Spaceflight motor whole machine on-orbit self-adapting minimum starting torque measurement method and device

Abstract

The application is suitable for the technical field of space scanning stepping motor control, and provides a space motor whole machine on-orbit adaptive minimum starting torque measurement method and device, a storage medium and a processor. The measurement method is based on a Q algorithm to optimize the on-orbit adaptive minimum starting torque measurement of the space motor whole machine, and solves the adaptation problem of the traditional Q algorithm in the motor parameter measurement scene through the collaborative design of the discretization parameters and the multi-dimensional reward / punishment items. The application can quickly and real-timely measure the minimum starting current meeting the error requirement. In the normal operation state of the motor, the minimum starting current of the whole machine can be determined by the estimated current at the minimum error, and then the minimum starting torque of the whole machine is obtained through calculation. Compared with the estimated value of the prior art, the application is more accurate and applicable to any environment and scene in which the motor can operate. The device adopting the method, the storage medium storing the method and the processor running the method also have the same technical effects.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of space scanning stepper motor control, and particularly relates to a space motor on-orbit adaptive minimum starting torque measurement method and device, a storage medium and a processor. BACKGROUND

[0002] The selection of the driving torque of the current space scanning stepper motor is usually based on the minimum starting torque plus the torque margin, and the existing minimum starting torque measurement can only be estimated according to the system design parameters through calculation, or roughly tested on a customized tooling using a counterweight, and it is difficult to accurately measure the actual minimum starting torque of the whole machine under normal load.

[0003] Especially in harsh conditions such as on-orbit vacuum and low temperature, the minimum starting torque of the scanning stepper motor will change, and it is difficult to simulate and measure on the ground.

[0004] The Q-learning algorithm (hereinafter referred to as Q algorithm) is a model-free reinforcement learning algorithm that learns the optimal policy by iteratively updating the Q function (state-action value function), and is suitable for decision-making problems with discrete states and actions. However, the general Q-learning algorithm has the characteristic of "single target orientation", such as minimizing error for maximum revenue. When applied to calculating motor parameters, it is prone to over-pursuit of error minimization, leading to uncontrolled increase of current; or frequent fluctuations in current, affecting motor stability and life; and inability to identify the "error-current" critical point. Therefore, the prior art has deficiencies. SUMMARY

[0005] The purpose of the present application is to provide a space motor on-orbit adaptive minimum starting torque measurement method and device, which solves the problems of inaccurate measurement of traditional motor minimum starting torque and inability to measure in extreme difficult conditions through the cooperative design of discretization and multi-dimensional reward / punishment items. In the normal state of the motor, the measurement can be implemented. It is suitable for various harsh extreme experimental conditions and has important significance for the system design of scanning motors.

[0006] In one aspect, the present application provides a space motor on-orbit adaptive minimum starting torque measurement method, which comprises the following steps:

[0007] s1. Initialize the Q table from the discretized motor operating state parameters and the discretized current adjustment action parameters, set the corresponding learning rate, discount factor, and initialize the moving average filter;

[0008] s2. The actual rotation angle of the motor is collected in real time by the encoder, the deviation from the target angle is calculated, and the deviation state and the corresponding motor current state are mapped into the discretized motor operation state parameters; the current adjustment amplitude of the current state is mapped into the discretized current adjustment action parameters by changing the PWM value to control the adjustment of the current size to change the motor output torque.

[0009] s3. Design a reward function based on the association between the deviation state and the corresponding current state, and the current adjustment amplitude taken to avoid the Q algorithm from falling into a single target guide;

[0010] s4. Select actions to perform using an epsilon-greedy strategy and converge the Q table;

[0011] s5. Calculate the new state and the reward, and update the Q value;

[0012] s6. If the convergence condition is met, output the minimum starting current, otherwise repeat steps s2-s5;

[0013] s7. Based on the linear relationship between current and torque, calculate the minimum starting torque according to the minimum starting current.

[0014] On the other hand, the present application also provides a spacecraft motor whole machine on-orbit adaptive minimum starting torque measuring device, which adopts the above measuring method; the device comprises a main control CPU, an interface expansion logic FPGA, a scanning motor and an encoder connected in sequence; the encoder collects the angle of the scanning motor and feeds back to the main control CPU;

[0015] The main control CPU iteratively calculates and updates the motor control current parameters by determining the mechanism rotation angle information fed back by the encoder, and updates the control current by controlling the PWM duty cycle through the interface expansion logic FPGA; finally determines the minimum starting current of the whole machine, and then calculates the minimum starting torque of the whole machine.

[0016] On the other hand, the present application also provides a storage medium, which stores a program file capable of realizing the above-mentioned spacecraft motor whole machine on-orbit adaptive minimum starting torque measuring method.

[0017] On the other hand, the present application also provides a processor for running a program, wherein the program performs the above-mentioned spacecraft motor whole machine on-orbit adaptive minimum starting torque measuring method when running.

[0018] In the measurement method of the present application, the overall minimum starting torque of the motor based on the Q algorithm is discretized by parameters, and is designed in cooperation with the multi-dimensional reward / punishment item, solving the difficulty of adapting the traditional Q algorithm to the motor parameter measurement scene. Based on the Q algorithm, the minimum starting current meeting the error requirement can be measured. In the normal state of the motor, the measurement can be implemented.

[0019] The measurement device of the present application updates the PWM parameters of the interface expansion logic FPGA by running the measurement method, and the interface expansion logic FPGA controls the driving chip to send control signals to the scanning motor to control the scanning motor to rotate an angle. The encoder collects the angle of the scanning motor and feeds back to the host CPU. The host CPU determines whether the difference between the scanning motor rotation angle and the step angle meets the error standard. The estimated current at the minimum error is used to determine the overall minimum starting current, and the overall minimum starting torque is calculated. Compared with the estimated value of the prior art, the value is more accurate and applicable in any environment and scene where the motor can operate. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is the flow chart of the on-orbit adaptive minimum starting torque measurement method of the space motor provided by the first embodiment of the present application;

[0021] Figure 2 is the module schematic diagram of the on-orbit adaptive minimum starting torque measurement device of the space motor provided by the second embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0023] The specific implementation of the present application is described in detail below in combination with specific embodiments:

[0024] Embodiment one:

[0025] Figure 1 The implementation process of the minimum starting torque measurement method of the scanning stepper motor provided by the first embodiment of the present application is shown. For the convenience of description, only the part related to the embodiment of the present application is shown, and the details are as follows:

[0026] In one aspect, the application provides a method for measuring the minimum starting torque of a space motor in orbit, comprising the following steps:

[0027] s1. initializing a Q table from discretized motor operating state parameters and discretized current adjustment action parameters, setting corresponding learning rate and discount factor, and initializing a moving average filter;

[0028] s2. collecting the actual rotation angle of the motor in real time through an encoder, calculating the deviation from the target angle, and then mapping the deviation state and the corresponding motor current state into the discretized motor operating state parameters; changing the PWM value to control the adjustment current size to change the motor output torque, and mapping the current adjustment amplitude generating the current state into the discretized current adjustment action parameters;

[0029] s3. designing a reward function based on the association between the deviation state and the corresponding current state, and the current adjustment amplitude taken, to avoid the Q algorithm from falling into a single target guide;

[0030] s4. selecting actions for execution using an epsilon-greedy strategy, and converging the Q table;

[0031] s5. calculating the new state and the reward, and updating the Q value;

[0032] s6. if the convergence condition is met, outputting the minimum starting current, otherwise repeating steps s2-s5;

[0033] s7. calculating the minimum starting torque based on the minimum starting current and the linear relationship between current and torque.

[0034] Generally, the formula for calculating the minimum torque of the moving part is:

[0035] ;

[0036] In the formula:

[0037] - the minimum driving torque of the moving part, in units of Newton-meters (N·m); );

[0038] - the moment of inertia of the moving part, in units of Newton-meters (N·m); );

[0039] - the maximum resistance torque of the moving part, in units of Newton-meters (N·m).

[0040] 1. The formula for the load inertia resistance torque is:

[0041] ​ ;

[0042] wherein:

[0043] total moment of inertia of the rotating mechanism;

[0044] maximum angular acceleration;

[0045] The friction resistance moment of the bearing is ;

[0046] The rotating resistance moment of the wire harness is ;

[0047] Therefore: ;

[0048] In actual measurement, the friction resistance moment of the bearing and the rotating resistance moment of the wire harness are difficult to accurately measure, especially in extreme harsh conditions, these resistance moments fluctuate greatly, according to the maximum output torque of the motor given by the motor manufacturer , we believe that the output torque is linearly related to the current, that is:

[0049] .

[0050] Therefore, the minimum output torque of the space scanning stepping motor can be indirectly measured based on the minimum starting current.

[0051] In the present application, based on the Q-learning algorithm of the table, the motor operating state parameters (deviation of actual rotation angle and target angle, current size) and the action space parameters (control PWM change) are discretized, divided into a limited number of state intervals, and a complete state space is constructed.

[0052] Further, in the step s1, the state space of the Q table is defined by the discretized motor operating state parameters, constituting the rows of the Q table;

[0053] The action space of the Q table is defined by the discretized current adjustment action parameters, constituting the columns of the Q table.

[0054] Further, the number of rows of the Q table is the product of the number of deviation states and the number of current states;

[0055] The number of columns of the Q table is the corresponding number of current adjustment amplitudes;

[0056] The dimension of the Q table is the product of the number of rows and the number of columns.

[0057] Specifically, the rows (states) and columns (actions) of the Q table are completely defined by discretized parameters, and the granularity of the parameters directly determines the size of the Q table. The state space is "5x5" (a total of 25 states), and the action space is 5 discrete actions, so the Q table has a dimension of 25x5 (25 rows of states and 5 columns of actions), and the initial Q value is set to 0. By dividing the angle deviation, current and other continuous quantities into discrete intervals, and quantizing the current adjustment action to a fixed amplitude, the algorithm can learn the mapping relationship of "state - action - reward" in a limited Q table. The discretization setting helps to balance accuracy and computational load, and supports the closed-loop control of "continuous input → discrete decision → continuous output" throughout the algorithm initialization, state perception, action selection, reward calculation, Q value update and convergence judgment steps.

[0058] Further, in the process of designing the reward function in step s3, the deviation state is divided into the following 5 intervals:

[0059] 1. When the angle error percentage of the actual rotation angle and the target angle > 20%, it is determined that the actual rotation angle is much larger than the target angle.

[0060] 2. When the angle error percentage of the actual rotation angle and the target angle > 10% and ≤ 20%, it is determined that the actual rotation angle is slightly larger than the target angle.

[0061] 3. When the angle error percentage of the actual rotation angle and the target angle ≥ -10% and ≤ 10%, it is determined that the actual rotation angle is close to the target angle.

[0062] 4. When the angle error percentage of the actual rotation angle and the target angle < -10% and ≥ -20%, it is determined that the actual rotation angle is slightly smaller than the target angle.

[0063] 5. When the angle error percentage of the actual rotation angle and the target angle < -20%, it is determined that the actual rotation angle is much smaller than the target angle.

[0064] The current state is divided into the following 5 intervals:

[0065] 1. When the measured current < 20% of the rated current, it is determined to be the extremely low current interval.

[0066] 2. When the measured current ≥ 20% and < 25% of the rated current, it is determined to be the low current interval.

[0067] 3. When the measured current ≥ 25% and ≤ 30% of the rated current, it is determined to be the medium current interval.

[0068] 4. When the measured current > 30% and ≤ 35% of the rated current, it is determined to be the high current interval.

[0069] 5. When the measured current is greater than 35% of the rated current, it is determined to be in the extremely high current range;

[0070] The current adjustment range is divided into the following 5 intervals:

[0071] 1. If the current increases by 1% of the PWM value based on the current, it is considered a slight increase in current;

[0072] 2. A 1% reduction in the PWM value from the current level is considered a minor reduction in current.

[0073] 3. A PWM value that increases the current by 10% from the current level is considered a significant increase in current.

[0074] 4. A PWM value that reduces the current by 10% from the current level is considered a significant reduction in current.

[0075] 5. If the current PWM value remains unchanged, it is determined to be a sustaining current.

[0076] Furthermore, in step s3, during the design of the reward function, the reward function is constructed using a combination of a basic reward and a penalty term; the reward function is:

[0077] r t = Basic reward - Angle error penalty - Current penalty - Fluctuation penalty + Critical point exploration reward;

[0078] The basic reward is used to encourage the angle error to approach the target, and is generally limited to within ±2°. The angle error penalty is designed based on the idea that the larger the error, the heavier the penalty, in order to suppress large errors. The current penalty is also designed based on the idea that the higher the current, the heavier the penalty, in order to encourage the realization of a low current state. The fluctuation penalty adopts the design idea that the more drastic the current change, the heavier the penalty, in order to suppress frequent current adjustments. The critical point exploration reward is used to provide additional rewards when the current decreases and the error does not increase significantly, in order to guide the search for a balance point of low current and small error.

[0079] In practice, the following scheme is used to implement the reward function:

[0080] 1. Basic Rewards

[0081] When the motor is in the state where "the actual rotation angle is close to the target angle", a positive reward of +10 is given;

[0082] When the motor is in a state where the actual rotation angle is slightly greater than the target angle or the actual rotation angle is slightly less than the target angle, a positive reward of +5 is given.

[0083] When the motor is in the state of "actual rotation angle is much larger than target angle" or "actual rotation angle is much smaller than target angle", it indicates that the current torque is far from enough to make the motor rotate to the target angle, and the current needs to be increased significantly, giving a basic negative reward of -5.

[0084] 2. Angle error penalty

[0085] Angle error penalty = absolute value of error percentage * penalty coefficient

[0086] The penalty coefficient can be set according to specific conditions. In the present application, it is set to 0.5.

[0087] 3. Current penalty

[0088] The higher the current level, the greater the penalty (encouraging to reduce current):

[0089] Very low current: 0;

[0090] Low current: 0.1 * PWM duty cycle;

[0091] Medium current: 0.2 * PWM duty cycle;

[0092] High current: 0.5 * PWM duty cycle;

[0093] Very high current: 1.0 * PWM duty cycle;

[0094] 4. Fluctuation penalty

[0095] Calculate the absolute difference between the current current and the last period current, i.e.:

[0096] Fluctuation penalty = (current period PWM duty cycle - last period PWM duty cycle) * fluctuation penalty coefficient

[0097] In the present application, the fluctuation penalty coefficient is set to 0.3.

[0098] 5. Critical point exploration reward

[0099] When the current is reduced and the error does not increase significantly, give an additional reward:

[0100] If the current is less than the last period current, and the angle error change ≤5%: reward +3;

[0101] If the current is less than the last period current, and the angle error change ≤10%: reward +1;

[0102] When the current increases but the error decreases significantly, give a moderate reward:

[0103] If the current is greater than the last period current, and the angle error decreases ≥15%: reward +2;

[0104] Furthermore, in step s5, the Q value represents the combined benefit of the minimum starting current and the minimum angle error that can be obtained after taking a certain action in a certain state; the calculation formula for updating the Q value is as follows:

[0105] ;

[0106] in: t For the current period; This is the current state; The action to be taken at present; The learning rate; The immediate reward calculated based on the reward function; The reward after an action is performed k cycles is used to assess the long-term effects of the action more accurately, thus avoiding policy mislearning caused by time delays or single errors. The time delay compensation factor is set to 0.8 in this invention; The discount factor represents the degree of influence of future decisions on current decisions; in this invention, it is set to 0.95. To perform the action The new state afterwards; For the new state The maximum value of Q among all possible actions.

[0107] Specifically, in the Q-value update mechanism, the discretization parameter defines "what the algorithm needs to learn" (the fineness of the state, the granularity of the action, and the boundaries of the physical constraints); the Q-value parameter defines "how to learn" (learning speed, the degree of attention to future impact, and the intensity of exploration). The state discretization granularity (number of intervals) corresponds to the learning rate α; the larger the state space, the smaller α. The action adjustment amplitude (step size) corresponds to the discount factor γ; the larger the action adjustment amplitude, the larger the corresponding γ. The number of actions corresponds to the exploration rate ε; the more actions, the higher the initial ε, and the slower the decay rate.

[0108] Meanwhile, this calculation process incorporates a time delay compensation factor. This helps to more accurately assess the long-term effects of actions and avoid mislearning of strategies due to time delays or single errors.

[0109] Furthermore, in step s6, the exploration rate is set. When the test starts The value is initially set to 0.5, then decays to 0.1, using a greedy strategy. The probability of randomly selecting an action, in order to The probability of choosing the current state The action with the largest Q value; gradually reduce the exploration rate according to the state and action feedback , until the angle error percentage and current change indicators meet the standards, the Q value table converges, and the motor can operate stably at the minimum current.

[0110] Specifically, during the operation of step s6, when the algorithm converges (i.e., the Q value changes by less than 0.01 for 10 consecutive periods), the "critical point fine-tuning" mode is started: that is, the "slightly reduce current" action is forced to be executed with a probability of 5%. If the error does not significantly increase after executing this action for 5 consecutive times, the basic current level is permanently reduced. If the absolute value of the angle error percentage increases again, stop exploring, calculate the minimum starting current based on the PWM duty cycle value before the absolute value of the angle error percentage increased, and then determine the minimum starting torque.

[0111] During the convergence process, the essence of Q value is to quantify the "minimum starting current and minimum angle error that can be obtained after taking a certain action in a certain state", and its core role is reflected in three aspects, directly supporting the exploration and determination of the minimum starting current:

[0112] 1. As a decision criterion for "current-error" trade-off;

[0113] 2. Accumulate learning of the mapping relationship of "starting critical point";

[0114] 3. Suppress "over-adjustment" and "conservative strategy" during the starting process.

[0115] The above measurement method of the present application also has the following technical effects:

[0116] 1. It is suitable for accurately measuring the overall load torque of the moving mechanism in extreme harsh conditions such as vacuum and low temperature. Under the premise of ensuring the normal operation of the moving mechanism in extreme harsh conditions, the minimum starting torque and current of the motor are accurately measured, the power consumption and heat consumption are reduced, resource waste is avoided, and the safety and stability of the motor operation are improved.

[0117] 2. Without estimating the initial test current, the current increase and decrease strategy can be dynamically adjusted to achieve rapid convergence of the algorithm. Through continuous interaction with the motor environment, the Q value table is optimized autonomously, and the optimal strategy is learned directly from the state-action feedback without prior knowledge of the physical model or load characteristics of the motor. Even if the motor parameters drift slightly over time, the algorithm can update the strategy through continuous iteration to maintain the optimal control effect.

[0118] 3. The Q value table is continuously updated through iterative optimization, and as the number of learning increases, the algorithm's optimal action judgment for various states will become more and more accurate. Even if a single test angle value is abnormal, it can quickly respond based on historical learning results to avoid over-adjustment.

[0119] 4. Strong anti-interference ability, set fluctuation penalty, reduce mechanical vibration in the determination process.

[0120] 5. Critical point exploration fine adjustment mode, actively explore the limit of current reduction, find the optimal current value that minimizes the operating angle error.

[0121] Example two:

[0122] On the other hand, as shown in the accompanying Figure 2 The application also provides a whole minimum starting torque measurement device for scanning stepper motor, which adopts the above method; the device comprises a main control CPU, an interface expansion logic FPGA, a scanning motor and an encoder which are electrically connected in sequence; the encoder collects the angle of rotation of the scanning motor and feeds back to the main control CPU;

[0123] The main control CPU iteratively calculates and updates the motor control current parameters by determining the motor rotation angle information fed back by the encoder, and updates the control current by controlling the PWM duty cycle through the interface expansion logic FPGA; finally determines the minimum starting current of the whole machine, and then calculates the minimum starting torque of the whole machine.

[0124] Based on the above measurement device, taking a two-phase hybrid stepper motor as an example, a two-phase double four-tap control method is adopted, the motor rotation angle is observed by the encoder, and the stepper motor accurately walks one step angle as the standard, the minimum starting torque of the whole machine of the scanning stepper motor in the whole machine state is accurately measured by iteratively calculating and updating the motor control current, and the measurement accuracy of the minimum starting torque of the whole machine is effectively improved.

[0125] Example three:

[0126] On the other hand, the application also provides a storage medium, which stores a program file capable of realizing the above-mentioned on-orbit adaptive minimum starting torque measurement method of the whole machine of the aerospace motor.

[0127] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by programs instructing related hardware, and the programs can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc.

[0128] Example four:

[0129] On the other hand, the application also provides a processor for running a program, wherein the program executes the above-mentioned on-orbit adaptive minimum starting torque measurement method of the whole machine of the aerospace motor.

[0130] In the embodiments of the present application, the on-orbit self-adaptive minimum starting torque measurement method of the aerospace motor can be implemented by corresponding hardware or software units. Each unit can be an independent software or hardware unit, or can be integrated into a software or hardware unit, which does not limit the present application. The specific implementation of each unit can refer to the description of Embodiment 1, which will not be repeated here.

[0131] The above merely describes the preferred embodiments of the present application and is not used to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for measuring the adaptive minimum starting torque of an aerospace motor in orbit, characterized in that, The measurement method includes the following steps: s1. Initialize the Q-table with discretized motor operating state parameters and discretized current adjustment action parameters, set the corresponding learning rate and discount factor, and initialize the moving average filter; s2. The actual rotation angle of the motor is collected in real time by the encoder. After calculating the deviation from the target angle, the deviation state and the corresponding motor current state are mapped to the discretized motor operating state parameters. The output torque of the motor is changed by controlling the adjustment current by changing the PWM value. The current adjustment amplitude that generates the current state is mapped to the discretized current adjustment action parameters. s3. Design a reward function based on the correlation between the deviation state and the corresponding current state, as well as the magnitude of the current adjustment taken, to avoid the Q algorithm getting stuck in a single goal. s4. Use an ε-greedy strategy to select actions to execute, and converge the Q-table; s5. Calculate the new state and reward, and update the Q value; s6. If the convergence condition is met, output the minimum starting current; otherwise, repeat steps s2-s5. s7. Based on the linear relationship between current and torque, calculate the minimum starting torque according to the minimum starting current.

2. The measurement method as described in claim 1, characterized in that, In step s1, the state space of the Q table is defined by discretized motor operating state parameters, which constitute the rows of the Q table; The action space of the Q-table is defined by discrete current adjustment action parameters, which constitute the columns of the Q-table.

3. The measurement method as described in claim 2, characterized in that, The The number of rows in the Q table is the product of the number of deviation states and the number of current states; The number of columns in the Q table corresponds to the number of times the current adjustment range is calculated. The dimension of the Q table is the product of the number of rows and the number of columns.

4. The measurement method as described in claim 3, characterized in that, In the process of designing the reward function in step s3, the deviation state is divided into the following 5 intervals: (1) When the percentage of the angular error between the actual rotation angle and the target angle is greater than 20%, it is determined that the actual rotation angle is much larger than the target angle; (2) When the percentage of the angular error between the actual rotation angle and the target angle is >10% and ≤20%, it is determined that the actual rotation angle is slightly larger than the target angle: (3) When the percentage of the angular error between the actual rotation angle and the target angle is ≥-10% and ≤10%, it is determined that the actual rotation angle is close to the target angle; (4) When the percentage of the angular error between the actual rotation angle and the target angle is less than -10% and greater than or equal to -20%, it is determined that the actual rotation angle is slightly smaller than the target angle. (5) When the percentage of the angular error between the actual rotation angle and the target angle is less than -20%, it is determined that the actual rotation angle is much smaller than the target angle; The current state is divided into the following 5 intervals: (1) If the measured current is less than 20% of the rated current, it is determined to be in the extremely low current range; (2) When the measured current is ≥20% and <25% of the rated current, it is determined to be in the low current range; (3) When the measured current is ≥25% and ≤30% of the rated current, it is judged as the medium current range; (4) When the measured current is greater than 30% and less than or equal to 35% of the rated current, it is determined to be in the high current range; (5) When the measured current is greater than 35% of the rated current, it is determined to be in the extremely high current range; The current adjustment range is divided into the following 5 intervals: (1) If the current increases by 1% of the PWM value based on the current, it is determined to be a slight increase in current; (2) A PWM value that reduces the current by 1% from the current value is considered a slight reduction in current; (3) If the current increases by 10% of the PWM value based on the current, it is judged as a significant increase in current; (4) A PWM value that reduces the current by 10% from the current value is considered a significant reduction in current; (5) Keep the current PWM value unchanged and determine it as a sustaining current.

5. The measurement method as described in claim 4, characterized in that, In step s3, during the design of the reward function, the reward function is constructed using a combination of basic reward and penalty term; the reward function is: r t = Basic reward - Angle error penalty - Current penalty - Fluctuation penalty + Critical point exploration reward; The basic reward is used to encourage the angle error to approach the target; the angle error penalty is used to suppress large errors; the current penalty is used to encourage the achievement of a low current state; the fluctuation penalty is used to suppress frequent adjustments; and the critical point exploration reward is used to provide additional rewards when the current decreases and the error does not increase significantly, in order to guide the search for a balance point of low current and small error.

6. The measurement method as described in claim 1, characterized in that, In step s5, the Q value represents the combined benefit of the minimum starting current and the minimum angle error that can be obtained after taking a certain action in a certain state; the calculation formula for updating the Q value is as follows: ； in: This is the current state; The action to be taken at present; The learning rate; The immediate reward calculated based on the reward function; The reward after an action is performed for k cycles is used to assess the long-term effects of the action more accurately, thus avoiding policy mislearning caused by time delays or single errors. The time delay compensation factor is set to 0.8 in this invention; The discount factor represents the degree of influence of future decisions on current decisions; in this invention, it is set to 0.

95. To perform the action The new state afterwards; For the new state The maximum value of Q among all possible actions.

7. The measurement method as described in claim 1, characterized in that, In step s6, the exploration rate is set. When the test starts The value is initially set to 0.5, then decays to 0.1, using a greedy strategy. The probability of randomly selecting an action, in order to The probability of choosing the current state The action with the highest Q value is selected; based on state and action feedback, the exploration rate is gradually reduced. Until the percentage of angular error and the current change index meet the standards, the Q value table converges, and the motor can operate stably with the minimum current.

8. A device for measuring the adaptive minimum starting torque of an aerospace motor in orbit, characterized in that, The device employs the measurement method described in any one of claims 1-7; the device comprises a main control CPU, an interface expansion logic FPGA, a scanning motor, and an encoder connected in sequence; the encoder collects the rotation angle of the scanning motor and feeds it back to the main control CPU. The main control CPU determines the rotation angle information of the mechanism fed back by the encoder, iteratively calculates and updates the motor control current parameters, and updates the control current by controlling the PWM duty cycle through the interface extended logic FPGA; finally, it determines the minimum starting current of the whole machine, and then calculates the minimum starting torque of the whole machine.

9. A storage medium, characterized in that, The storage medium stores program files capable of implementing the on-orbit adaptive minimum starting torque measurement method for aerospace motors as described in any one of claims 1 to 7.

10. A processor, characterized in that, The processor is used to run a program, wherein the program executes the on-orbit adaptive minimum starting torque measurement method for aerospace motors as described in any one of claims 1 to 7.

Images