Satellite orbiting attitude and on-orbit computing power task scheduling joint optimization method and device based on thermoelectric combined regulation, computer equipment and storage medium

Abstract

The embodiment of the application relates to the field of satellite control, and provides a satellite orbit-entry attitude and on-orbit computing power task scheduling joint optimization method and device based on thermoelectric joint regulation, computer equipment and a storage medium, the method comprising: for any satellite attitude control period, collecting multiple-source state data of a satellite to be controlled in real time; performing solar vector calculation to obtain a solar vector prediction sequence; performing thermal state evaluation processing according to the multiple-source state data and the solar vector prediction sequence to obtain a temperature prediction curve; performing energy state evaluation processing to obtain a battery state of charge prediction curve; performing multi-target joint optimization decision calculation according to main task pointing constraint parameters, the temperature prediction curve and the battery state of charge prediction curve of each candidate attitude fine adjustment scheme to determine a target scheme; and the satellite to be controlled performing attitude adjustment operation and task rearrangement operation. The implementation of the method improves the stability of the satellite system in executing tasks.

Description

Technical Field

[0001] This application relates to the field of satellite control technology, and in particular to a method, apparatus, computer equipment, and storage medium for joint optimization of satellite orbital attitude and on-orbit computing power task scheduling based on thermoelectric joint regulation. Background Technology

[0002] With the rapid deployment of low-Earth orbit satellite constellations and the continuous growth in demand for onboard intelligent processing, on-orbit computing power has become a core indicator for measuring satellite information processing capabilities. An increasing number of satellites are equipped with high-performance computing chips to perform tasks such as on-orbit preprocessing of remote sensing data, target identification, and inter-satellite collaborative computing, reducing reliance on downlink bandwidth from ground stations.

[0003] However, high-performance chips face severe challenges in both thermal management and energy supply during on-orbit operation. The space environment in which satellites are located alternates between high-temperature radiation during the sunshine period and extremely low temperatures during the shadow period, and the chip's working status is severely affected by thermal cycling.

[0004] Current satellite system designs generally employ an architecture where attitude control and computing power management are independent. The attitude control system focuses on meeting payload pointing requirements, such as the alignment accuracy of Earth observation cameras or the beam direction of communication antennas; the computing power scheduling system independently orchestrates task execution plans based on task queues and computing resource status. The lack of information exchange and collaborative optimization mechanisms between these two subsystems completely ignores the impact of attitude changes on chip thermal environment and solar panel power generation efficiency, leading to frequent overheating, frequency throttling, or power-down shutdowns of computing chips, resulting in poor stability of the satellite system's mission execution. Summary of the Invention

[0005] This application provides a method, apparatus, computer equipment, and storage medium for joint optimization of satellite orbital attitude and on-orbit computing task scheduling based on thermoelectric joint control. More specifically, this application provides a method, apparatus, computer equipment, computer storage medium, and computer program product for joint optimization of satellite orbital attitude and on-orbit computing task scheduling based on thermoelectric joint control, thereby improving the stability of satellite system mission execution.

[0006] In a first aspect, embodiments of this application provide a method for joint optimization of satellite injection attitude and on-orbit computing power task scheduling based on thermoelectric joint control, including:

[0007] For any given satellite attitude control cycle, multi-source state data of the satellite to be controlled is acquired in real time.

[0008] The orbital data of the satellite to be controlled is obtained, and the solar vector is calculated based on the orbital data to obtain the solar vector prediction sequence;

[0009] Based on the multi-source state data and the solar vector prediction sequence, thermal state assessment is performed to determine the current temperature change rate of the satellite to be controlled; based on the current temperature change rate, temperature prediction is performed on each candidate attitude fine-tuning scheme to obtain a temperature prediction curve;

[0010] Based on the multi-source state data and solar vector forecast sequence, energy state assessment is performed to determine the net power data of the satellite to be controlled; based on the net power data, battery state of charge prediction is performed on each of the candidate attitude fine-tuning schemes to obtain battery state of charge prediction curves;

[0011] Based on the main mission direction constraint parameters, the temperature prediction curves and battery state of charge prediction curves of each candidate attitude fine-tuning scheme, a multi-objective joint optimization decision calculation is performed to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes.

[0012] The satellite to be controlled is controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme; the satellite to be controlled is controlled to perform task rescheduling operations according to the task scheduling command corresponding to the target scheme.

[0013] Optionally, in some embodiments of this application, before the step of acquiring multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle, the method further includes: performing initialization configuration after the satellite to be controlled completes initial attitude acquisition and orbit determination after entering orbit; specifically including:

[0014] Input the satellite body geometric model parameters of the satellite to be controlled, determine the normal direction of the computing chip mounting surface in the satellite body coordinate system, and determine the normal direction of the solar panel in the satellite body coordinate system;

[0015] Set the constraint threshold for the main task pointing constraint parameter;

[0016] Set the upper and lower limits of the suitable operating temperature range and the temperature change rate limit for the computing chip;

[0017] Set the upper and lower limits of the safe operating range for the battery's state of charge and the minimum reserve threshold;

[0018] Set the control cycle length and prediction time window length for any of the satellite attitude control cycles;

[0019] Initialize the model parameters in the thermal state assessment process and the energy state assessment process.

[0020] Optionally, in some embodiments of this application, the step of acquiring multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle includes:

[0021] The attitude sensor component outputs attitude data; the temperature sensor component outputs temperature status data; the power monitoring component outputs energy status data; and the light sensor component outputs solar irradiance measurement values ​​for each panel.

[0022] The data to be labeled is obtained by summarizing the attitude data, temperature status data, energy status data, and solar irradiance measurements of each panel;

[0023] The data to be labeled is processed according to a unified timestamp and then aggregated to the onboard data bus to obtain the multi-source status data.

[0024] Optionally, in some embodiments of this application, the step of acquiring the orbital data of the satellite to be controlled and calculating the solar vector based on the orbital data to obtain a solar vector prediction sequence includes:

[0025] The latest orbital position and velocity data are read as the orbital data, and the orbital propagation model is run based on the orbital data to generate an orbital position time series covering the prediction time window from the current moment.

[0026] Based on the generated orbital position time series, according to Earth epoch data and solar position model, the direction of the solar vector in the orbital coordinate system corresponding to each time step within the prediction time window is calculated to determine the direction sequence of the solar vector;

[0027] The parameters of the Earth's shadow cone are determined based on the direction sequence of the solar vector. The start and end times of the sunshine and shadow segments within the prediction time window are calculated based on the geometric relationship between the satellite orbit represented by the orbit data and the Earth's shadow cone, thus obtaining the time series of the sunshine and shadow segments.

[0028] The time series of the orbital position, the direction sequence of the solar vector, and the time series of the sunshine and shadow segments are used as forecast results and stored in time series data format to obtain the solar vector forecast sequence.

[0029] Optionally, in some embodiments of this application, the step of performing thermal state assessment processing based on the multi-source state data and the solar vector forecast sequence to determine the current temperature change rate of the satellite to be controlled includes:

[0030] The solar irradiance heat flux density received by the chip surface under the current orientation is obtained by multiplying the solar constant, the solar absorptivity of the panel, and the cosine of the solar incident angle, so as to determine the external irradiance heat input.

[0031] Calculate the chip's own heat generation power based on the chip's current computing load to determine the chip's own heat generation.

[0032] The net heat input is calculated based on external irradiation heat input, chip self-generated heat and heat conduction and radiation heat dissipation.

[0033] The current temperature change rate is obtained by dividing the net heat input by the heat capacity of the chip and its substrate.

[0034] Optionally, in some embodiments of this application, the step of performing energy state assessment processing based on the multi-source state data and solar vector forecast sequence to determine the net power data of the satellite to be controlled includes:

[0035] The effective irradiated area of ​​the solar panel under the current orientation is obtained by multiplying the total area of ​​the solar panel by the cosine of the solar incidence angle.

[0036] Estimate the current power generation based on the effective irradiated area and the conversion efficiency of the solar cell;

[0037] Based on the power consumption of the computing chip, the power consumption of the attitude control system, the power consumption of the communication system, and the power consumption of other payloads, the total power consumption of all satellite payloads at the current moment is calculated.

[0038] Based on the current power generation and the total power consumption of all satellite loads at the current moment, net power data is determined to characterize whether the power is in surplus or deficit.

[0039] Optionally, in some embodiments of this application, after the steps of controlling the satellite to be controlled to perform attitude adjustment operation according to the attitude target instruction corresponding to the target scheme and controlling the satellite to be controlled to perform task rescheduling operation according to the task scheduling instruction corresponding to the target scheme, the method further includes:

[0040] When any of the satellite attitude control cycles ends, the actual acquired values ​​of the satellite to be controlled are obtained;

[0041] Determine the predicted values ​​in the temperature prediction curve and the battery state of charge prediction curve;

[0042] The numerical comparison results are obtained by comparing the actual collected values ​​with the predicted values.

[0043] The model parameters in the thermal state assessment and energy state assessment processes are adjusted based on the numerical comparison results.

[0044] Secondly, embodiments of this application provide a joint optimization device for satellite orbital attitude and on-orbit computing task scheduling based on thermoelectric joint control, which has the function of implementing the joint optimization method for satellite orbital attitude and on-orbit computing task scheduling based on thermoelectric joint control provided in the first aspect above. The function can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above function, and the modules can be software and / or hardware.

[0045] In one possible design, the device includes:

[0046] The sensor acquisition module is used to acquire multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle.

[0047] The orbit prediction module is used to acquire the orbit data of the satellite to be controlled, and to calculate the solar vector based on the orbit data to obtain the solar vector prediction sequence;

[0048] The thermal state assessment module is used to perform thermal state assessment processing based on the multi-source state data and the solar vector prediction sequence to determine the current temperature change rate of the satellite to be controlled; and to perform temperature prediction on each candidate attitude fine-tuning scheme based on the current temperature change rate to obtain a temperature prediction curve.

[0049] The energy state assessment module is used to perform energy state assessment processing based on the multi-source state data and the solar vector forecast sequence to determine the net power data of the satellite to be controlled; and to predict the battery state of charge for each of the candidate attitude fine-tuning schemes based on the net power data to obtain the battery state of charge prediction curve.

[0050] The joint optimization decision module is used to perform multi-objective joint optimization decision calculations based on the main mission direction constraint parameters, the temperature prediction curves of each candidate attitude fine-tuning scheme, and the battery state of charge prediction curve, and to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes.

[0051] The execution module is used to control the satellite to be controlled to perform attitude adjustment operations according to the attitude target instructions corresponding to the target scheme; and to control the satellite to be controlled to perform task rescheduling operations according to the task scheduling instructions corresponding to the target scheme.

[0052] In another aspect, this application provides a computer device including at least one connected processor and a memory, wherein the memory is used to store program code, and the processor is used to call the program code in the memory to execute the methods described in the above aspects.

[0053] In another aspect, embodiments of this application provide a computer storage medium including instructions that, when executed on a computer, cause the computer to perform the methods described in the above aspects.

[0054] In another aspect, this application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the methods described in the above aspects.

[0055] Compared to traditional technologies, the technical solution of this application introduces chip thermal state and battery SOC as secondary optimization targets in attitude control decision-making. Within the range allowed by the primary mission pointing constraint parameters, the attitude is finely adjusted, actively improving the thermal environment of the chip and the power generation conditions of the solar panel. This eliminates the generation of adverse thermoelectric conditions from the physical source, enabling the chip to operate for a longer period of time under suitable temperature and sufficient power supply conditions, thereby improving the stability of the satellite system in performing its mission. Attached Figure Description

[0056] Figure 1 This is a flowchart of one embodiment.

[0057] Figure 2 This is a flowchart illustrating the overall technical process in one embodiment.

[0058] Figure 3 This is a diagram showing the relationship between key technical features in one embodiment.

[0059] Figure 4 This is a flowchart illustrating the specific implementation steps in one embodiment.

[0060] Figure 5 This is a structural block diagram of the device in one embodiment.

[0061] Figure 6 This is a system overall structure diagram of one embodiment.

[0062] Figure 7 This is a schematic diagram illustrating the working principle of sensor data acquisition and orbit prediction in one embodiment.

[0063] Figure 8 This is a schematic diagram illustrating the working principle of thermal state assessment and energy state assessment in one embodiment.

[0064] Figure 9 This is a schematic diagram illustrating the working principle of joint optimization decision-making and execution feedback in one embodiment.

[0065] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0066] The terms "first," "second," etc., used in the embodiments of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. The division of modules appearing in the embodiments of this application is only a logical division. In actual applications, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not performed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interface, and the indirect coupling or communication connection between modules may be electrical or other similar forms. None of these are limited in the embodiments of this application. Furthermore, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed among multiple circuit modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of this application.

[0067] Figure 1 This is a flowchart illustrating one embodiment, such as... Figure 1 As shown in the embodiments of this application, the satellite orbital attitude and on-orbit computing task scheduling joint optimization method based on thermoelectric joint control refers to the simultaneous management of the satellite's heating environment and on-orbit power supply energy conditions; satellite orbital attitude refers to the satellite's orientation in space after entering its operational orbit; and on-orbit computing task scheduling refers to the allocation and adjustment of the runtime and load of on-orbit computing tasks. This method specifically includes:

[0068] The S1100 acquires multi-source status data of the satellite to be controlled in real time for any satellite attitude control cycle.

[0069] Among them, the satellite attitude control cycle, or simply the control cycle, refers to the complete time interval of one round of attitude acquisition, calculation, and adjustment of the satellite.

[0070] Among them, multi-source state data refers to the aggregated measurement information from various sensors, including satellite attitude, temperature, illumination, and power supply.

[0071] S1200 acquires the orbital data of the satellite to be controlled, calculates the solar vector based on the orbital data, and obtains the solar vector prediction sequence.

[0072] Among them, orbital data refers to the measurement parameters related to the satellite's real-time position and operating speed in orbit.

[0073] Among them, solar vector calculation refers to solving the spatial direction vector of the satellite pointing to the sun in the orbital coordinate system, and solar vector forecast sequence refers to the collection of various time series data at each moment within the forecast window.

[0074] S1300 performs thermal state assessment processing based on multi-source state data and solar vector forecast sequence to determine the current temperature change rate of the satellite to be controlled; and performs temperature prediction on each candidate attitude fine-tuning scheme based on the current temperature change rate to obtain temperature prediction curves.

[0075] Among them, thermal state assessment processing refers to calculating the temperature evolution process of the satellite chip as it generates and dissipates heat; the temperature change rate is used to characterize the magnitude of the chip temperature increase or decrease per unit time.

[0076] Among them, the candidate attitude fine-tuning scheme, also known as the candidate attitude scheme, refers to the multiple satellite orientation angles that can be selected within the allowable deviation of the main mission.

[0077] Among them, the temperature prediction curve is used to reflect the trend of chip temperature change over time during the prediction period.

[0078] S1400 performs energy state assessment processing based on multi-source state data and solar vector forecast sequence to determine the net power data of the satellite to be controlled; based on the net power data, it performs battery state of charge prediction for each candidate attitude fine-tuning scheme to obtain battery state of charge prediction curves.

[0079] Among them, energy status assessment processing refers to calculating the power balance relationship between the power consumption of the onboard power generation load.

[0080] Among them, net power data is used to characterize the difference between the on-board power generation and the overall load power consumption, which can be characterized as net power surplus or net power deficit.

[0081] Among them, the battery state of charge (SOC) prediction refers to the process of estimating the change of the remaining battery power during the prediction period; the battery state of charge prediction curve, i.e. the SOC prediction curve, is used to reflect the time sequence change trend of the battery power during the prediction period.

[0082] S1500 performs multi-objective joint optimization decision calculations based on the main mission pointing constraint parameters, the temperature prediction curves of each candidate attitude fine-tuning scheme, and the battery state of charge prediction curves, and determines the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes.

[0083] Among them, the main mission pointing constraint parameter refers to the limited value that ensures the alignment accuracy of the satellite's core main mission.

[0084] Among them, the multi-objective joint optimization decision calculation refers to the optimal solution of synchronous balance temperature control power supply attitude constraints.

[0085] The target scheme refers to the optimal attitude scheduling combination for comprehensive operating conditions selected within a single control cycle.

[0086] S1600 controls the satellite to be controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme; and controls the satellite to be controlled to perform task rescheduling operations according to the task scheduling command corresponding to the target scheme.

[0087] Among them, attitude target command refers to the control execution signal that drives the satellite to adjust to a specified orientation angle; attitude adjustment operation refers to the action of changing the satellite's spatial orientation by relying on the flywheel magnetic torque device.

[0088] Among them, the task scheduling instruction refers to the signal that changes the running window and computing load of the onboard computing task; the task reordering operation refers to the reassignment of the start time and running frequency of the queued computing tasks.

[0089] Compared to traditional technologies, in this embodiment, multi-source state data of the satellite to be controlled is first collected in real time for any satellite attitude control cycle. Then, the orbital data of the satellite to be controlled is acquired, and solar vector prediction sequence is obtained by calculating the solar vector based on the orbital data. Then, thermal state assessment processing is carried out to determine the current temperature change rate, and temperature prediction curves are generated for each candidate attitude fine-tuning scheme. Simultaneously, energy state assessment processing is carried out to determine net power data and obtain battery state of charge prediction curves. Then, multi-objective joint optimization decision calculation is performed by combining the main mission pointing constraint parameters, temperature prediction curves, and battery state of charge prediction curves to select the target scheme. Finally, attitude adjustment operation is performed according to the attitude target command of the target scheme, and task reordering operation is performed according to the task scheduling command. The technical solution of this embodiment introduces chip thermal state and battery SOC as secondary optimization objectives in attitude control decision-making. Within the range allowed by the main mission pointing constraint parameters, the attitude is fine-tuned, actively improving the thermal environment of the chip and the power generation conditions of the solar panel. The adverse thermoelectric conditions are eliminated from the physical source, enabling the chip to operate for a longer time under suitable temperature and sufficient power supply conditions, thereby improving the stability of the satellite system in performing its mission.

[0090] Optionally, in some embodiments of this application, before the step of acquiring multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle, the method further includes: performing initialization configuration after the satellite to be controlled has completed initial attitude acquisition and orbit determination after entering orbit; specifically including: inputting the satellite body geometric model parameters of the satellite to be controlled, determining the normal direction of the computing chip mounting surface in the satellite body coordinate system, determining the normal direction of the solar panel in the satellite body coordinate system; setting the constraint threshold of the main task pointing constraint parameters; setting the upper and lower limits of the suitable operating temperature range and the temperature change rate limit of the computing chip; setting the upper and lower limits of the safe operating range and the minimum reserve threshold of the battery state of charge; setting the control cycle length and prediction time window length of any satellite attitude control cycle; initializing the model parameters in the thermal state assessment processing and energy state assessment processing.

[0091] Initial attitude acquisition refers to the satellite autonomously locking its initial spatial orientation angle after entering orbit; orbit determination refers to calculating and locking the satellite's precise on-orbit position and operating speed parameters.

[0092] Initialization configuration, also known as system initialization and parameter configuration, refers to the loading of basic parameters at system startup to define the safety boundaries of various operating conditions.

[0093] Among them, the satellite body geometric model parameters refer to the position, size, and structural values ​​of the components on the satellite body panel; the satellite body coordinate system refers to the spatial reference coordinate system established with the center of the satellite body as the origin; and the normal direction refers to the spatial direction perpendicular to the chip panel or the solar panel facing outward.

[0094] Among them, the constraint threshold of the main task-oriented constraint parameter refers to the maximum limit value that each working condition index is allowed to fluctuate, including the maximum allowable value of the three-axis attitude angle deviation.

[0095] The prediction time window length, which is the length of any satellite attitude control cycle, refers to the time span of a single forward projection of thermoelectric condition changes.

[0096] The model parameters include fixed coefficient constant values ​​involved in the calculation within the thermoelectric assessment model, model parameters of the thermal state assessment model in the thermal state assessment process, and model parameters of the energy state assessment model in the energy state assessment process.

[0097] In this embodiment, the safety boundary is defined by initializing all-dimensional parameters before orbit insertion, which improves the accuracy of subsequent thermoelectric attitude prediction calculation and enhances the adaptability of the system startup and operation.

[0098] Optionally, in some embodiments of this application, for any satellite attitude control cycle, multi-source state data of the satellite to be controlled is collected in real time, including: attitude data output from the attitude sensor component; temperature status data output from the temperature sensor component; energy status data output from the power monitoring component; solar irradiance measurement values ​​of each panel output from the illumination sensor component; data to be labeled is obtained by summarizing the attitude data, temperature status data, energy status data and solar irradiance measurement values ​​of each panel; and the data to be labeled is annotated according to a unified timestamp and then aggregated to the onboard data bus to obtain multi-source state data.

[0099] The attitude sensor assembly refers to a combination of sensor components that measure the satellite's spatial orientation and rotational speed; the attitude data includes the current three-axis attitude quaternion and angular rate vector.

[0100] Among them, the temperature sensor assembly refers to the combination of sensor components that collect the real-time temperature of the chip's surface; the temperature status data includes the core temperature of the computing chip, the temperature of the chip mounting panel, the temperature of the heat dissipation panel, and the temperatures of other panels.

[0101] Among them, the power monitoring component refers to the combination of monitoring devices that collect the voltage and current of photovoltaic cells and batteries; the energy status data includes the voltage and current values ​​of each branch of the solar panel and the terminal voltage and charging and discharging current values ​​of the battery.

[0102] Among them, the light sensor component refers to the sensor device that measures the intensity of sunlight irradiance received by the surface of a star; the solar irradiance measurement value refers to the measured value of the solar energy received per unit area of ​​the surface.

[0103] Among them, data to be labeled refers to the raw data set that has not yet been bound with time stamps after being collected by various sensors; timestamp refers to the time sequence mark information that marks the precise collection time corresponding to each piece of collected data; labeling processing refers to the operation of binding the corresponding collection time time time sequence mark to the raw measurement data; and onboard data bus refers to the unified transmission channel for transmitting and exchanging data between various modules on the satellite.

[0104] In this embodiment, by using multi-sensor synchronous acquisition and time-series unified labeling and transmission, the synchronization of multi-source operating condition data acquisition and transmission efficiency are improved, thereby enhancing the data foundation quality for subsequent evaluation and calculation.

[0105] Optionally, in some embodiments of this application, acquiring the orbital data of the satellite to be controlled and calculating the solar vector based on the orbital data to obtain a solar vector forecast sequence includes: reading the latest orbital position and velocity data as orbital data; running an orbital propagation model based on the orbital data to generate an orbital position time series covering the prediction time window from the current moment; based on the generated orbital position time series, calculating the direction of the solar vector in the orbital coordinate system corresponding to each time step within the prediction time window according to Earth epoch data and the solar position model, and determining the direction sequence of the solar vector; determining the parameters of the Earth's shadow cone based on the direction sequence of the solar vector; calculating the start and end times of the sunshine segment and the shadow segment within the prediction time window based on the geometric relationship between the satellite orbit and the Earth's shadow cone represented by the orbital data, and obtaining the time series of the sunshine segment and the shadow segment; and storing the orbital position time series, the direction sequence of the solar vector, and the time series of the sunshine segment and the shadow segment as forecast results in a time series data format to obtain a solar vector forecast sequence.

[0106] The orbital data includes orbital position and velocity data, which refers to the satellite's current spatial coordinates and its speed of movement in orbit; the orbital propagation model refers to the dynamic calculation model that predicts the changes in the satellite's position and velocity over future periods; and the orbital position time series refers to the time-series array composed of the satellite's spatial coordinates at each moment of the prediction window.

[0107] Among them, Earth epoch data refers to astronomical reference parameters that record the changes in Earth's spatial position over time; solar position model refers to a calculation model that extrapolates the spatial orientation of the sun relative to Earth at any given moment; time step refers to the smallest equal time unit used to divide the prediction window for point-by-point calculation of the working conditions; orbital coordinate system refers to a spatial orientation reference system established with the satellite orbital plane as the reference; and solar vector direction sequence refers to the time sequence set of the satellite's orientation toward the sun at each moment in the prediction window.

[0108] Among them, the Earth's shadow cone refers to the cone-shaped dark space region formed on the opposite side of the sun when the Earth blocks sunlight; the geometric relationship refers to the positional correspondence between the satellite's orbit and the spatial range of the shadow cone.

[0109] Among them, the start and end times of the sunshine segment refer to the time points when the satellite enters and leaves the sunlit area; the start and end times of the shadow segment refer to the time points when the satellite enters and leaves the shadow area; the time series of the sunshine and shadow segments refer to the time series of the marked sunshine and shadow periods within the prediction window.

[0110] Among them, time series data format refers to a standardized format in which data is stored in chronological order.

[0111] In this embodiment, the accuracy of predicting solar shading conditions is improved by synchronously calculating the sun's azimuth and the time of shadow through orbital extrapolation, thereby enhancing the precision of the environmental input before thermoelectric assessment.

[0112] Optionally, in some embodiments of this application, thermal state assessment processing is performed based on multi-source state data and solar vector prediction sequences to determine the current temperature change rate of the satellite to be controlled. This includes: obtaining the solar irradiance heat flux density received by the chip surface under the current attitude by multiplying the solar constant, the solar absorptivity of the panel, and the cosine of the solar incidence angle, to determine the external irradiance heat input; calculating the chip's own heat generation power based on the chip's current computing power load to determine the chip's own heat generation; calculating the net heat input based on the external irradiance heat input, the chip's own heat generation, and heat conduction and radiation; and obtaining the current temperature change rate by dividing the net heat input by the heat capacity of the chip and its substrate.

[0113] Among them, the solar constant refers to the basic energy of sunlight received per unit area at Earth's orbit; the panel solar absorptivity refers to the fixed proportional coefficient of solar radiation energy absorbed by the panel surface; the solar incident angle refers to the size of the spatial angle formed between the direction of solar irradiation and the normal to the panel surface; the solar irradiance heat flux density refers to the amount of solar thermal energy received per unit area per unit time; and the external irradiance heat input refers to the total amount of external heat transferred to the chip panel surface by sunlight.

[0114] Among them, computing load refers to the magnitude of the computational load undertaken by the chip in the current computing task; self-generated heat power refers to the amount of heat released by the chip per unit time during the operation of the computing process.

[0115] Among them, thermal conduction heat dissipation refers to the portion of the chip's heat that is conducted outward through the contact substrate; radiative heat dissipation refers to the portion of the chip's heat that is dissipated into space in the form of thermal radiation; and net heat input refers to the difference between the total heat input and the total heat dissipation remaining on the chip.

[0116] Among them, the chip substrate refers to the satellite structure supporting plate components that carry the fixed computing power chips; heat capacity is used to characterize the amount of heat absorbed by the chip substrate to raise the overall temperature by a unit.

[0117] In this embodiment, the net heat input is accurately calculated by using multiple heat sources and multiple heat dissipation channels, which improves the accuracy of the calculation of the chip's real-time temperature change rate and enhances the reliability of subsequent temperature trend prediction.

[0118] In another embodiment, prior to the thermal state assessment step, the method further includes: receiving the current chip temperature and the temperatures of each panel from the temperature state data; wherein, the current chip temperature, also known as the core temperature of the computing chip, is used as the starting point for temperature time-series prediction, and the initial temperature superposition rate is used to accumulate the change over time to generate the entire temperature prediction curve; the temperatures of each panel include the chip mounting panel temperature, the heat dissipation panel temperature, and the temperatures of other panels, which are used to accurately calculate the amount of heat dissipated by the chip through conduction and radiation, so as to determine the heat dissipation through conduction and radiation.

[0119] Specifically, temperature prediction is performed on each candidate attitude fine-tuning scheme based on the current temperature change rate to obtain a temperature prediction curve, including: temperature prediction is performed on each candidate attitude fine-tuning scheme based on the current temperature change rate and the current chip temperature to obtain a temperature prediction curve.

[0120] In another embodiment, temperature prediction is performed on each candidate attitude fine-tuning scheme based on the current rate of temperature change to obtain a temperature prediction curve. This includes: calculating the change in the solar incident angle on the surface where the chip is located for each candidate attitude fine-tuning scheme, and then estimating the chip temperature change trajectory within the future prediction time window under that scheme to obtain a temperature prediction curve.

[0121] Optionally, in some embodiments of this application, an energy state assessment is performed based on multi-source state data and solar vector forecast sequences to determine the net power data of the satellite to be controlled. This includes: obtaining the effective irradiated area of ​​the solar panels under the current attitude by multiplying the total area of ​​the solar panels by the cosine of the solar incidence angle; estimating the current power generation based on the effective irradiated area and the conversion efficiency of the solar cells; calculating the total power consumption of all satellite loads at the current moment based on the power consumption of the computing chip, the power consumption of the attitude control system, the power consumption of the communication system, and the power consumption of other payloads; and determining the net power data used to characterize whether the power is in surplus or deficit based on the current power generation and the total power consumption of all satellite loads at the current moment.

[0122] The total area of ​​the solar panel refers to the physical area of ​​the photovoltaic panel without any obstructions; the effective irradiated area refers to the actual area of ​​the photovoltaic panel that can be converted into electrical energy when sunlight shines vertically on it.

[0123] Among them, solar cell conversion efficiency refers to the fixed energy conversion ratio coefficient of light energy to electrical energy; power generation refers to the amount of electrical energy that a photovoltaic panel can convert and output per unit time.

[0124] Among them, the power consumption of the computing chip is the power consumption of the chip when performing tasks; the power consumption of the attitude control system refers to the electrical power consumed by the flywheel torque motor to regulate the satellite's attitude; the power consumption of the communication system refers to the electrical power consumed by the satellite's signal transmission and reception equipment; the power consumption of other payloads refers to the power consumption of the other on-board equipment except for calculating the communication attitude; and the total power consumption refers to the sum of the power consumed by all the satellite's electrical equipment at the same time.

[0125] Among them, power surplus refers to the state where the power generation is greater than the total power consumption of the whole machine, resulting in excess rechargeable electrical energy; power deficit refers to the state where the total power consumption of the whole machine is greater than the power generation, requiring the battery to discharge to replenish energy.

[0126] In this embodiment, the power difference is comprehensively calculated by comparing the photovoltaic irradiated area with the total multi-load power consumption of the entire system, which improves the accuracy of on-board net power measurement and enhances the reliability of predicting changes in battery power.

[0127] In another embodiment, before performing energy state assessment processing, the method further includes: receiving current solar panel voltage and current data (voltage and current values ​​of each branch of the solar panel) and estimated battery SOC values.

[0128] Among them, the current voltage and current data of the solar panels are used to directly calculate the real-time photovoltaic power generation. The current power generation estimated based on the effective irradiated area and the conversion efficiency of the solar cells is verified to ensure the accuracy of the data.

[0129] The battery SOC estimate is calculated based on the battery terminal voltage and charging / discharging current values ​​collected by the power monitoring component, combined with battery capacity and charging / discharging loss models. The battery SOC estimate is used to determine the current remaining battery capacity, identify the battery's safe charging / discharging range, predict the trend of capacity increase / decrease in conjunction with power balance results, and assist in generating a complete battery state of charge prediction curve.

[0130] In another embodiment, the battery state of charge (SOC) prediction is performed on each candidate attitude fine-tuning scheme based on net power data to obtain a battery SOC prediction curve. This includes: executing charging and discharging strategies under each candidate attitude fine-tuning scheme and candidate computing power load scheme based on net power data and time period characteristics, calculating the power generation sequence and SOC change trajectory within the future prediction time window to obtain the battery SOC prediction curve; wherein, executing the charging and discharging strategy specifically includes: during the sunshine period, the net power surplus is used to charge the battery, controlling the SOC to rise; during the shade period or when the load power consumption exceeds the power generation power, the battery discharges to supply energy, controlling the SOC to fall.

[0131] In another embodiment, based on the main mission pointing constraint parameters, the temperature prediction curves of each candidate attitude fine-tuning scheme, and the battery state of charge prediction curve, a multi-objective joint optimization decision calculation is performed to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes. This includes: determining the upper and lower limits of the three-axis attitude angle offset based on the current main mission state to determine the main mission pointing constraint parameters; discretizing and generating a set of candidate attitude schemes within the range corresponding to the main mission pointing constraint parameters; the set of candidate attitude schemes includes each candidate attitude fine-tuning scheme; for each candidate attitude fine-tuning scheme, reading the corresponding temperature prediction curve and battery state of charge prediction curve, and calculating the comprehensive evaluation index of each candidate attitude fine-tuning scheme; after traversing all candidate attitude fine-tuning schemes, selecting the scheme with the best comprehensive evaluation index as the target scheme for the optimal decision in the current cycle.

[0132] In another embodiment, for each candidate attitude fine-tuning scheme, the corresponding temperature prediction curve and battery state of charge prediction curve are read respectively, and a comprehensive evaluation index for each candidate attitude fine-tuning scheme is calculated, including: based on the index evaluation rules, calculating the degree of temperature deviation from the suitable range of each candidate attitude fine-tuning scheme according to the temperature prediction curve; calculating the degree of SOC deviation from the safe range of each candidate attitude fine-tuning scheme according to the battery state of charge prediction curve; and obtaining a comprehensive evaluation index for each candidate attitude fine-tuning scheme by weighted combination of the degree of temperature deviation from the suitable range and the degree of SOC deviation from the safe range.

[0133] The evaluation rules include: the closer the temperature in the temperature prediction curve is to the center value of the suitable range, the higher the comprehensive evaluation index; the closer the SOC in the battery state of charge prediction curve is to the center value of the safe range, the higher the comprehensive evaluation index; and a penalty is given for schemes whose temperature or SOC prediction trajectories touch the upper or lower limits of the constraints within the prediction window.

[0134] In another embodiment, the method further includes: determining the attitude target instruction corresponding to the target scheme; and adjusting the computing power task scheduling parameters according to the task scheduling rules based on the temperature prediction curve and battery state of charge prediction curve under the target scheme to generate the task scheduling instruction corresponding to the target scheme.

[0135] The task scheduling rules include: scheduling high-power tasks during the most favorable periods of temperature and SOC conditions, and scheduling low-power tasks during the most unfavorable periods of conditions.

[0136] In this embodiment, by using attitude prediction-driven task scheduling, high-computing tasks are pre-arranged during the most favorable thermoelectric conditions, which significantly improves the continuous computing power output of a single satellite under the same orbital environment and hardware configuration, effectively improving the utilization efficiency of onboard computing resources and the reliability of task completion.

[0137] In another embodiment, controlling the satellite to be controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme includes: calculating the angular offset from the current attitude to the target attitude according to the attitude target command corresponding to the target scheme, and generating a flywheel speed change command or a magnetic torquer current command through the attitude tracking controller; driving the satellite to smoothly adjust to the target attitude according to the flywheel speed change command or the magnetic torquer current command; continuously monitoring the main mission pointing deviation during the attitude adjustment operation, and stopping the attitude adjustment operation if the main mission pointing deviation is outside the allowable range.

[0138] In another embodiment, controlling the satellite to be controlled to perform task rescheduling operations according to the task scheduling instructions corresponding to the target scheme includes: adjusting the execution time period and operation frequency of each task in the current task queue according to the task scheduling instructions corresponding to the target scheme; specifically, it includes: reallocating execution windows for tasks that have been queued but have not yet started, and adjusting the operation frequency of running tasks as needed.

[0139] Optionally, in some embodiments of this application, after controlling the satellite to be controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme and controlling the satellite to be controlled to perform task rescheduling operations according to the task scheduling command corresponding to the target scheme, the method further includes: when any satellite attitude control cycle ends, acquiring the actual collected values ​​of the satellite to be controlled; determining the predicted values ​​in the temperature prediction curve and the battery state of charge prediction curve; performing a numerical comparison based on the actual collected values ​​and the predicted values ​​to obtain a numerical comparison result; and adjusting the model parameters in the thermal state assessment processing and the energy state assessment processing based on the numerical comparison result.

[0140] Among them, the actual collected value refers to the true temperature and power values ​​measured by the sensor after the control cycle ends; the predicted value refers to the estimated temperature and power values ​​at the corresponding time derived by the model when the cycle starts; the numerical comparison refers to the calculation process of comparing the difference between the actual measured value and the model estimated value; the numerical comparison result refers to the conclusion of the comparison between the magnitude of the deviation between the measured value and the predicted value.

[0141] In this embodiment, the accuracy of long-term prediction of thermoelectric conditions under multiple iterations is improved by comparing and correcting the model's calculation parameters through periodic closed-loop measurements, and the impact of calculation deviations caused by on-orbit environmental aging is reduced.

[0142] In another embodiment, adjusting the model parameters in the thermal state assessment process and the energy state assessment process based on the numerical comparison results includes: if the numerical comparison results indicate that the deviation between the actual temperature and the predicted temperature exceeds a set temperature correction threshold, then fine-tuning the thermal parameters in the thermal state assessment model of the thermal state assessment process; if the numerical comparison results indicate that the deviation between the actual SOC and the predicted SOC exceeds a set SOC correction threshold, then fine-tuning the efficiency parameters in the energy state assessment model of the energy state assessment process.

[0143] The technical research process and other technical details of this application are described below with reference to a specific embodiment.

[0144] In traditional technologies, with the rapid deployment of low-Earth orbit satellite constellations and the continuous growth in demand for onboard intelligent processing, on-orbit computing power has become a core indicator for measuring satellite information processing capabilities. More and more satellites are equipped with high-performance computing chips. However, the on-orbit operation of these high-performance chips faces severe challenges in both thermal management and energy supply.

[0145] Current satellite system designs generally employ an architecture where attitude control and computing power management are independent. The lack of information exchange and collaborative optimization mechanisms between the attitude control system and the computing power scheduling system leads to frequent problems such as overheating, frequency throttling, or power failure of computing chips.

[0146] In one embodiment, a separate satellite attitude control and computing power management scheme is adopted. Its working principle is as follows: the attitude control system calculates the target attitude angle based on the main mission's pointing requirements and drives the actuators to complete attitude adjustments; the computing power scheduling system independently acquires the task queue and allocates tasks according to priority and computing resource occupancy. Thermal control relies on fixed designs such as passive heat dissipation coatings and active heaters, while energy management relies on a power management module that executes a simple load cutoff strategy based on battery SOC. The main drawbacks of this scheme are: first, attitude control does not consider the chip's thermal state; when the satellite's attitude causes the chip's surface to face the sun for an extended period, the chip temperature rises rapidly, leading to frequency reduction or triggering overheat protection, resulting in a significant drop in continuous computing power; second, computing power scheduling does not refer to attitude forecast information and cannot predict future trends in the thermal environment and power supply conditions, leading to a severe mismatch between task scheduling and actual available resources, with high-computing-power tasks being scheduled during periods of insufficient power and forced to be interrupted.

[0147] However, the separate satellite attitude control and computing power management scheme has the following drawbacks.

[0148] First, attitude control and computing power thermal management are completely disconnected. Existing attitude control systems only optimize the pointing accuracy of the primary task, neglecting the impact of attitude angle changes on the solar illumination conditions of the satellite surface where the computing chip is located. Specifically, when the primary task's attitude causes the chip's mounting surface to consistently face the sun, the chip receives intense radiation, leading to a rapid temperature rise, triggering overheat protection mechanisms, forcing a reduction in computing frequency or even shutdown. This results in computing task interruptions, reduced data processing timeliness, and severely impacts the satellite's on-orbit intelligent processing capabilities.

[0149] Second, computing power scheduling lacks attitude prediction information support. Existing computing power scheduling systems statically orchestrate tasks based on the current task queue and computing resources, without acquiring or utilizing information about future attitude changes and the resulting changes in thermal environment and power supply conditions. Specifically, high-power computing tasks may be scheduled to be executed when the system is about to enter the shadow period and the battery SOC is low, causing the task to be forced to terminate midway due to insufficient power supply, resulting in lost computing results and wasted resources from repeated calculations.

[0150] Third, energy utilization efficiency is low. Because attitude control does not take into account the optimization of the solar panel's irradiated area, and the computing load is not dynamically adjusted according to power generation conditions, the satellite may generate insufficient power during the solar panel's irradiated segment due to poor solar panel irradiation angle, while the shadowed segment cannot support continuous computing power operation due to insufficient energy storage in the early stage. As a result, the continuous computing power that a single satellite can support is far lower than the chip design capability.

[0151] Based on this, this application provides a method for joint optimization of satellite attitude control and on-orbit computing task scheduling based on thermoelectric joint control, also known as a method for joint optimization of attitude control and computing task scheduling. This method aims to solve the problem of independent attitude control and computing management lacking coordination in existing satellite systems; for ease of description in the embodiments, it can also be simply referred to as the optimization method of this application. The details are as follows.

[0152] The optimization method of this application introduces the thermal state of the computing chip and the state of charge (SOC) of the battery as secondary optimization objectives in the attitude control decision. Under the premise of satisfying the primary mission pointing constraints, the satellite attitude is finely adjusted to achieve a balance between moderate heat dissipation of the chip in avoiding light and sufficient illumination of the solar panel. At the same time, the computing power scheduling system uses attitude prediction information to dynamically schedule future tasks, ultimately achieving the goal of significantly improving the continuous computing power level that a single satellite can support.

[0153] The overall technical solution of the optimization method in this application is summarized as follows:

[0154] The core idea of ​​the optimization method proposed in this application is to break down the information barrier between the attitude control subsystem and the computing power scheduling subsystem in existing satellite systems, and establish a two-way coupled joint optimization closed loop. The overall technical solution includes five key components: multi-source state perception, joint optimization decision-making, attitude fine-tuning execution, computing power task orchestration, and closed-loop feedback correction.

[0155] In the multi-source state perception stage, the system collects real-time multi-dimensional state information such as the satellite's current attitude parameters, computing chip temperature, battery SOC, solar panel output power, and orbital position. This information constitutes the input dataset for joint optimization decision-making, providing a complete state foundation for subsequent coordinated control. Attitude parameters include the satellite's three-axis Euler angles and angular rates in the orbital coordinate system; computing chip temperature is obtained through temperature sensors mounted on the chip package surface and its heat dissipation substrate; battery SOC is estimated by the power management unit based on the integration of charging and discharging current and battery terminal voltage; solar panel output power is collected in real-time by the photovoltaic array's current and voltage monitoring circuit; and orbital position information is calculated from the onboard navigation receiver or the orbital elements recorded on the ground.

[0156] In the joint optimization decision-making stage, the system inputs the aforementioned multi-source state information into the joint optimization model. This model prioritizes minimizing the primary task pointing deviation, with secondary optimization objectives being maintaining the computing chip temperature within a suitable operating range and the battery SOC within a safe operating range. It comprehensively calculates the optimal attitude fine-tuning amount and the corresponding computing task orchestration scheme. The key to joint optimization decision-making lies in establishing the mapping relationship between attitude angle and chip thermal environment, as well as the mapping relationship between attitude angle and solar panel power generation efficiency. The solar radiation intensity received by the satellite surface where the chip is located varies at different attitude angles, directly affecting the chip's external heat input; similarly, the effective irradiated area of ​​the solar panel changes with the attitude angle, affecting power generation. The joint optimization model integrates these two sets of mapping relationships with the primary task pointing constraints into a multi-objective optimization framework, seeking the attitude solution that optimizes the combined benefits of thermal and electronic control within a constraint space that satisfies pointing accuracy requirements.

[0157] In the attitude fine-tuning execution phase, the attitude control system, based on the target attitude angle command output by the joint optimization decision, fine-tunes the satellite's attitude through actuators such as reaction wheels. The magnitude of the fine-tuning is limited by the pointing accuracy constraints of the primary mission, and is typically performed within the allowable pointing deviation margin of the primary mission. The core of this phase is to ensure that attitude fine-tuning does not affect the normal execution of the primary mission, meaning that the imaging quality of the Earth observation camera or the link performance of the communication antenna is not degraded due to fine-tuning. While executing the fine-tuning command, the attitude control system continuously monitors the pointing deviation of the primary mission. Once the deviation is detected to be close to the allowable upper limit, the primary mission attitude is immediately restored first.

[0158] In the task orchestration phase, the computing power scheduling system receives attitude prediction information from the joint optimization decision output. Based on this, it calculates the future trends of the chip's thermal environment and the solar panel's power generation, and then dynamically adjusts the task execution plan. Specifically, when the attitude prediction indicates that the chip will be in a high-temperature environment and the battery SOC is sufficient for a certain period, the scheduling system can appropriately reduce the computing load during that period to avoid overheating. When the prediction indicates that a period of abundant sunshine and good chip heat dissipation is approaching, the scheduling system can schedule high-computing tasks to be executed in advance during that period. This task orchestration method based on attitude prediction ensures a high degree of match between the temporal distribution of computing load and the temporal distribution of thermoelectric conditions.

[0159] In the closed-loop feedback correction stage, at the end of each control cycle, the system feeds back the actual chip temperature, battery SOC, and solar panel power to the joint optimization decision module for comparison with the expected values. If the actual state deviates from the expected value by more than a set threshold, the parameters of the optimization model are corrected, updating the attitude fine-tuning and task scheduling scheme for the next control cycle. This closed-loop mechanism can effectively cope with model deviations caused by uncertainties such as changes in the space environment and aging of onboard equipment, ensuring the robustness of the joint optimization scheme in long-term operation.

[0160] Figure 2 This is an overall technical flow diagram of one embodiment. Figure 2 The presentation demonstrates five key components and their logical relationships. The multi-source state perception module inputs collected attitude, temperature, SOC, power, and orbital information into the joint optimization decision-making module. Simultaneously, the joint optimization decision-making module outputs attitude fine-tuning instructions and computing power task orchestration schemes, which are executed by the attitude fine-tuning execution module and the computing power task orchestration module, respectively. The actual operating results of the two execution modules are returned to the joint optimization decision-making module through a closed-loop feedback correction module, forming a continuously iterative optimization loop. The core feature of the overall scheme is that the joint optimization decision-making module simultaneously drives both the attitude and computing power subsystems, achieving their collaborative optimization.

[0161] The optimization method of this application has the following technical features:

[0162] (1) Thermoelectric dual-target coupling characteristics of attitude fine-tuning

[0163] This application simultaneously applies attitude fine-tuning to two objectives: thermal management of the computing chip and optimization of solar panel power generation. The satellite's geometry dictates that the surface housing the computing chip and the surface housing the solar panel are typically located in different orientations. When the satellite's attitude is fine-tuned, the angles of both the chip surface and the solar panel surface relative to the solar vector direction change simultaneously. However, the direction and magnitude of these changes are determined by the satellite's geometry, resulting in a coupling effect. For example, when attitude fine-tuning causes the chip surface to deviate from the solar direction to reduce irradiance input, the angle of illumination on the solar panel surface may also change, potentially increasing or decreasing the illuminated area. The joint optimization model needs to accurately describe this coupling relationship, minimizing or maximizing the effective illuminated area of ​​the solar panel while reducing the chip's heat input. This characteristic is the physical basis for achieving combined thermoelectric control and is the key difference between this invention and single-objective optimization methods.

[0164] (2) Characteristics of attitude prediction-driven look-ahead mission orchestration

[0165] The computing task orchestration in this application is driven by attitude prediction information, achieving look-ahead scheduling. Traditional computing scheduling only makes decisions based on currently available resources, which is a reactive scheduling approach and cannot predict upcoming changes in resource conditions. In this invention, the joint optimization decision module generates an attitude prediction sequence for several future control cycles while calculating the current attitude fine-tuning scheme, and uses this to infer future trends in chip temperature and power generation. The computing scheduling execution module uses this prediction information to schedule high-computing tasks during periods when the prediction indicates good heat dissipation and sufficient power supply, and schedule low-power standby or data preprocessing tasks during periods when the prediction indicates unfavorable thermal environment or tight power supply. This look-ahead orchestration feature ensures a high degree of matching between the temporal distribution of computing load and the temporal distribution of thermoelectric conditions, avoiding resource waste and task interruptions.

[0166] (3) The hierarchical constraint characteristic of prioritizing the main task

[0167] This invention maintains the primary task's pointing priority throughout the joint optimization process. The feasible domain for attitude fine-tuning is strictly limited to the pointing deviation margin allowed by the primary task. When the primary task is performing observation or communication tasks requiring high-precision pointing, the available fine-tuning space shrinks, and the joint optimization model automatically reduces the magnitude of attitude fine-tuning, relying more on task rearrangement on the computing power scheduling side to address thermoelectric challenges. When the primary task is idle or in periods of low precision requirements, the degrees of freedom for attitude fine-tuning increase, and the joint optimization model can fully utilize attitude adjustments to improve thermoelectric conditions. This hierarchical constraint feature ensures that the performance of the primary task is unaffected by joint optimization, while maximizing the continuous operation capability of computing power within the constraints allowed by the primary task.

[0168] Figure 3 This is a diagram showing the relationship between key technical features in one embodiment. Figure 3 This demonstrates the intrinsic connection between three key technical features. The satellite's geometry determines the dual mapping relationship between attitude angles and chip surface irradiance and solar panel irradiance; these two mappings together construct a thermoelectric dual-objective coupled model. This coupled model generates attitude prediction sequences under the hierarchical constraints of the main mission orientation, which in turn drive look-ahead mission orchestration. The three features are progressively layered and mutually supportive, jointly achieving the overall effect of joint optimization.

[0169] The specific implementation steps of the optimization method in this application are as follows:

[0170] Step 1: System initialization and parameter configuration.

[0171] After the satellite enters orbit and completes initial attitude acquisition and orbit determination, the joint optimization system is initialized and configured. Configuration includes: inputting satellite geometric model parameters; defining the normal direction of the computing chip mounting surface and the solar panel in the satellite coordinate system; setting the constraint threshold for the main task pointing accuracy, i.e., the maximum allowable value of the three-axis attitude angle deviation; setting the upper and lower limits of the suitable operating temperature range and temperature change rate limit for the computing chip; setting the upper and lower limits of the safe operating range and minimum reserve threshold for the battery SOC; setting the control cycle length and prediction time window length for joint optimization decision-making; and initializing the parameters of the thermal state assessment model and the energy state assessment model. After system initialization, it enters the normal operation cycle.

[0172] Step 2: Real-time acquisition of multi-source status data.

[0173] At the start of each control cycle, the sensor acquisition module initiates the data acquisition process. The attitude sensor component outputs the current three-axis attitude quaternion and angular rate vector; the temperature sensor component outputs the core temperature of the computing chip, the chip mounting panel temperature, the heat dissipation panel temperature, and the temperatures of other panels; the power monitoring component outputs the voltage and current values ​​of each branch of the solar panel and the terminal voltage and charging / discharging current values ​​of the battery; the illumination sensor component outputs the measured solar irradiance of each panel. All data is timestamped and then aggregated to the onboard data bus.

[0174] Step 3: Orbit prediction and solar vector calculation.

[0175] The orbit prediction module reads the latest orbital position and velocity data, runs the orbital propagation model, and generates an orbital position time series covering the prediction time window from the current moment. Based on this, according to Earth epoch data and the solar position model, it calculates the direction of the solar vector in the orbital coordinate system for each time step within the prediction time window; and according to the geometric relationship between the satellite orbit and the Earth's shadow cone, it calculates the start and end times of the sunshine and shadow segments within the prediction time window. The prediction results are stored in time series data format and submitted to the thermal state assessment module and the energy state assessment module.

[0176] Step 4: Thermal condition assessment and temperature prediction.

[0177] The thermal state assessment module receives the current chip temperature, the temperatures of each panel, and the solar vector prediction sequence, and performs a thermal balance analysis. First, it calculates the solar irradiance heat flux density received by the chip's surface under the current attitude; this value is obtained by multiplying the solar constant, the panel's solar absorptivity, and the cosine of the solar incidence angle. Then, it calculates the chip's own heat generation power based on the chip's current computing load. The difference between the external irradiance heat input, the chip's own heat generation, and the heat dissipation through conduction and radiation is taken as the net heat input and divided by the heat capacity of the chip and its substrate to obtain the current temperature change rate. Based on this, the joint optimization decision module provides several candidate attitude fine-tuning schemes. The thermal state assessment module calculates the change in the solar incidence angle of the chip's surface for each candidate attitude scheme, and then predicts the chip temperature change trajectory within the future prediction time window under that scheme. The temperature prediction curves of each candidate scheme are submitted to the joint optimization decision module in time series form.

[0178] Step 5: Energy Status Assessment and SOC Prediction.

[0179] The Energy State Assessment module receives current solar panel voltage and current data, battery SOC estimates, and solar vector forecast sequences, and performs power balance analysis. First, it calculates the effective irradiated area of ​​the solar panels under the current attitude, obtained by multiplying the total area of ​​the solar panels by the cosine of the solar incidence angle. Based on the effective irradiated area and the conversion efficiency of the solar cells, it estimates the current power generation. Then, it calculates the total power consumption of all satellite payloads at the current moment, including computing chip power consumption, attitude control system power consumption, communication system power consumption, and other payload power consumption. Subtracting the total load power consumption from the power generation yields the net power surplus or deficit. During sunny periods, the net power surplus is used to charge the batteries, increasing the SOC; during shadow periods or when load power consumption exceeds power generation, the batteries discharge to provide energy, decreasing the SOC. For each candidate attitude fine-tuning scheme and candidate computing load scheme, the Energy State Assessment module calculates the power generation sequence and SOC change trajectory within the future prediction time window. The SOC prediction curves for each candidate scheme are submitted to the joint optimization decision module.

[0180] Step Six: Joint Optimization Decision Calculation.

[0181] The joint optimization decision module gathers the main task's constraint parameters, temperature prediction curves, and SOC prediction curves for each candidate scheme, and performs multi-objective optimization calculations. First, it determines the feasible region for attitude fine-tuning: based on the current main task state, it determines the upper and lower limits of the three-axis attitude angle offset, and discretizes and generates a set of candidate attitude schemes within this range. Then, for each candidate attitude scheme, it reads its corresponding temperature prediction curve and SOC prediction curve, and calculates the scheme's comprehensive evaluation index. The comprehensive evaluation index is a weighted combination of the degree to which the temperature deviates from the suitable range and the degree to which the SOC deviates from the safe range; the closer the temperature is to the center value of the suitable range and the closer the SOC is to the center value of the safe range, the better the evaluation index. For schemes whose predicted temperature or SOC trajectories touch the upper or lower limits of the constraints within the prediction window, a penalty term is applied. After traversing all candidate schemes, the scheme with the best comprehensive evaluation index is selected as the optimal decision for the current cycle. Simultaneously, based on the thermoelectric condition prediction under the optimal attitude scheme, it adjusts the computing power task scheduling parameters, scheduling high-power tasks during the most favorable temperature and SOC periods and low-power tasks during unfavorable periods. Finally, it outputs the attitude target instruction and task scheduling instruction for the current cycle.

[0182] Step 7: Fine-tuning of posture and reordering of tasks.

[0183] The attitude control execution module receives attitude target commands, calculates the angular offset from the current attitude to the target attitude, and generates flywheel speed change commands or magnetometer current commands through the attitude tracking controller to drive the satellite to smoothly adjust to the target attitude. During attitude adjustment, the attitude control execution module continuously monitors the main task pointing deviation to ensure that the deviation remains within the allowable range. The computing power scheduling execution module receives task scheduling commands, adjusts the execution time and computing frequency of each task in the current task queue, reallocates execution windows for queued but not yet started tasks, and adjusts the computing frequency of running tasks as needed.

[0184] Step 8: Closed-loop feedback and model calibration.

[0185] At the end of the current control cycle, the closed-loop feedback module collects the actual chip temperature, battery SOC, and solar panel power, and compares them with the predicted values ​​from steps four and five for that cycle. If the deviation between the actual and predicted temperatures exceeds the set temperature correction threshold, the thermal parameters in the thermal state assessment model are fine-tuned. If the deviation between the actual and predicted SOC exceeds the set SOC correction threshold, the efficiency parameters in the energy state assessment model are fine-tuned. The corrected model parameters are used for prediction calculations in the next control cycle, enabling the system to gradually adapt to actual changes in the on-orbit environment and the degradation of equipment performance. After completing the feedback correction, the system enters the next control cycle, repeating the process from step two, forming a continuously operating joint optimization closed loop.

[0186] Figure 4 This is a flowchart illustrating the specific implementation steps in one embodiment. Figure 4 This paper demonstrates the complete execution flow and logical relationship of the eight implementation steps of this invention. After system initialization, a cyclical process starting from step two is entered. After state data acquisition, orbit prediction calculation is triggered. The orbit prediction results are simultaneously supplied to two parallel analysis processes: thermal state assessment and energy state assessment. The results of the two assessments converge to the joint optimization decision module for comprehensive optimization. The decision results drive the synchronous execution of attitude fine-tuning and task rescheduling. After closed-loop feedback correction, the execution results are fed back to the state data acquisition step, completing a full optimization cycle. The entire process embodies the closed-loop control concept from data acquisition to decision execution and feedback correction, ensuring that joint optimization plays a continuous and effective role in the long-term on-orbit operation of the satellite.

[0187] The key technical point of the optimization method in this application is:

[0188] (1) Joint optimization framework for attitude control and computing power thermal management.

[0189] This approach expands satellite attitude control from simple primary mission pointing adjustment to a multi-objective control method that simultaneously serves the thermal management of computing chips. While meeting the primary mission pointing constraints, it uses attitude fine-tuning to alter the sunlight conditions on the surface where the chip is located, achieving appropriate light-shielding and heat dissipation for the chip. This framework breaks down the barriers of traditional separate designs for attitude control and computing management, possessing groundbreaking protective value.

[0190] (2) Coupled optimization model of attitude angle and thermoelectric dual objective.

[0191] A coupled mapping relationship is established whereby attitude angle changes simultaneously affect the irradiated heat input of the chip surface and the power generation efficiency of the solar panel. This allows for the simultaneous optimization of both thermal and electrical control objectives within a single attitude fine-tuning action, achieving joint thermoelectric regulation. This coupled model is the core algorithmic support for the joint optimization method and possesses significant technical protection value.

[0192] (3) A forward-looking computing task orchestration mechanism based on attitude prediction.

[0193] The computing power scheduling system utilizes attitude prediction sequences generated by the joint optimization decision module to predict future trends in thermal environment and power supply conditions. Based on this, it proactively schedules the execution time, frequency, and duration of computing tasks, achieving a high degree of matching between the temporal distribution of computing load and the temporal distribution of thermal and electrical conditions. This mechanism differs from traditional responsive scheduling and has independent protective value.

[0194] (4) Hierarchical constraint strategy with priority given to primary tasks.

[0195] A hierarchical constraint structure is adopted in the joint optimization, prioritizing the accuracy of the primary task's pointer as a hard constraint, while optimizing the computing chip temperature and battery SOC as secondary soft constraints. When the primary task's pointer constraint tightens, the optimization strategy is automatically adjusted to favor the computing power scheduling side. This hierarchical strategy ensures that joint optimization does not impair the performance of the primary task, and has practical engineering value in protecting its performance.

[0196] (5) Closed-loop feedback driven model adaptive correction mechanism.

[0197] By periodically comparing the deviations between actual operating conditions and predicted values, the parameters of the thermal state assessment model and the energy state assessment model are automatically corrected. This enables the joint optimization system to adapt to environmental changes and equipment aging during long-term on-orbit operation, maintaining the robustness of the optimization effect. This mechanism has protective value in enhancing the long-term reliability of the system.

[0198] The advantages of the optimization method in this application are:

[0199] Compared to existing separate satellite attitude control and computing power management schemes, the fundamental advantage of this application lies in upgrading attitude control from a simple orientation task to a thermoelectric joint regulation method that simultaneously serves computing power. This transforms two previously independent and unresponsive subsystems into a bidirectionally coupled, collaborative optimization closed loop. In existing schemes, the attitude control system is unaware that its attitude selection will lead to overheating of the computing chip or insufficient power generation from the solar panels, and the computing power scheduling system is also unaware of the thermoelectric changes that future attitude changes will bring. The result of these two subsystems operating independently is that the computing chip frequently experiences overheating, frequency reduction, or power shortages, and the continuous computing power that a single satellite can support is far lower than the chip's design capacity. This invention introduces the chip's thermal state and battery SOC as secondary optimization targets into the attitude control decision-making process. Within the constraints allowed by the primary mission, it fine-tunes the attitude, actively improving the chip's thermal environment and the solar panel's power generation conditions. This eliminates the generation of adverse thermoelectric conditions from a physical source, enabling the chip to operate for a longer period under suitable temperature and sufficient power supply conditions.

[0200] Meanwhile, the forward-looking computing task orchestration mechanism of this invention has a significant advantage in resource utilization efficiency compared to the reactive scheduling of existing solutions. Existing solutions only trigger frequency reduction protection when the chip overheats or load cutoff when the SOC is insufficient. This reactive approach leads to a significant waste of invested computing resources and poor continuity and determinism in task execution. This invention, through attitude prediction-driven task orchestration, schedules high-computing tasks in advance during periods of most favorable thermoelectric conditions. Under the same orbital environment and hardware configuration, this significantly improves the continuous computing power output achievable by a single satellite, effectively enhancing the utilization efficiency of onboard computing resources and the reliability of task completion.

[0201] In addition, in some embodiments, the optimization method of this application may also include a unidirectional optimization scheme that prioritizes computing power scheduling and supplements it with attitude fine-tuning. This scheme does not change the design of the existing attitude control system, still using the main task orientation as the sole objective to determine the satellite attitude, but introduces a unidirectional reading channel for attitude information on the computing power scheduling side. The computing power scheduling system passively acquires current and future attitude prediction information, calculates thermoelectric conditions based on this, and adjusts the task scheduling, but does not issue fine-tuning requests to the attitude control system. The advantage of this alternative scheme is that it does not require modification to the existing attitude control system, has lower engineering implementation difficulty, and lower system coupling; the disadvantage is that it abandons the ability to actively improve thermoelectric conditions brought about by attitude fine-tuning, and can only passively adapt to the thermoelectric environment brought about by attitude. When the main task attitude happens to expose the chip to strong radiation continuously, simply reducing the task frequency cannot fundamentally solve the overheating problem, and the continuous increase in computing power is limited.

[0202] In addition, in some embodiments, the optimization method of this application may also include an enhanced heat dissipation scheme by adding independent thermal control hardware. This scheme does not improve the chip's thermal environment through attitude fine-tuning, but instead adds active thermal control hardware such as deployable heat dissipation fins or variable heat pipes to the computing chip mounting area to cope with high-temperature environments by increasing the heat dissipation area or enhancing heat conduction capabilities. Computing power scheduling still uses the existing independent scheduling method. The advantage of this alternative is that the thermal control effect is direct and deterministic, independent of attitude state; the disadvantages are that it increases the satellite's mass, volume, and cost, and the deployable heat dissipation fins may also affect the satellite's aerodynamic characteristics and attitude stability. Furthermore, it does not solve the optimization problem on the energy supply side, and still faces the risk of computing power interruption due to SOC decline when solar panel power generation is insufficient. Compared to the scheme of this invention, which achieves dual improvement in thermoelectricity through joint optimization at the software level, the hardware enhancement scheme has a lower cost-effectiveness.

[0203] The optimization method of this application can be applied to the following scenarios:

[0204] Typical Application Scenario 1: On-orbit Intelligent Processing of Low-Earth Orbit Remote Sensing Satellites. Low-Earth orbit remote sensing satellites equipped with high-performance AI inference chips need to perform computationally intensive tasks such as target detection and change detection after completing Earth imaging. The orbital period of these satellites is approximately 90 to 100 minutes, with about one-third of that time spent in Earth's shadow. During the sunshine period, the chip may overheat due to strong radiation and reduce its frequency; after entering the shadow period, power generation interruption leads to rapid battery discharge. Using the method of this invention, the attitude control system fine-tunes the attitude during the intervals between Earth imaging tasks to ensure the chip surface is adequately protected from sunlight, while simultaneously optimizing the solar panel's exposure angle for full charging. The computing power scheduling system concentrates inference tasks during periods of good heat dissipation and sufficient SOC, significantly improving the effective computing power output within a single orbital cycle.

[0205] Typical Application Scenario 2: Edge Computing for Low-Earth Orbit (LEO) Communication Satellite Constellations. Satellites in LEO communication constellations, while fulfilling their primary communication relay tasks, also undertake some edge computing tasks, such as user data preprocessing and traffic analysis. Communication antenna pointing constraints provide a certain margin for attitude fine-tuning. The method of this invention can optimize attitude within the communication antenna pointing constraints to improve the thermoelectric conditions of the computing chip, and rationally schedule the execution time of computing tasks based on the tidal characteristics of communication services and attitude prediction information, so that edge computing capabilities are fully utilized during off-peak communication periods.

[0206] Technical Implementation Environment Description: The joint optimization decision-making algorithm of this application runs on the onboard main control computer. Multi-objective optimization calculations can be completed using the remaining computing power of the existing onboard processor, without requiring additional dedicated hardware. The sensors required by the system (star sensor, gyroscope, temperature sensor, power monitoring circuit) are all standard configurations on existing satellite platforms. This invention only adds data fusion, state assessment, and joint optimization decision-making functions at the software level, requiring minimal modification to the satellite platform hardware, and exhibits good engineering feasibility and portability on existing satellite platforms.

[0207] In the optimization method of this application, on-orbit computing power refers to the data processing and computing capabilities provided by the onboard computing chip during satellite operation, including tasks such as remote sensing image processing, edge inference, and inter-satellite data fusion. The sustainability and stability of on-orbit computing power are limited by the dynamic balance between chip temperature, power supply, and task load, and are one of the core indicators for the development of current onboard intelligent processing technology.

[0208] In the optimization method of this application, the State of Charge (SOC) of the battery represents the percentage of the satellite's current remaining battery capacity relative to its rated capacity. It is a key parameter for measuring the available energy storage level of the onboard energy system. SOC directly determines the length of time that the satellite can maintain the normal operation of computing chips and other payloads during shadow periods or high-power missions, and is a core input for energy scheduling decisions.

[0209] In the optimization method of this application, the attitude control system (ACS) is responsible for controlling the satellite's spatial orientation in the orbital coordinate system. It achieves precise adjustment and maintenance of the satellite's pointing direction through actuators such as reaction wheels, magnetic torquers, and thrusters. Traditional attitude control systems focus solely on achieving accurate payload pointing. This invention extends its functionality to a multi-objective control method that also considers computing chip thermal management and solar panel irradiation optimization.

[0210] In the optimization method of this application, Thermal-Electrical Joint Regulation refers to a coordinated adjustment strategy that integrates the thermal state management of the computing chip and the power generation efficiency optimization of the solar panel into the attitude control decision. By fine-tuning the satellite attitude angle, it simultaneously achieves both light-shielding heat dissipation on the surface where the computing chip is located and the maximum irradiated area of ​​the solar panel, thereby creating favorable conditions for the continuous operation of computing power in both thermal and electrical control dimensions.

[0211] In the optimization method of this application, task scheduling refers to the planning process of prioritizing, allocating time windows, and adjusting load intensity of computing tasks to be executed based on the current and future resource status of the satellite. The task scheduling in this invention no longer relies solely on computing resources and communication windows, but incorporates the thermal environment and power supply conditions reflected in attitude prediction as key constraints into the scheduling decision.

[0212] It should be noted that any technical feature in any of the above embodiments provided in this application is also applicable to any of the following embodiments provided in this application; the technical features in the relevant embodiments of the various methods provided in this application are also applicable to the relevant embodiments of the various devices, systems or equipment provided in this application; the same, related or corresponding technical features in the various embodiments provided in this application can be referenced and explained to each other, and similarities will not be repeated.

[0213] Figure 5 Here is a structural block diagram of the device in one embodiment, with reference to Figure 5 The satellite orbital attitude and on-orbit computing power task scheduling joint optimization device based on thermoelectric joint control includes:

[0214] The sensor acquisition module 501 is used to acquire multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle.

[0215] The orbit prediction module 502 is used to acquire the orbit data of the satellite to be controlled, calculate the solar vector based on the orbit data, and obtain the solar vector prediction sequence.

[0216] The thermal state assessment module 503 is used to perform thermal state assessment processing based on multi-source state data and solar vector forecast sequence to determine the current temperature change rate of the satellite to be controlled; and to perform temperature prediction on each candidate attitude fine-tuning scheme based on the current temperature change rate to obtain temperature prediction curves.

[0217] The energy state assessment module 504 is used to perform energy state assessment processing based on multi-source state data and solar vector forecast sequence to determine the net power data of the satellite to be controlled; and to predict the battery state of charge for each candidate attitude fine-tuning scheme based on the net power data to obtain the battery state of charge prediction curve.

[0218] The joint optimization decision module 505 is used to perform multi-objective joint optimization decision calculations based on the main mission direction constraint parameters, the temperature prediction curves of each candidate attitude fine-tuning scheme, and the battery state of charge prediction curves, and to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes.

[0219] The execution module 506 is used to control the satellite to be controlled to perform attitude adjustment operations according to the attitude target instructions corresponding to the target scheme; and to control the satellite to be controlled to perform task rescheduling operations according to the task scheduling instructions corresponding to the target scheme.

[0220] In this embodiment of the application, based on, as follows Figure 5 The connections between the various modules or units shown in the diagram enhance the stability of the satellite system in performing its missions through their cooperation.

[0221] The composition and technical principle of the satellite orbital attitude and on-orbit computing power task scheduling joint optimization device based on thermoelectric joint control are described below from the perspective of other technical descriptions.

[0222] The satellite orbital attitude and on-orbit computing power task scheduling joint optimization device based on thermoelectric joint control provided in this application, also known as the satellite orbital attitude and on-orbit computing power task scheduling joint optimization system based on thermoelectric joint control, can be simply referred to as the device or system. The overall structure of the system / device is as follows:

[0223] The system's overall structure consists of eight functional modules: sensor acquisition module, orbit prediction module, thermal status assessment module, energy status assessment module, joint optimization decision-making module, attitude control execution module, computing power scheduling execution module, and closed-loop feedback module. These modules interact in real-time via an onboard data bus, collectively forming a complete attitude and computing power joint optimization system.

[0224] The sensor acquisition module is responsible for acquiring all the raw data required for system operation. This module includes attitude sensor components, temperature sensor components, power monitoring components, and illumination sensor components. The attitude sensor components include star sensors and gyroscopes, providing real-time measurements of the satellite's three-axis attitude angles and angular rates. The temperature sensor components have temperature measurement points deployed on the surface of the computing chip package, the heat dissipation panel, and key surfaces of the satellite body, providing real-time data on the chip core temperature, heat dissipation surface temperature, and the radiation temperature of each panel. The power monitoring component monitors the output voltage and current of the solar panel array, as well as the terminal voltage and charging / discharging current of the battery pack in real time, providing raw electrical parameters for SOC estimation and power generation calculation. The illumination sensor components are installed on multiple surfaces of the satellite body to detect the solar irradiance intensity on each surface, assisting in determining the current sunshine and shadow status.

[0225] Based on real-time orbital position and velocity information provided by the onboard navigation receiver, and combined with an orbital dynamics model, the orbital prediction module predicts the satellite's orbital position, the timing of alternation between sunlight and shadow, and the direction of the solar vector over several future orbital periods. The prediction results from this module provide a temporal environmental forecast basis for joint optimization decision-making, enabling optimization decisions to extend from reactive control at the current moment to forward-looking planning for future periods.

[0226] The thermal state assessment module receives temperature sensor data and orbit prediction data to establish a thermal state assessment model for the computing chip. This model comprehensively considers four heat flux components: heat generated by the chip's own computational power consumption, external solar irradiance heat input, heat conduction from the chip to the heat dissipation panel, and radiative heat dissipation from the heat dissipation panel to space. It estimates the chip's current temperature change rate and predicts the temperature change trajectory under different attitude and computing load conditions over a future period. The thermal state assessment module outputs the chip's current temperature, temperature change trend, and temperature prediction curves for each candidate attitude, serving as thermal constraint inputs for joint optimization decision-making.

[0227] The Energy State Assessment module receives power monitoring data and orbit forecast data to establish a state assessment model for the satellite's energy system. Based on the current output power of the solar panels and the state of charge (SOC) of the battery, combined with future sunshine and shadow periods obtained from orbit forecasts and the variation patterns of the solar panel's illuminated area under different attitudes, this model predicts the SOC trajectory under various candidate attitudes and computing load conditions over a future period. The Energy State Assessment module outputs the current SOC, power generation, load power consumption, and SOC prediction curves for each candidate scenario, serving as the electrical constraint input for joint optimization decision-making.

[0228] The joint optimization decision module is the core of the entire system. It receives the outputs from the thermal state assessment module and the energy state assessment module, as well as the main task's target parameters. It then runs a multi-objective joint optimization algorithm to calculate the optimal attitude fine-tuning angle and computing power task orchestration scheme that simultaneously satisfies the main task's target constraints, chip temperature constraints, and battery SOC constraints. The joint optimization decision module completes a full optimization calculation within each decision cycle and outputs the attitude target command and task scheduling command for the current cycle.

[0229] The attitude control execution module receives attitude target commands from the joint optimization decision module and converts them into control signals for attitude actuators such as reaction wheels and magnetic torquers, driving the satellite to complete attitude fine-tuning. This module contains an attitude tracking controller to ensure the smoothness and accuracy of the attitude adjustment process, while continuously monitoring the main mission pointing deviation to ensure it does not exceed the allowable range.

[0230] The computing power scheduling and execution module receives task scheduling instructions from the joint optimization decision module and rearranges the computing task queues of the onboard computing chip. Based on these instructions, the module adjusts the execution time, computation frequency, and parallelism of each task to match the actual computing load with current and future thermoelectric conditions.

[0231] The closed-loop feedback module collects actual operating status data at the end of each control cycle, compares it with the expected value of the joint optimization decision module, calculates the deviation, and feeds it back to the joint optimization decision module for updating the initial conditions and correcting the model parameters for the next cycle of optimization calculation.

[0232] Figure 6 This is a system overall structure diagram in one embodiment. Figure 6 The system's eight functional modules and their data flow relationships are illustrated. The sensor acquisition module and orbit prediction module, located in the data input layer, provide raw data and forecast information to the thermal state assessment module and energy state assessment module, respectively. The thermal state assessment module and energy state assessment module, located in the state analysis layer, input the processed thermal and electrical constraints into the joint optimization decision module. The joint optimization decision module, located in the core decision layer, simultaneously outputs instructions to the attitude control execution module and the computing power scheduling execution module. The results from the two execution modules are aggregated by the closed-loop feedback module and then fed back to the joint optimization decision module, forming a complete closed-loop control structure.

[0233] The working principle of the system in this application is as follows:

[0234] (1) Working principle of sensor data acquisition and track prediction

[0235] The sensor acquisition module continuously acquires multi-dimensional data on the satellite's operational status at a fixed sampling frequency. The star sensor in the attitude sensor assembly determines the satellite's absolute attitude in the inertial coordinate system by capturing images of the starry sky and matching them with the satellite's surface; the gyroscope measures the satellite's three-axis angular rates, providing high-frequency dynamic information on attitude changes. The data from both types of sensors are fused using an attitude determination algorithm to output high-precision attitude quaternions or Euler angles. The temperature sensor assembly deploys thermistors or thermocouples on the computing chip package surface, various surfaces of the satellite, and the heat sink. Each sampling cycle reads the temperature value at each measurement point and marks the corresponding spatial location, forming a temperature distribution map of the satellite's surface. The power monitoring assembly measures the output current and voltage of each series branch of the solar panel, as well as the charging and discharging current and terminal voltage of the battery pack, using Hall current sensors and precision voltage sampling circuits to calculate instantaneous power generation and cumulative charging and discharging capacity.

[0236] After acquiring the orbital six-root numbers output by the onboard navigation receiver, the orbit prediction module uses an orbital propagation model to extrapolate the satellite's time-position sequence over several future orbital periods. Based on this, the module calculates the direction angle of the solar vector in the satellite's orbital coordinate system for each time point, as well as the precise times when the satellite enters and leaves the Earth's shadow. This prediction information provides temporal environmental boundary conditions for thermal and energy state assessments.

[0237] Figure 7 This is a schematic diagram illustrating the working principle of sensor data acquisition and orbit prediction in one embodiment. Figure 7 The demonstration showcased the data processing flow within the sensor acquisition module and its connection to the orbit prediction module. Raw data from the star sensor and gyroscope were fused to generate precise attitude information; multi-point temperature measurements from the temperature sensor array generated surface temperature distribution information; and electrical parameter data from the power monitoring circuit were calculated to generate power and SOC information. These three data streams converged at the sensor data aggregation node and were then transmitted to the orbit prediction and solar vector calculation module, providing complete input for subsequent thermal and energy assessments.

[0238] (2) Working principles of thermal state assessment and energy state assessment

[0239] The thermal state assessment module works by analyzing the thermal balance of the satellite surface and the computing chip. The chip's temperature change depends on the dynamic balance of four heat flux components: Joule heat generated by the chip's own computation (heat input), heat from solar irradiance conducted to the chip via the chip's mounting panel (external heat input), heat transferred from the chip to the heat dissipation panel via a heat conduction path (conductive heat dissipation), and heat lost by the heat dissipation panel to the space background via infrared radiation (radiative heat dissipation). Based on the current chip temperature and panel temperatures provided by the sensor acquisition module, combined with solar vector direction information from the orbit prediction module, the thermal state assessment module calculates the solar irradiance heat flux density received by the chip's mounting surface at the current attitude. Furthermore, for multiple candidate attitude angles proposed by the joint optimization decision module, the module calculates the change in irradiance heat flux density of the chip's mounting surface under each candidate attitude, thereby predicting the chip temperature change trajectory corresponding to each candidate scheme over a future period.

[0240] The working principle of the Energy State Assessment module is based on power balance analysis of the satellite energy system. The power generation of the solar panels depends on the solar irradiance and the incident angle between the sunlight and the normal to the solar panel. The closer the incident angle is to zero degrees, the larger the effective irradiated area and the higher the power generation efficiency. This module calculates the effective irradiated area of ​​the solar panels for each candidate attitude angle based on the solar vector direction and shading period information provided by the orbit prediction module, and then estimates the power generation time series under each candidate scheme. Simultaneously, this module combines the current computing power load power consumption fed back by the computing power scheduling execution module, as well as the baseline power consumption of other onboard equipment, to calculate the total load power consumption. By comparing the power generation power with the total load power consumption time-by-time, and overlaying the battery charging and discharging model, the SOC change trajectory under each candidate scheme can be predicted.

[0241] Figure 8 This is a schematic diagram illustrating the working principle of thermal state assessment and energy state assessment in one embodiment. Figure 8 The parallel workflow of the thermal state assessment and energy state assessment modules is demonstrated. Chip temperature, panel temperature data, solar vector, and candidate attitude information are input into the thermal balance calculation process to generate temperature prediction curves for each candidate scenario. Solar vector, candidate attitude information, SOC, and load power consumption data are input into the calculation process for the irradiated area and power generation to generate SOC prediction curves for each candidate scenario. Both sets of prediction curves are simultaneously submitted to the joint optimization decision module.

[0242] (3) Working principle of joint optimization decision-making and execution feedback

[0243] The core working principle of the joint optimization decision module is to find the optimal attitude fine-tuning amount and computing task orchestration scheme within a multi-objective optimization framework. This module first determines the constraints and objective function of the optimization problem. The primary constraint is that the main task pointing deviation does not exceed a specified threshold, ensuring that the main task performance is not affected. Secondary optimization objectives include two aspects: first, maintaining the computing chip temperature within a suitable operating temperature range within the prediction time window to avoid overheating-induced frequency throttling or overcooling-induced startup delays; second, maintaining the battery SOC within a safe operating range within the prediction time window to avoid deep discharge leading to battery life loss or power interruption.

[0244] The joint optimization decision module's decision variables include two sets: one set is the attitude fine-tuning angle, which is the allowable three-axis angular offset based on the primary task's baseline attitude; the other set is the computing power task orchestration parameters, namely the start time, computation frequency, and duration of each task to be executed. The optimization model uses the prediction curves provided by the thermal state assessment module and the energy state assessment module as the evaluation basis for each candidate scheme, traversing or searching the decision variable space to find the combination of schemes that optimizes the comprehensive evaluation index of the secondary objective.

[0245] In a complete decision-making cycle, the joint optimization decision module first obtains the current state value and prediction curve from the thermal state assessment module and the energy state assessment module. Then, based on the requirements of the main task, it determines the feasible region for attitude fine-tuning, generates several candidate attitude schemes within the feasible region, evaluates the corresponding temperature prediction and SOC prediction for each candidate attitude scheme, comprehensively ranks the candidate schemes, and selects the optimal scheme as the output command for the current cycle. After receiving the attitude target command, the attitude control execution module drives the actuator to complete the attitude adjustment through the attitude tracking controller. After receiving the task orchestration command, the computing power scheduling execution module rearranges the task queue and sets the parameters. At the end of the cycle, the closed-loop feedback module collects the actual state, calculates the deviation, and sends it back for model correction in the next cycle.

[0246] Figure 9 This is a schematic diagram illustrating the working principle of joint optimization decision-making and execution feedback in one embodiment. Figure 9 This demonstrates the internal working logic of the joint optimization decision-making module and its connection with the execution and feedback stages. Temperature prediction and SOC prediction data, along with the main task's guiding constraints, are input into the candidate solution generation and evaluation process. After comprehensive ranking, the optimal solution is selected, and attitude and task commands are output. Upon completion, actual state acquisition and deviation feedback transmit the results back, updating the prediction basis for the next cycle, forming a closed loop of continuous optimization.

[0247] In yet another embodiment, a computer device is provided, such as a terminal, whose internal structure diagram may be as follows: Figure 10As shown, it includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. The computer program can be executed by the processor to implement the various methods described in the above embodiments.

[0248] Those skilled in the art will understand that Figure 10 The structure shown is only a block diagram of a part of the structure related to the present application and does not constitute a limitation on the computer device on which the present application is applied. It may also include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements, in order to realize the function of the computer device.

[0249] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0250] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the systems, devices, equipment, modules or units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0251] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, devices, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0252] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0253] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0254] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0255] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium, or a semiconductor medium (e.g., a solid-state drive), etc.

[0256] The technical solutions provided by the embodiments of this application have been described in detail above. Specific examples have been used in the embodiments of this application to illustrate the principles and implementation methods of the embodiments of this application. The description of the above embodiments is only for the purpose of helping to understand the methods and core ideas of the embodiments of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments of this application. Therefore, the content of this specification should not be construed as a limitation on the embodiments of this application.

Claims

1. A joint optimization method for satellite injection attitude and on-orbit computing power task scheduling based on thermoelectric joint control, characterized in that, The method includes: For any given satellite attitude control cycle, multi-source state data of the satellite to be controlled is acquired in real time. The orbital data of the satellite to be controlled is obtained, and the solar vector is calculated based on the orbital data to obtain the solar vector prediction sequence; Based on the multi-source state data and the solar vector prediction sequence, thermal state assessment is performed to determine the current temperature change rate of the satellite to be controlled; based on the current temperature change rate, temperature prediction is performed on each candidate attitude fine-tuning scheme to obtain a temperature prediction curve; Based on the multi-source state data and solar vector forecast sequence, energy state assessment is performed to determine the net power data of the satellite to be controlled; based on the net power data, battery state of charge prediction is performed on each of the candidate attitude fine-tuning schemes to obtain battery state of charge prediction curves; Based on the main mission direction constraint parameters, the temperature prediction curves and battery state of charge prediction curves of each candidate attitude fine-tuning scheme, a multi-objective joint optimization decision calculation is performed to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes. The satellite to be controlled is controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme; the satellite to be controlled is controlled to perform task rescheduling operations according to the task scheduling command corresponding to the target scheme.

2. The method according to claim 1, characterized in that, Before the step of acquiring multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle, the method further includes: performing initialization configuration after the satellite to be controlled has completed initial attitude acquisition and orbit determination after entering orbit; specifically including: Input the geometric model parameters of the satellite body to be controlled, determine the normal direction of the computing chip mounting surface in the satellite body coordinate system, and determine the normal direction of the solar panel in the satellite body coordinate system; Set the constraint threshold for the main task pointing constraint parameter; Set the upper and lower limits of the suitable operating temperature range and the temperature change rate limit for the computing chip; Set the upper and lower limits of the safe operating range for the battery's state of charge and the minimum reserve threshold; Set the control cycle length and prediction time window length for any of the satellite attitude control cycles; Initialize the model parameters in the thermal state assessment process and the energy state assessment process.

3. The method according to claim 1, characterized in that, For any given satellite attitude control cycle, the multi-source state data of the satellite to be controlled is acquired in real time, including: The attitude sensor component outputs attitude data; the temperature sensor component outputs temperature status data; the power monitoring component outputs energy status data; and the light sensor component outputs solar irradiance measurement values ​​for each panel. The data to be labeled is obtained by summarizing the attitude data, temperature status data, energy status data, and solar irradiance measurements of each panel; The data to be labeled is processed according to a unified timestamp and then aggregated to the onboard data bus to obtain the multi-source status data.

4. The method according to claim 1, characterized in that, The process of acquiring the orbital data of the satellite to be controlled, and calculating the solar vector based on the orbital data to obtain a solar vector prediction sequence includes: The latest orbital position and velocity data are read as the orbital data, and the orbital propagation model is run based on the orbital data to generate an orbital position time series covering the prediction time window from the current moment. Based on the generated orbital position time series, according to Earth epoch data and solar position model, the direction of the solar vector in the orbital coordinate system corresponding to each time step within the prediction time window is calculated to determine the direction sequence of the solar vector; The parameters of the Earth's shadow cone are determined based on the direction sequence of the solar vector. The start and end times of the sunshine and shadow segments within the prediction time window are calculated based on the geometric relationship between the satellite orbit represented by the orbit data and the Earth's shadow cone, thus obtaining the time series of the sunshine and shadow segments. The time series of the orbital position, the direction sequence of the solar vector, and the time series of the sunshine and shadow segments are used as forecast results and stored in time series data format to obtain the solar vector forecast sequence.

5. The method according to claim 1, characterized in that, The step of performing thermal state assessment processing based on the multi-source state data and the solar vector forecast sequence to determine the current temperature change rate of the satellite to be controlled includes: The solar irradiance heat flux density received by the chip surface under the current orientation is obtained by multiplying the solar constant, the solar absorptivity of the panel, and the cosine of the solar incident angle, so as to determine the external irradiance heat input. Calculate the chip's own heat generation power based on the chip's current computing load to determine the chip's own heat generation. The net heat input is calculated based on external irradiation heat input, chip self-generated heat and heat conduction and radiation heat dissipation. The current temperature change rate is obtained by dividing the net heat input by the heat capacity of the chip and its substrate.

6. The method according to claim 1, characterized in that, The step of performing energy state assessment processing based on the multi-source state data and solar vector forecast sequence to determine the net power data of the satellite to be controlled includes: The effective irradiated area of ​​the solar panel under the current orientation is obtained by multiplying the total area of ​​the solar panel by the cosine of the solar incidence angle. Estimate the current power generation based on the effective irradiated area and the conversion efficiency of the solar cell; Based on the power consumption of the computing chip, the power consumption of the attitude control system, the power consumption of the communication system, and the power consumption of other payloads, the total power consumption of all satellite payloads at the current moment is calculated. Based on the current power generation and the total power consumption of all satellite loads at the current moment, net power data is determined to characterize whether the power is in surplus or deficit.

7. The method according to claim 1, characterized in that, The satellite to be controlled is controlled to perform attitude adjustment operations according to the attitude target command corresponding to the target scheme. After controlling the satellite to be controlled to perform a task rescheduling operation according to the task scheduling instruction corresponding to the target scheme, the method further includes: When any of the satellite attitude control cycles ends, the actual acquired values ​​of the satellite to be controlled are obtained; Determine the predicted values ​​in the temperature prediction curve and the battery state of charge prediction curve; The numerical comparison results are obtained by comparing the actual collected values ​​with the predicted values. The model parameters in the thermal state assessment and energy state assessment processes are adjusted based on the numerical comparison results.

8. A joint optimization device for satellite injection attitude and on-orbit computing power task scheduling based on thermoelectric joint control, characterized in that, The device includes: The sensor acquisition module is used to acquire multi-source state data of the satellite to be controlled in real time for any satellite attitude control cycle. The orbit prediction module is used to acquire the orbit data of the satellite to be controlled, and to calculate the solar vector based on the orbit data to obtain the solar vector prediction sequence; The thermal state assessment module is used to perform thermal state assessment processing based on the multi-source state data and the solar vector prediction sequence to determine the current temperature change rate of the satellite to be controlled; and to perform temperature prediction on each candidate attitude fine-tuning scheme based on the current temperature change rate to obtain a temperature prediction curve. The energy state assessment module is used to perform energy state assessment processing based on the multi-source state data and the solar vector forecast sequence to determine the net power data of the satellite to be controlled; and to predict the battery state of charge for each of the candidate attitude fine-tuning schemes based on the net power data to obtain the battery state of charge prediction curve. The joint optimization decision module is used to perform multi-objective joint optimization decision calculations based on the main mission direction constraint parameters, the temperature prediction curves of each candidate attitude fine-tuning scheme, and the battery state of charge prediction curve, and to determine the target scheme for any satellite attitude control cycle from the candidate attitude fine-tuning schemes. The execution module is used to control the satellite to be controlled to perform attitude adjustment operations according to the attitude target instructions corresponding to the target scheme; and to control the satellite to be controlled to perform task rescheduling operations according to the task scheduling instructions corresponding to the target scheme.

9. A computer device, characterized in that, The computer device includes: At least one processor and memory; The memory is used to store program code, and the processor is used to call the program code stored in the memory to execute the method as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, It includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images