate execution paths for different mission sequences; or an evolutionary algorithm is executed to iteratively optimize mission sequences using genetic operators. The simulation process outputs the optimal mission execution plan. The autonomous mission decision-making module converts the simulation results into a satellite attitude control command sequence and a set of payload operating parameters.

[0090] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the multi-objective optimization decision-making simultaneously considers multiple optimization objectives, including maximizing long-term scientific returns, minimizing emergency mission response time, optimizing energy and fuel consumption, and avoiding high uncertainty areas;

[0091] The autonomous task decision-making module performs collaborative optimization of at least two optimization objectives during digital twin simulation.

[0092] This invention's multi-objective optimization decision-making process simultaneously considers four optimization objectives within a digital twin: maximizing long-term scientific returns to assess the scientific value of observational data; minimizing emergency mission response time to calculate mission execution delays; optimizing energy and fuel consumption to statistically measure resource consumption; and avoiding high-uncertainty regions by referencing the uncertainty map output by the onboard weather prediction processing module. The autonomous mission decision-making module employs a multi-objective optimization algorithm in simulation to handle at least two objectives: establishing the objective function weight matrix; calculating the Pareto optimal solution set; and selecting the optimal solution that satisfies the constraints. Collaborative optimization ensures that the decision-making scheme balances the needs of multiple missions.

[0093] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the autonomous mission decision module exchanges observation intentions with other satellites through inter-satellite links and performs distributed collaborative mission planning in a digital twin;

[0094] For all instruction sequences generated by autonomous decision-making, a hash digest is generated using lightweight blockchain technology and stored in an append-only log.

[0095] The autonomous mission decision-making module of this invention receives observation intention data packets sent by neighboring satellites via an inter-satellite laser communication link. These packets contain the target area coordinates and the observation time window. The autonomous mission decision-making module imports the external observation intentions into a digital twin collaborative simulation environment. The digital twin executes a distributed collaborative mission planning algorithm: coordinating the allocation of observation areas across multiple satellites; avoiding imaging time conflicts; and optimizing the overall constellation coverage efficiency. The autonomous mission decision-making module performs blockchain notarization on all generated instruction sequences: invoking a lightweight blockchain smart contract; generating a hash digest of the instruction sequence; and writing it to an append-only distributed log file. This notarization process provides audit trail capabilities for onboard autonomous decision-making.

[0096] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: the instruction execution verification module, while imaging, uses the payload's fast preview mode to perform real-time on-orbit image quality assessment and generate actual imaging utility values.

[0097] The actual imaging utility value is compared with the expected value of the prediction module to generate deviation data;

[0098] The deviation data, together with the actual meteorological conditions, is fed back as a feedback data stream to the on-board weather forecasting and processing module in real time.

[0099] The instruction execution verification module of this invention enables a fast preview mode during the payload imaging phase. This fast preview mode achieves millisecond-level preview frame generation by reducing image resolution. The instruction execution verification module analyzes the preview frames to calculate image sharpness and cloud coverage, generating an actual imaging utility value. The module then retrieves the expected utility value pre-provided by the onboard weather prediction processing module. Finally, the module calculates the numerical deviation between the actual and expected imaging utility values. This deviation data, along with meteorological parameters at the actual imaging time, is encapsulated into a feedback data stream. The module transmits this feedback data stream in real-time to the onboard weather prediction processing module via the onboard internal network.

[0100] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention further includes: when the imaging quality is lower than a preset threshold, the instruction execution verification module immediately interrupts the current mission and sends an interruption signal and a resource release signal to the autonomous mission decision module;

[0101] The autonomous task decision-making module invokes the digital twin to replan and utilize the saved time and resources to execute backup tasks.

[0102] The instruction execution verification module of this invention continuously monitors the imaging quality assessment results. When the imaging quality assessment result falls below a preset quality threshold, the instruction execution verification module immediately suspends the current imaging task. The instruction execution verification module sends an interrupt signal containing the task number to the autonomous task decision module. The interrupt signal includes a resource release signal indicating the reclaimable storage space and remaining imaging time. Upon receiving the signal, the autonomous task decision module activates the digital twin replanning module. The digital twin calls upon the resource parameters released by the interrupted task. The digital twin generates a backup task execution sequence to fill the resource gap. The autonomous task decision module outputs a new satellite attitude control instruction sequence and payload operating parameter set.

[0103] Specifically, the satellite autonomous mission planning system based on weather forecasting described in this invention is characterized by further comprising: the mission execution feedback log being transmitted back to the ground system via a satellite-to-ground link, used for post-event analysis of the accuracy of weather forecasts and the effectiveness of autonomous decision-making, and used to support continuous optimization of the model.

[0104] The instruction execution verification module of this invention transmits task execution feedback logs via a satellite-to-ground data transmission link. After receiving the task execution feedback logs, the ground system performs data analysis: extracting actual meteorological data and comparing it with forecast data to determine weather forecast accuracy; and calculating the task completion rate to assess the effectiveness of autonomous decision-making. The ground system generates model optimization instructions: updating neural network weight parameters; and adjusting the target weights of the optimization algorithm. The ground system sends the model optimization instructions to the satellite via the telemetry, tracking, and command (TT&C) station uplink. The onboard weather forecasting and processing module and the autonomous task decision-making module receive and apply the updated parameters. This closed loop supports the satellite system's ability to continuously optimize task planning.

[0105] This invention addresses the command delay issue by constructing a full-link onboard autonomous decision-making system. The meteorological data receiving module directly acquires meteorological forecast products and converts them into standardized data formats, eliminating the need for manual command uploading. The onboard weather forecasting and processing module processes data locally on the satellite, generating executable quantifiable maps, avoiding time-consuming data transmission from the ground control center. In disaster emergency scenarios, the satellite autonomously initiates high-priority tasks based on a pre-injected list of observation task priorities, achieving minute-level response and significantly shortening the traditional command transmission cycle of several hours.

[0106] This invention employs a multi-source fusion and real-time correction mechanism to address weather interference issues. The meteorological data receiving module integrates ground forecasts, multi-satellite real-time cloud images, and a historical weather model library, generating high-precision on-board analysis field products through data assimilation technology. The on-board weather prediction and processing module's observation condition assessment model, based on multi-band data cubes and observation geometry information, outputs imaging probability and uncertainty maps to accurately identify cloud-covered areas. When the predicted imaging probability of the target area is below a threshold, the autonomous mission decision module automatically switches to a backup observation area or adjusts the imaging band to reduce the invalid image capture rate.

[0107] This invention improves resource utilization and emergency response efficiency through closed-loop optimization. The command execution verification module records actual meteorological conditions and imaging quality assessment results in real time, generating deviation data that is fed back to the prediction module to trigger online model fine-tuning. In disaster scenarios, when the actual imaging quality falls below a preset threshold, the command execution verification module immediately interrupts the task and releases resources. The autonomous task decision module then invokes the digital twin to replan a backup task within seconds. The digital twin synchronizes with the satellite's real-time resource status, dynamically balancing scientific returns, emergency response speed, and fuel consumption in multi-objective optimization, enabling the satellite to maintain efficient data acquisition capabilities even under complex meteorological conditions.

[0108] The satellite-ground collaboration mechanism enhances system robustness. Mission execution feedback logs are transmitted back to the ground system for in-depth analysis. The ground system then optimizes model parameters and updates them via the satellite-ground link. During continuous operation, the federated learning framework enables multiple satellite models to share optimization experience, gradually improving the overall meteorological forecasting accuracy and mission decision reliability of the constellation, forming a complete solution from single-satellite autonomy to constellation collaboration.

[0109] Implementation

[0110] The meteorological data receiving module receives global gridded weather forecast products transmitted from ground-based meteorological data services via a satellite-to-ground data link. It performs parsing and decoding operations on the weather forecast products, extracting cloud cover and atmospheric transmittance information for the satellite's transit area. The module converts the weather forecast information into standardized data recognizable by the onboard processing unit. It also receives real-time cloud images and water vapor data broadcast from neighboring meteorological satellites via inter-satellite links. Furthermore, it constructs a historical weather model library by accessing locally stored remote sensing historical data on the satellite. The module employs a data assimilation algorithm to fuse three types of data sources: global gridded weather forecast products as the background field, real-time cloud images and water vapor data as the observation field, and the historical weather model library providing a statistical correction benchmark. Finally, the module generates an onboard analysis field product with enhanced spatiotemporal resolution and transmits it to the onboard weather prediction and processing module via the onboard bus.

[0111] The onboard weather forecasting and processing module uses a lightweight convolutional neural network as the core of its observation condition assessment model. This model receives onboard analysis field products and extracts multi-band spectral features and multi-element meteorological parameters to form a three-dimensional data cube. Simultaneously, it acquires orbital parameters from the satellite orbit prediction system and sensor parameters from the payload control system, fusing them to generate an observation geometric information matrix. The model outputs three sets of data products: an imaging probability map with a grid-based annotation of the successful imaging probability for each observation area; an uncertainty map quantifying the confidence level of the prediction results; and optimal observation band suggestions providing spectral band selection options under different meteorological conditions. The onboard weather forecasting and processing module integrates these three sets of output data into an executable, quantifiable map in a unified coordinate system, which is then transmitted to the autonomous mission decision-making module.

[0112] The autonomous mission decision-making module deploys a lightweight digital twin on the onboard computing platform. The digital twin synchronously displays propellant balance (representing fuel status), battery charge status (representing electrical charge status), memory remaining capacity (representing storage status), and payload health status codes (representing equipment status) in real time. The autonomous mission decision-making module receives executable quantification maps and inputs them into the digital twin environment. The digital twin executes a Monte Carlo tree search algorithm to simulate execution paths for different mission sequences, or uses an evolutionary algorithm to iteratively optimize mission sequences through genetic operators. The multi-objective optimization process simultaneously handles four objectives: maximizing long-term scientific returns to assess the scientific value of observational data; minimizing emergency mission response time to calculate mission execution delays; optimizing energy and fuel consumption to statistically measure resource consumption; and avoiding high-uncertainty areas by referencing the uncertainty map output by the onboard weather prediction processing module. The autonomous mission decision-making module outputs satellite attitude control command sequences and payload operating parameter sets.

[0113] The command execution verification module sends the attitude control command sequence and parameter set to the satellite attitude control system and payload control system. The satellite attitude control system performs Earth orientation and attitude adjustment, while the payload control system configures imaging parameters and initiates imaging. During imaging, the command execution verification module enables the payload's fast preview mode, generating low-resolution preview frames to analyze image sharpness and cloud coverage, and outputs the actual imaging utility value. The command execution verification module retrieves the expected utility value provided by the onboard weather prediction processing module and calculates the numerical deviation. When the imaging quality assessment result is lower than the preset quality threshold, the command execution verification module immediately interrupts the mission and sends a mission interruption signal and resource release signal to the autonomous mission decision module. The autonomous mission decision module calls the digital twin weight to plan a backup mission and uses the released resources to generate a new command sequence. The command execution verification module encapsulates the actual meteorological parameters, imaging quality assessment results, and deviation data into a feedback data stream and transmits it back to the onboard weather prediction processing module in real time.

[0114] The onboard weather forecasting and processing module updates the model based on feedback data: it adjusts the weight parameters of the convolutional neural network using online incremental learning, or it aggregates and updates parameters with other satellite models through a federated learning framework. Mission execution feedback logs are transmitted to the ground system via the satellite-to-ground link. The ground system analyzes the actual meteorological data in the logs, compares the forecast accuracy with the predicted data, calculates the mission completion rate to assess the effectiveness of autonomous decision-making, and generates model optimization instructions to be sent back to the satellite. In disaster emergency scenarios, when the satellite detects a sudden earthquake event, the autonomous mission decision-making module immediately initiates disaster area observation based on a preset mission priority list. The onboard weather forecasting and processing module analyzes cloud cover in real time and dynamically adjusts the imaging bands. The instruction execution verification module interrupts invalid imaging and redirects satellite resources when it detects sudden changes in cloud cover, forming a complete closed-loop response mechanism.

Claims

1. A weather forecast based autonomous mission planning system for satellites, characterized in that, The method comprises the following steps: a meteorological data receiving module is used to obtain global gridded weather forecast products provided by a ground meteorological data service through a satellite-ground data link, analyze and decode the weather forecast products, extract meteorological forecast information in a satellite transit area, and convert the meteorological forecast information into standardized format data for calling by a satellite processing unit; the meteorological data receiving module is connected to a satellite weather prediction processing module, and is used to provide the standardized format data to the satellite weather prediction processing module; a satellite weather prediction processing module is used to receive the standardized format data, fuse the standardized format data with satellite orbit prediction data and payload parameters, calculate an effective imaging probability in a satellite observation window according to an observation condition evaluation model, and generate an executable quantification atlas of each observation target in a future orbit period; the satellite weather prediction processing module is connected to an autonomous task decision module, and is used to provide the executable quantification atlas to the autonomous task decision module; an autonomous task decision module is used to receive the executable quantification atlas, combine a pre-injected observation task priority list and a current satellite resource state, adopt a dynamic rolling optimization strategy for planning and decision-making, and output a satellite attitude control instruction sequence and a payload working parameter set; the autonomous task decision module is connected to an instruction execution verification module, and is used to provide the satellite attitude control instruction sequence and the payload working parameter set to the instruction execution verification module; an instruction execution verification module is used to receive the satellite attitude control instruction sequence and the payload working parameter set, send the satellite attitude control instruction sequence and the payload working parameter set to a satellite platform attitude control system and an effective payload control system to complete satellite ground orientation, attitude adjustment and imaging, record meteorological conditions and imaging quality evaluation results at an actual imaging time, and generate a task execution feedback log.

2. The weather forecast based satellite autonomous mission planning system of claim 1, wherein, Further comprising: the meteorological data receiving module obtains global gridded weather forecast products provided by a ground meteorological data service through a satellite-ground data link; the meteorological data receiving module also obtains real-time cloud images and water vapor data of other meteorological satellites through an inter-satellite link or a direct receiving mode; the meteorological data receiving module also accesses satellite self-remote sensing historical data to construct a historical weather pattern library; the meteorological data receiving module fuses data in the global gridded weather forecast products, the real-time cloud images and water vapor data, and the historical weather pattern library by using a data assimilation technology, takes a ground forecast as a background field, corrects by using observation data received in real time, and generates a satellite analysis field product; the meteorological data receiving module provides the satellite analysis field product to the satellite weather prediction processing module.

3. The weather forecast based satellite autonomous mission planning system of claim 2, wherein, Further comprising: the satellite weather prediction processing module adopts a lightweight convolutional neural network or a graph neural network as an observation condition evaluation model; input of the observation condition evaluation model includes a multi-band, multi-element data cube in the satellite analysis field product from the meteorological data receiving module, and observation geometry information fused from satellite orbit prediction data and payload parameters; The output of the observation condition evaluation model includes an imaging probability atlas, an uncertainty atlas, and an optimal observation band recommendation; The on-board weather prediction processing module integrates the imaging probability atlas, the uncertainty atlas, and the optimal observation band recommendation to generate the executable metric quantification atlas.

4. The weather forecast based satellite autonomous mission planning system of claim 3, wherein, Further comprising: The instruction execution verification module records the actual meteorological condition at the imaging time as actual weather data; The instruction execution verification module records the actual imaging quality evaluation result as actual imaging quality data; The instruction execution verification module calculates the deviation between the actual weather data and the predicted weather data; The instruction execution verification module provides the deviation data and the actual imaging quality data to the on-board weather prediction processing module; The on-board weather prediction processing module performs online incremental learning or model fine-tuning under federated learning on the neural network model weight according to the deviation data and the actual imaging quality data.

5. The weather forecast based satellite autonomous mission planning system of claim 4, wherein, Further comprising: The autonomous task decision module constructs a lightweight digital twin of the satellite platform and the task on board; The digital twin synchronizes the fuel state, the power state, the storage state, and the load health state of the satellite in real time; The autonomous task decision module receives the executable metric quantification atlas from the on-board weather prediction processing module; The autonomous task decision module inputs the executable metric quantification atlas into the digital twin; The autonomous task decision module performs Monte Carlo tree search or evolutionary algorithm simulation and deduction in the digital twin; The autonomous task decision module outputs a satellite attitude control instruction sequence and a load working parameter set based on the simulation and deduction result.

6. The weather forecast based satellite autonomous mission planning system of claim 5, wherein, Further comprising: The multi-objective optimization decision considers multiple optimization objectives, including maximizing long-term scientific return, minimizing emergency task response time, optimizing energy and fuel consumption, and avoiding high-uncertainty regions; The autonomous task decision module cooperatively optimizes at least two optimization objectives in the digital twin simulation and deduction.

7. The weather forecast based satellite autonomous mission planning system of claim 6, wherein, Further comprising: The autonomous task decision module exchanges observation intentions with other satellites through inter-satellite links and performs distributed cooperative task planning in the digital twin; And for all instruction sequences generated by autonomous decision, a hash digest is generated using lightweight blockchain technology and stored in an append-only log.

8. The weather forecast based satellite autonomous mission planning system of claim 7, wherein, Further comprising: The instruction execution verification module performs on-orbit real-time image quality evaluation using the fast preview mode of the load during imaging to generate an actual imaging utility value; The actual imaging utility value is compared with the expected value of the prediction module to generate deviation data; The deviation data and the actual meteorological condition are used as feedback data flow and are fed back to the on-board weather prediction processing module in real time.

9. The weather forecast based satellite autonomous mission planning system of claim 8, wherein, Further comprising: When the imaging quality is lower than the preset threshold, the instruction execution verification module immediately interrupts the current task and sends an interruption signal and a resource release signal to the autonomous task decision module; The autonomous task decision module calls the digital twin to re-plan a backup task using the saved time and resources.

10. The weather forecast based satellite autonomous mission planning system of claim 9, wherein, Further comprising: The task execution feedback log is transmitted back to the ground system through the satellite-ground link for post-analysis of the weather forecast accuracy and the effectiveness of autonomous decision, which supports the continuous optimization of the model.

Images