An active suppression method for on-orbit hand-transmitted vibration based on parameter adaptive learning

Abstract

A method for actively suppressing hand-transmitted vibration in orbit based on parameter adaptive learning. The present invention relates to a method for actively suppressing hand-transmitted vibration in an on-orbit service process. The present invention is to solve the problems that the existing traditional command shaping vibration suppression method has poor robustness and is difficult to overcome the possible interference and uncertainty in the actual on-orbit service process hand-transmitted vibration dynamics model; and when the external interference exceeds a certain range, the output of the on-orbit service process hand-transmitted vibration dynamics model will be in a divergent state and out of control, which will bring great danger. The process is: 1: obtain the state equation of the on-orbit service process hand-transmitted vibration dynamics model; 2: solve the parameters of the command shaper; 3: obtain the hand-transmitted vibration force; 4: integrate the hand-transmitted vibration force to obtain a new 5: based on the calculated state variables, determine the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration command shaping in the on-orbit service process. The present invention belongs to the field of aerospace and automation.

Description

Technical Field

[0001] The invention belongs to the field of aerospace and automation, and in particular relates to a method for actively suppressing hand-transmitted vibration during an on-orbit service process. Background Art

[0002] With the smooth progress of my country's space station project development mission, the demand for space extravehicular activities (EVA) such as on-orbit assembly, maintenance and replacement, inspection and maintenance, extravehicular scientific experiments, and space release and recovery is becoming increasingly urgent. EVA missions are becoming increasingly diverse and uncertain, which puts higher requirements on the safety of astronauts and the reliability of operations. The demand for fine operation capabilities of EVA missions is imminent. During space extravehicular activities, astronauts, spacecraft, extravehicular maintenance auxiliary facilities and tools, and the space environment they are in constitute a complex man-machine-environment system. Among them, extravehicular maintenance auxiliary facilities and tools are an important part of astronauts' extravehicular activities. Their main task is to ensure that astronauts can repair and replace on-orbit replaceable units outside the space station, and ensure the long-term stable operation of the space station in orbit and the smooth completion of various tasks.

[0003] Most of the on-orbit assembly, maintenance and replacement, and precise operation and efficient disassembly and assembly of various screws / bolts outside the cabin by astronauts are carried out by extravehicular power tools. The internal vibration and external transmission of extravehicular power tools are directly related to the success or failure of extravehicular activities in space, and even affect the health and on-orbit safety of astronauts. Hand-transmitted vibration in space refers to the mechanical vibration or impact of the tool body and pressurized space suit gloves transmitted to the fingers, palms, arms and shoulders of astronauts when astronauts use hand-held vibration tools or touch vibrating workpieces in the suit state. Long-term exposure to a strong vibration environment, or contact with the vibration source after hand fatigue, can cause obvious dullness and fatigue in the hands and arms, and even produce hand-transmitted vibration syndrome such as vibration white finger disease. When astronauts are in suits outside the cabin, the characteristics of the space suit may lead to further amplification of the vibration signal, causing great safety hazards. Therefore, the study of the suppression of hand-transmitted vibration in space is an urgent problem to be solved in the execution of space missions. At present, the traditional command shaping vibration suppression method has poor robustness and is difficult to overcome the possible interference and uncertainty in the actual system. In particular, when external interference exceeds a certain range, the system will be in a divergent state and out of control, which will bring great danger. Summary of the invention

[0004] The purpose of the present invention is to solve the problems that the existing traditional command shaping vibration suppression methods have poor robustness and are difficult to overcome possible interference and uncertainty in the hand-transmitted vibration dynamics model of the actual on-orbit service process; and when the external interference exceeds a certain range, the output of the hand-transmitted vibration dynamics model of the on-orbit service process will be in a divergent state and out of control, which will bring great danger. A method for active suppression of on-orbit hand-transmitted vibration based on parameter adaptive learning is proposed.

[0005] A method for active suppression of on-orbit hand-transmitted vibration based on parameter adaptive learning. The specific process is as follows:

[0006] Step 1: Obtain the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process;

[0007] Step 2: Determine the type of instruction shaper ZVDD, and solve the parameters of the instruction shaper ZVDD, the parameters of the instruction shaper ZVDD include an amplitude ratio parameter and a self-oscillation period;

[0008] Step 3: The simplified hand-transmitted vibration force F during the on-orbit service process is regarded as a continuous pulse signal, and each pulse signal of the simplified hand-transmitted vibration force F during the on-orbit service process is shaped to obtain a new simplified hand-transmitted vibration force F′ during the on-orbit service process;

[0009] Step 4: Integrate the new simplified on-orbit service process hand-transmitted vibration force F′ obtained in step 3 to obtain the new shaped signal force

[0010] Step 5: Apply force based on the new shaped signal The state variables of the dynamic model of hand-transmitted vibration in the on-orbit service process are calculated, and the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping in the on-orbit service process is determined based on the state variables.

[0011] The beneficial effects of the present invention are:

[0012] The present invention proposes a method for actively suppressing hand-transmitted vibration during an on-orbit service process based on parameter adaptive learning, designs a shaping parameter adaptive rate, and improves and optimizes the final shaping result according to the actual situation, so that the system has high robustness and does not affect the stability of the closed-loop system while having the ability to overcome the model error and system uncertainty.

[0013] (1) The continuous signal reconstructed by the instruction shaper has the same effect as the original signal, the shaping logic is simple to implement, and the oscillation suppression effect is good.

[0014] (2) Adaptive reinforcement learning of parameters has a stronger tolerance for model errors and system uncertainties, and has extremely high robustness and stability.

[0015] (3) The input value is adjusted by means of energy instruction shaping, which once again ensures the practical operability and flexibility of the scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The flow chart of hand-transmitted vibration suppression for command shaping;

[0017] Figure 2 Flowchart for determining parameters for adaptive reinforcement learning;

[0018] Figure 3 This is the input signal diagram before and after shaping;

[0019] Figure 4 This is a graph showing the input signal changing with time before and after shaping. DETAILED DESCRIPTION

[0020] Specific implementation method 1: The specific process of this implementation method of an on-orbit hand-transmitted vibration active suppression method based on parameter adaptive learning is as follows:

[0021] Step 1: Obtain the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process;

[0022] Step 2: Determine the type of instruction shaper ZVDD, and solve the parameters of the instruction shaper ZVDD, the parameters of the instruction shaper ZVDD include an amplitude ratio parameter and a self-oscillation period;

[0023] Step 3: Consider the simplified on-orbit service process hand-transmitted vibration force F (input signal) as a continuous pulse signal, and shape each pulse signal of the simplified on-orbit service process hand-transmitted vibration force F to obtain a new simplified on-orbit service process hand-transmitted vibration force F′ (there are many pulses in F, and F′ is recorded as a collection of pulses at each moment);

[0024] Step 4: Integrate the new simplified on-orbit service process hand-transmitted vibration force F′ obtained in step 3 to obtain the new shaped signal force

[0025] Step 5: Apply force based on the new shaped signal The state variables of the dynamic model of hand-transmitted vibration in the on-orbit service process are calculated, and the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping in the on-orbit service process is determined based on the state variables.

[0026] Specific implementation method 2: This implementation method is different from the specific implementation method 1 in that the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process is obtained in step 1; the specific process is:

[0027] Ignoring the nonlinear effect of hand-transmitted vibration during on-orbit service, the hand-transmitted vibration during on-orbit service is simplified into a linear structure, and the simplified dynamic model of hand-transmitted vibration during on-orbit service is written as follows:

[0028]

[0029] Where M represents the simplified mass matrix of the hand-transmitted vibration in the on-orbit service process, C represents the simplified damping matrix of the hand-transmitted vibration in the on-orbit service process, K represents the simplified stiffness matrix of the hand-transmitted vibration in the on-orbit service process, F represents the simplified force (input) of the hand-transmitted vibration in the on-orbit service process, q represents the simplified state variable (output) of the dynamic model of the hand-transmitted vibration in the on-orbit service process, and They represent the first and second order derivatives of q with respect to time respectively;

[0030] The state variables of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process are the column vectors composed of displacements in the simplified dynamic model of hand-transmitted vibration in the on-orbit service process. The equivalent expression of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process is:

[0031]

[0032] Where I represents the unit matrix with the same dimension as q;

[0033] Write equation (2) in the form of state equation, the expression is:

[0034]

[0035] The other steps and parameters are the same as those in the first embodiment.

[0036] Specific implementation method three: This implementation method is different from specific implementation method one or two in that, in step 2, the instruction shaper type ZVDD is determined, and the parameters of the instruction shaper ZVDD are solved. The parameters of the instruction shaper ZVDD include the amplitude ratio parameter and the self-oscillation period. The specific process is:

[0037] Step 21: Use adaptive reinforcement learning to obtain the damping ratio ξ and natural frequency ω of the hand-transmitted vibration model (state equation of Formula 3) during the on-orbit service process n The best value of

[0038] Step 22: Damping ratio ξ and natural frequency ω based on the hand-transmitted vibration model (state equation of Formula 3) during on-orbit service n The optimal value of the command shaper ZVDD is obtained. and the self-oscillation period T of the hand-transmitted vibration model during on-orbit service d The best value.

[0039] The other steps and parameters are the same as those in the first or second embodiment.

[0040] Specific implementation method 4: This implementation method is different from any one of the specific implementation methods 1 to 3 in that the damping ratio ξ and the natural frequency ω of the hand-transmitted vibration model during the on-orbit service process are obtained by using adaptive reinforcement learning in step 21. n The optimal value of; the specific process is:

[0041] Step 211: Set the system error penalty function R(ξ,ω n );

[0042] Where ξ represents the damping ratio, ω n represents the natural frequency;

[0043] Step 212: Initialize the learning rate α, the first-order moment control parameter β1, the second-order moment control parameter β2, and the parameter vector x0=[ξ0,ω n0 ], the first-order moment vector m0, the second-order moment vector v0 and the threshold ε (usually small) that determines the convergence of the parameter vector;

[0044] Step 213: Set k = 1, gradient

[0045] in represents the gradient operator;

[0046] Parameter vector x k =[ξ k ,ω nk ]、ξ k Represents the damping ratio of the hand-transmitted vibration model during the on-orbit service process of the kth cycle, ω nk represents the natural frequency of the hand-transmitted vibration model during the on-orbit servicing process during the kth cycle;

[0047] Step 214: based on the first-order moment vector, the second-order moment vector, the first-order moment control parameter β1, the second-order moment control parameter β2, and the gradient, update the first-order moment vector and the second-order moment vector;

[0048] m k =β1m k-1 +(1-β1)g k

[0049]

[0050] Where m k 、v k Represent the first-order moment vector and second-order moment vector of the kth cycle, m k-1 、v k-1They represent the first-order moment vector and the second-order moment vector of the k-1th cycle, respectively. β1 and β2 represent the first-order moment control parameter and the second-order moment control parameter of the hand-transmitted vibration during the on-orbit service process, respectively.

[0051] Step 215: Calculate the corrected first-order moment vector and second-order moment vector;

[0052]

[0053]

[0054] in represent the first-order moment vector and second-order moment vector of the k-th cycle after correction respectively;

[0055] Step 216: updating the parameter vector based on the corrected first-order moment vector and second-order moment vector;

[0056]

[0057] Where α represents the learning rate of the adaptive algorithm;

[0058] ε is the threshold (usually small, such as 1e-6, to ensure that the denominator is not 0);

[0059] Step 217: Let k = k + 1, and repeat steps 213 to 216 until the parameter vector x appears. k Convergence, that is, |x k -x k-1 |<ε; take the current x k =[ξ,ω n ] is the optimal value of damping ratio and natural frequency.

[0060] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.

[0061] Specific implementation method 5: This implementation method is different from any one of the specific implementation methods 1 to 4 in that the system error penalty function R(ξ,ω) in step 211 n ) is expressed as:

[0062]

[0063] where y e (ξ,ω n ,t) represents the error output of hand-transmitted vibration during the on-orbit service process, and t represents the current operating time of the hand-transmitted vibration during the on-orbit service process;

[0064] The hand-transmitted vibration error output y during the on-orbit service process e =He;

[0065] Where e represents the difference between the column vector of hand-transmitted vibration displacement in the simplified on-orbit service process and the expected value, and H is the weight coefficient;

[0066] The difference e between the column vector of the hand-transmitted vibration displacement in the simplified on-orbit service process and the expected value is expressed as:

[0067]

[0068] in represents the derivative of e with respect to time, A′ and B are the correlation matrix parameters respectively;

[0069]

[0070]

[0071] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.

[0072] Specific implementation method 6: This implementation method is different from any one of the specific implementation methods 1 to 5 in that the damping ratio ξ and the natural frequency ω of the hand-transmitted vibration model based on the on-orbit service process in step 22 are n The optimal value of the shaper amplitude ratio parameter is obtained and the self-oscillation period T of the simplified hand-transmitted vibration model during on-orbit service d The optimal value of; the specific process is:

[0073]

[0074] The other steps and parameters are the same as those in Specific Implementations 1 to 5.

[0075] Specific implementation method 7: This implementation method is different from any one of specific implementation methods 1 to 6 in that in step 3, the simplified on-orbit service process hand-transmitted vibration force F (input signal) is regarded as a continuous pulse signal, and each pulse signal of the simplified on-orbit service process hand-transmitted vibration force F is shaped to obtain a new simplified on-orbit service process hand-transmitted vibration force F′ (there are many pulses in F); the specific process is:

[0076] The shaper ZVDD is used to shape each pulse signal of the hand-transmitted vibration force F in the simplified on-orbit service process. The instruction shaper ZVDD expression is:

[0077]

[0078] Among them A ZVDD represents the row vector composed of the amplitudes of the pulses of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, t irepresents the row vector of the time delay of each pulse of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, K represents the amplitude ratio parameter of the shaper, T d It represents the self-oscillation period of the simplified dynamic model of hand-transmitted vibration during on-orbit service.

[0079] The other steps and parameters are the same as those in Specific Embodiments 1 to 6.

[0080] Specific implementation eight: This implementation differs from specific implementations one to seven in that in step 5, the force is applied to the newly shaped signal. The state variables of the dynamic model of hand-transmitted vibration in the on-orbit service process are calculated, and the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping in the on-orbit service process is determined based on the state variables; the specific process is:

[0081] Based on the new shaped signal force The new simplified state variable q of the hand-transmitted vibration dynamics model of the on-orbit service process is determined by formula (3) (for example, the load torque in the case of screwing);

[0082] Determine whether the state variable q meets the requirements (for example, the load torque required to tighten the screw is 4-4.5Nm):

[0083] If the simplified on-orbit service process meets the conditions, the amplitude of each pulse of the hand-transmitted vibration instruction shaping constitutes the row vector A ZVDD For the final result;

[0084] If it does not meet the requirement, continue to increase the amplitude of each pulse of the instruction shaper ZVDD expression (5) in proportion and slowly, obtain a new shaped signal force (for example, screwing is controlled by the control current), and determine the new simplified state variable of the hand-transmitted vibration dynamic model of the on-orbit service process through formula (3) (for example, the load torque in the case of screwing); continue to judge whether the state variable q meets the requirement until the state variable q meets the requirement. At this time, the row vector A composed of the amplitudes of each pulse of the simplified on-orbit service process hand-transmitted vibration instruction shaping is ZVDD For the final result.

[0085] Taking the hand-transmitted vibration during the on-orbit service process with 9 degrees of freedom as an example, Figure 3 and Figure 4 The input signal and vibration amplitude before and after the shaping transformation are shown respectively.

[0086] The other steps and parameters are the same as those in Specific Embodiments 1 to 7.

[0087] The invention includes three parts: command shaping of input signal, adaptive updating of parameters and optimization of command shaping. The goal is to reduce hand-transmitted vibration as much as possible.

[0088] The pulses and delays of the instruction shaping can be written in the following matrix form (i.e., the shaping expression):

[0089]

[0090] Among them A i ,T i represents the amplitude and delay of the ith pulse of the hand-transmitted vibration command shaping in the simplified on-orbit service process, i = 1, 2, ..., n, n represents the total number of pulses;

[0091] A means A i The row vector is composed of i=1,2,…,n; T represents T i The row vector is i=1,2,…,n.

[0092] Instruction shaping optimization:

[0093] Next, we will use the energy-based method to correct the value of each pulse of the instruction shaping:

[0094] The input value is the control current, and the output value is the load torque. When the load torque reaches the requirement, the command is considered to have been completed.

[0095] The input signal after instruction shaping is input as the control current. When it is detected that the output load torque reaches the requirement, the current input is turned off. If the final output load torque can meet the requirement by relying on inertia, it means that the original instruction shaping result is feasible; otherwise, if the final output load torque is less than the requirement, the input signal of the instruction shaping will be continuously adjusted through multiple iterations.

[0096] In each iteration, the pulse amplitudes of the shaper expression (5) are increased slowly and proportionally, and the state vector before and after the shaping transformation is calculated again using the state equation expressed by equation (3). At this time, the interval between each pulse is kept unchanged. After each slow increase in the pulse amplitude in proportion, a test is performed. If the final load torque meets the requirements, the input pulse result at this time is determined as the final result. Otherwise, the pulse amplitude continues to be increased slowly and proportionally.

[0097] Taking the task of tightening screws as an example, the input value is the control current, and the output value is the screw load torque. When the load torque reaches 4-4.5Nm, the screw is considered to be tightened. Input according to the result after instruction shaping. When the load torque is detected to be 3-3.5Nm, turn off the current input. If the final load torque of the screw can reach 4-4.5Nm by relying on inertia, it means that the original instruction shaping result is feasible; otherwise, if the final load torque is less than 4Nm, we will slowly increase the amplitude of each pulse in proportion and keep the interval time of each pulse unchanged. Tests are carried out after each slight increase in the pulse amplitude. If the final load torque reaches 4-4.5Nm, the input pulse result at this time is determined to be the final result, otherwise, the pulse amplitude continues to be slightly increased.

[0098] The following examples are used to verify the beneficial effects of the present invention:

[0099] Embodiment 1:

[0100] Step 1: Obtain the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process (taking the screw tightening task as an example, the goal is to make the screw load torque 4 to 4.5 Nm); the specific process is:

[0101] Ignoring the nonlinear effect of hand-transmitted vibration during on-orbit service, the hand-transmitted vibration during on-orbit service is simplified into a linear structure, and the simplified dynamic model of hand-transmitted vibration during on-orbit service is written as follows:

[0102]

[0103] Where M represents the simplified mass matrix of the hand-transmitted vibration in the on-orbit service process, C represents the simplified damping matrix of the hand-transmitted vibration in the on-orbit service process, K represents the simplified stiffness matrix of the hand-transmitted vibration in the on-orbit service process, F represents the simplified input current of the hand-transmitted vibration in the on-orbit service process, q represents the load torque of the screw in the simplified hand-transmitted vibration model of the on-orbit service process, and They represent the first and second order derivatives of q with respect to time respectively;

[0104] The state variables of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process are the column vectors composed of displacements in the simplified dynamic model of hand-transmitted vibration in the on-orbit service process. The equivalent expression of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process is:

[0105]

[0106] Where I represents the unit matrix with the same dimension as q;

[0107] Write equation (2) in the form of state equation, the expression is:

[0108]

[0109] Step 2: Determine the instruction shaper type ZVDD and solve the parameters of the instruction shaper ZVDD. The parameters of the instruction shaper ZVDD include the amplitude ratio parameter and the self-oscillation period. The specific process is:

[0110] Step 21: Use adaptive reinforcement learning to obtain the damping ratio ξ and natural frequency ω of the hand-transmitted vibration model (state equation of Formula 3) during the on-orbit service process n The optimal value of; the specific process is:

[0111] Step 211: Set the system error penalty function R(ξ,ω n );

[0112] Where ξ represents the damping ratio, ω n represents the natural frequency;

[0113] Step 212: Initialize the learning rate α, the first-order moment control parameter β1, the second-order moment control parameter β2, and the parameter vector x0=[ξ0,ω n0 ], the first-order moment vector m0, the second-order moment vector v0 and the threshold ε (usually small) that determines the convergence of the parameter vector;

[0114] Step 213: Set k = 1, gradient

[0115] in represents the gradient operator;

[0116] Parameter vector x k =[ξ k ,ω nk ]、ξ k Represents the damping ratio of the hand-transmitted vibration model during the on-orbit service process of the kth cycle, ω nk represents the natural frequency of the hand-transmitted vibration model during the on-orbit servicing process during the kth cycle;

[0117] Step 214: based on the first-order moment vector, the second-order moment vector, the first-order moment control parameter β1, the second-order moment control parameter β2, and the gradient, update the first-order moment vector and the second-order moment vector;

[0118] m k =β1m k-1 +(1-β1)g k

[0119]

[0120] Where m k 、v k Represent the first-order moment vector and second-order moment vector of the kth cycle, m k-1 、v k-1 They represent the first-order moment vector and the second-order moment vector of the k-1th cycle, respectively. β1 and β2 represent the first-order moment control parameter and the second-order moment control parameter of the hand-transmitted vibration during the on-orbit service process, respectively.

[0121] Step 215: Calculate the corrected first-order moment vector and second-order moment vector;

[0122]

[0123]

[0124] in represent the first-order moment vector and second-order moment vector of the k-th cycle after correction respectively;

[0125] Step 216: updating the parameter vector based on the corrected first-order moment vector and second-order moment vector;

[0126]

[0127] Where α represents the learning rate of the adaptive algorithm;

[0128] ε is the threshold (usually small, such as 1e-6, to ensure that the denominator is not 0);

[0129] Step 217: Let k = k + 1, and repeat steps 213 to 216 until the parameter vector x appears. k Convergence, that is, |x k -x k-1 |<ε; take the current x k =[ξ,ω n ] is the optimal value of damping ratio and natural frequency.

[0130] Step 22: Damping ratio ξ and natural frequency ω based on the hand-transmitted vibration model (Equation 3) during on-orbit service n The optimal value of the command shaper ZVDD is obtained. and the self-oscillation period T of the hand-transmitted vibration model during on-orbit service d The optimal value of; the specific process is:

[0131]

[0132] Step 3: Consider the simplified on-orbit service process hand-transmitted vibration input current F as a continuous pulse signal, and shape each pulse signal of the simplified on-orbit service process hand-transmitted vibration input current F to obtain a new simplified on-orbit service process hand-transmitted vibration input current F′ (there are many pulses in F, and F′ is recorded as a collection of pulses at each moment); the specific process is:

[0133] The shaper ZVDD is used to shape each pulse signal of the hand-transmitted vibration input current F during the simplified on-orbit service process. The instruction shaper ZVDD expression is:

[0134]

[0135] Among them A ZVDD represents the row vector composed of the amplitude of each pulse of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, t i represents the row vector of the time delay of each pulse of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, K represents the amplitude ratio parameter of the shaper, T d It represents the self-oscillation period of the simplified dynamic model of hand-transmitted vibration during on-orbit service.

[0136] Step 4: Integrate the new simplified on-orbit service process hand-transmitted vibration input current F′ obtained in step 3 to obtain a new shaped signal input current

[0137] Step 5: Input current based on the new shaped signal The state variables of the dynamic model of hand-transmitted vibration in the on-orbit service process are calculated, and the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping in the on-orbit service process is determined based on the state variables; the specific process is:

[0138] Based on the new shaped signal input current The new simplified load moment q of the hand-transmitted vibration dynamics model of the on-orbit service process is determined by formula (3);

[0139] Determine whether the load torque q of the screw is in the range of 4 to 4.5 Nm:

[0140] If the simplified on-orbit service process meets the conditions, the amplitude of each pulse of the hand-transmitted vibration instruction shaping constitutes the row vector A ZVDD For the final result;

[0141] If it does not meet the requirement, continue to increase the amplitude of each pulse of the instruction shaper ZVDD expression (5) in proportion and slowly to obtain a new shaped input current, and determine the new simplified load torque of the hand-transmitted vibration dynamic model of the on-orbit service process through formula (3); continue to judge whether the load torque q meets the requirement until the load torque q meets the requirement. At this time, the row vector A composed of the amplitudes of each pulse of the simplified on-orbit service process hand-transmitted vibration instruction shaping is ZVDD For the final result.

[0142] Embodiment 2:

[0143] Energy-based instruction shaping and correction optimization (taking the task of tightening screws as an example, the goal is to make the screw load torque 4 to 4.5 Nm)

[0144] Step 1: Initialize the number of loops k = 0;

[0145] Step 2: Set a suitable multiple step size ∈;

[0146] Step 3: Use speed control until the load torque reaches 3 to 3.5 Nm;

[0147] Step 4: Transform to current control, the input current signal is k∈ times of each pulse of the instruction shaping (the current signal of the first cycle is 0), and the interval time of each pulse is kept unchanged;

[0148] Step 5: Repeat steps 3 to 4 until the screw stops (angular velocity is 0) and the load torque reaches 4 to 4.5 Nm;

[0149] Step 6: Determine the corrected instruction shaping signal.

[0150] The present invention may also have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for actively suppressing on-orbit hand-transmitted vibration based on parameter adaptive learning, characterized in that: The specific process of the method is: Step 1: Obtain the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process; Step 2: Determine the type of instruction shaper ZVDD, and solve the parameters of the instruction shaper ZVDD, the parameters of the instruction shaper ZVDD include an amplitude ratio parameter and a self-oscillation period; Step 3: The simplified hand-transmitted vibration force F during the on-orbit service process is regarded as a continuous pulse signal, and each pulse signal of the simplified hand-transmitted vibration force F during the on-orbit service process is shaped to obtain a new simplified hand-transmitted vibration force F′ during the on-orbit service process; Step 4: Integrate the new simplified on-orbit service process hand-transmitted vibration force F′ obtained in step 3 to obtain the new shaped signal force Step 5: Apply force based on the new shaped signal Calculate the state variables of the dynamic model of hand-transmitted vibration during the on-orbit service process, and determine the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping during the on-orbit service process based on the state variables; In step 1, the state equation of the hand-transmitted vibration dynamics model during the on-orbit service process is obtained; the specific process is: Ignoring the nonlinear effect of hand-transmitted vibration during on-orbit service, the hand-transmitted vibration during on-orbit service is simplified into a linear structure, and the simplified dynamic model of hand-transmitted vibration during on-orbit service is written as follows: Where M represents the simplified mass matrix of the hand-transmitted vibration in the on-orbit service process, C represents the simplified damping matrix of the hand-transmitted vibration in the on-orbit service process, K represents the simplified stiffness matrix of the hand-transmitted vibration in the on-orbit service process, F represents the simplified force of the hand-transmitted vibration in the on-orbit service process, q represents the simplified state variable of the dynamic model of the hand-transmitted vibration in the on-orbit service process, and They represent the first and second order derivatives of q with respect to time respectively; The state variables of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process are the column vectors composed of displacements in the simplified dynamic model of hand-transmitted vibration in the on-orbit service process. The equivalent expression of the simplified dynamic model of hand-transmitted vibration in the on-orbit service process is: Where I represents the unit matrix with the same dimension as q; Write equation (2) in the form of state equation, the expression is: In the step 2, the type of instruction shaper ZVDD is determined, and the parameters of the instruction shaper ZVDD are solved. The parameters of the instruction shaper ZVDD include an amplitude ratio parameter and a self-oscillation period. The specific process is: Step 21: Use adaptive reinforcement learning to obtain the damping ratio ξ and natural frequency ω of the hand-transmitted vibration model during on-orbit service n The best value of Step 22: Damping ratio ξ and natural frequency ω of the hand-transmitted vibration model based on the on-orbit service process n The optimal value of the command shaper ZVDD is obtained. and the self-oscillation period T of the hand-transmitted vibration model during on-orbit service d The best value of In step 21, adaptive reinforcement learning is used to obtain the damping ratio ξ and natural frequency ω of the hand-transmitted vibration model during the on-orbit service process. n The optimal value of; the specific process is: Step 211: Set the system error penalty function R(ξ,ω n ); Where ξ represents the damping ratio, ω n represents the natural frequency; Step 212: Initialize the learning rate α, the first-order moment control parameter β1, the second-order moment control parameter β2, and the parameter vector x0=[ξ0,ω n0 ], the first-order moment vector m0, the second-order moment vector v0 and the threshold ε for determining the convergence of the parameter vector; Step 213: Set k = 1, gradient in represents the gradient operator; Parameter vector x k =[ξ k ,ω nk ]、ξ k Represents the damping ratio of the hand-transmitted vibration model during the on-orbit service process of the kth cycle, ω nk represents the natural frequency of the hand-transmitted vibration model during the on-orbit servicing process during the kth cycle; Step 214: based on the first-order moment vector, the second-order moment vector, the first-order moment control parameter β1, the second-order moment control parameter β2, and the gradient, update the first-order moment vector and the second-order moment vector; m k =β1m k-1 +(1-β1)g k Where m k 、v k Represent the first-order moment vector and second-order moment vector of the kth cycle, m k-1 、v k-1 They represent the first-order moment vector and the second-order moment vector of the k-1th cycle, respectively. β1 and β2 represent the first-order moment control parameter and the second-order moment control parameter of the hand-transmitted vibration during the on-orbit service process, respectively. Step 215: Calculate the corrected first-order moment vector and second-order moment vector; in represent the first-order moment vector and second-order moment vector of the k-th cycle after correction respectively; Step 216: updating the parameter vector based on the corrected first-order moment vector and second-order moment vector; Where α represents the learning rate of the adaptive algorithm; ε is the threshold Step 217: Let k = k + 1, and repeat steps 213 to 216 until the parameter vector x appears. k Convergence, that is, |x k -x k-1 |<ε; take the current x k =[ξ,ω n ] is the optimal value of damping ratio and natural frequency; In step 211, the system error penalty function R(ξ,ω n ) is expressed as: where y e (ξ,ω n ,t) represents the error output of hand-transmitted vibration during the on-orbit service process, and t represents the current operating time of the hand-transmitted vibration during the on-orbit service process; The hand-transmitted vibration error output y during the on-orbit service process e =He; Where e represents the difference between the column vector of hand-transmitted vibration displacement in the simplified on-orbit service process and the expected value, and H is the weight coefficient; The difference e between the column vector of the hand-transmitted vibration displacement in the simplified on-orbit service process and the expected value is expressed as: in represents the derivative of e with respect to time, A′ and B are the correlation matrix parameters respectively; The damping ratio ξ and natural frequency ω of the hand-transmitted vibration model in the on-orbit service process in step 22 are n The optimal value of the shaper amplitude ratio parameter is obtained and the self-oscillation period T of the simplified hand-transmitted vibration model during on-orbit service d The optimal value of; the specific process is:

2. The method for actively suppressing on-orbit hand-transmitted vibration based on parameter adaptive learning according to claim 1, characterized in that: In the step 3, the simplified on-orbit service process hand-transmitted vibration force F is regarded as a continuous pulse signal, and each pulse signal of the simplified on-orbit service process hand-transmitted vibration force F is shaped to obtain a new simplified on-orbit service process hand-transmitted vibration force F′; the specific process is: The command shaper ZVDD is used to shape each pulse signal of the hand-transmitted vibration force F in the simplified on-orbit service process. The expression of the shaper ZVDD is: Among them A ZVDD represents the row vector composed of the amplitudes of the pulses of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, t i represents the row vector of the time delay of each pulse of the hand-transmitted vibration command shaping in the simplified on-orbit service process under the ZVDD method, K represents the amplitude ratio parameter of the shaper, T d It represents the self-oscillation period of the simplified dynamic model of hand-transmitted vibration during on-orbit service.

3. The method for actively suppressing on-orbit hand-transmitted vibration based on parameter adaptive learning according to claim 2, characterized in that: In step 5, the force is applied based on the new shaped signal The state variables of the dynamic model of hand-transmitted vibration in the on-orbit service process are calculated, and the row vector composed of the amplitudes of each pulse of the hand-transmitted vibration instruction shaping in the on-orbit service process is determined based on the state variables; the specific process is: Based on the new shaped signal force The new simplified state variable q of the hand-transmitted vibration dynamic model of the on-orbit service process is determined by formula (3); Determine whether the state variable q meets the requirements: If the simplified on-orbit service process meets the conditions, the amplitude of each pulse of the hand-transmitted vibration instruction shaping constitutes the row vector A ZVDD For the final result; If it does not meet the requirement, continue to increase the amplitude of each pulse of the instruction shaper ZVDD expression (5) in proportion to obtain a new shaped signal force, and determine the new simplified state variable of the hand-transmitted vibration dynamic model of the on-orbit service process through formula (3); continue to judge whether the state variable q meets the requirement until the state variable q meets the requirement. At this time, the row vector A composed of the amplitudes of each pulse of the simplified on-orbit service process hand-transmitted vibration instruction shaping is ZVDD For the final result.

Images