Automatic vibration isolation method and device

Abstract

The invention provides an automatic vibration isolation method and device which can be applied to the technical field of vibration isolation. The method comprises the steps that a sensor is used for collecting a vibration signal, and the vibration signal is obtained by conducting vibration interference on a controlled platform through an external vibration interference source; comparing the vibration signal with the pre-sampling signal to obtain a comparison result; under the condition that the comparison result represents that the vibration signal is similar to the pre-sampling signal, a control parameter corresponding to the pre-sampling signal is determined, and the control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampling signal through an optimization model; generating a control signal according to the control parameter and the real-time vibration signal; an actuator is used for responding to the control signal to apply vibration isolation force to the controlled platform, and the vibration isolation force is used for restraining vibration interference generated by an external vibration interference source to the controlled platform.

Description

Technical Field

[0001] The present disclosure relates to the field of vibration isolation technology, and more particularly to an automatic vibration isolation method and device. Background Art

[0002] Active vibration isolation primarily consists of four components: a sensor, a controller, an actuator, and a controlled platform. The sensor's vibration signal is processed by the controller and then output to the actuator. The actuator then applies a corresponding force to the controlled platform, counteracting the original vibration.

[0003] In related technologies, it is usually necessary to obtain a response function associated with the controlled platform in order to design a corresponding vibration isolation control algorithm for control. However, due to the complexity of the theoretical model, the output of the theoretical model deviates from reality, resulting in a discrepancy between the design of the vibration isolation control algorithm and the actual vibration isolation effect. Furthermore, when the theoretical model has a large number of parameters, excessive human resources are required to integrate and process the parameters. Furthermore, it is difficult to achieve the optimal vibration isolation effect based on the theoretical model, which limits the improvement of the vibration isolation effect. Summary of the Invention

[0004] In view of the above problems, the present disclosure provides an automatic vibration isolation method and device.

[0005] According to a first aspect of the present disclosure, an automatic vibration isolation method is provided, comprising: collecting a vibration signal using a sensor, wherein the vibration signal is obtained by causing an external vibration interference source to interfere with a controlled platform; comparing the vibration signal with a pre-sampled signal to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determining a control parameter corresponding to a preset signal, wherein the control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampled signal using an optimization model; generating a control signal based on the control parameter and a real-time vibration signal; and applying a vibration isolation force to the controlled platform using an actuator in response to the control signal, wherein the vibration isolation force is used to suppress vibration interference caused by the external vibration interference source to the controlled platform.

[0006] According to an embodiment of the present disclosure, the initial control parameters are optimized using an optimization model to obtain the pre-sampling control parameters, which is performed based on the following operations: constructing a Gaussian distribution based on the initial control parameters, the variable control parameters and the initial control effect, the initial control effect is associated with the initial control parameters, and the variable control parameters are associated with the initial control parameters; based on the Gaussian distribution, predicting the variable control parameters to obtain a predicted control effect; and when the predicted control effect meets the first preset optimization condition, determining the variable control parameters as the pre-sampling control parameters.

[0007] According to an embodiment of the present disclosure, a Gaussian distribution is constructed based on initial control parameters, variable control parameters and initial control effects, including: obtaining an initial tabletop residual vibration signal corresponding to the initial control parameters, wherein the initial tabletop residual vibration signal is obtained by applying an initial vibration isolation force to the controlled platform based on the initial control parameters, and the initial vibration isolation force is generated according to the initial control parameters; determining the initial control effect corresponding to the initial control parameters based on the Allan deviation of the initial tabletop residual vibration signal; generating a kernel function associated with the Gaussian distribution based on the initial control parameters and the variable control parameters; and constructing a Gaussian distribution based on the kernel function, the variable control parameters, the initial control parameters and the initial control effect.

[0008] According to an embodiment of the present disclosure, a Gaussian distribution is constructed based on the kernel function, variable control parameters, initial control parameters and initial control effects, including: constructing a correlation matrix of the variable control parameters and the initial control parameters to obtain a first kernel function; constructing an autocorrelation matrix of the variable control parameters to obtain a second kernel function, wherein the first kernel function and the second kernel function have the same hyperparameters; optimizing the hyperparameters to obtain target hyperparameters; updating the first kernel function and the second kernel function according to the target hyperparameters to obtain updated first kernel function and second kernel function; determining the kernel function associated with the Gaussian distribution based on the updated first kernel function and second kernel function.

[0009] According to an embodiment of the present disclosure, hyperparameters are optimized to obtain target hyperparameters, including: constructing an autocorrelation matrix of initial control parameters to obtain a third kernel function; constructing a likelihood function based on the hyperparameters, the third kernel function and the initial control effect; optimizing the hyperparameters to obtain optimized hyperparameters; calculating the likelihood function value of the likelihood function based on the optimized hyperparameters; and determining the optimized hyperparameters as target hyperparameters when the likelihood function value meets a second preset optimization condition.

[0010] According to an embodiment of the present disclosure, when the predicted control effect satisfies a first preset optimization condition, the variable control parameter is determined as a pre-sampling control parameter, including: analyzing the predicted control effect using an expected improvement function to obtain an analysis result; when the analysis result indicates that the predicted control effect satisfies a third preset optimization condition, the variable control parameter is determined as a pre-sampling control parameter.

[0011] According to an embodiment of the present disclosure, a control signal is generated based on pre-sampled control parameters and vibration signals, including: using a controller to analyze the pre-sampled control parameters and vibration signals to determine a control strategy; and generating a control signal suitable for an actuator based on the control strategy and the vibration signal.

[0012] According to an embodiment of the present disclosure, the above method also includes: using a target identification model to perform parameter identification processing on the vibration signal to obtain a target vibration signal, wherein the parameters of the target identification model are obtained by optimizing the initial parameters of the target identification model using an optimization model; wherein, comparing the vibration signal and the pre-sampled signal to obtain a comparison result includes: comparing the target vibration signal and the pre-sampled signal to obtain a comparison result.

[0013] According to an embodiment of the present disclosure, the above method further includes:

[0014] A sensor is used to collect a residual vibration signal of the table top, which is obtained by applying a vibration isolation force to the controlled platform; the residual vibration signal of the table top is analyzed to obtain a table top analysis result; when the table top analysis result indicates a vibration isolation anomaly or the comparison result indicates that the vibration signal and the pre-sampled signal are not similar, an optimization model is used to optimize the pre-sampled control parameters to obtain target control parameters; a target control signal is generated based on the target control parameters and the vibration signal; an actuator is used to apply a target vibration isolation force to the controlled platform, wherein the actuator generates the target vibration isolation force in response to the target control signal.

[0015] A second aspect of the present disclosure provides an automatic vibration isolation device, comprising: a controlled platform; a sensor for collecting a vibration signal, wherein the vibration signal is obtained by an external vibration interference source causing vibration interference to the controlled platform; a controller for comparing the vibration signal with a pre-sampled signal to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determining a control parameter corresponding to the pre-sampled signal, wherein the control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampled signal using an optimization model; generating a control signal based on the pre-sampled control parameter and the vibration signal; and an actuator for generating a vibration isolation force in response to the control signal, and applying the vibration isolation force to the controlled platform, wherein the vibration isolation force is used to suppress vibration interference caused by the external vibration interference source to the controlled platform.

[0016] According to an embodiment of the present disclosure, a sensor is used to collect a vibration signal; the vibration signal and the pre-sampled signal are compared to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, a control parameter corresponding to the pre-sampled signal is determined, wherein the control parameter is obtained by optimizing the initial control parameter corresponding to the pre-sampled signal using an optimization model; a control signal is generated based on the control parameter; and an actuator is used to apply a vibration isolation force to the controlled platform in response to the control signal. Since the initial control parameter corresponding to the pre-sampled signal is optimized using an optimization model, there is no need to consider the difference in complexity between the theoretical model and the actual response function associated with the controlled platform, and the mapping relationship between the control parameter and the optimization effect can be automatically obtained based on the data of the training process. In addition, in actual use, the optimized preset control parameters can be called directly by comparing the vibration signal and the pre-sampled signal, avoiding the waste of resources caused by the need to debug the parameters, thereby achieving the effect of improving the efficiency of automatic vibration isolation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0018] Figure 1 Schematically shows a flow chart of an automatic vibration isolation method according to an embodiment of the present disclosure;

[0019] Figure 2 Schematically shows an optimization diagram based on a back-end sensor according to an embodiment of the present disclosure;

[0020] Figure 3 Schematically shows an optimization diagram based on a back-end sensor and a front-end sensor according to an embodiment of the present disclosure;

[0021] Figure 4 Schematically shows an optimization diagram of a target recognition model according to an embodiment of the present disclosure;

[0022] Figure 5 Schematically shows a structural block diagram of an automatic vibration isolation device according to an embodiment of the present disclosure;

[0023] Figure 6 A schematic diagram of an automatic vibration isolation device according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0024] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0025] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0026] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0027] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0028] An embodiment of the present disclosure provides an automatic vibration isolation method, comprising: collecting a vibration signal using a sensor, wherein the vibration signal is obtained by an external vibration interference source causing vibration interference to a controlled platform; comparing the vibration signal with a pre-sampled signal to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determining a control parameter corresponding to the pre-sampled signal, wherein the control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampled signal using an optimization model; generating a control signal based on the control parameter; and applying a vibration isolation force to the controlled platform using an actuator in response to the control signal, wherein the vibration isolation force is used to suppress the vibration interference caused to the controlled platform by the external vibration interference source.

[0029] Figure 1 A flow chart of an automatic vibration isolation method according to an embodiment of the present disclosure is schematically shown.

[0030] like Figure 1 As shown, the automatic vibration isolation method of this embodiment includes operations S110 to S150.

[0031] In operation S110 , a sensor is used to collect a vibration signal, wherein the vibration signal is obtained by an external vibration interference source causing vibration interference on a controlled platform.

[0032] According to an embodiment of the present disclosure, the sensor may be a vibration sensor, such as a displacement meter, a speedometer, an accelerometer or other vibration measuring equipment, for collecting a vibration signal generated by an external vibration interference source causing vibration interference on a controlled platform.

[0033] According to an embodiment of the present disclosure, the vibration interference caused by the external vibration interference source to the controlled platform may be caused by the natural environment or may be artificially imposed on the controlled platform.

[0034] According to the embodiments of the present disclosure, the controlled platform can be a platform of any material and any size, can be affected by external vibration interference sources to produce targeting, and can also receive the force applied by the actuator.

[0035] In operation S120 , the vibration signal and the pre-sampled signal are compared to obtain a comparison result.

[0036] According to an embodiment of the present disclosure, the pre-sampled signal may be at least one training vibration signal selected and recorded during a training process.

[0037] According to an embodiment of the present disclosure, the vibration signal and the pre-sampled signal are compared to determine whether the currently collected vibration signal has consistent or similar spectral characteristics with the recorded pre-sampled signal.

[0038] In operation S130 , when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, a control parameter corresponding to the pre-sampled signal is determined, wherein the control parameter is obtained by optimizing the initial control parameter corresponding to the pre-sampled signal using an optimization model.

[0039] According to an embodiment of the present disclosure, if the comparison result indicates that the vibration signal and the pre-sampled signal are similar, that is, the currently collected vibration signal can be used to call the control parameters corresponding to the pre-sampled signal obtained during the training process. According to an embodiment of the present disclosure, the optimization model can be a Gaussian process regression model, which is used to fully utilize the historical control parameters during the training process based on the Bayesian optimization algorithm, and can achieve relatively good optimization results with a smaller number of optimization times.

[0040] In operation S140 , a control signal is generated according to the pre-sampled control parameter.

[0041] According to an embodiment of the present disclosure, the pre-sampled control parameters may be obtained by training an optimization model, so that the actuator can apply a vibration isolation force to the controlled platform based on the control parameters to suppress vibration interference from external vibration interference sources.

[0042] According to an embodiment of the present disclosure, the initial control parameter may be a control parameter randomly generated according to a pre-sampled signal during a training process.

[0043] According to an embodiment of the present disclosure, the pre-sampled control signal may be obtained based on the pre-sampled control parameter, so that the actuator can receive and obtain an output strategy for adjusting the voltage or current corresponding to the pre-sampled control parameter.

[0044] In operation S150 , a vibration isolation force is applied to the controlled platform by an actuator in response to the pre-sampled control signal, wherein the vibration isolation force is used to suppress vibration interference generated by an external vibration interference source on the controlled platform.

[0045] According to an embodiment of the present disclosure, the actuator may be a voice coil motor, a Lorentz motor, a piezoelectric ceramic, a hydraulic device, a shape memory alloy, or the like, which is a device that outputs a corresponding force or displacement controlled by voltage or current.

[0046] According to an embodiment of the present disclosure, the actuator applies a vibration isolation force to the controlled platform to suppress vibration interference from an external vibration interference source.

[0047] According to an embodiment of the present disclosure, a sensor is used to collect a vibration signal; the vibration signal and the pre-sampled signal are compared to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, the control parameters corresponding to the pre-sampled signal are determined, wherein the pre-sampled control parameters are obtained by optimizing the initial control parameters corresponding to the pre-sampled signal using an optimization model; a control signal is generated based on the pre-sampled control parameters; and an actuator is used to apply a vibration isolation force to the controlled platform in response to the control signal. Since the initial control parameters corresponding to the pre-sampled signal are optimized using an optimization model, there is no need to consider the difference in complexity between the theoretical model and the actual response function associated with the controlled platform, and the mapping relationship between the control parameters and the optimization effect can be automatically obtained based on the data of the training process. In addition, in actual use, the optimized control parameters can be called directly by comparing the vibration signal and the pre-sampled signal, avoiding the waste of resources caused by the need to debug the parameters, thereby achieving the effect of improving the efficiency of automatic vibration isolation.

[0048] According to an embodiment of the present disclosure, the optimization model is used to optimize the initial control parameters to obtain the pre-sampling control parameters, which is performed based on the following operations: constructing a Gaussian distribution based on the initial control parameters, the variable control parameters and the initial control effect, the initial control effect is associated with the initial control parameters, and the variable control parameters are associated with the initial control parameters; based on the Gaussian distribution, predicting the variable control parameters to obtain a predicted control effect; when the predicted control effect meets the first preset optimization condition, determining the variable control parameters as the pre-sampling control parameters

[0049] According to an embodiment of the present disclosure, the initial control parameter may be a preset control parameter for adjusting the actuator.

[0050] According to an embodiment of the present disclosure, the variable control parameter may be at least one control parameter obtained by adjusting the initial control parameter.

[0051] According to an embodiment of the present disclosure, the initial control effect is an effect obtained according to vibration feedback of the controlled platform after adjusting the actuator based on the initial control parameters.

[0052] According to an embodiment of the present disclosure, a Gaussian distribution is constructed based on the initial control parameters, the variable control parameters, and the initial control effect, so as to obtain a distribution relationship between the control parameters and the control effect.

[0053] According to an embodiment of the present disclosure, the control effect of the variable control parameter is derived based on Gaussian distribution to obtain a predicted control effect.

[0054] According to an embodiment of the present disclosure, the optimization model may be configured to perform multiple rounds of optimization, each round having an intermediate control parameter and an intermediate control effect corresponding to the intermediate control parameter.

[0055] According to an embodiment of the present disclosure, the first preset optimization condition may be to compare the intermediate control effects of all rounds, select an optimal intermediate control effect, and determine the intermediate control parameters corresponding to the optimal intermediate control effect as pre-sampling control parameters.

[0056] According to the embodiments of the present disclosure, by pre-optimizing the control parameters during the training process, the obtained pre-sampled control parameters can be directly called in actual applications, thereby reducing the response time of the system and achieving a rapid vibration isolation effect.

[0057] According to an embodiment of the present disclosure, a Gaussian distribution is constructed based on initial control parameters, variable control parameters and initial control effects, including: obtaining an initial tabletop residual vibration signal corresponding to the initial control parameters, wherein the initial tabletop residual vibration signal is obtained by applying an initial vibration isolation force to the controlled platform based on the initial control parameters, and the initial vibration isolation force is generated according to the initial control parameters; determining the initial control effect corresponding to the initial control parameters based on the Allan deviation of the initial tabletop residual vibration signal; generating a kernel function associated with the Gaussian distribution based on the initial control parameters and the variable control parameters; and constructing a Gaussian distribution based on the kernel function, the variable control parameters, the initial control parameters and the initial control effect.

[0058] According to an embodiment of the present disclosure, obtaining an initial feedback signal corresponding to the initial control parameters may be to generate an initial control signal based on the initial control parameters, so that the actuator applies an initial vibration isolation force to the controlled platform based on the initial control signal. Under the action of an external interference source and the initial vibration isolation force, the controlled platform will generate a new vibration state, and the sensor will collect the vibration state to obtain the initial feedback signal.

[0059] According to an embodiment of the present disclosure, the Allan deviation may be calculated by calculating the deviation between vibration signals of the controlled platform every 1 second after applying the initial vibration isolation force to the controlled platform, and the deviation may be determined as the initial control effect corresponding to the initial control parameter.

[0060] According to the embodiment of the present disclosure, a Gaussian distribution is constructed based on the kernel function, the variable control parameter, the initial control parameter and the initial control effect. The initial control parameter can be recorded as , and get the corresponding initial control effect Wherein, the Gaussian distribution is shown in formula (1).

[0061] (1);

[0062] in, represents a Gaussian distribution, Indicates variable control parameters, represents the predictive control effect, All indicate that the parameters are controlled by variables and initial control parameters Generated kernel function.

[0063] According to an embodiment of the present disclosure, a Gaussian distribution is constructed based on the kernel function, variable control parameters, initial control parameters and initial control effects, including: constructing a correlation matrix of the variable control parameters and the initial control parameters to obtain a first kernel function; constructing an autocorrelation matrix of the variable control parameters to obtain a second kernel function, wherein the first kernel function and the second kernel function have the same hyperparameters; optimizing the hyperparameters to obtain target hyperparameters; updating the first kernel function and the second kernel function according to the target hyperparameters to obtain updated first kernel function and second kernel function; and determining the kernel function associated with the Gaussian distribution based on the updated first kernel function and second kernel function.

[0064] According to an embodiment of the present disclosure, the correlation matrix of the variable control parameter and the initial control parameter can be constructed by using multiple variable sub-control parameters of the variable control parameter and multiple initial sub-control parameters of the initial control parameter to obtain the first kernel function , used to analyze the correlation between variable control parameters and initial control parameters.

[0065] According to an embodiment of the present disclosure, the autocorrelation matrix of the variable control parameters can be constructed by using multiple sets of variable control parameters to obtain the second kernel function , used to analyze the correlation between multiple groups of variable control parameters.

[0066] According to an embodiment of the present disclosure, the hyperparameters of the kernel function are used to control the size of the kernel function, and the Gaussian distribution can be adjusted by adjusting the hyperparameters.

[0067] For example, the second kernel function may be as shown in formula (2).

[0068] (2);

[0069] in, Represents the second kernel function ,in, is the control parameter of the i-th group of variables, is the jth group of variable control parameters, and They are all hyperparameters. It can be seen that the size of the second kernel function depends on the size of the hyperparameters.

[0070] According to an embodiment of the present disclosure, the hyperparameters can be optimized by adjusting the numerical values ​​of the hyperparameters to update the first kernel function and the second kernel function, and then the kernel function constructed based on the updated first kernel function and the second kernel function can output an accurate predictive control effect through the above-mentioned Gaussian distribution to obtain the optimal control parameters.

[0071] According to an embodiment of the present disclosure, hyperparameters are optimized to obtain target hyperparameters, including: constructing an autocorrelation matrix of initial control parameters to obtain a third kernel function; constructing a likelihood function based on the hyperparameters, the third kernel function and the initial control effect; optimizing the hyperparameters to obtain optimized hyperparameters; calculating the likelihood function value of the likelihood function based on the optimized hyperparameters; and determining the optimized hyperparameters as target hyperparameters when the likelihood function value meets a second preset optimization condition.

[0072] According to an embodiment of the present disclosure, the autocorrelation matrix of the initial control parameters can be constructed by using multiple sets of initial control parameters to obtain the third kernel function , which is used to analyze the autocorrelation between multiple sets of initial control parameters.

[0073] According to an embodiment of the present disclosure, the likelihood function constructed based on the hyperparameters, the third kernel function and the initial control effect is It can be shown as formula (3).

[0074] (3);

[0075] Where N represents the normal distribution, and by optimizing the value of the hyperparameter, The likelihood function value of is maximized to meet the second preset optimization condition, and the hyperparameter corresponding to the maximum likelihood function value is determined as the target hyperparameter.

[0076] According to the embodiments of the present disclosure, by optimizing the hyperparameters based on the likelihood function, the constructed kernel function can more accurately analyze the correlation between the various parameters, thereby enabling the Gaussian distribution to obtain the best preset control parameters.

[0077] According to an embodiment of the present disclosure, when the predicted control effect satisfies a first preset optimization condition, the variable control parameter is determined as a pre-sampling control parameter, including: analyzing the predicted control effect using an expected improvement function to obtain an analysis result; when the analysis result indicates that the predicted control effect satisfies a third preset optimization condition, the variable control parameter is determined as a pre-sampling control parameter.

[0078] According to an embodiment of the present disclosure, multiple rounds of optimization may be performed before determining the predictive control parameters. In order to obtain new intermediate control parameters in the new round, the control parameters may be collected using an acquisition function, and the acquisition function includes at least one of a random sampling function, an expected improvement function, a probability improvement function, and a Thompson sampling function.

[0079] The expected improvement (EI) function can be shown as formula (4).

[0080] (4);

[0081] in, Represents expectations, It is the optimal predictive control effect in the current round and historical rounds. It is the variable control parameter corresponding to the optimal predictive control effect in the current round and historical rounds.

[0082] Further, based on the expectation of improvement The variable control parameters at the maximum value, determine the new variable control parameters , and based on the new variable control parameters, determine the new predictive control effect of the controlled platform measured by the sensor .

[0083] The variable control data set consists of Expanded to According to the expanded variable control data set, the hyperparameters of the kernel function can be updated, and then the Gaussian distribution can be updated. Then, according to the updated Gaussian distribution, a new round of variable control parameters can be obtained by the acquisition function, and then a new round of predictive control effects can be obtained, until the variable control parameters that meet the preset optimization conditions are found and determined as the predictive control parameters. Among them, the third preset optimization condition can be that after completing the optimization of a fixed round, the variable control parameters corresponding to the optimal predictive control effect in all rounds are selected as the predictive control parameters.

[0084] Figure 2 Schematically shows an optimization diagram based on a back-end sensor according to an embodiment of the present disclosure; Figure 3 The diagram schematically shows an optimization diagram based on a back-end sensor and a front-end sensor according to an embodiment of the present disclosure.

[0085] According to an embodiment of the present disclosure, the sensor may include a front-end sensor disposed at the front end and a rear-end sensor disposed at the rear end. Taking the rear-end sensor as an example, Figure 2 As shown in the figure, the initial control parameters are converted into initial control signals by the controller and input into the actuator. The actuator applies initial vibration isolation force to the controlled platform, and the initial control effect is obtained by the back-end sensor. Subsequently, the initial control parameters in the controller are adjusted by the optimization model and acquisition function to obtain variable control parameters.

[0086] If the front-end sensor is also included, then Figure 3 As shown in the figure, the vibration signal is collected by the front-end sensor, and the initial control signal is randomly generated according to the vibration signal and input into the actuator. The actuator applies the initial vibration isolation force to the controlled platform, and then the initial control effect is obtained by the back-end sensor. Subsequently, the initial control parameters are optimized by the optimization model and the acquisition function to obtain the variable control parameters.

[0087] According to an embodiment of the present disclosure, generating a corresponding control signal according to pre-sampled control parameters includes: analyzing the pre-sampled control parameters using a controller to determine a preset control strategy; and generating a control signal suitable for the actuator according to the preset control strategy.

[0088] According to an embodiment of the present disclosure, a controller is used to analyze pre-sampled control parameters to determine the parameters or current and voltage that need to be adjusted for the actuator in the current state, and to generate a control strategy for the actuator parameters.

[0089] According to an embodiment of the present disclosure, the control strategy is converted into the form of a signal, i.e., a control signal, so as to be transmitted to the actuator, so that the actuator can adjust its own parameters in response to the control signal, or modify the state of current and voltage, etc., to achieve the suppression of vibration interference from external interference sources.

[0090] According to an embodiment of the present disclosure, the above method also includes: using a target identification model to perform parameter identification processing on the vibration signal to obtain a target vibration signal, wherein the parameters of the target identification model are obtained by optimizing the initial parameters of the target identification model using an optimization model; wherein, comparing the vibration signal with a pre-sampled preset signal to obtain a comparison result includes: comparing the target vibration signal with a pre-sampled preset signal to obtain a comparison result.

[0091] According to the embodiments of the present disclosure, the types of vibration signals are different due to different external interference sources.

[0092] According to the embodiments of the present disclosure, a target recognition model is used to identify and process vibration signals. Since the type of vibration signal is determined, the comparison between the target vibration signal and the preset signal, or the construction of the Gaussian distribution during the training process, can be more accurate, reducing the errors caused by different types of vibration signals.

[0093] According to embodiments of the present disclosure, when a theoretical model is needed to approximate the actual system response, it is necessary to determine the transfer function (i.e., the target identification model) corresponding to the vibration signal and / or pre-sampled signal to the sensor, but the parameters of this transfer function cannot be determined. By optimizing the initial parameters of the target identification model using an optimization model, the parameters of the transfer function can be automatically optimized to obtain the optimal parameters.

[0094] Figure 4 The figure schematically shows an optimization diagram of the target recognition model according to an embodiment of the present disclosure.

[0095] According to the embodiments of the present disclosure, Figure 4 As shown, the optimization model is used to optimize the initial parameters of the target identification model, which can be the predicted value of the vibration signal after the target identification model is calculated. Compared with the actual output value y of the sensor, the predicted value is obtained by the target recognition model, and the actual output value y is the signal collected by the sensor after applying vibration interference to the controlled platform. The difference between the two is the loss value, that is, The initial parameters of the target recognition model are optimized based on the loss value. After multiple rounds of optimization, the parameters of the target recognition model are obtained.

[0096] According to an embodiment of the present disclosure, the above method also includes: using a sensor to collect a feedback signal, the feedback signal is obtained by applying a vibration isolation force to the controlled platform; analyzing the feedback signal to obtain a feedback analysis result; when the feedback analysis result indicates that the vibration isolation is abnormal or the comparison result indicates that the vibration signal and the pre-sampled signal are not similar, using an optimization model to optimize the pre-sampled control parameters to obtain target control parameters; generating a target control signal based on the target control parameters; using an actuator to apply a target vibration isolation force to the controlled platform, wherein the actuator generates the target vibration isolation force in response to the target control signal.

[0097] According to an embodiment of the present disclosure, after the vibration isolation force is applied to the controlled platform, the controlled platform is detected by a sensor to obtain a vibration signal in the current state as a feedback signal.

[0098] According to an embodiment of the present disclosure, when the feedback analysis result indicates that the vibration isolation is abnormal or the comparison result indicates that the vibration signal and the pre-sampled signal are not similar, it indicates that the preset control parameters obtained in the training process corresponding to the current external interference source environment are not applicable, and the control parameters need to be re-optimized using the optimization model to adapt to the current external interference source environment.

[0099] According to an embodiment of the present disclosure, the optimized control parameters are used as target control parameters, so that the actuator can apply a target vibration isolation force to the controlled platform based on the target control parameters, thereby suppressing vibration interference from external interference sources.

[0100] Figure 5 The structural block diagram of the automatic vibration isolation device according to an embodiment of the present disclosure is schematically shown.

[0101] A second aspect of the present disclosure provides an automatic vibration isolation device, comprising: a controlled platform 510; a sensor 520 for collecting a vibration signal, wherein the vibration signal is obtained by superimposing the vibration interference of an external vibration interference source on the controlled platform 510 and the original vibration of the table; a controller 530 for comparing the vibration signal and a pre-sampled signal to obtain a comparison result; when the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determining a control parameter corresponding to the pre-sampled signal, wherein the control parameter is obtained by optimizing the initial control parameter corresponding to the pre-sampled signal using an optimization model; generating a corresponding control signal according to the control parameter; an actuator 540 for generating a vibration isolation force in response to the control signal and applying the vibration isolation force to the controlled platform 510, wherein the vibration isolation force is used to suppress the vibration interference generated by the external vibration interference source on the controlled platform 510.

[0102] Figure 6 A schematic diagram of an automatic vibration isolation device according to an embodiment of the present disclosure is schematically shown.

[0103] The external interference source 1 causes vibration interference to the controlled platform 2. At this time, the front-end sensor 3 collects the vibration signal, which is used to compare with the pre-sampled signal to obtain a comparison result; when the controller 6 determines that the comparison result indicates that the vibration signal and the pre-sampled signal are similar, it determines the control parameters corresponding to the pre-sampled signal, generates a corresponding control signal, and transmits it to the actuator 7; the actuator 7 applies a vibration isolation force to the controlled platform 2 in response to the control signal, and the rear-end sensor 5 collects the residual vibration signal after the vibration isolation force is applied; wherein, if the controller 6 analyzes the feedback signal and the feedback analysis result obtained indicates that the vibration isolation is abnormal or the comparison result indicates that the vibration signal and the pre-sampled signal are not similar, the optimization model 4 optimizes the pre-sampled control parameters to obtain the target control parameters and inputs them to the controller 6.

[0104] Those skilled in the art will appreciate that the features described in the various embodiments of the present disclosure may be combined and / or coupled in various ways, even if such combinations or couplings are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of the present disclosure may be combined and / or coupled in various ways without departing from the spirit and teachings of the present disclosure. All such combinations and / or couplings fall within the scope of the present disclosure.

[0105] The above describes the embodiments of the present disclosure. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. Without departing from the scope of the present disclosure, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. An automatic vibration isolation method, characterized in that: The method comprises: Using a sensor to collect a vibration signal, wherein the vibration signal is obtained by an external vibration interference source causing vibration interference on the controlled platform; Comparing the vibration signal with the pre-sampled signal to obtain a comparison result; If the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determining a pre-sampled control parameter corresponding to the pre-sampled signal, wherein the pre-sampled control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampled signal using an optimization model; generating a control signal according to the pre-sampled control parameter and the vibration signal; An actuator is used to apply a vibration isolation force to the controlled platform in response to the control signal, wherein the vibration isolation force is used to suppress the vibration interference generated by the external vibration interference source on the controlled platform.

2. The method according to claim 1, characterized in that Optimizing the initial control parameters using the optimization model to obtain the pre-sampling control parameters is performed based on the following operations: Constructing a Gaussian distribution according to the initial control parameter, the variable control parameter, and the initial control effect, wherein the initial control effect is associated with the initial control parameter, and the variable control parameter is associated with the initial control parameter; Based on the Gaussian distribution, the variable control parameters are predicted to obtain a predicted control effect; In a case where the predictive control effect satisfies a first preset optimization condition, the variable control parameter is determined as the pre-sampling control parameter.

3. The method according to claim 2, characterized in that The constructing of Gaussian distribution according to the initial control parameters, the variable control parameters and the initial control effect includes: Acquiring an initial tabletop residual vibration signal corresponding to the initial control parameters, wherein the initial tabletop residual vibration signal is obtained by applying an initial vibration isolation force to the controlled platform based on the initial control parameters, and the initial vibration isolation force is generated according to the initial control parameters; determining an initial control effect corresponding to the initial control parameter according to the Allan deviation of the initial table residual vibration signal; generating a kernel function associated with the Gaussian distribution according to the initial control parameter and the variable control parameter; The Gaussian distribution is constructed according to the kernel function, the variable control parameters, the initial control parameters and the initial control effect.

4. The method according to claim 3, characterized in that The constructing of the Gaussian distribution according to the kernel function, the variable control parameters, the initial control parameters and the initial control effect includes: Constructing a correlation matrix between the variable control parameters and the initial control parameters to obtain a first kernel function; constructing an autocorrelation matrix of the variable control parameters to obtain a second kernel function, wherein the first kernel function and the second kernel function have the same hyperparameters; Optimizing the hyperparameters to obtain target hyperparameters; updating the first kernel function and the second kernel function according to the target hyperparameter to obtain updated first kernel function and second kernel function; A kernel function associated with the Gaussian distribution is determined according to the updated first kernel function and the second kernel function.

5. The method according to claim 4, characterized in that Optimizing the hyperparameters to obtain target hyperparameters includes: Constructing an autocorrelation matrix of the initial control parameters to obtain a third kernel function; Constructing a likelihood function according to the hyperparameters, the third kernel function, and the initial control effect; Optimizing the hyperparameters to obtain optimized hyperparameters; Calculating a likelihood function value of the likelihood function based on the optimized hyperparameters; When the likelihood function value satisfies a second preset optimization condition, the optimized hyperparameter is determined as the target hyperparameter.

6. The method according to claim 2, characterized in that When the predictive control effect satisfies a first preset optimization condition, determining the variable control parameter as the pre-sampling control parameter includes: Analyzing the predictive control effect using an expected lift function to obtain an analysis result; In a case where the analysis result indicates that the predictive control effect satisfies a third preset optimization condition, the variable control parameter is determined as the pre-sampling control parameter.

7. The method according to claim 1, characterized in that Generating a control signal according to the pre-sampled control parameter and the vibration signal includes: Utilizing a controller to analyze the pre-sampled control parameters and the vibration signal to determine a control strategy; The control signal applied to the actuator is generated according to the braking strategy and the vibration signal.

8. The method according to claim 1, further comprising: Using a target identification model, performing parameter identification processing on the vibration signal to obtain a target vibration signal, wherein the parameters of the target identification model are obtained by optimizing initial parameters of the target identification model using the optimization model; The comparing the vibration signal and the pre-sampled signal to obtain a comparison result includes: The target vibration signal and the pre-sampled signal are compared to obtain the comparison result.

9. The method according to claim 1, characterized in that The method further comprises: collecting a residual vibration signal of a tabletop using the sensor, wherein the residual vibration signal of the tabletop is obtained by applying the vibration isolation force to the controlled platform; Analyzing the residual vibration signal of the table to obtain a table analysis result; When the table analysis result indicates that the vibration isolation is abnormal or the comparison result indicates that the vibration signal is not similar to the pre-sampled signal, Optimizing the pre-sampling control parameters using the optimization model to obtain target control parameters; generating a target control signal according to the target control parameter and the vibration signal; A target vibration isolation force is applied to the controlled platform by using an actuator, wherein the actuator generates the target vibration isolation force in response to the target control signal.

10. An automatic vibration isolation device, characterized in that: The device comprises: the accused platform; A sensor for collecting a vibration signal, wherein the vibration signal is obtained by an external vibration interference source causing vibration interference on the controlled platform; a controller configured to compare the vibration signal and a pre-sampled signal to obtain a comparison result; if the comparison result indicates that the vibration signal and the pre-sampled signal are similar, determine a pre-sampled control parameter corresponding to the pre-sampled signal, wherein the pre-sampled control parameter is obtained by optimizing an initial control parameter corresponding to the pre-sampled signal using an optimization model; and generate a control signal based on the pre-sampled control parameter and the vibration signal; The actuator is used to generate a vibration isolation force in response to the control signal and apply the vibration isolation force to the controlled platform, wherein the vibration isolation force is used to suppress the vibration interference generated by the external vibration interference source on the controlled platform.

Images