Satellite solar panel sun pointing attitude regulation method and device and storage medium

Abstract

The application discloses a satellite solar panel sun-pointing attitude regulation method and device and a storage medium, and relates to the technical field of satellite attitude control. The method comprises the following steps: determining a first output torque according to a roll axis feedforward target angle and a pitch axis feedforward target angle of a solar panel; determining a roll angle error and a pitch angle error of the solar panel according to the roll axis feedforward target angle and the pitch axis feedforward target angle respectively, and determining an error vector according to the roll angle error and the pitch angle error; determining a second output torque according to a recurrent neural network which is trained in advance based on the error vector; and determining a third output torque according to the first output torque and the second output torque, and driving the solar panel to complete sun-pointing orientation according to the third output torque. Thus, the application realizes high-precision, self-adaptive, non-accumulative error and low-vibration sun-pointing orientation control of a low-orbit satellite solar panel under dynamic atmospheric disturbance and mechanical load drift.

Description

Technical Field

[0001] This application relates to the field of satellite attitude control technology, and in particular to a method, device, and storage medium for adjusting the solar attitude of a satellite solar panel. Background Technology

[0002] As a core component of satellite energy supply, the solar panels of low-Earth orbit (LEO) satellites directly determine the stability and endurance of the satellite's on-orbit operation through their sun-alignment accuracy. LEO satellites operate in a thin atmosphere, and are susceptible to dynamic disturbances such as atmospheric molecule collisions and non-uniform atmospheric density distribution, which can easily cause the solar panels to deviate in attitude. Therefore, precise control strategies are needed to achieve attitude adjustment and sun alignment.

[0003] For example, the invention disclosed in CN120482380A, entitled "Optimal Attitude and Solar Panel Control Method and System for Non-Sun-Synchronous Orbit Satellites," includes the following steps: Step S1: Calculate the illumination angle for each orbit during the satellite's lifetime based on the collected orbital parameters; Step S2: Calculate the half-angle of the illumination zone and the illumination duration under each orbital illumination angle; Step S3: Calculate the illumination energy for each orbit under different combinations of the solar panel drive shaft's swing position and the satellite's different flight attitudes; Step S4: Compare the illumination energy for each orbit and select the optimal combination of attitude and solar panel drive position with the highest energy in each cycle; Step S5: Obtain the satellite attitude and solar panel drive shaft position control strategy for all orbital cycles throughout the satellite's entire lifetime.

[0004] For example, the invention with publication number CN112937923A, entitled "Control of a Dual-Axis Solar Array Drive Mechanism for a Near-Earth Inclined Orbit Satellite," includes: a first axis mechanism, a satellite body, a second axis mechanism, and a solar array; the first axis mechanism is mounted on the satellite body, the second axis mechanism is connected to the first axis mechanism via a connecting rod, and the solar array is connected to the second axis mechanism.

[0005] In the field of low-Earth orbit (LEO) satellite solar panel control, traditional solutions often employ proportional-integral-derivative (PID) control algorithms. These algorithms, due to their simple structure and ease of engineering implementation, have been widely used in various attitude control scenarios. However, considering the specific requirements of solar panel orientation for LEO satellites, PID control presents the following problems: (1) Difficult to adjust parameters: The atmospheric disturbances and mechanical loads faced by the solar panels of low-orbit satellites will drift with the time of on-orbit operation. The fixed-parameter PID controller cannot adaptively adjust and it is difficult to maintain the best control performance under all operating conditions. It is easy to cause adjustment lag or overshoot due to parameter mismatch.

[0006] (2) Weak anti-disturbance capability: PID control lacks the ability to predict and adaptively compensate for dynamic disturbances. The PID controller cannot quickly capture the changes in atmospheric disturbances experienced by the low-orbit satellite solar panels, and can only make passive adjustments after the disturbance occurs, which can easily lead to attitude deviations exceeding the allowable range (≤0.5), affecting the accuracy of sun pointing.

[0007] (3) Cumulative error: The SADA drive system of low-orbit satellite solar panels has nonlinear factors such as motor nonlinearity and mechanical assembly deviation. PID control cannot effectively compensate for such nonlinear errors. At the same time, the integral link is prone to error accumulation. During long-term on-orbit operation, small errors will continue to accumulate, which will cause the attitude deviation of the solar panels to gradually increase, requiring frequent manual intervention and adjustment, increasing on-orbit maintenance costs.

[0008] (4) Poor timing correlation and insufficient adjustment continuity: The changes in the attitude deviation of the solar panel have historical patterns, but PID control only focuses on the current and recent error signals and cannot use timing information to predict the trend of deviation changes. Oscillations are likely to occur during the adjustment process, making it difficult to achieve a smooth transition of attitude. Especially when atmospheric disturbances change abruptly, high-frequency vibrations of the solar panel are likely to occur, affecting the lifespan of the solar panel structure.

[0009] The existing PID control algorithms are ill-suited to the dynamic atmospheric disturbances and mechanical load drift of low-orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration. No effective solution has yet been proposed. Summary of the Invention

[0010] The embodiments of this disclosure provide a method, apparatus, and storage medium for controlling the solar attitude of a satellite solar panel, in order to at least solve the technical problems existing in the prior art, such as the difficulty of PID control algorithms adapting to dynamic atmospheric disturbances and mechanical load drift of low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration.

[0011] According to one aspect of the present disclosure, a method for adjusting the attitude of a satellite solar panel towards the sun is provided, comprising: determining a first output torque for driving a motor of the solar panel to overcome atmospheric drag based on a target angle for the flip axis feedforward and a target angle for the pitch axis feedforward of the solar panel, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel; determining a flip angle error and a pitch angle error of the solar panel based on the target angle for the flip axis feedforward and the target angle for the pitch axis feedforward, and determining an error vector based on the flip angle error and the pitch angle error; determining a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second flip axis output torque and a second pitch axis output torque for compensating for the flip angle error and the pitch angle error; and determining a third output torque based on the first output torque and the second output torque, and driving the solar panel to complete the orientation towards the sun based on the third output torque.

[0012] According to another aspect of the present disclosure, a storage medium is also provided, the storage medium including a stored program, wherein, when the program is executed, a processor performs any of the methods described above.

[0013] According to another aspect of the present disclosure, a satellite solar panel attitude control device for solar orientation is also provided, comprising: a first output torque determination module, configured to determine a first output torque for driving a motor of the solar panel to overcome atmospheric drag based on a target angle for the flip axis feedforward and a target angle for the pitch axis feedforward of the solar panel, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel; an error vector determination module, configured to determine the flip angle error and the pitch angle error of the solar panel based on the target angle for the flip axis feedforward and the target angle for the pitch axis feedforward, respectively, and determine an error vector based on the flip angle error and the pitch angle error; a second output torque determination module, configured to determine a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second flip axis output torque and a second pitch axis output torque for compensating for the flip angle error and the pitch angle error; and a solar orientation module, configured to determine a third output torque based on the first output torque and the second output torque, and drive the solar panel to complete solar orientation based on the third output torque.

[0014] According to another aspect of the present disclosure, a satellite solar panel attitude control device for solar orientation is also provided, comprising: a processor; and a memory connected to the processor, for providing the processor with instructions to perform the following processing steps: determining a first output torque for driving a motor of the solar panel to overcome atmospheric drag based on a target angle for the flip axis feedforward and a target angle for the pitch axis feedforward of the solar panel, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel; determining a flip angle error and a pitch angle error of the solar panel based on the target angle for the flip axis feedforward and the target angle for the pitch axis feedforward, respectively, and determining an error vector based on the flip angle error and the pitch angle error; determining a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second flip axis output torque and a second pitch axis output torque for compensating for the flip angle error and the pitch angle error; and determining a third output torque based on the first output torque and the second output torque, and driving the solar panel to complete solar orientation based on the third output torque.

[0015] This application first calculates the first output torque for driving the solar panel motor to overcome atmospheric drag based on the target angles of the solar panel's flip axis and pitch axis. This first output torque includes the first flip axis output torque corresponding to the solar panel's flip axis and the first pitch axis output torque corresponding to the solar panel's pitch axis. Next, based on the target angles of the flip axis and pitch axis, the flip angle error and pitch angle error are determined respectively, and these errors are combined to form a two-dimensional error vector. This error vector is then used as input to a recurrent neural network, which performs time-series recursive calculations and updates the hidden state using the previous hidden state. The network outputs a second flip axis output torque to compensate for the flip angle error and a second pitch axis output torque to compensate for the pitch angle error. Finally, the first and second output torques are summed to obtain a third output torque, which is then sent to the flip axis motor and the pitch axis motor, thereby driving the solar panel to rotate around the flip axis and pitch axis, making the actual rotation angle of the solar panel approach the target angle, thus achieving solar orientation.

[0016] Therefore, this application adopts a feedforward-feedback composite control strategy based on recurrent neural networks, utilizing the temporal memory capability and nonlinear fitting characteristics of recurrent neural networks to adaptively compensate for dynamic atmospheric disturbances and mechanical load drift. This solves the technical problems of existing PID control algorithms, which are unable to adapt to the dynamic atmospheric disturbances and mechanical load drift of low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this disclosure and form part of this application, illustrate exemplary embodiments of this disclosure and are used to explain this disclosure, but do not constitute an undue limitation of this disclosure. In the drawings: Figure 1 This is a hardware structure block diagram of a computing device for implementing the method described in Embodiment 1 of this disclosure; Figure 2 This is a schematic diagram of the overall structure of a satellite equipped with a dual-degree-of-freedom solar panel according to Embodiment 1 of this disclosure; Figure 3 This is a schematic diagram illustrating the definition of the satellite body coordinate system according to Embodiment 1 of this disclosure; Figure 4 This is a schematic diagram illustrating the definition of a solar panel in the solar panel coordinate system according to Embodiment 1 of this disclosure; Figure 5 This is a flowchart illustrating the satellite solar panel attitude control method according to Embodiment 1 of this disclosure; Figure 6 This is a schematic diagram of the structure of a recurrent neural network according to the method described in Embodiment 1 of this disclosure; Figure 7 This is a schematic diagram of the satellite solar panel attitude control device according to Embodiment 2 of this disclosure; Figure 8 This is a schematic diagram of the satellite solar panel attitude control device according to Embodiment 3 of this disclosure. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this disclosure.

[0019] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure 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 of this disclosure described herein can be implemented in orders other than those 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 apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] Example 1

[0021] According to this embodiment, a method for adjusting the solar attitude of a satellite solar panel is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0022] Figure 1 This is a schematic diagram of the hardware architecture of satellite system 10. (Reference) Figure 1 As shown, satellite system 10 includes an integrated electronic system, which includes a processor, a memory, a bus management module, and a communication interface. The memory is connected to the processor, allowing the processor to access the memory, read program instructions stored in the memory, read data from the memory, or write data to the memory. The bus management module is connected to the processor and also to a bus such as a CAN bus. Thus, the processor can communicate with onboard peripherals connected to the bus through the bus managed by the bus management module. Furthermore, the processor also communicates with devices such as cameras, star sensors, telemetry and command transponders, and data transmission equipment via the communication interface. Those skilled in the art will understand that… Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, a satellite may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0023] It is worth noting that the aforementioned spaceborne peripherals connected to the CAN bus can be one or multiple. These spaceborne peripherals include, but are not limited to, GNSS modules, fiber optic gyroscopes, and high-torque flywheels. Further details will not be elaborated upon here.

[0024] It should be noted that, Figure 1One or more processors and / or other data processing circuits shown herein may generally be referred to as "data processing circuitry". This data processing circuitry may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuitry may be a single, independent processing module, or may be integrated, in whole or in part, into any other element in a computing device. As involved in embodiments of this disclosure, the data processing circuitry serves as processor control (e.g., selection of a variable resistor termination path connected to an interface).

[0025] Figure 1 The memory shown can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for determining the communication frequency band corresponding to the beam in the embodiments of this disclosure. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, to implement the above-mentioned method for determining the communication frequency band corresponding to the beam in the application program. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory.

[0026] It should be noted here that, in some optional embodiments, the above... Figure 1 The device shown may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that... Figure 1 This is only one instance of a specific particular instance, and is intended to illustrate the types of components that may exist in the aforementioned devices.

[0027] Figure 2 This is a schematic diagram of the overall structure of the satellite equipped with a dual-degree-of-freedom solar panel as described in this embodiment. Figure 2 As shown, the satellite includes a satellite body 100, and solar panels 210 and 220 connected to both sides of the satellite body 100 via a dual-degree-of-freedom solar panel drive mechanism. The drive mechanism is equipped with independent tilt and pitch axes to achieve dual-degree-of-freedom rotational adjustment of the solar panels 210 and 220. The solar panels include solar panels 210 and 220.

[0028] Specifically, refer to Figure 2 As shown, along the flight direction of the satellite body 100 ( Figure 2 The horizontal dashed line indicates that the solar panel 210 and 220 are set as a flip axis, allowing them to rotate 360° around the flip axis. Figure 2 (As shown by the arc-shaped arrows on the left and right sides), used to adjust the flip angle of solar panels 210 and 220.

[0029] The pitch axis is set perpendicular to the azimuth axis. Figure 2 (As shown by the vertical dashed line in the middle), solar panels 210 and 220 can pitch and rotate around the pitch axis. Figure 2 (As shown by the two sets of arc-shaped arrows in the middle), used to adjust the pitch angle of solar panels 210 and 220. Figure 1 The satellite system 10 shown is deployed in Figure 2 The satellite body 100 shown.

[0030] Figure 3 The satellite body coordinate system O- of the satellite body 100 described in this embodiment A diagram illustrating the definition. (See reference.) Figure 3 As shown, solar panel 210 along + Extending along the axial direction, the solar panel 220 is along - Extending along the axial direction. Furthermore... Figure 3 A schematic diagram is shown of solar panels 210 and 220 rotating to their zero-position. (Reference) Figure 3 As shown, when returning to zero, the normals of solar panels 210 and 220 are perpendicular to the normals of the satellite body 100. The axial direction is consistent. This allows the satellite to further achieve whole-satellite solar orbit, making the satellite body 100- With the axis pointing towards the sun, solar panels 210 and 220 can be operated towards the sun.

[0031] It should be noted that, as Figure 3 As shown, in the satellite body coordinate system O- Below, the rotation axis corresponds to the forward direction (windward direction) of the satellite body 100. shaft (i.e.) shaft and - (The pitch axis is set perpendicular to the tilt axis). shaft (i.e.) shaft and - axis). shaft (i.e.) shaft and - The azimuth axis points towards the Earth's center and serves as the reference axis for attitude control.

[0032] To facilitate the explanation of the technical solution of this embodiment, the solar panel 210 will be used as an example in the following description. The structure, coordinate system definition and control principle of the solar panel 220 are completely the same as those of the solar panel 210, and can be directly applied by those skilled in the art.

[0033] Figure 4The solar panel 210 described in this embodiment is in the solar panel coordinate system O- The following is a definition diagram. (See attached diagram.) Figure 4 As shown, in the windsurfing coordinate system O- Below, along the horizontal axis extending from the solar panel 210 ( The axis () is the flip axis of the solar panel coordinate system, which is related to the pitch axis of the satellite body 100. Parallel, corresponding to the flip angle, can drive the solar panel 210 around this flip axis. Perform a flipping and rotation. Correspondingly, the vertical axis perpendicular to the horizontal axis ( The axis (x) is the pitch axis of the solar panel coordinate system, which is related to the azimuth axis of the satellite body 100. Parallel, corresponding to the pitch angle, can drive the solar panel 210 around the pitch axis. Perform pitch and rotation.

[0034] Under the aforementioned operating environment, according to the first aspect of this embodiment, a method for adjusting the solar attitude of a satellite solar panel is provided. This method comprises... Figure 1 The integrated electronic system shown is implemented. Figure 5 A flowchart illustrating the method is shown below. (Refer to...) Figure 5 As shown, the method includes: S502: Based on the target angle of the solar panel's flip axis feedforward and the target angle of the pitch axis feedforward, determine the first output torque for the motor that drives the solar panel to overcome atmospheric resistance, wherein the first output torque includes the first flip axis output torque corresponding to the solar panel's flip axis and the first pitch axis output torque corresponding to the solar panel's pitch axis. S504: Based on the target angles of the flip axis and pitch axis, determine the flip angle error and pitch angle error of the solar panel respectively, and determine the error vector based on the flip angle error and pitch angle error; S506: Based on the error vector, determine the second output torque according to the pre-trained recurrent neural network, wherein the second output torque includes the second roll axis output torque and the second pitch axis output torque for compensating for roll angle error and pitch angle error; S508: Determine the third output torque based on the first output torque and the second output torque, and drive the solar panel to complete the sun orientation based on the third output torque.

[0035] In the embodiments of this application, the integrated electronic system needs to define three coordinate systems, namely the inertial coordinate system O- Satellite body coordinate system O- and the windsurfing coordinate system O- Among them, the inertial coordinate system O- It serves as a space inertial reference for describing solar ephemeris and satellite orbital ephemeris.

[0036] Furthermore, in the satellite body coordinate system O- middle: The axis is the flip axis, pointing in the direction the satellite is moving (i.e., the windward direction), corresponding to the flip angle. The axis is the pitch axis, perpendicular to A plane corresponds to a pitch angle. The axis is the azimuth axis, pointing towards the Earth's center, corresponding to the azimuth angle. Additionally, the windsurf coordinate system O- The specific definition has been explained above and will not be repeated here.

[0037] Specifically, the integrated electronic system can feed forward the target angle based on the solar panel's flip axis at time t. target angle fed forward with pitch axis The first flip shaft output torque is obtained to drive the solar panel motor to counteract atmospheric drag. and the output torque of the first pitch axis (Corresponding to step S502). The first tilting shaft outputs torque. The first pitch axis output torque is used to counteract atmospheric drag during rotation around the tilt axis. Used to counteract atmospheric drag during rotation around the pitch axis. Regarding how to determine the target feed angle for the tilt axis. target angle fed forward with pitch axis The specific operation, and the target angle fed forward based on the flip axis. target angle fed forward with pitch axis The method for calculating the first output torque will be explained in detail later.

[0038] Furthermore, the integrated electronic system feeds forward the target angle based on the determined flip axis. and pitch axis feedforward target angle Calculate the target angle fed forward from the flip axis respectively. Corresponding flip angle error and the target angle fed forward with the pitch axis Corresponding pitch angle error Then, the integrated electronic system will adjust the flip angle error. With pitch angle error The two components, arranged column-wise, form a two-dimensional vector, which is the error vector. The error vector is used to characterize the degree to which the solar panel deviates from the target attitude at time t (corresponding to step S504).

[0039] Next, combined Figure 6As shown, the integrated electronic system uses the constructed error vector as input to a recurrent neural network (RNN), and then uses this RNN to perform time-series recursive calculations and hidden state updates, finally outputting a two-dimensional vector. The first component of this two-dimensional vector is the output torque of the second flip shaft. The second component is used to compensate for the roll angle error, and it is the output torque of the second pitch axis. This is used to compensate for pitch angle errors (corresponding to step S506). The recurrent neural network is an artificial neural network with temporal memory capabilities. It contains hidden states that store the network output information from the previous moment, which, together with the current input, determine the current output. The method for determining the second output torque using the recurrent neural network based on the error vector will be explained later.

[0040] Next, the integrated electronic system adds the first output torque to the second output torque to obtain the flip axis component corresponding to the solar panel's flip axis and the pitch axis component corresponding to the solar panel's pitch axis. Then, the integrated electronic system sends the flip axis component of the third output torque to the flip axis motor and the pitch axis component to the pitch axis motor. The flip axis motor then outputs the third flip axis torque corresponding to the flip axis at time t. The solar panels are driven to rotate around the flip axis, and the pitch axis motor outputs torque based on the received time t corresponding to the third pitch axis. The solar panel is driven to rotate around the pitch axis. Through this driving mechanism, the actual rotation angle of the solar panel continuously approaches the feedforward target angle (the feedforward target angle of the flip axis). target angle fed forward with pitch axis This completes the orientation to the sun (corresponding to step S508).

[0041] As mentioned in the background technology above: For the special requirements of solar panels of low-orbit satellites for sun orientation, PID control has the following problems: (1) Difficult to adjust parameters: The atmospheric disturbances and mechanical loads faced by the solar panels of low-orbit satellites will drift with the on-orbit operation time. The PID controller with fixed parameters cannot adaptively adjust, making it difficult to maintain the best control performance under all operating conditions. It is easy to cause adjustment lag or overshoot due to parameter mismatch. (2) Weak anti-disturbance capability: PID control lacks the ability to predict and adaptively compensate for dynamic disturbances. The PID controller cannot quickly capture the change law of atmospheric disturbances encountered by the low-orbit satellite panels. It can only make passive adjustments after the disturbance occurs, which can easily lead to attitude deviation exceeding the allowable range (≤0.5), affecting the sun orientation accuracy. (3) Cumulative error: The SADA drive system of low-orbit satellite panels has nonlinear factors such as motor nonlinearity and mechanical assembly deviation. PID control cannot effectively compensate for such nonlinear errors. At the same time, the integral link is prone to error accumulation. During long-term on-orbit operation, small errors are constantly superimposed, which will cause the attitude deviation of the panels to gradually increase, requiring frequent manual intervention and adjustment, increasing the on-orbit maintenance cost. (4) Poor timing correlation and insufficient adjustment continuity: The changes in the attitude deviation of the solar panel have historical patterns, but PID control only focuses on the current and recent error signals and cannot use timing information to predict the trend of deviation changes. Oscillations are likely to occur during the adjustment process, making it difficult to achieve a smooth transition of attitude. Especially when atmospheric disturbances change abruptly, high-frequency vibrations of the solar panel are likely to occur, affecting the lifespan of the solar panel structure.

[0042] In view of this, this application first calculates the first output torque for the motor driving the solar panel to overcome atmospheric drag based on the target feedforward angles of the solar panel's flip axis and pitch axis. This first output torque includes the first flip axis output torque corresponding to the solar panel's flip axis and the first pitch axis output torque corresponding to the solar panel's pitch axis. Next, based on the target feedforward angles of the flip axis and pitch axis, the flip angle error and pitch angle error are determined respectively, and these errors are combined to form a two-dimensional error vector. Then, this error vector is used as input to a recurrent neural network, which performs time-series recursive calculations and hidden state updates in conjunction with the hidden state from the previous moment, outputting the second flip axis output torque for compensating for the flip angle error and the second pitch axis output torque for compensating for the pitch angle error. Finally, the first output torque and the second output torque are summed to obtain the third output torque, which is then sent to the flip axis motor and the pitch axis motor to drive the solar panel to rotate around the flip axis and the pitch axis, so that the actual rotation angle of the solar panel approaches the feedforward target angle, thereby completing the orientation to the sun.

[0043] Therefore, this application adopts a feedforward-feedback composite control strategy based on recurrent neural networks, utilizing the temporal memory capability and nonlinear fitting characteristics of recurrent neural networks to adaptively compensate for dynamic atmospheric disturbances and mechanical load drift. This solves the technical problems of existing PID control algorithms, which are unable to adapt to the dynamic atmospheric disturbances and mechanical load drift of low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration.

[0044] Optionally, the operation of determining the second output torque based on the error vector and a pre-trained recurrent neural network includes: concatenating the error vector with the feature vector output by the recurrent neural network at the previous time step to determine the input vector; inputting the input vector into the recurrent neural network to determine the feature vector corresponding to the error vector; and inputting the feature vector into a fully connected layer, whereby the fully connected layer outputs the second flip axis output torque corresponding to the flip axis and the second pitch axis output torque corresponding to the pitch axis, wherein the second output torque includes the second flip axis output torque and the second pitch axis output torque.

[0045] Specifically, in order to improve the ability of recurrent neural networks (RNNs) to capture temporal changes in errors and enhance the continuity of perturbation suppression, reference Figure 6 As shown, the integrated electronic system will use the error vector at the current moment. The feature vector output by the RNN at the previous time step (i.e., time step (t-1)) The vectors are concatenated and used as the input vector for the current round of the RNN. Its splicing method is as follows: This allows the recurrent neural network to utilize both current error information and historical time-series information simultaneously. The feature vector is the hidden state stored in the previous time step of the recurrent neural network, and it is also a multi-dimensional vector (the dimension is determined by the network structure).

[0046] Next, the integrated electronic system will input vector After being fed into a pre-trained recurrent neural network, the network will combine the current input vector with its internally trained weight parameters and activation function. The feature vector from the previous time step Calculate the new feature vector at the current time. The new feature vector at the current moment. Including the flip angle error at the current moment and pitch angle error and flip angle error and pitch angle error The temporal characteristics of the changing trend.

[0047] Furthermore, the integrated electronic system will use the feature vector at time t. The input is fed into a fully connected layer. This fully connected layer is a feedforward neural network layer where each neuron is connected to all dimensions of the input. This fully connected layer processes the feature vector... Perform a linear weighted summation and add a bias term, then use an identity mapping (e.g., f(o)=o) as the activation function to output the second output torque. Among them, the second output torque Includes compensation for flip angle error The second flip shaft output torque and for compensating for pitch angle errors Second pitch axis output torque .

[0048] Thus, through the above methods, recurrent neural networks can simultaneously utilize current deviations and historical time series information to have the ability to predict the changing trends of atmospheric disturbances and achieve proactive compensation for dynamic disturbances.

[0049] Optionally, the operation of inputting the input vector into a recurrent neural network to determine the feature vector corresponding to the error vector includes: inputting the input vector into the input layer of the recurrent neural network; and using the hidden layer and output layer of the recurrent neural network to determine the feature vector corresponding to the error vector.

[0050] Specifically, the integrated electronic system first inputs the vector The input vector is passed to the input layer of the recurrent neural network. Subsequently, the integrated electronic system utilizes the hidden layer of the recurrent neural network to process the input vector. The hidden layer contains recurrently connected neurons, each storing the feature vector from the previous time step. and the input vector at the current time. They participate in the computation together. Next, the hidden layer outputs the original hidden state at the current time step. Finally, the integrated electronic system passes the output of the hidden layer to the output layer; the output layer then outputs the feature vector at the current time step. .

[0051] In this way, an accurate timing representation is provided for the generation of the second output torque in the subsequent fully connected layer, thereby enhancing the network's ability to remember and predict dynamic atmospheric disturbances.

[0052] Optionally, the operation of determining the third output torque based on the first output torque and the second output torque includes: adding the first tilt axis output torque of the first output torque to the second tilt axis output torque of the second output torque to determine the third tilt axis output torque of the third output torque; and adding the first pitch axis output torque of the first output torque to the second pitch axis output torque of the second output torque to determine the third pitch axis output torque of the third output torque.

[0053] Specifically, for the direction of the flip axis, the integrated electronic system will output the first flip axis output torque from the first output torque. The second flip shaft output torque in the second output torque Adding them together, we obtain the third flip shaft output torque from the third output torque. The calculation formula is as follows: (1) For the pitch axis direction, the integrated electronic system will output the first pitch axis output torque from the first output torque. The second pitch axis output torque in the second output torque Adding them together yields the third pitch axis output torque in the third output torque. The calculation formula is as follows: (2) In the above manner, the integrated electronic system combines the feedforward open-loop control quantity (first output torque) and the feedback closed-loop compensation quantity (second output torque) to obtain the final drive torque (third output torque).

[0054] Optionally, the method further includes: determining a second solar unit vector in the satellite body coordinate system based on a first attitude transformation matrix from the inertial coordinate system to the satellite body coordinate system and a first solar unit vector in the inertial coordinate system; determining a second solar panel normal direction vector based on a first solar panel normal direction vector in the solar panel's solar panel coordinate system and a second attitude transformation matrix from the solar panel coordinate system to the satellite body coordinate system; and determining the solar panel's flip axis feedforward target angle and pitch axis feedforward target angle based on the second solar unit vector and the first solar panel normal direction vector, provided that the second solar unit vector and the second solar panel normal direction vector are preset to be equal.

[0055] Specifically, the integrated electronic system first acquires the inertial coordinate system O- at time t. To the satellite body coordinate system O- First attitude transformation matrix and the inertial coordinate system O- The first solar unit vector below (This vector represents the direction of the Sun in the inertial frame of reference). Then, the integrated electronic system calculates the second solar unit vector in the satellite's body coordinate system according to the following formula. : (3) Next, the integrated electronic system acquires the coordinate system O- of the solar panel at time t. The first normal direction vector of the windshield (The normal direction vector of the first sail) Fixed as ), and the windsurfing coordinate system O- To the satellite body coordinate system O- The second attitude transformation matrix The integrated electronic system then multiplies the two values ​​to calculate the normal vector of the second solar panel in the satellite's body coordinate system. The specific calculation formula is as follows: (4) In the formula, the second attitude transformation matrix at time t It can be determined by the coordinate system around the satellite body at the corresponding time. The first rotation matrix of the axis (pitch axis) and the coordinate system around the satellite body The second rotation matrix of the axis (azimuth axis) The specific expression is as follows: (5) Wherein, the first rotation matrix It can accurately reflect the attitude relationship of the solar panel rotating around the flip axis, and its deployment form is as follows: Furthermore, the second rotation matrix It can accurately reflect the attitude relationship of the solar panel rotating around the pitch axis, and its deployment form is as follows: Furthermore, since the solar panel's normal needs to be precisely aligned with the sun, the integrated electronic system pre-sets a second solar unit vector. It should be in line with the vector of the normal direction of the second sail. They are equal. That is: (6) Based on this, the integrated electronic system can be derived from formulas (3) to (6) as follows: (7) Based on formula (7), the integrated electronic system can, according to the obtained second solar unit vector, and the known first solar panel normal direction vector Combined with the second attitude transformation matrix The correspondence between the angle and the solar panel's rotation axis feedforward target angle at time t can be directly calculated. target angle fed forward with pitch axis .

[0056] Thus, through the above methods, the integrated electronic system can directly calculate the target angles of the flip axis and pitch axis, enabling the solar panels to be aligned with the sun in advance based on ephemeris information.

[0057] Optionally, determining the first output torque for driving the motor to overcome atmospheric drag based on the target angles of the solar panel's flip axis and pitch axis includes: acquiring a pre-set flip axis output torque as the first flip axis output torque; determining the solar panel attitude based on the target angles of the flip axis and pitch axis; determining the angle between the solar panel's normal direction and the satellite's windward direction based on the solar panel attitude; determining the effective windward area of ​​the solar panel based on the angle and the fixed area of ​​the solar panel; determining the atmospheric disturbance force on the solar panel based on the effective windward area; and determining the first pitch axis output torque based on the atmospheric disturbance force.

[0058] Specifically, due to the symmetrical structure of the solar panels on both sides of the flip shaft, the atmospheric disturbance torques cancel each other out. Therefore, the integrated electronic system first obtains the pre-set flip shaft output torque as the first flip shaft output torque. The first flip shaft outputs torque. It is a fixed value used to overcome mechanical friction and assembly deviations.

[0059] Next, the integrated electronic system feeds forward the target angle based on the flip axis. target angle fed forward with pitch axis The system determines the current attitude of the solar panel (i.e., the rotation angle of the panel around the flip axis and pitch axis). Then, based on this attitude, the integrated electronic system calculates the angle between the solar panel's normal direction (perpendicular to the front of the solar panel) and the satellite's windward direction (the satellite's forward direction, i.e., the direction the flip axis points). (i.e., the normal vector of the second windshield) and (The included angle between axes).

[0060] Subsequently, the integrated electronic system adjusts the angle and the fixed area of ​​the windsurfing board accordingly. Calculate the effective windward area of ​​the solar panel (based on the inherent area of ​​the panel itself). The specific calculation formula is as follows: (8) Among them, the effective windward area This reflects the projected area of ​​atmospheric molecules actually impacting the solar panel at time t. The integrated electronic system then utilizes this effective windward area. Combining parameters such as atmospheric density, satellite orbital velocity, and aerodynamic drag coefficient, the atmospheric disturbance force is calculated using the following formula. : (9) In the formula, Let be the atmospheric disturbance force at time t. Let be the atmospheric density at low Earth orbit at time t. Let be the satellite's orbital velocity at time t. Let be the effective windward area at time t. This is the starting resistance coefficient (which can be set to a fixed value).

[0061] Furthermore, the integrated electronic system will analyze the atmospheric disturbance at time t. Decomposed to the effective perturbation components along the normal direction of the solar panel at time t. The calculation formula is as follows: (10) Then, the integrated electronic system utilizes this effective disturbance component. The primary torque used to actively counteract atmospheric disturbances on the pitch axis, i.e., the first pitch axis output torque, is determined by the following formula. : (11) In the formula, The first pitch axis output torque at time t. Atmospheric disturbance force at time t The component of the pitch axis corresponding to the direction of the normal to the sail. This is the distance between the center of the solar panel and the pitch axis.

[0062] Thus, through the above methods, the first output torque can be dynamically adjusted according to the direction of the solar panel, thereby achieving active pre-compensation for atmospheric drag.

[0063] Alternatively, the flip angle error of the solar panel can be determined according to the following formula. : ;in, The target angle for the flip axis corresponding to the flip axis of the solar panel; The actual rotation angle of the flip axis corresponding to the flip axis of the solar panel; and the pitch angle error of the solar panel determined according to the following formula. : ;in, The target pitch angle is the pitch axis feedforward angle corresponding to the pitch axis of the solar panel; The actual pitch axis rotation angle corresponding to the pitch axis of the solar panel, and the operation of determining the error vector based on the flip angle error and the pitch angle error, includes: constructing the flip angle error and the pitch angle error into a two-dimensional column vector as the error vector.

[0064] Specifically, the integrated electronic system utilizes the flip axis to feed forward the target angle. and the actual rotation angle of the flip axis The flip angle error is determined according to the following formula. : (12) Furthermore, the integrated electronic system utilizes the pitch axis to feed forward the target angle. and pitch axis actual rotation angle The pitch angle error is determined according to the following formula. : (13) Then, the integrated electronic system will adjust the flip angle error. and pitch angle error As two components, arranged in top-bottom order, they form a two-dimensional column vector. , which is the error vector at the current moment, is used to characterize the degree of combined attitude deviation of the sail in both the roll axis and pitch axis directions.

[0065] In addition, refer to Figure 1 As shown, according to a second aspect of this embodiment, a storage medium is provided. The storage medium includes a stored program, wherein, when the program is executed, a processor performs any of the methods described above.

[0066] Therefore, according to this embodiment, this application adopts a feedforward-feedback composite control strategy based on recurrent neural networks. It utilizes the temporal memory capability and nonlinear fitting characteristics of recurrent neural networks to adaptively compensate for dynamic atmospheric disturbances and mechanical load drift. This solves the technical problems in existing technologies where PID control algorithms are difficult to adapt to dynamic atmospheric disturbances and mechanical load drift in low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration.

[0067] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0068] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0069] Example 2

[0070] Figure 7 A satellite solar panel attitude control device according to this embodiment is shown, which corresponds to the method described according to Embodiment 1. (Reference) Figure 7 As shown, the device includes: a first output torque determination module 710, used to determine a first output torque for driving the motor of the solar panel to overcome atmospheric drag based on the target angles of the solar panel's flip axis and pitch axis, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel; an error vector determination module 720, used to determine the flip angle error and pitch angle error of the solar panel based on the target angles of the flip axis and pitch axis, respectively, and to determine an error vector based on the flip angle error and pitch angle error; a second output torque determination module 730, used to determine a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second flip axis output torque and a second pitch axis output torque for compensating for the flip angle error and pitch angle error; and a solar orientation module 740, used to determine a third output torque based on the first output torque and the second output torque, and to drive the solar panel to complete solar orientation based on the third output torque.

[0071] Optionally, the operation of determining the second output torque based on the error vector and a pre-trained recurrent neural network includes: concatenating the error vector with the feature vector output by the recurrent neural network at the previous time step to determine the input vector; inputting the input vector into the recurrent neural network to determine the feature vector corresponding to the error vector; and inputting the feature vector into a fully connected layer, whereby the fully connected layer outputs the second flip axis output torque corresponding to the flip axis and the second pitch axis output torque corresponding to the pitch axis, wherein the second output torque includes the second flip axis output torque and the second pitch axis output torque.

[0072] Optionally, the operation of inputting the input vector into a recurrent neural network to determine the feature vector corresponding to the error vector includes: inputting the input vector into the input layer of the recurrent neural network; and using the hidden layer and output layer of the recurrent neural network to determine the feature vector corresponding to the error vector.

[0073] Optionally, the operation of determining the third output torque based on the first output torque and the second output torque includes: adding the first tilt axis output torque of the first output torque to the second tilt axis output torque of the second output torque to determine the third tilt axis output torque of the third output torque; and adding the first pitch axis output torque of the first output torque to the second pitch axis output torque of the second output torque to determine the third pitch axis output torque of the third output torque.

[0074] Optionally, the device further includes a feedforward target angle determination module, used to determine a second solar unit vector in the satellite body coordinate system based on a first attitude transformation matrix from the inertial coordinate system to the satellite body coordinate system and a first solar unit vector in the inertial coordinate system; to determine a second solar panel normal direction vector based on a first solar panel normal direction vector in the solar panel's solar panel coordinate system and a second attitude transformation matrix from the solar panel coordinate system to the satellite body coordinate system; and, if the second solar unit vector is preset to be equal to the second solar panel normal direction vector, to determine the flip axis feedforward target angle and the pitch axis feedforward target angle of the solar panel based on the second solar unit vector and the first solar panel normal direction vector.

[0075] Optionally, determining the first output torque for driving the motor to overcome atmospheric drag based on the target angles of the solar panel's flip axis and pitch axis includes: acquiring a pre-set flip axis output torque as the first flip axis output torque; determining the solar panel attitude based on the target angles of the flip axis and pitch axis; determining the angle between the solar panel's normal direction and the satellite's windward direction based on the solar panel attitude; determining the effective windward area of ​​the solar panel based on the angle and the fixed area of ​​the solar panel; determining the atmospheric disturbance force on the solar panel based on the effective windward area; and determining the first pitch axis output torque based on the atmospheric disturbance force.

[0076] Alternatively, the flip angle error of the solar panel can be determined according to the following formula. : ;in, The target angle for the flip axis corresponding to the flip axis of the solar panel; The actual rotation angle of the flip axis corresponding to the flip axis of the solar panel; and the pitch angle error of the solar panel determined according to the following formula. : ;in, The target pitch angle is the pitch axis feedforward angle corresponding to the pitch axis of the solar panel; The actual pitch axis rotation angle corresponding to the pitch axis of the solar panel, and the operation of determining the error vector based on the flip angle error and the pitch angle error, includes: constructing the flip angle error and the pitch angle error into a two-dimensional column vector as the error vector.

[0077] Therefore, according to this embodiment, this application adopts a feedforward-feedback composite control strategy based on recurrent neural networks. It utilizes the temporal memory capability and nonlinear fitting characteristics of recurrent neural networks to adaptively compensate for dynamic atmospheric disturbances and mechanical load drift. This solves the technical problems in existing technologies where PID control algorithms are difficult to adapt to dynamic atmospheric disturbances and mechanical load drift in low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration.

[0078] Example 3

[0079] Figure 8 A satellite solar panel attitude control device according to this embodiment is shown, which corresponds to the method described according to Embodiment 1. (Reference) Figure 8 As shown, the device includes: a processor 810; and a memory 820 connected to the processor 810, for providing the processor 810 with instructions to process the following steps: determining a first output torque for driving the motor of the solar panel to overcome atmospheric drag based on the target angles of the solar panel's flip axis feedforward and pitch axis feedforward, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel; determining the flip angle error and pitch angle error of the solar panel based on the target angles of the flip axis feedforward and pitch axis feedforward, respectively, and determining an error vector based on the flip angle error and pitch angle error; determining a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second flip axis output torque and a second pitch axis output torque for compensating for the flip angle error and pitch angle error; and determining a third output torque based on the first output torque and the second output torque, and driving the solar panel to complete solar orientation based on the third output torque.

[0080] Optionally, the operation of determining the second output torque based on the error vector and a pre-trained recurrent neural network includes: concatenating the error vector with the feature vector output by the recurrent neural network at the previous time step to determine the input vector; inputting the input vector into the recurrent neural network to determine the feature vector corresponding to the error vector; and inputting the feature vector into a fully connected layer, whereby the fully connected layer outputs the second flip axis output torque corresponding to the flip axis and the second pitch axis output torque corresponding to the pitch axis, wherein the second output torque includes the second flip axis output torque and the second pitch axis output torque.

[0081] Optionally, the operation of inputting the input vector into a recurrent neural network to determine the feature vector corresponding to the error vector includes: inputting the input vector into the input layer of the recurrent neural network; and using the hidden layer and output layer of the recurrent neural network to determine the feature vector corresponding to the error vector.

[0082] Optionally, the operation of determining the third output torque based on the first output torque and the second output torque includes: adding the first tilt axis output torque of the first output torque to the second tilt axis output torque of the second output torque to determine the third tilt axis output torque of the third output torque; and adding the first pitch axis output torque of the first output torque to the second pitch axis output torque of the second output torque to determine the third pitch axis output torque of the third output torque.

[0083] Optionally, the memory 820 is also configured to provide the processor 810 with instructions to process the following steps: determining a second solar unit vector in the satellite body coordinate system based on a first attitude transformation matrix from the inertial coordinate system to the satellite body coordinate system and a first solar unit vector in the inertial coordinate system; determining a second solar panel normal direction vector based on a first solar panel normal direction vector in the solar panel's solar panel coordinate system and a second attitude transformation matrix from the solar panel coordinate system to the satellite body coordinate system; and, if the second solar unit vector is preset to be equal to the second solar panel normal direction vector, determining the solar panel's flip axis feedforward target angle and pitch axis feedforward target angle based on the second solar unit vector and the first solar panel normal direction vector.

[0084] Optionally, determining the first output torque for driving the motor to overcome atmospheric drag based on the target angles of the solar panel's flip axis and pitch axis includes: acquiring a pre-set flip axis output torque as the first flip axis output torque; determining the solar panel attitude based on the target angles of the flip axis and pitch axis; determining the angle between the solar panel's normal direction and the satellite's windward direction based on the solar panel attitude; determining the effective windward area of ​​the solar panel based on the angle and the fixed area of ​​the solar panel; determining the atmospheric disturbance force on the solar panel based on the effective windward area; and determining the first pitch axis output torque based on the atmospheric disturbance force.

[0085] Alternatively, the flip angle error of the solar panel can be determined according to the following formula. : ;in, The target angle for the flip axis corresponding to the flip axis of the solar panel; The actual rotation angle of the flip axis corresponding to the flip axis of the solar panel; and the pitch angle error of the solar panel determined according to the following formula. : ;in, The target pitch angle is the pitch axis feedforward angle corresponding to the pitch axis of the solar panel; The actual pitch axis rotation angle corresponding to the pitch axis of the solar panel, and the operation of determining the error vector based on the flip angle error and the pitch angle error, includes: constructing the flip angle error and the pitch angle error into a two-dimensional column vector as the error vector.

[0086] Therefore, according to this embodiment, this application adopts a feedforward-feedback composite control strategy based on recurrent neural networks. It utilizes the temporal memory capability and nonlinear fitting characteristics of recurrent neural networks to adaptively compensate for dynamic atmospheric disturbances and mechanical load drift. This solves the technical problems in existing technologies where PID control algorithms are difficult to adapt to dynamic atmospheric disturbances and mechanical load drift in low-Earth orbit satellites, resulting in low control accuracy, error accumulation, and high-frequency vibration.

[0087] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0088] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0092] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0093] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for adjusting the solar attitude of a satellite solar panel, characterized in that, include: Based on the target angles of the solar panel's flip axis feedforward and pitch axis feedforward, a first output torque for driving the motor to overcome atmospheric resistance is determined, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel. Based on the target angle of the flip axis feedforward and the target angle of the pitch axis feedforward, the flip angle error and pitch angle error of the solar panel are determined respectively, and the error vector is determined based on the flip angle error and pitch angle error; Based on the error vector, a second output torque is determined according to a pre-trained recurrent neural network, wherein the second output torque includes a second roll axis output torque and a second pitch axis output torque for compensating for the roll angle error and the pitch angle error; and Based on the first output torque and the second output torque, a third output torque is determined, and based on the third output torque, the solar panel is driven to complete the sun orientation.

2. The method according to claim 1, characterized in that, Based on the error vector, the operation of determining the second output torque according to the pre-trained recurrent neural network includes: The error vector is concatenated with the feature vector output by the recurrent neural network at the previous time step to determine the input vector; The input vector is fed into the recurrent neural network to determine the feature vector corresponding to the error vector; and The feature vector is input to the fully connected layer, and the fully connected layer outputs the second flip axis output torque corresponding to the flip axis and the second pitch axis output torque corresponding to the pitch axis, wherein the second output torque includes the second flip axis output torque and the second pitch axis output torque.

3. The method according to claim 2, characterized in that, The operation of inputting the input vector into the recurrent neural network to determine the feature vector corresponding to the error vector includes: The input vector is fed into the input layer of the recurrent neural network; and The hidden layer and output layer of the recurrent neural network are used to determine the feature vector corresponding to the error vector.

4. The method according to claim 1, characterized in that, The operation of determining the third output torque based on the first output torque and the second output torque includes: The first flip-axis output torque of the first output torque is added to the second flip-axis output torque of the second output torque to determine the third flip-axis output torque of the third output torque; and The first pitch axis output torque of the first output torque is added to the second pitch axis output torque of the second output torque to determine the third pitch axis output torque of the third output torque.

5. The method according to claim 1, characterized in that, Also includes: Based on the first attitude transformation matrix from the inertial coordinate system to the satellite body coordinate system and the first solar unit vector in the inertial coordinate system, the second solar unit vector in the satellite body coordinate system is determined. The second solar panel normal direction vector is determined based on the first solar panel normal direction vector in the solar panel coordinate system and the second attitude transformation matrix from the solar panel coordinate system to the satellite body coordinate system. as well as Assuming the second solar unit vector is equal to the second solar panel normal direction vector, the target angles for the flip axis and pitch axis of the solar panel are determined based on the second solar unit vector and the first solar panel normal direction vector.

6. The method according to claim 5, characterized in that, Based on the target feedforward angles of the solar panel's tilt axis and pitch axis, the operation of determining the first output torque of the motor driving the solar panel to overcome atmospheric drag includes: Obtain the preset output torque of the flip shaft and use it as the first output torque of the flip shaft; The attitude of the solar panel is determined based on the target angle of the flip axis feedforward and the target angle of the pitch axis feedforward. Based on the solar panel's attitude, determine the angle between the normal direction of the solar panel and the satellite's windward direction; The effective windward area of ​​the solar panel is determined based on the included angle and the fixed area of ​​the solar panel. Based on the effective windward area, determine the atmospheric disturbance force experienced by the solar panel; and The output torque of the first pitch axis is determined based on the atmospheric disturbance force.

7. The method according to claim 1, characterized in that, The flip angle error of the solar panel is determined according to the following formula. : ； in, The target angle for the flip axis corresponding to the flip axis of the solar panel; The actual rotation angle of the flip axis corresponding to the flip axis of the solar panel; and The pitch angle error of the solar panel is determined using the following formula. : ； in, The pitch axis feedforward target angle corresponding to the pitch axis of the solar panel; The actual pitch axis rotation angle corresponding to the pitch axis of the solar panel, and wherein... The operation of determining the error vector based on the roll angle error and the pitch angle error includes: constructing a two-dimensional column vector from the roll angle error and the pitch angle error as the error vector.

8. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, the method described in any one of claims 1 to 7 is performed by a processor.

9. A satellite solar panel attitude control device, characterized in that, include: The first output torque determination module is used to determine the first output torque for driving the motor of the solar panel to overcome atmospheric resistance based on the target feed angle of the flip axis and the target feed angle of the pitch axis of the solar panel. The first output torque includes the first flip axis output torque corresponding to the flip axis of the solar panel and the first pitch axis output torque corresponding to the pitch axis of the solar panel. The error vector determination module is used to determine the flip angle error and pitch angle error of the solar panel based on the flip axis feedforward target angle and the pitch axis feedforward target angle, respectively, and to determine the error vector based on the flip angle error and the pitch angle error. The second output torque determination module is configured to determine a second output torque based on the error vector and a pre-trained recurrent neural network, wherein the second output torque includes a second roll axis output torque and a second pitch axis output torque for compensating for the roll angle error and the pitch angle error; and The solar orientation module is used to determine a third output torque based on the first output torque and the second output torque, and drive the solar panel to complete the solar orientation based on the third output torque.

10. A satellite solar panel attitude control device, characterized in that, include: processor; as well as A memory, connected to the processor, for providing the processor with instructions to perform the following processing steps: Based on the target angles of the solar panel's flip axis feedforward and pitch axis feedforward, a first output torque for driving the motor to overcome atmospheric resistance is determined, wherein the first output torque includes a first flip axis output torque corresponding to the flip axis of the solar panel and a first pitch axis output torque corresponding to the pitch axis of the solar panel. Based on the target angle of the flip axis feedforward and the target angle of the pitch axis feedforward, the flip angle error and pitch angle error of the solar panel are determined respectively, and the error vector is determined based on the flip angle error and pitch angle error; Based on the error vector, a second output torque is determined according to a pre-trained recurrent neural network, wherein the second output torque includes a second roll axis output torque and a second pitch axis output torque for compensating for the roll angle error and the pitch angle error; and Based on the first output torque and the second output torque, a third output torque is determined, and based on the third output torque, the solar panel is driven to complete the sun orientation.

Images