Satellite attitude adaptive control algorithm and system based on DSAC-T

Abstract

The application relates to the field of satellite communication and discloses a satellite attitude adaptive control algorithm and system based on DSAC-T, solve the problem that the existing algorithm can make the satellite angular velocity swing between positive and negative values, so that the smooth running of the satellite can be better controlled, compared with the prior art, the DSAC-T algorithm has faster control speed, the algorithm converges faster, the real-time performance of the satellite attitude control is improved, the anti-interference degree is improved, good control stability is achieved in different interference experimental environments, good robustness is achieved for interference not seen in the training process, the trained agent also has good adaptability to satellites with different moments of inertia, so that the agent can be conveniently migrated to a new satellite, and better universality is achieved.

Description

Technical Field

[0001] This invention relates to the field of satellite communication, specifically a satellite attitude adaptive control algorithm and system based on DSAC-T. Background Technology

[0002] Attitude control is crucial for the safe and normal operation of space satellites. When performing different missions, satellites need to be controlled to change their attitude direction accordingly. In the existing technology, satellite attitude control schemes include the DDPG-based nanosatellite attitude control algorithm, the PPO-based satellite attitude control policy algorithm, and the TD3 (Twin Delayed Deep Deterministic Policy Gradient)-based satellite attitude adaptive continuous control algorithm.

[0003] Existing methods have slow convergence speeds, and the control accuracy and stability are subject to certain instabilities. They are also prone to back-and-forth jitter during control, which is detrimental to the attitude stability control of satellites. Moment of inertia affects the ease with which an object can change its rotation. Smaller moment of inertia makes it easier for an object to change its rotation state. Existing schemes based on the TD3 algorithm are not stable enough and tend to apply larger control torques in directions where the satellite's moment of inertia is small, causing the satellite to jitter back and forth. Summary of the Invention

[0004] The purpose of this invention is to provide a satellite attitude adaptive control algorithm and system based on DSAC-T to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A satellite attitude adaptive control algorithm based on DSAC-T includes:

[0007] The parameters and environment are initialized, and the desired control quantity is calculated based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle.

[0008] The system responds to and executes the desired control quantity, and calculates a reward value based on the current state, the reward value being used to characterize the satellite's approach to the desired target;

[0009] The intermediate data is packaged and cached in the revisit cache, and the storage of the revisit cache is monitored. If the amount of data reaches a preset value, a training identifier is generated.

[0010] In response to the training identifier, the system retrieves batch data from the cache for training to update its own parameters and evaluates the results. The training and evaluation process is repeated for a predetermined number of training steps to confirm the optimal model.

[0011] As a further aspect of the present invention: the step of calculating and obtaining the desired control quantity based on the current attitude specifically includes:

[0012] The satellite state is modeled in a simulation environment, and the modeling model of the state is characterized as follows:

[0013]

[0014] Where the error quaternion q e Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω is used to characterize the angular velocity of the satellite, and e is used to characterize the axis of rotation;

[0015] Based on the error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as:

[0016] Wherein, φ is the desired rotation angle.

[0017] As a further embodiment of the present invention: in the step of performing fitting calculation on the desired control quantity to obtain the rotation axis and the corresponding desired rotation angle, the specific fitting calculation method is as follows:

[0018] For the error quaternion q e q e The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e =[0,0,0,1], ω = 0, meaning the fitted satellite reaches its optimal attitude, and the desired rotation angle at this point is calculated.

[0019] As a further aspect of the present invention: the step of responding to and executing the desired control quantity, and calculating a reward value based on the current state specifically includes:

[0020] Receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation;

[0021] When the satellite returns to its new state, a reward value is calculated and given, which is characterized as follows:

[0022]

[0023] in, q s,t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

[0024] As a further aspect of the present invention, the step of confirming the optimal model specifically includes:

[0025] Based on the cumulative sum of the rewards obtained, the indicators are evaluated. If the optimal model and the optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result.

[0026] If the optimal model and optimal result are already stored, the current result will be compared with the optimal result. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

[0027] This invention aims to provide a satellite attitude adaptive control system based on DSAC-T, comprising:

[0028] The initial calculation module is used to initialize the parameters and environment, and calculate the desired control quantity based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle.

[0029] The reward assignment module is used to respond to and execute the desired control quantity, and calculate a reward value based on the current state, the reward value being used to characterize the degree of approach of the satellite to the desired target;

[0030] The data monitoring module is used to package intermediate data and cache it in the revisit cache area, and monitor the storage of the revisit cache area. If the amount of data reaches a preset value, a training identifier is generated.

[0031] The iterative evaluation module is used to respond to the training identifier, retrieve batch data from the cache for training to update its own parameters, and evaluate the results. The training evaluation process is repeated for a predetermined number of training steps to confirm the optimal model.

[0032] As a further aspect of the present invention: the initial calculation module specifically includes:

[0033] A state modeling unit is used to model the satellite state in a simulation environment. The state modeling model is characterized as follows:

[0034]

[0035] Where the error quaternion qe Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω is used to characterize the angular velocity of the satellite, and e is used to characterize the axis of rotation;

[0036] Parameter fitting unit for using error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as:

[0037] Wherein, φ is the desired rotation angle.

[0038] As a further embodiment of the present invention, the parameter fitting unit specifically includes:

[0039] The expected assignment subunit is used to assign values ​​to the error quaternion q. e q e The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e =[0,0,0,1], ω = 0, meaning the fitted satellite reaches its optimal attitude, and the desired rotation angle at this point is calculated.

[0040] As a further embodiment of the present invention: the reward granting module specifically includes:

[0041] The command response unit is used to receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation.

[0042] The reward feedback unit is used to calculate and provide a reward value when the satellite returns to a new state. The reward value is characterized as follows:

[0043]

[0044] in, q s,t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

[0045] As a further embodiment of the present invention: the cyclic evaluation module includes:

[0046] The optimal setting unit is used to evaluate indicators based on the cumulative sum of the obtained rewards. If the optimal model and optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result.

[0047] The optimal decision unit is used to compare the current result with the optimal result if the optimal model and the optimal result are already stored. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

[0048] Compared with existing technologies, the beneficial effects of this invention are: it solves the problem that existing algorithms cause satellite angular velocity to constantly fluctuate between positive and negative values, thus enabling better control of the satellite's stable operation. Furthermore, compared with existing technologies, the DSAC-T algorithm has a faster control speed and faster convergence, improving the real-time performance of satellite attitude control, enhancing anti-interference capabilities, exhibiting good control stability under different interference experimental environments, and demonstrating good robustness to interference not encountered during training. The trained agent also shows good adaptability to satellites with different moments of inertia, making it easy to migrate to new satellites and providing better versatility. Attached Figure Description

[0049] Figure 1 This is a flowchart of a satellite attitude adaptive control algorithm based on DSAC-T.

[0050] Figure 2 This is a flowchart of a satellite attitude adaptive control algorithm based on DSAC-T.

[0051] Figure 3 This is a block diagram of a satellite attitude adaptive control system based on DSAC-T. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0053] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0054] like Figure 1 and Figure 2 The satellite attitude adaptive control algorithm based on DSAC-T, as described in one embodiment of the present invention, includes the following steps:

[0055] S10, initialize the parameters and environment, and calculate the desired control quantity based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle.

[0056] S20, respond to and execute the desired control quantity, and calculate a reward value based on the current state, the reward value being used to characterize the degree of approach of the satellite to the desired target;

[0057] S30: Pack the intermediate data and cache it in the revisit cache area, and monitor the storage of the revisit cache area. If the amount of data reaches the preset value, generate a training identifier.

[0058] S40, in response to the training identifier, retrieve batch data from the back-visit cache for training to update its own parameters, and perform result evaluation, repeating the training evaluation process for a predetermined number of training steps to confirm the optimal model.

[0059] This embodiment presents a satellite attitude adaptive control algorithm based on DSAC-T, which solves the problem of existing algorithms causing the satellite's angular velocity to oscillate between positive and negative values. This allows for better control of the satellite's stable operation. Compared to existing technologies, the DSAC-T algorithm has a faster control speed and faster convergence, improving the real-time performance of satellite attitude control and enhancing its anti-interference capabilities. It exhibits good control stability under different interference experimental environments and demonstrates good robustness to interferences not encountered during training. The trained agent also shows good adaptability to satellites with different moments of inertia, making it easy to migrate to new satellites and providing better versatility. Specifically, the DSAC-T (Distributive Soft Actor-Critic with Three refinements)-based satellite attitude adaptive continuous control algorithm achieves faster and more stable control while possessing a certain degree of resistance to external disturbances. DSAC-T is currently the best-performing deep reinforcement learning algorithm, learning and training based on artificial intelligence and neural network methods. By learning the state-action return distribution, it achieves good control capabilities and also introduces expected value substitution. The three improvements—substituting, twin-value distribution learning, and variance-based critic gradient adjusting—enhanced learning stability, gave it a degree of adaptability, reduced the tedium and uncertainty of manual design, achieved better results, and demonstrated a degree of generalization, making it more resistant to unseen situations during training.

[0060] In another preferred embodiment of the present invention, the step of calculating and obtaining the desired control quantity based on the current attitude specifically includes:

[0061] The satellite state is modeled in a simulation environment, and the modeling model of the state is characterized as follows:

[0062]

[0063] Where the error quaternion q e Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω is used to characterize the angular velocity of the satellite, and e is used to characterize the axis of rotation;

[0064] Based on the error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as:

[0065] Wherein, φ is the desired rotation angle.

[0066] Furthermore, in the step of fitting and calculating the desired control quantity to obtain the rotation axis and the corresponding desired rotation angle, the specific fitting and calculation method is as follows:

[0067] For the error quaternion q e q e The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e =[0,0,0,1], ω = 0, meaning the fitted satellite reaches its optimal attitude, and the desired rotation angle at this point is calculated.

[0068] In this embodiment, at the beginning, the parameters required by the algorithm and the simulation environment are initialized first, and then the training loop iteration process begins; in the simulation environment, the state is modeled as Where the error quaternion q e Used to represent the satellite's current attitude, e represents the rotation axis, and φ represents the desired rotation angle. q e The derivative with respect to time t, ω represents the satellite's angular velocity; therefore, here we set it to be ω when q e = [0, 0, 0, 1], When ω = 0, the current state is considered to have been reached. In each iteration, the environment is first reset, and the satellite's rotation speed is reset to 0. Then, a desired rotation angle φ ∈ [30°, 150°] and rotation axis e are randomly generated as the target for the agent to control the satellite's rotation in this round. The environment is set with two termination conditions: when the maximum allowed number of steps is reached (here, 500 steps are set during training) or when the maximum allowed angular velocity (0.25 rad / s) is reached, the simulation will terminate and the next round of simulation will begin. In each step, the agent uses the satellite's current state as the input to the neural network, calculates an output through the network, and applies this output as a control torque to the satellite to control its rotation.

[0069] In another preferred embodiment of the present invention, the step of responding to and executing the desired control quantity, and calculating a reward value based on the current state specifically includes:

[0070] Receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation;

[0071] When the satellite returns to its new state, a reward value is calculated and given, which is characterized as follows:

[0072]

[0073] in, q s，t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

[0074] In this embodiment, after receiving the control command from the agent, the satellite begins to rotate. Since the control torque is generated at intervals, the rotation will continue for a period of time before the next control torque is generated. Then, the satellite returns to its new state and provides a reward value. The reward is calculated according to the following formula, where, q s，t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s Value, if the current q s A smaller value indicates that the spacecraft is closer to the desired goal than in the previous step, so it is given a positive reward; conversely, a larger value is given a negative reward. This encourages the agent to continuously control the spacecraft to rotate in the desired direction and prevents the agent from regressing. m represents the maximum number of steps, n represents the current number of steps, and in each step, if q... s,tFor a value ≥0.99999762 (i.e., the error angle φ ≤ 0.25°), an additional reward of 9 will be given. When the simulation ends, if the spacecraft can still maintain this state, a reward of 50 will be given directly. However, if the spacecraft's angular velocity exceeds the maximum allowable angular velocity (0.5 rad / s) at the end, a penalty of -25 will be imposed. By setting the rewards, the agent can be guided to learn how to continuously maintain the spacecraft's attitude within the expected range, thus preventing the agent from falling into a bad environment.

[0075] In another preferred embodiment of the present invention, the step of confirming the optimal model specifically includes:

[0076] Based on the cumulative sum of the rewards obtained, the indicators are evaluated. If the optimal model and the optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result.

[0077] If the optimal model and optimal result are already stored, the current result will be compared with the optimal result. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

[0078] In this embodiment, as in the previous embodiment, the data generated will be packaged into a tuple and stored in the replay buffer. If the replay buffer contains enough data, the agent will extract a batch of data from the buffer for training and update its own parameters. The algorithm will save the current model every certain number of steps.

[0079] After several training iterations, the algorithm evaluates the current model several times. During evaluation, a new environment is created, and the latest saved model parameters are used to initialize a new model, instead of using the training environment and model. The reward obtained by the agent (i.e., the cumulative sum of rewards) is used as the evaluation metric. If the optimal model and optimal result have not yet been saved, the current model is used as the optimal model, and the current result is used as the optimal result. If the optimal model and optimal result already exist, the current result is compared with the optimal result. If the current evaluation result is better than the existing optimal evaluation result, the currently used model is saved as the optimal model, and the current result is saved as the optimal result. By adding an evaluation step during training, the algorithm helps to better judge the training effect and save the model parameters in a timely manner. This ensures that even if unexpected performance degradation occurs during subsequent training, the algorithm still maintains the model parameters with the best performance. This training-evaluation loop continues until a specified number of training steps are reached. Because the optimal model is saved during the evaluation phase, it is very convenient to apply the model to downstream tasks and subsequent applications.

[0080] like Figure 3 As shown, the present invention also provides a satellite attitude adaptive control system based on DSAC-T, which includes:

[0081] The initial calculation module 100 is used to initialize the parameters and environment, and calculate the desired control quantity based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle.

[0082] The reward assignment module 200 is used to respond to and execute the desired control quantity, and calculate a reward value based on the current state, the reward value being used to characterize the degree of approach of the satellite to the desired target;

[0083] The data monitoring module 300 is used to package intermediate data and cache it in the revisit cache area, and monitor the storage of the revisit cache area. If the amount of data reaches a preset value, a training identifier is generated.

[0084] The iterative evaluation module 400 is used to respond to the training identifier, obtain batch data from the revisit cache for training to update its own parameters, and perform result evaluation. The training evaluation process is repeated for a predetermined number of training steps to confirm the optimal model.

[0085] In another preferred embodiment of the present invention, the initial calculation module 100 specifically includes:

[0086] A state modeling unit is used to model the satellite state in a simulation environment. The state modeling model is characterized as follows:

[0087]

[0088] Where the error quaternion q e Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω is used to characterize the angular velocity of the satellite, and e is used to characterize the axis of rotation;

[0089] Parameter fitting unit for using error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as:

[0090] Wherein, φ is the desired rotation angle.

[0091] In another preferred embodiment of the present invention, the parameter fitting unit further includes:

[0092] The expected assignment subunit is used to assign values ​​to the error quaternion q. e qe The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e = [0, 0, 0, 1], ω = 0, meaning the fitted satellite reaches its optimal attitude, and the desired rotation angle at this point is calculated.

[0093] In another preferred embodiment of the present invention, the reward-granting module specifically includes:

[0094] The command response unit is used to receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation.

[0095] The reward feedback unit is used to calculate and provide a reward value when the satellite returns to a new state. The reward value is characterized as follows:

[0096] in, q s,t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

[0097] In another preferred embodiment of the present invention, the cyclic evaluation module includes:

[0098] The optimal setting unit is used to evaluate indicators based on the cumulative sum of the obtained rewards. If the optimal model and optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result.

[0099] The optimal decision unit is used to compare the current result with the optimal result if the optimal model and the optimal result are already stored. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

[0100] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0101] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0102] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A satellite attitude adaptive control method based on DSAC-T, characterized in that, Include: The parameters and environment are initialized, and the desired control quantity is calculated based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle. The system responds to and executes the desired control quantity, and calculates a reward value based on the current state, the reward value being used to characterize the satellite's approach to the desired target; The intermediate data is packaged and cached in the revisit cache, and the storage of the revisit cache is monitored. If the amount of data reaches a preset value, a training identifier is generated. In response to the training identifier, the system retrieves batch data from the cache for training to update its own parameters and evaluates the results. The training and evaluation process is repeated for a predetermined number of training steps to confirm the optimal model. The step of calculating and obtaining the desired control quantity based on the current attitude specifically includes: The satellite state is modeled in a simulation environment, and the modeling model of the state is characterized as follows: Where the error quaternion q e Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω, is used to characterize the angular velocity of a satellite; Based on the error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as: Wherein, φ is the desired rotation angle; In the step of fitting and calculating the desired control quantity to obtain the rotation axis and the corresponding desired rotation angle, the specific fitting and calculation method is as follows: For the error quaternion q e q e The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e =[0,0,0,1], ω = 0, meaning the fitted satellite has reached its optimal attitude, and the desired rotation angle at this point is calculated. The steps of responding to and executing the desired control quantity, and calculating a reward value based on the current state, specifically include: Receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation; When the satellite returns to its new state, a reward value is calculated and given, which is characterized as follows: Wherein, 'e' is used to characterize the rotation axis. q s,t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

2. The satellite attitude adaptive control method based on DSAC-T according to claim 1, characterized in that, The steps for confirming the optimal model specifically include: Based on the cumulative sum of the rewards obtained, the indicators are evaluated. If the optimal model and the optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result. If the optimal model and optimal result are already stored, the current result will be compared with the optimal result. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

3. A satellite attitude adaptive control system based on DSAC-T, characterized in that, Include: The initial calculation module is used to initialize the parameters and environment, and calculate the desired control quantity based on the current attitude. The desired control quantity includes the rotation axis and the corresponding desired rotation angle. The reward assignment module is used to respond to and execute the desired control quantity, and calculate a reward value based on the current state, the reward value being used to characterize the degree of approach of the satellite to the desired target; The data monitoring module is used to package intermediate data and cache it in the revisit cache area, and monitor the storage of the revisit cache area. If the amount of data reaches a preset value, a training identifier is generated. The iterative evaluation module is used to respond to the training identifier, retrieve batch data from the cache for training to update its own parameters, and evaluate the results. The training evaluation process is repeated for a predetermined number of training steps to confirm the optimal model. The initial calculation module specifically includes: A state modeling unit is used to model the satellite state in a simulation environment. The state modeling model is characterized as follows: Where the error quaternion q e Used to characterize the satellite's current attitude. Used to characterize q e The derivative with respect to time, ω is used to characterize the angular velocity of the satellite, and e is used to characterize the axis of rotation; Parameter fitting unit for using error quaternion q e The desired control quantity is fitted and calculated to obtain the rotation axis and the corresponding desired rotation angle, wherein the error quaternion q e Characterized as: Wherein, φ is the desired rotation angle; The parameter fitting unit specifically includes: The expected assignment subunit is used to assign values ​​to the error quaternion q. e q e The derivative with respect to time t And assign a value to the satellite's angular velocity ω, so that the q e =[0,0,0,1], ω = 0, meaning the fitted satellite has reached its optimal attitude, and the desired rotation angle at this point is calculated. The reward assignment module specifically includes: The command response unit is used to receive a control command containing the desired control quantity, execute the control command, and generate a corresponding control torque at a preset time interval to control the satellite rotation. The reward feedback unit is used to calculate and provide a reward value when the satellite returns to a new state. The reward value is characterized as follows: Wherein, 'e' is used to characterize the rotation axis. q s,t Indicates the current q s value, and q s,t-1 Indicates the q of the previous step s The values ​​are m, which represents the maximum number of steps, and n, which represents the current number of steps.

4. A satellite attitude adaptive control system based on DSAC-T according to claim 3, characterized in that, The cyclic evaluation module includes: The optimal setting unit is used to evaluate indicators based on the cumulative sum of the obtained rewards. If the optimal model and optimal result are not currently stored, the current model is set as the optimal model and the current result is taken as the optimal result. The optimal decision unit is used to compare the current result with the optimal result if the optimal model and the optimal result are already stored. If the current evaluation result is better than the existing optimal evaluation result, the currently used model will be saved as the optimal model and the current result will be saved as the optimal result.

Images