A residual vibration suppression method, device and shaper

Abstract

The present invention discloses a residual vibration suppression method, device and input shaper to solve the problem in the prior art that different types of shapers have different processing methods and are difficult to solve. The method includes: determining the target function of residual vibration; generating an initial pulse sequence group; performing an evolutionary operation on any initial pulse to obtain an evolved pulse sequence group; when the evolutionary operation termination condition is met, the result of the evolutionary operation is used as the target pulse sequence group of the input shaper, and vice versa, the pulse sequence continues to be evolved. By adopting a penalty function, the constraint problem of residual vibration solution can be converted into an unconstrained solution problem, and the specific input shaper is not distinguished, which has universality. Secondly, multiple groups of mutation strategies are adopted to implement the evolution of pulse sequences. By reasonably setting multiple groups of mutation strategies, the complementarity of the characteristics of multiple groups of mutation strategies is achieved, and the group diversity, global search capability and group convergence speed of the pulse sequence solution process are taken into consideration.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a residual vibration suppression method, device and shaper. Background Art

[0002] There is a residual vibration phenomenon in the equipment positioning in the field of motion control. The residual vibration will affect the positioning accuracy of the equipment, so it must be suppressed, especially for systems that require precise positioning, such as medical devices that have strict requirements on the anti-shake and stability of the equipment. The input signal will produce a certain amount of residual vibration when passing through the medical device, affecting the accuracy of the equipment output signal.

[0003] Input shaping is a commonly used method to effectively suppress residual vibration. The main principle is to suppress residual vibration by shaping the system input signal. The input shaper contains a series of pulse sequences. By selecting a suitable pulse sequence, the residual vibration can be well suppressed. At present, in the vibration suppression based on input shapers, there are three typical input shapers, namely, zero vibration (Zero Vibration, ZV) input shaper, zero vibration differential (Zero Vibration and Derivation, ZVD) input shaper and extra insensitivity (Extra Insensitivity, EI) input shaper.

[0004] For the input shaper, how to obtain a suitable pulse sequence is the key to suppressing residual vibration. Based on the existing technology, at present, the EI input shaper has no analytical solution when solving the pulse sequence, and can only rely on mathematical software for numerical solution. There is currently no unified method for solving and obtaining a suitable pulse sequence for single-peak and multi-peak EI input shapers, which limits the vibration suppression effect of the EI input shaper. The ZVD input shaper can solve the analytical solution when solving the pulse sequence, but when it is extended to multiple differentials (also called ZVD multiple differential input shaper ZVDq, q represents q-fold differential), as the order of differentials increases, the difficulty of obtaining its analytical solution becomes increasingly greater. Summary of the invention

[0005] The present invention provides a residual vibration suppression method, device and shaper to solve the following problems existing in the prior art: 1. For the ZVD multiple differential input shaper, as the differential order increases, the difficulty of solving becomes increasingly greater, which limits the vibration suppression effect. 2. For the EI shaper, it is impossible to obtain an analytical solution. At present, there is no unified method to solve the pulse sequence of the single-peak and multi-peak EI input shaper, which limits the vibration suppression effect of the EI shaper.

[0006] In a first aspect, in order to solve the above technical problems, an embodiment of the present invention provides a residual vibration suppression method, the method comprising:

[0007] Step A: Penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration;

[0008] Step B: generating an initial pulse sequence group including at least four initial pulse sequences;

[0009] Step C: traverse the initial pulse sequence group, perform an evolution operation on any initial pulse, and obtain an evolved pulse sequence group;

[0010] The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse;

[0011] Step D: Determine whether the termination condition of the evolution operation is triggered. If so, use the evolved pulse sequence group obtained in step C as the target pulse sequence group of the input shaper; if not, use the evolved pulse sequence group obtained in step C as the generated initial sequence group and execute step C.

[0012] Preferably, the evolution operation termination condition includes that the execution times of step C reaches a preset number of times.

[0013] Preferably, the evolution operation termination condition includes that the objective function value of the evolution pulse sequence obtained in step C is less than or equal to a preset value.

[0014] Preferably, the residual vibration expression is:

[0015]

[0016] Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, cos is the cosine calculation, sin is the sine calculation, ∑ is the sum calculation, and n is the number of pulses contained in the input shaper; A i is the amplitude of the pulse train, t i is the time delay corresponding to the pulse train, where t i<t i+1 .

[0017] Preferably, when the input shaper is a ZVD multiple differential input shaper, the constraint equation of the residual vibration is:

[0018]

[0019] Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, n is the number of pulses contained in the input shaper, q represents the derivative order, q=n-2, n≥2.

[0020] Preferably, when the input shaper is a single-peak EI input shaper, the constraint equation of the residual vibration is:

[0021] Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, n is the number of pulses contained in the input shaper, V exp is the preset limit value of residual vibration, ω h is greater than ω n The preset frequency, ω l is less than ω n preset frequency.

[0022] Preferably, when the input shaper is a multi-peak EI input shaper, the constraint equation of the residual vibration is:

[0023]

[0024] Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, n is the number of pulses contained in the input shaper, and K is the number of peaks of the sensitivity curve of the EI shaper.

[0025] In a second aspect, an embodiment of the present invention further provides a residual vibration suppression device, the device comprising:

[0026] An acquisition unit, used for penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration;

[0027] An initialization unit, used for generating an initial pulse sequence group including at least four initial pulse sequences;

[0028] An evolution unit, used for traversing the initial pulse sequence group, performing an evolution operation on any initial pulse, and obtaining an evolved pulse sequence group;

[0029] The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse;

[0030] A judgment unit, used to judge whether the evolution operation termination condition is triggered;

[0031] a determination unit, configured to use the evolved pulse sequence group obtained in the evolution unit as the target pulse sequence group of the input shaper when the evolution operation termination condition is triggered;

[0032] The cyclic unit is used to use the evolution pulse sequence group obtained by the evolution unit as the generated initial sequence group to trigger the evolution unit to perform the evolution operation when the evolution operation termination condition is not triggered.

[0033] In a third aspect, an embodiment of the present invention provides an input shaper, including: a memory, a transceiver, and a processor:

[0034] A memory for storing a computer program; a transceiver for transmitting and receiving data under the control of the processor; and a processor for reading the computer program in the memory and executing any one of the above methods.

[0035] Through the technical solutions in one or more of the above embodiments of the embodiments of the present invention, the embodiments of the present invention have at least the following technical effects:

[0036] In the embodiment provided by the present invention, the constraint equation of the residual vibration corresponding to the input shaper is penalized, and the constraint problem of solving the residual vibration is converted into an unconstrained solution problem, and the specific input shaper is not distinguished, which has universality. Secondly, an evolution operation is performed on the pulse sequence of the input shaper, and multiple groups of mutation strategies are used in the evolution operation (each group of mutation strategies includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR). By reasonably setting each group of mutation strategies, the characteristics of each group of mutation strategies can be complemented to each other, so as to achieve a balance between the group diversity, global search capability and group convergence speed in the pulse sequence solution process. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flow chart of a residual vibration suppression method provided by an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of the mutation strategy in the CoDE algorithm in an embodiment of the present invention;

[0039] Figure 3 A schematic diagram of a vibration curve of a display in an OCT system without vibration suppression provided in an embodiment of the present invention;

[0040] Figure 4 A schematic diagram of a display vibration curve after adding vibration suppression in an OCT system provided by an embodiment of the present invention;

[0041] Figure 5 A schematic structural diagram of a residual vibration suppression device provided by an embodiment of the present invention;

[0042] Figure 6 A schematic diagram of the structure of an input shaper provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In the embodiments of the present invention, the term "and / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0044] In the embodiments of the present application, the term "plurality" refers to two or more than two, and other quantifiers are similar.

[0045] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0046] Please refer to Figure 4 , which is a residual vibration suppression method provided by an embodiment of the present invention, the processing process of the method is as follows.

[0047] Step A: Penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration;

[0048] For most control systems, it can be simplified to a second-order system. When the pulse sequence acts on the second-order system, the corresponding residual vibration expression can be obtained as follows:

[0049]

[0050] Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object (system). For a given controlled object, the vibration frequency is known and fixed. ξ is the damping ratio of the controlled object. cos is the cosine calculation, sin is the sine calculation, ∑ is the sum calculation, and n is the number of pulses contained in the input shaper. A i is the amplitude of the pulse train, t i is the time delay corresponding to the pulse train, where t i <t i+1 To minimize the time delay, t1 can be set to 0. To suppress the residual vibration, that is, to obtain different A i and t i , which limits the value of formula (1).

[0051] For different input shapers, the constraint equations of the residual vibration are different. Based on the constraint equations, the process of using the penalty function to construct the objective function will be described in detail in the subsequent embodiments.

[0052] Step B: generating an initial pulse sequence group including at least four initial pulse sequences; Step C: traversing the initial pulse sequence group, performing an evolution operation on any initial pulse, and obtaining an evolved pulse sequence group;

[0053] The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence after the current pulse sequence is mutated by each mutation strategy; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence;

[0054] Among them, the current pulse sequence corresponds to multiple groups of mutation strategies, so after the current pulse sequence undergoes the mutation strategy, it will mutate into multiple mutated pulse sequences (one group of mutation strategies corresponds to one mutated pulse sequence). During calculation, it is necessary to calculate the objective function of the current pulse sequence and the objective function of each group of pulse sequences in the multiple mutated pulse sequences corresponding to the current pulse sequence, and then compare the obtained objective functions, and select the pulse sequence corresponding to the minimum value of the objective function as the evolved pulse sequence of the current pulse sequence. It can be understood that for the current pulse sequence, the evolved pulse sequence may be the current pulse sequence itself, or it may be any mutated pulse sequence.

[0055] Each group of mutation strategies includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR. The current pulse sequence is any initial pulse sequence in the initial pulse sequence group. The evolution pulse sequence group is composed of the evolution pulses of all pulse sequences in the initial pulse;

[0056] Step D: Determine whether the evolution operation termination condition is triggered. If so, use the evolved pulse sequence group obtained in step C as the target pulse sequence group of the input shaper; if not, use the evolved pulse sequence group obtained in step C as the initial sequence group and execute step C.

[0057] The target pulse sequence group obtained in step D is the optimal solution of equation (1).

[0058] In the embodiment, the constraint equation of the residual vibration corresponding to the input shaper is penalized, and the constraint problem of solving the residual vibration is converted into an unconstrained solution problem, and the specific input shaper is not distinguished, which has universality. Secondly, the differential evolution algorithm DE is used in the pulse sequence evolution operation of the input shaper, but it is different from the existing DE algorithm. In the present application, when calculating the evolved pulse sequence, multiple groups of mutation strategies are adopted (each group of mutation strategies includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR). By reasonably setting the mutation strategy, the characteristics of each group of mutation strategies can be complemented to each other, so as to achieve the balance of population diversity, global search capability and population convergence speed in the pulse sequence solution process.

[0059] The termination condition of the evolution operation includes that the execution times of step C reaches a preset number of cycles. Preferably, the termination condition of the evolution operation may also be that the objective function value of the evolved pulse sequence obtained in step C is less than or equal to a preset value.

[0060] In order to facilitate the understanding of the embodiments provided by the present application, the above steps are described in detail below. First, for different types of input shapers (hereinafter referred to as shapers), the constraint equations of residual vibration are described in detail.

[0061] For the ZVD multiple differential (or ZVDq) input shaper, the residual vibration and the q-order differential, q-1, order differential, ..., 1-order differential of the residual vibration are required to be 0, q = n-2, n ≥ 2, the number of pulses n = q + 2, q = 0, 1, 2, .... The corresponding constraint equations are as follows:

[0062]

[0063] When q = 0 and q = 1, the analytical solution of equation (2) is easy to obtain. However, as q increases, it becomes increasingly difficult to obtain the analytical solution. However, the larger q is, the better the vibration suppression effect that the ZVDq input shaper can achieve in theory.

[0064] For a single peak EI input shaper, the residual vibration expression is required to be a non-zero value V exp , and above ω n The frequency ω h Its value is 0, and below ω n The frequency ω l Its value is also 0. Therefore, in [ω l ,ω h ], it can ensure that the input shaper makes the residual vibration expression of the system less than or equal to V exp . Its constraint equation is as follows:

[0065]

[0066] It should be noted that there is no analytical solution for equation (3). h and frequency ω l is the default value.

[0067] For the multi-peak EI shaper, the K-peak shaper requires a m The residual vibration at the frequency (ω m +ω m+1 ) / 2 is 0, that is, the constraint equation is as follows:

[0068]

[0069] Where K is the number of peaks in the sensitivity curve of the EI shaper. Equation (4) also has no analytical solution and is relatively difficult to solve.

[0070] Based on the above description, for different types of shapers, the constraint equations corresponding to the residual vibration are determined. Under the condition of satisfying the constraint equations, the pulse sequence that minimizes the residual vibration can be solved by constructing a penalty function to obtain the objective function. The following examples illustrate the process of constructing penalty functions for several shapers.

[0071] The residual vibration expression of the second-order system is shown in equation (1). There are three main types of typical input shapers, namely, zero vibration (ZV) input shaper, zero vibration differential (ZVD) input shaper and extra insensitivity (EI) input shaper.

[0072] The ZV shaper is the simplest input shaper, consisting of two pulses. The ZV shaper requires the residual vibration expression to be 0. At the same time, in order to minimize the pulse delay, t1=0, Then the constraint equation of the ZV shaper is

[0073]

[0074] According to constraint equation (5), time t1 and t2 are known, and unknowns are A1 and A2. The penalty function for solving the ZV shaper pulse sequence can be constructed as:

[0075]

[0076] The objective function is:

[0077] value=g(1,1)+g(1,2) (7)

[0078] The ZVD input shaper requires a natural frequency ω n At the damping ratio ξ, the residual vibration expression is 0. In order to enhance the robustness of the shaper to the system vibration suppression, it is required that equation (1) is n The first-order derivative at is 0. The ZVD input shaper consists of three pulses. To minimize the pulse delay, t1=0. The unknowns are A1, A2, and A3. The constraint equation corresponding to the ZVD shaper is:

[0079]

[0080] According to formula (8), the penalty function of the ZVD pulse sequence can be constructed as follows:

[0081]

[0082] The objective function is:

[0083] value=g(1,1)+g(1,2)+g(1,3) (10)

[0084] The EI input shaper requires the residual vibration expression to be a non-zero value V exp , and above ω n The frequency ω h Its value is 0, and below ω n The frequency ω l Its value is also 0. Therefore, in [ω l ,ω h ], it can ensure that the shaper makes the residual vibration expression of the system less than or equal to V exp The EI shaper contains three pulse sequences. To minimize the time-delay pulse, we can get t1=0. The unknowns are A1, A2, A3, and t2. The constraint equation corresponding to the EI shaper is:

[0085]

[0086] According to formula (11), the penalty function can be constructed as follows:

[0087]

[0088] After constructing the penalty function, set the penalty factor σ (which can be customized. In this example, the value of σ is much larger than the weight 1 of g(1,5)). 10 For example, the objective function expression is as follows:

[0089]

[0090] In the penalty function formulas corresponding to the above types of shapers, abs[] represents the absolute value. It should be noted that, according to formula (1), the amplitude and time delay of the pulse sequence are also constraints that need to be satisfied for the residual vibration, so they should be taken into consideration when constructing the penalty function.

[0091] The objective function value here is also called the augmented objective function, which transforms the residual vibration suppression problem of each shaper from a constrained condition to an unconstrained condition solution. According to the principle of penalty function, the pulse sequence that can minimize the value of the objective function can also minimize the residual vibration in equation (1).

[0092] After obtaining the objective function, this application adopts an evolutionary algorithm based on a combination of composite vectors and control parameters (Differential Evolution with Composite Trial Vector Generation Strategies and Control Parameters, CoDE, hereinafter referred to as CoDE) to solve the pulse sequence of the input shaper. First, an initial pulse sequence group, also called a population, is generated from a feasible solution space by random sampling. The number of pulse sequences is greater than or equal to 4.

[0093] The CoDE algorithm is an improved genetic algorithm, which mainly includes mutation operations, crossover operations and selection operations on parent pulses. It should be noted that a pulse sequence is also called an individual in a population, and each pulse sequence can be represented as a vector vector. Therefore, an individual or a vector can be used to describe a pulse sequence. If the initial pulse sequence group generated by random sampling is recorded as the 0th generation population, the evolved pulse sequence group obtained by the evolutionary operation is the 1st generation population, and the evolved pulse sequence group obtained by the evolutionary operation of the 1st generation is the 2nd generation population, and so on. For the 1st generation, the pulse sequence in the 0th generation is the parent generation, and the 1st generation is the offspring of the 0th generation. For the 2nd generation, the pulse sequence in the 1st generation is the parent generation, and the 2nd generation is the offspring of the 1st generation, and so on.

[0094] The principle of the CoDE algorithm is explained in detail below.

[0095] Step 1: Mutation Operation

[0096] The most basic mutation component of DE (Differential Evolution) is the differential vector of the parent generation. There are five common DE mutation operators:

[0097]

[0098] Among them, r1, r2, r3, r4, r5∈{1, 2, ..., N} are different integers, and r1, r2, r3, r4, r5 are different from the current vector (in this application, the current pulse sequence) index i, the population size is greater than or equal to 4, and F is the scaling factor.

[0099] The name of the mutation operator is related to its function and characteristics. Rand means randomly selecting individuals in the population to participate in the mutation, best means selecting the best individual in the population to participate in the mutation, 1 means selecting one individual, and 2 means selecting two individuals. The above 5 mutation operators are understood as follows:

[0100] The rand / 1 mutation operator randomly selects three individuals in the population that are different from the target vector for differential mutation. Each individual in the population has the same probability of being selected, and the solution space near the selected individuals is likely to be explored in subsequent steps. This operator is one of the mutation operators with the largest search range and has a strong perturbation ability, but the convergence speed is relatively slow.

[0101] The best / 1 operator is an operator that mutates the best individual in history. The explored space is closely around the best individual, which makes the population gather faster. Therefore, this operator has a faster convergence speed and stronger local search ability, but poorer global search ability.

[0102] The current-to-best / 1 operator is an operator that uses the current individual as the basis vector for perturbation. The operator consists of a vector pointing to the optimal solution and a perturbation vector. It enables the current individual to explore the space around the best. The convergence speed of this operator is fast, but it loses a certain amount of search ability.

[0103] The best / 2 operator uses the best two individuals to participate in the mutation operation. When the scaling factor F remains unchanged, the optimal movement range is enlarged, and the global search capability is increased by sacrificing the convergence speed, but the search range is still limited to the vicinity of the optimal individual.

[0104] The rand / 2 operator adds a disturbance variable, which increases the individual's range of motion while keeping the scaling factor F unchanged, giving it a stronger search capability, but at the same time a slower convergence speed.

[0105] Step 2: Crossover Operation

[0106] For the target vector in the population (in this application, it refers to the current pulse sequence. When processing the current pulse sequence, it is understood to be the target vector currently being processed) (If the number of pulses contained in the input shaper is n, then is an n-dimensional vector, The superscript t in The subscript t+1 in the t+1 represents the parent and child generations), which will be compared with the mutation vector Perform crossover operation to generate experimental individuals (In this application, it refers to the evolved pulse sequence obtained by calculating the current pulse sequence through the mutation strategy). The evolution of is first performed by random selection, so that At least one For other bits, a crossover probability factor CR can be used to determine Which one of Contribution, who Contribution. The equation for the crossover operation is as follows:

[0107]

[0108] Where rand(j)∈[0,1] is a uniformly distributed random number, j represents the jth variable, CR is the crossover probability factor, the value range is [0,1], and the size is predetermined. randn(i)∈[1,2,…,D], where D is the dimension of the solution, that is, the number of numbers contained in a set of solutions. is the index of the randomly selected variable to ensure At least one variable is represented by Contribute, otherwise It is possible to The same cannot generate new individuals (referring to evolved individuals). From the formula, we can see that if CR is larger, then right The more contribution, when CR=1, It is beneficial to maintain the diversity of the population and global search; if CR is smaller, right The more contribution, when CR=0, It is beneficial to local search and accelerate the convergence rate. In actual operation, the size of CR can be reasonably selected according to the solution needs.

[0109] Step 3: Select an operation

[0110] DE adopts a "greedy" search strategy, and the individuals generated after mutation and crossover operations compete. Taking minimization optimization as an example, the evolution pulse sequence with a small objective function value is selected, and the equation for the selection operation is (f() in formula (16) is the objective function value corresponding to the shaper:

[0111]

[0112] From the above content, it can be seen that different mutation operators, scaling factors and crossover probability factors will produce different offspring individuals. Therefore, this application adopts the method of setting mutation strategies (each group of mutation strategies includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR) to improve the genetic algorithm. For ease of understanding, please refer to Figure 2 As shown, it is a schematic diagram of the mutation strategy, and the combination of the scaling factor and the crossover probability factor is recorded as the control parameter. Different mutation operators and control parameter settings have different characteristics, so they can complement each other to solve different types of problems. That is, the mutation strategy in this application can reasonably set the mutation operator and control parameter combination according to the actual solution needs to meet the solution requirements. For example, for three groups of mutation strategies, the first group of control parameters can be set to solve separable problems; the second group of control parameter settings mainly maintain the diversity of the group, and make the three experimental individual generation strategies have stronger global search capabilities; the third group of control parameter settings encourage the mutation operator to mine in the search space to speed up the convergence of the group.

[0113] For example, three mutation operators and three sets of control parameter settings are selected to construct knowledge bases respectively. Figure 2 The r and s of are both equal to 3. Formula (17) is an example of a scheme in which the mutation operator randomly couples three control parameter settings to generate experimental individuals.

[0114]

[0115] Based on the CoDE algorithm, each target vector individual generates three test individuals (mutated pulse sequences), compares the three test individuals with the parent individual, and selects the individual with the smallest objective function to enter the next generation. Among them, the individual with the smallest objective function may be the parent individual or the test individual.

[0116] The following is an example of the steps to solve the shaper pulse sequence based on CoDE:

[0117] 1) G = 0; / / Evolutionary algebra

[0118] 2) Generate N from a feasible solution space by random sampling P = 30 initial pulse sequences N P Group pulse sequence, denoted as: P0;

[0119] 3) Use the penalty function to penalize each inequality constraint and make V(ω n ,ξ) reaches the minimum value, and the objective function value is obtained by constructing the objective function:

[0120] 4) FES = N P ;

[0121] 5) while FES<Max_FES do / / Max_FES indicates the maximum number of pulse sequences tried;

[0122] 6) / / Initially, the first generation pulse sequence group is an empty set

[0123] 7) for i = 1: N P do

[0124] 8) For the pulse sequence with index i, a set of control parameters is selected from [F = 1.0, CR = 0.1], [F = 1.0, CR = 0.9], [F = 0.8, CR = 0.2], and the pulse sequence is respectively calculated through the three selected mutation operators "DE / rand / 1", "DE / rand / 2", and "DE / current-to-rand / 1" to calculate the experimental individuals (pulse sequences) to generate three sets of evolutionary pulse sequences

[0125] 9) Calculate these three pulse sequences The objective function value of

[0126] 10) Select the best pulse sequence from the three pulse sequences in 9), denoted as This is the target vector for the next generation;

[0127] 11)

[0128] 12) FES = FES + 3

[0129] 13) end for

[0130] 14) G = G + 1

[0131] 15) end while

[0132] 16)Output: A set of pulse sequences with the minimum objective function value

[0133] First, the CoDE algorithm randomly generates the initial generation of pulse sequence populations and records them as generation 0. In the evolution to the first generation, a set of control parameters is randomly selected, and each set of pulse sequences can generate three evolved vectors (three sets of evolved pulse sequences) through three selected mutation operators. The three newly generated vectors are compared with the objective function of the original target vector (i.e., the original set of pulse sequences). The set of pulse sequences with the minimum objective function is taken as the first generation of pulse sequences. All 30 sets of pulse sequences in generation 0 need to complete the evolution to the first generation. In this way, the optimal solution can be obtained.

[0134] In each operation cycle of the CoDE algorithm, multiple sets of control parameters and multiple mutation operators are comprehensively applied to obtain multiple evolutionary pulse sequences. The performance of the two complements each other, effectively accelerating the convergence speed while ensuring the accuracy of the solution.

[0135] A practical application of the solution provided by the present application is described below. When the solution of the present application is applied to vibration suppression of an OCT (Optical Coherence Tomography) imaging system, a significant suppression effect is achieved.

[0136] In the OCT system, the monitor fixed on the desktop has multiple degrees of freedom to move, and its vibration is particularly obvious. The vibration of the monitor is selected to evaluate the vibration suppression effect. During the use of OCT, the vibration is most severe during the continuous small displacement movement of the focusing process. The motor micro-feed working condition is as follows: the motor acceleration time is 30ms, the deceleration time is 30ms, the uniform speed time is 90ms, the running distance is 0.05 circles, and it runs in the same direction for 20 times continuously. Under the working conditions of the above parameters, if Figure 3 This is the vibration curve of the display without EI input shaping. The vibration frequency is about 7Hz, and the damping ratio is estimated to be 0.28. At this time, the vibration peaks of the display space in the three directions of XYZ are X=0.1001, Y=0.0727, and Z=0.0830.

[0137] Solving ω using CoDE n=7,ξ=0.28, the EI shaper pulse sequence is obtained:

[0138]

[0139] Figure 4 is the vibration curve of the display after adding the EI shaper. After adding the EI shaper, the peak-to-peak values ​​of the vibration in the three directions of the display space XYZ are respectively: EI =0.0254, Y EI =0.0264, Z EI = 0.0288. It can be seen that the vibration of the display is significantly suppressed, and the vibration attenuation ratio is about 60%.

[0140] Based on the same inventive concept, an embodiment of the present invention provides a residual vibration suppression device. The specific implementation of the target function processing and pulse sequence evolution method in the device can be found in the description of the method embodiment part, and the repeated parts will not be repeated. Figure 5 FIG. 1 is a schematic diagram of a residual vibration suppression device provided by an embodiment of the present invention, comprising:

[0141] An acquisition unit 51 is used to penalize the constraint equation of the residual vibration to obtain the objective function of the residual vibration;

[0142] An initialization unit 52, configured to generate an initial pulse sequence group including at least four initial pulse sequences;

[0143] An evolution unit 53, used for traversing the initial pulse sequence group, performing an evolution operation on any initial pulse, and obtaining an evolved pulse sequence group;

[0144] The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse;

[0145] A judging unit 54, used to judge whether the evolution operation termination condition is triggered;

[0146] A determination unit 55, configured to use the evolved pulse sequence group obtained in the evolution unit as a target pulse sequence group for an input shaper when an evolution operation termination condition is triggered;

[0147] The circulation unit 56 is used for, when the evolution operation termination condition is not triggered, using the evolution pulse sequence group obtained by the evolution unit as the initial sequence group to trigger the evolution unit to perform the evolution operation.

[0148] In the embodiment, the constraint equation of the residual vibration is penalized, and the constraint problem of solving the residual vibration is converted into an unconstrained solution problem, and the specific input shaper is not distinguished, which has universality. Secondly, an evolution operation is performed on the pulse sequence of the input shaper, and the differential evolution algorithm DE is used in the evolution operation, but it is different from the existing DE algorithm. In this application, when calculating the evolved pulse sequence, multiple groups of mutation strategies are used (each group of mutation strategies includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR). By reasonably setting each group of mutation strategies, the characteristics of each group of mutation strategies can be complemented to each other, so as to achieve a balance between the group diversity, global search capability and group convergence speed in the pulse sequence solution process.

[0149] refer to Figure 6 As shown, an embodiment of the present invention also provides an input shaper, including a memory 61, a transceiver 62, and a processor 63:

[0150] The memory 61 is used to store computer programs; the transceiver 62 is used to send and receive data under the control of the processor 63; the processor 63 is used to perform the following steps:

[0151] Step A: Penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration;

[0152] Step B: generating an initial pulse sequence group including at least four initial pulse sequences;

[0153] Step C: traverse the initial pulse sequence group, perform an evolution operation on any initial pulse, and obtain an evolved pulse sequence group;

[0154] The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse;

[0155] Step D: Determine whether the evolution operation termination condition is triggered. If so, use the evolved pulse sequence group obtained in step C as the target pulse sequence group of the input shaper; if not, use the evolved pulse sequence group obtained in step C as the initial sequence group and execute step C.

[0156] The processor 63 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or a complex programmable logic device (CPLD). The processor may also adopt a multi-core architecture.

[0157] Based on the same inventive concept, an embodiment of the present invention further provides a processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, and the computer program is used to enable the processor to execute the implementation method of the input shaper as described above.

[0158] It should be noted here that the above-mentioned input shaper provided in the embodiment of the present invention can implement all the method steps implemented in the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as those in the method embodiment will not be described in detail here.

[0159] It should be noted that the division of units in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional unit in each embodiment of the present application may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0160] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present application. The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic storage (such as floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)), etc.

[0161] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0162] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A residual vibration suppression method, characterized in that: The method includes: Step A: Penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration; Step B: generating an initial pulse sequence group including at least four initial pulse sequences; Step C: traverse the initial pulse sequence group, perform an evolution operation on any initial pulse, and obtain an evolved pulse sequence group; The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse; Step D: judging whether the evolution operation termination condition is triggered, if so, taking the evolved pulse sequence group obtained in step C as the target pulse sequence group of the input shaper; if not, taking the evolved pulse sequence group obtained in step C as the generated initial sequence group and executing step C; The step A comprises: Penalize the constraint equation of the residual vibration corresponding to the input shaper, and transform the constraint problem of residual vibration solution into an unconstrained solution problem; Each set of mutation strategies includes: mutation operation, crossover operation and selection operation; The mutation operation is configured with a DE mutation operator, and the DE mutation operator includes: Among them, r1, r2, r3, r4, r5∈{1, 2, ..., N} are different integers, and r1, r2, r3, r4, r5 are different from the index i of the current pulse sequence, the population size is greater than or equal to 4, F is the scaling factor, Rand means randomly selecting individuals in the population to participate in the mutation, best means selecting the best individual in the population to participate in the mutation, 1 means selecting one individual, and 2 means selecting two individuals; The crossover operation configuration has the crossover operation equation: Among them, rand(j)∈[0,1] is a uniformly distributed random number, j represents the jth variable, CR is the crossover probability factor, the value range is [0,1], the size is predetermined, randn(i)∈[1,2,…,D], D is the dimension of the solution, refers to the evolved pulse sequence obtained by calculating the mutation strategy of the current pulse sequence. is the mutation vector, is the target vector; The selection operation configuration has a selection operation equation: Among them, f() is the objective function corresponding to the shaper.

2. The method according to claim 1, characterized in that The evolution operation termination condition includes that the execution times of step C reaches a preset number of times.

3. The method according to claim 1, characterized in that The evolution operation termination condition includes that the objective function value of the evolution pulse sequence obtained in step C is less than or equal to a preset value.

4. The method according to claim 1, characterized in that The residual vibration expression is: Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, cos is the cosine calculation, sin is the sine calculation, ∑ is the sum calculation, and n is the number of pulses contained in the input shaper; A i is the amplitude of the pulse train, t i is the time delay corresponding to the pulse train, where t i <t i+1 .

5. The method according to claim 4, characterized in that When the input shaper is a ZVD multiple differential input shaper, the constraint equation of the residual vibration is: Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, n is the number of pulses contained in the input shaper, q represents the derivative order, q=n-2, n≥2.

6. The method according to claim 4, characterized in that When the input shaper is a single-peak EI input shaper, the constraint equation of the residual vibration is: Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, n is the number of pulses contained in the input shaper, V exp is the preset limit value of residual vibration, ω h is greater than ω n The preset frequency, ω l is less than ω n preset frequency.

7. The method according to claim 4, characterized in that When the input shaper is a multi-peak EI input shaper, the constraint equation of the residual vibration is: in, is the residual vibration, ω m is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, m is the number of pulses contained in the input shaper, and K is the number of peaks of the sensitivity curve of the EI shaper.

8. A residual vibration suppression device, characterized in that: The device includes: An acquisition unit, used for penalizing the constraint equation of the residual vibration to obtain the objective function of the residual vibration; An initialization unit, used for generating an initial pulse sequence group including at least four initial pulse sequences; An evolution unit, used for traversing the initial pulse sequence group, performing an evolution operation on any initial pulse, and obtaining an evolved pulse sequence group; The evolution operation includes: according to a plurality of preset mutation strategies, calculating the pulse sequence of the current pulse sequence after each mutation strategy is applied; respectively calculating the objective function of the current pulse sequence and the objective function of each mutated pulse sequence; selecting the pulse sequence with the smallest objective function as the evolved pulse sequence of the current pulse sequence; wherein each mutation strategy includes: a differential evolution algorithm DE mutation operator, a scaling factor F and a crossover probability factor CR; the current pulse sequence is any initial pulse sequence in the initial pulse sequence group; the evolved pulse sequence group is composed of the evolved pulses of all pulse sequences in the initial pulse; A judgment unit, used to judge whether the evolution operation termination condition is triggered; a determination unit, configured to use the evolved pulse sequence group obtained in the evolution unit as the target pulse sequence group of the input shaper when the evolution operation termination condition is triggered; A cyclic unit, used for triggering the evolution unit to perform the evolution operation by using the evolution pulse sequence group obtained by the evolution unit as the generated initial sequence group when the evolution operation termination condition is not triggered; The acquisition unit is also used to penalize the constraint equation of the residual vibration corresponding to the input shaper, and convert the constraint problem of residual vibration solution into an unconstrained solution problem; Each set of mutation strategies includes: mutation operation, crossover operation and selection operation; The mutation operation is configured with a DE mutation operator, and the DE mutation operator includes: Among them, r1, r2, r3, r4, r5∈{1, 2, ..., N} are different integers, and r1, r2, r3, r4, r5 are different from the index i of the current pulse sequence, the population size is greater than or equal to 4, F is the scaling factor, Rand means randomly selecting individuals in the population to participate in the mutation, best means selecting the best individual in the population to participate in the mutation, 1 means selecting one individual, and 2 means selecting two individuals; The crossover operation configuration has the crossover operation equation: Among them, rand(j)∈[0,1] is a uniformly distributed random number, j represents the jth variable, CR is the crossover probability factor, the value range is [0,1], the size is predetermined, randn(i)∈[1,2,…,D], D is the dimension of the solution, refers to the evolved pulse sequence obtained by calculating the mutation strategy of the current pulse sequence. is the mutation vector, is the target vector; The selection operation configuration has a selection operation equation: Among them, f() is the objective function corresponding to the shaper.

9. The device according to claim 8, characterized in that The residual vibration expression is: Among them, V(ω n ,ξ) is the residual vibration,ω n is the natural frequency of the controlled object, ξ is the damping ratio of the controlled object, cos is the cosine calculation, sin is the sine calculation, ∑ is the sum calculation, and n is the number of pulses contained in the input shaper; A i is the amplitude of the pulse train, t i is the time delay corresponding to the pulse train, where t i <t i+1 .

10. An input shaper, characterized in that Including memory, transceiver, processor: Memory for storing computer programs; a transceiver, for transmitting and receiving data under the control of the processor; A processor, configured to read the computer program in the memory and execute the method according to any one of claims 1 to 7.

11. A processor-readable storage medium, characterized in that: The processor-readable storage medium stores a computer program, and the computer program is used to enable the processor to execute the method according to any one of claims 1 to 7.

Images