A detection signal control method and system for a gear measurement center

Abstract

The present invention relates to the field of gear measurement technology, and discloses a detection signal control method and system for a gear measurement center. The original detection signal of the gear measurement center is collected in real time by multiple sensors, the collected original detection signal is filtered, the filtered detection signal is amplified by an operational amplifier, and the initial digital signal is linearly fitted by a least squares algorithm, a random forest model is constructed, and a genetic algorithm is used to optimize the hyperparameters of the random forest model. Based on the target digital signal, various characteristic parameters of the gear are extracted by a parameter identification model, and an improved chicken swarm optimization algorithm model is used to dynamically adjust the measurement parameters based on the error analysis results, and the measurement action of the gear measurement center is controlled by a signal control unit according to the corrected detection signal. The accuracy and stability of the detection signal in the gear measurement center are improved, the measurement error is reduced, and the accuracy of gear measurement is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of gear measurement, and in particular to a detection signal control method and system for a gear measurement center. Background Art

[0002] In modern mechanical manufacturing, gears are important transmission components, and their accuracy and quality directly affect the performance and reliability of the entire mechanical system. The gear measurement center is a high-precision measuring device used to measure various parameters of gears. The accuracy and stability of its detection signal directly affect the accuracy of the measurement results. However, the existing gear measurement center has some shortcomings in the detection signal control, such as large signal interference, low control accuracy, slow response speed, and the technology for gear measurement in the prior art usually first determines the category of the gear, and then measures the gear based on the determination result of the gear. This gear measurement method is not accurate enough in today's world where gear processing technology, gear parameters and gear morphology are changing with each passing day. These problems seriously affect the accuracy and efficiency of gear measurement. Therefore, how to automatically identify the measurement parameters of the gear, automatically detect and adjust the gear parameters according to the identification results, and ensure the stability and high efficiency of the gear detection signal are technical problems that need to be solved at this stage. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and to design a detection signal control method and system for a gear measurement center.

[0004] A first aspect of the present invention provides a detection signal control method for a gear measurement center, the detection signal control method for a gear measurement center comprising the following steps:

[0005] The original detection signal of the gear measurement center is collected in real time by multiple sensors, and the collected original detection signal is filtered by using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal;

[0006] Amplifying the filtered detection signal using an operational amplifier to obtain an amplified detection signal, converting the amplified detection signal into an initial digital signal, and performing linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal;

[0007] Constructing a random forest model, optimizing the hyperparameters of the random forest model using a genetic algorithm to obtain a parameter identification model, extracting various characteristic parameters of the gear through the parameter identification model based on the target digital signal, and outputting the parameter identification result;

[0008] Analyzing the parameter error of the gear based on the parameter identification result, and dynamically adjusting the measurement parameters based on the error analysis result by using an improved flock optimization algorithm model, and generating a compensation detection signal corresponding to the gear measurement center according to the measurement parameters;

[0009] The compensated detection signal is transmitted to the signal control unit of the gear measurement center, and the compensated detection signal and the original detection signal are fused through a difference filtering compensation algorithm to generate a corrected detection signal. The measurement action of the gear measurement center is controlled by the signal control unit according to the corrected detection signal.

[0010] Optionally, in a first implementation of the first aspect of the present invention, the original detection signal of the gear measurement center is collected in real time by multiple sensors, and the collected original detection signal is filtered using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal, including:

[0011] By setting up a sensor array, a comprehensive scan is performed on the running gear at the gear detection center to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor;

[0012] The Mallat fast wavelet transform algorithm is used to decompose the original detection signal in the wavelet domain to decompose the original detection signal into detail components at multiple scales, and threshold quantization processing is performed according to the soft threshold to obtain the wavelet coefficients:

[0013]

[0014]

[0015] In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points;

[0016] A Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain an observation signal. The compressed wavelet coefficients are reconstructed using the OMP algorithm based on the observation signal, and the reconstructed wavelet coefficients are inversely transformed to obtain a filtered detection signal.

[0017] Optionally, in a second implementation of the first aspect of the present invention, a Gaussian random matrix is ​​selected as an observation matrix, the wavelet coefficients obtained after decomposition are compressed to obtain an observation signal, the compressed wavelet coefficients are reconstructed based on the observation signal using an OMP algorithm, and the reconstructed wavelet coefficients are inversely transformed to obtain a filtered detection signal, including:

[0018] The Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain the observation signal;

[0019] Projecting the observed signal onto all atoms of an index dictionary, finding the atomic number of the index dictionary with the largest projection coefficient, and obtaining a low-resolution atomic number, wherein the index dictionary is a low-spectral-resolution Fourier dictionary;

[0020] Determine the spectrum range where the observation signal is located according to the low-resolution atomic number, and project the observation signal to the corresponding atomic number in a fine dictionary according to the corresponding range, wherein the fine dictionary is a high-spectrum-resolution Fourier dictionary;

[0021] The projected atomic numbers are added to the support set and the iterative process is started. In each iteration, the coefficients of the atoms in the support set are optimized using the least squares method. After each iteration, it is determined whether the residual between the reconstructed signal and the observed signal is less than a threshold. If so, the iteration ends and the reconstructed signal is obtained.

[0022] The wavelet coefficients reconstructed by the OMP algorithm are converted back to the time domain through inverse Mallat transform to obtain the filtered detection signal.

[0023] Optionally, in a third implementation of the first aspect of the present invention, the random forest model is constructed, and a genetic algorithm is used to optimize the hyperparameters of the random forest model to obtain a parameter identification model, and various characteristic parameters of the gear are extracted through the parameter identification model based on the target digital signal, and a parameter identification result is output, including:

[0024] Acquire a target digital signal of a gear measurement center, and perform standardization processing and principal component analysis processing on the target digital signal in sequence to obtain sample data;

[0025] Construct a random forest model, input the sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20;

[0026] The fitness function is constructed according to the objective function and the penalty function. The fitness of each individual in the population is calculated according to the fitness function. The expression of the fitness function is:

[0027]

[0028] In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint;

[0029] The group is selected, crossed, and mutated based on the fitness of the individual. The probability of crossover is 0.5 and the probability of mutation is 0.1. The new group is used as the population at the beginning of the next iteration.

[0030] Repeat the iteration until the number of iterations t equals the maximum number of iterations, and take the individual with the largest fitness in the population as the optimal solution. Assign the optimal solution to the random forest model to obtain the parameter identification model;

[0031] The target digital signal is input into the parameter identification model, and the mapping relationship between the characteristics in the target digital signal and the gear parameters is predicted by the parameter identification model, and various characteristic parameters of the gear are extracted, and the parameter identification result is output, wherein the parameter identification result at least includes the gear type, gear pitch, gear tooth shape, and gear tooth direction.

[0032] Optionally, in a fourth implementation of the first aspect of the present invention, the parameter error of the gear is analyzed based on the parameter identification result, and the measurement parameters are dynamically adjusted based on the error analysis result using an improved chicken swarm optimization algorithm model, and a compensation detection signal corresponding to the gear measurement center is generated according to the measurement parameters, including:

[0033] Extracting characteristic parameters related to the error from the parameter identification results, calculating the statistics of the characteristic parameters, analyzing the error size and the error fluctuation degree, and obtaining the error analysis results, wherein the statistics at least include the mean, variance, and standard deviation;

[0034] Constructing an improved swarm optimization algorithm model, taking the error analysis results as the input of the algorithm, taking the gear measurement parameters as the optimization variables, and dynamically adjusting the measurement parameters using the improved swarm optimization algorithm model, wherein the measurement parameters at least include sampling frequency and measurement point distribution;

[0035] The measurement parameters adjusted by the improved flock optimization algorithm are collected, and the optimized measurement parameters are converted into an electrical signal format to obtain a compensation detection signal.

[0036] Optionally, in a fifth implementation of the first aspect of the present invention, constructing an improved swarm optimization algorithm model, taking the error analysis result as the input of the algorithm, taking the gear measurement parameter as the optimization variable, and dynamically adjusting the measurement parameter using the improved swarm optimization algorithm model includes:

[0037] Initialize the chicken flock, set the total number of chickens to 200, the maximum number of iterations to 200, and calculate the fitness of the chicken flock:

[0038]

[0039] Where e represents the gear error vector, J represents the error weight matrix, represents the adjusted measurement parameter vector, represents the number of elements of e, () represents the Euclidean norm function;

[0040] Determine individual extreme values ​​and group extreme values, and sort the chickens in descending order of fitness to obtain roosters, hens and chicks, update the positions of the chicken members, and use an adaptive parameter adjustment mechanism to move the chickens in the search space;

[0041] Repeat the optimization iteration until the maximum number of iterations is reached, stop the iteration process, and output the optimized measurement parameters.

[0042] Optionally, in a sixth implementation of the first aspect of the present invention, the compensated detection signal is transmitted to a signal control unit of a gear measurement center, the compensated detection signal and the original detection signal are fused through a difference filtering compensation algorithm to generate a corrected detection signal, and the measurement action of the gear measurement center is controlled through the signal control unit according to the corrected detection signal, including:

[0043] Transmitting the compensated detection signal to a signal control unit of a gear measurement center, and the signal control unit synchronizes and aligns the received compensated detection signal with the collected original detection signal;

[0044] The compensation detection signal is converted into and the original detection signal Fusion, calculation of the compensated detection signal and the original detection signal The difference:

[0045]

[0046] The compensation detection signal and the original detection signal The difference Perform filtering:

[0047]

[0048] Calculate the compensation value, and based on the compensation value and the original detection signal Generate a modified detection signal :

[0049]

[0050]

[0051] In the formula, represents the compensation coefficient, Indicates the window size, represents the corrected detection signal, represents the difference after smoothing, is an index variable, indicating a specific position in the signal sequence, Indicates that in the summation process, it is used as an index variable The change of is used to traverse all positions in the window, Indicates at location The difference between

[0052] Identify the equipment parameters that need to be adjusted according to the corrected detection signal, generate corresponding control instructions, and send the control instructions to the measuring equipment of the gear measurement center to control and perform the measurement task through the signal control unit, wherein the equipment parameters at least include the probe path and the scanning speed;

[0053] The measurement process is monitored in real time, the measurement data and device status information are collected, and feedback data is generated to continuously iterate and optimize the signal control process based on the feedback data.

[0054] The second aspect of the present invention provides a detection signal control system for a gear measurement center, the detection signal control system for the gear measurement center comprises a signal filtering module, a signal conversion module, a parameter identification module, a dynamic adjustment module and a signal correction module, wherein:

[0055] A signal filtering module is used to collect the original detection signal of the gear measurement center in real time through a variety of sensors, and filter the collected original detection signal using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal;

[0056] A signal conversion module is used to amplify the filtered detection signal using an operational amplifier to obtain an amplified detection signal, convert the amplified detection signal into an initial digital signal, and perform linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal;

[0057] A parameter identification module is used to construct a random forest model, optimize the hyperparameters of the random forest model using a genetic algorithm to obtain a parameter identification model, extract various characteristic parameters of the gear based on the target digital signal through the parameter identification model, and output the parameter identification result;

[0058] A dynamic adjustment module, used to analyze the parameter error of the gear based on the parameter identification result, and dynamically adjust the measurement parameters based on the error analysis result by using an improved chicken swarm optimization algorithm model, and generate a compensation detection signal corresponding to the gear measurement center according to the measurement parameters;

[0059] A signal correction module is used to transmit the compensated detection signal to the signal control unit of the gear measurement center, fuse the compensated detection signal and the original detection signal through a difference filtering compensation algorithm to generate a corrected detection signal, and control the measurement action of the gear measurement center through the signal control unit according to the corrected detection signal.

[0060] Optionally, in a sixth implementation of the second aspect of the present invention, the signal filtering module includes a signal collection submodule, a wavelet decomposition submodule and a signal reconstruction submodule, wherein:

[0061] A signal collection submodule is used to perform a comprehensive scan of the running gears at the gear detection center through a set sensor array to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor;

[0062] The wavelet decomposition submodule is used to decompose the original detection signal in the wavelet domain by using the Mallat fast wavelet transform algorithm, so as to decompose the original detection signal into detail components at multiple scales, and perform threshold quantization processing according to the soft threshold to obtain the wavelet coefficients:

[0063]

[0064]

[0065] In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points;

[0066] The signal reconstruction submodule is used to select a Gaussian random matrix as the observation matrix, compress the wavelet coefficients obtained after decomposition to obtain an observation signal, reconstruct the compressed wavelet coefficients based on the observation signal using the OMP algorithm, and obtain the filtered detection signal by performing an inverse wavelet transform on the reconstructed wavelet coefficients.

[0067] Optionally, in a sixth implementation of the second aspect of the present invention, the parameter identification module includes a principal component analysis submodule, an initialization submodule, a calculation submodule, a genetic processing submodule, a repetition submodule and a prediction submodule, wherein:

[0068] A principal component analysis submodule is used to obtain a target digital signal of a gear measurement center, and perform standardization processing and principal component analysis processing on the target digital signal in sequence to obtain sample data;

[0069] The initialization submodule is used to construct a random forest model, input the sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20;

[0070] The calculation submodule is used to construct a fitness function based on the objective function and the penalty function, and calculate the fitness of each individual in the population according to the fitness function. The expression of the fitness function is:

[0071]

[0072] In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint;

[0073] The genetic processing submodule is used to select, crossover, and mutate the population based on the fitness of the individual. The probability of crossover is 0.5 and the probability of mutation is 0.1. The new population is used as the population at the beginning of the next iteration.

[0074] The repeated iteration submodule is used to repeat the iteration until the number of iterations t equals the maximum number of iterations, and the individual with the largest fitness in the population is taken as the optimal solution. The optimal solution is assigned to the random forest model to obtain a parameter identification model;

[0075] A prediction submodule is used to input the target digital signal into the parameter identification model, predict the mapping relationship between the characteristics in the target digital signal and the gear parameters through the parameter identification model, extract the various characteristic parameters of the gear, and output the parameter identification result, wherein the parameter identification result at least includes the gear type, gear pitch, gear tooth shape, and gear tooth direction.

[0076] The beneficial effects are that the original detection signal of the gear measurement center is collected in real time by multiple sensors, and the Mallat fast wavelet transform algorithm, compression processing and OMP algorithm are used to filter the collected original detection signal to obtain a filtered detection signal; the filtered detection signal is amplified by an operational amplifier to obtain an amplified detection signal, the amplified detection signal is converted into an initial digital signal, and the initial digital signal is linearly fitted by a least squares algorithm to obtain a target digital signal; a random forest model is constructed, and a genetic algorithm is used to optimize the hyperparameters of the random forest model to obtain a parameter identification model, and various characteristic parameters of the gear are extracted through the parameter identification model based on the target digital signal, and a parameter identification result is output; the parameter error of the gear is analyzed based on the parameter identification result, and the improved chicken swarm optimization algorithm model is used to dynamically adjust the measurement parameters based on the error analysis result, and a compensation detection signal corresponding to the gear measurement center is generated according to the measurement parameters; the compensation detection signal is transmitted to the signal control unit of the gear measurement center, and the compensation detection signal and the original detection signal are fused by a difference filter compensation algorithm to generate a corrected detection signal, and the measurement action of the gear measurement center is controlled by the signal control unit according to the corrected detection signal. The following can be achieved: 1. The accuracy and stability of the detection signal in the gear measurement center are improved, the measurement error is reduced, and the accuracy of gear measurement is improved. 2. The response speed of the detection signal in the gear measurement center is accelerated, the measurement efficiency is improved, and the needs of rapid measurement of various types of gears in modern manufacturing can be met. 3. The use of advanced digital filtering and data analysis algorithms improves the anti-interference ability and adaptability of the system, and it can operate stably in complex working environments. 4. The detection signal control method and system have good versatility and scalability, and can be applied to gear measurement centers of different types and specifications. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0078] Figure 1 A schematic diagram of a first embodiment of a detection signal control method for a gear measurement center provided by an embodiment of the present invention;

[0079] Figure 2 A schematic diagram of a second embodiment of a detection signal control method for a gear measurement center provided by an embodiment of the present invention;

[0080] Figure 3 A schematic diagram of a third embodiment of a detection signal control method for a gear measurement center provided by an embodiment of the present invention;

[0081] Figure 4 A schematic diagram of a first embodiment of a detection signal control system for a gear measurement center provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0082] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0083] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 A first embodiment schematic diagram of a detection signal control method for a gear measurement center provided by an embodiment of the present invention, the method specifically comprises the following steps:

[0084] Step 101, collecting original detection signals of the gear measurement center in real time through multiple sensors, filtering the collected original detection signals using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain filtered detection signals;

[0085] Specifically, in this embodiment, a sensor array is set up to perform a comprehensive scan on the running gears at the gear detection center to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor; the Mallat fast wavelet transform algorithm is used to decompose the original detection signal in the wavelet domain to decompose the original detection signal into detail components at multiple scales, and threshold quantization processing is performed according to the soft threshold to obtain the wavelet coefficient:

[0086]

[0087]

[0088] In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points; a Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain the observation signal. The compressed wavelet coefficients are reconstructed based on the observation signal using the OMP algorithm, and the reconstructed wavelet coefficients are inversely transformed to obtain the filtered detection signal.

[0089] Specifically, in this embodiment, a Gaussian random matrix is ​​selected as the observation matrix, the wavelet coefficients obtained after decomposition are compressed to obtain an observation signal, the compressed wavelet coefficients are reconstructed based on the observation signal using the OMP algorithm, and the reconstructed wavelet coefficients are inversely transformed to obtain the filtered detection signal. The implementation steps are as follows:

[0090] A Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain the observation signal; the observation signal is projected onto all atoms in the index dictionary, and the atomic number of the index dictionary with the largest projection coefficient is found to obtain the low-resolution atomic number, wherein the index dictionary is a low-spectrum resolution Fourier dictionary; the spectrum range where the observation signal is located is determined according to the low-resolution atomic number, and the observation signal is projected onto the corresponding atomic number in the fine dictionary according to the corresponding range, wherein the fine dictionary is a high-spectrum resolution Fourier dictionary; the projected atomic number is added to the support set, and an iterative process is entered. In each iteration, the coefficients of the atoms in the support set are optimized using the least squares method. After each iteration, it is determined whether the residual between the reconstructed signal and the observed signal is less than a threshold. If so, the iteration ends to obtain the reconstructed signal; the wavelet coefficients reconstructed by the OMP algorithm are converted back to the time domain through an inverse Mallat transform to obtain a filtered detection signal.

[0091] Step 102: amplify the filtered detection signal using an operational amplifier to obtain an amplified detection signal, convert the amplified detection signal into an initial digital signal, and perform linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal;

[0092] Specifically, in this embodiment, signal filtering is a preprocessing step for removing noise from the original detection signal, which may come from electromagnetic interference, the device itself or the surrounding environment. A low-pass filter, a high-pass filter or a bandpass / bandstop filter is used. After filtering, the quality of the signal is significantly improved, providing a purer basis for subsequent processing. Subsequently, the operational amplifier is used to amplify the filtered signal. The operational amplifier performs well in signal amplification due to its high gain, low noise and high input impedance. Through circuit configuration (such as in-phase amplifier, inverting amplifier, etc.), the amplification factor can be accurately controlled to ensure that the amplified signal is neither distorted nor introduces additional noise, thereby obtaining a detection signal with sufficient amplitude to meet the dynamic range requirements of subsequent AD conversion. Digital signal conversion: The amplified analog signal needs to be converted into a digital signal, which is completed by an analog-to-digital converter (ADC). Analog-to-digital conversion is the conversion of a continuous analog voltage or current into a discrete sequence of digital values. Least squares linear fitting: The least squares algorithm is used to perform linear fitting on the initial digital signal obtained. The least squares method finds the best fitting straight line by minimizing the sum of squared errors. The error refers to the vertical distance from each data point to the fitting straight line. The process can effectively eliminate the influence of random errors and extract the main trend of the signal, thereby obtaining a clearer target digital signal, which is extremely important for subsequent data analysis, pattern recognition or control strategy formulation. Through the processing flow, the original detection signal is extracted from the noise interference environment, amplified, digitized, and then optimized through mathematical models, and finally converted into a highly pure and representative target digital signal.

[0093] Step 103: construct a random forest model, use a genetic algorithm to optimize the hyperparameters of the random forest model, obtain a parameter identification model, extract various characteristic parameters of the gear through the parameter identification model based on the target digital signal, and output the parameter identification result;

[0094] Specifically, in this embodiment, a target digital signal of a gear measurement center is obtained, and the target digital signal is subjected to standardization processing and principal component analysis processing in sequence to obtain sample data;

[0095] Construct a random forest model, input sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20;

[0096] The fitness function is constructed according to the objective function and the penalty function. The fitness of each individual in the population is calculated according to the fitness function. The expression of the fitness function is:

[0097]

[0098] In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint;

[0099] The group is selected, crossed, and mutated based on the fitness of the individual. The probability of crossover is 0.5 and the probability of mutation is 0.1. The new group is used as the population at the beginning of the next iteration.

[0100] Repeat the iteration until the number of iterations t equals the maximum number of iterations, and take the individual with the largest fitness in the population as the optimal solution. Assign the optimal solution to the random forest model to obtain the parameter identification model;

[0101] The target digital signal is input into the parameter identification model, and the mapping relationship between the features in the target digital signal and the gear parameters is predicted by the parameter identification model, and various characteristic parameters of the gear are extracted, and the parameter identification results are output, wherein the parameter identification results at least include the gear type, gear pitch, gear tooth shape, and gear tooth direction.

[0102] Step 104: Analyze the parameter error of the gear based on the parameter identification result, and dynamically adjust the measurement parameters based on the error analysis result using the improved chicken swarm optimization algorithm model, and generate a compensation detection signal corresponding to the gear measurement center according to the measurement parameters;

[0103] Specifically, in this embodiment, characteristic parameters related to the error are extracted from the parameter identification results, the statistics of the characteristic parameters are calculated, the error size and the degree of error fluctuation are analyzed, and the error analysis results are obtained, wherein the statistics include at least the mean, variance, and standard deviation; an improved swarm optimization algorithm model is constructed, the error analysis results are used as the input of the algorithm, the gear measurement parameters are used as the optimization variables, and the measurement parameters are dynamically adjusted using the improved swarm optimization algorithm model, wherein the measurement parameters include at least the sampling frequency and the distribution of measurement points; the measurement parameters adjusted by the improved swarm optimization algorithm are collected, and the optimized measurement parameters are converted into an electrical signal format to obtain a compensated detection signal.

[0104] Specifically, in this embodiment, characteristic parameters related to the error are extracted from the parameter identification results, statistics of the characteristic parameters are calculated, the error size and the error fluctuation degree are analyzed, and the error analysis results are obtained, wherein the statistics at least include mean, variance, and standard deviation, and the following steps are included:

[0105] The parameter identification results are analyzed in depth to select key characteristic parameters that can directly or indirectly reflect the gear error. For example, the deviation of the tooth shape, the variation of the pitch, the runout of the gear, etc. By carefully comparing and screening a large amount of measurement data, the features that are significantly correlated with the error are determined. Data mining technology and feature engineering methods are used to accurately extract these key features from a complex set of parameters. This may involve operations such as data cleaning, feature selection, and feature construction to ensure that the extracted feature parameters have clear physical meanings and good characterization capabilities for errors. The statistics of the feature parameters are calculated, the error size and error fluctuation degree are analyzed, and the error analysis results are obtained, where the statistics include at least the mean, variance, and standard deviation: for each extracted feature parameter, its mean is calculated separately. The mean represents the average level of the feature parameter and reflects the central trend of the error in general. Next, the variance and standard deviation are calculated. The variance measures the dispersion of the feature parameter value relative to the mean, and the standard deviation is the square root of the variance, which more intuitively reflects the degree of dispersion of the data. Larger variance and standard deviation indicate larger fluctuations in the error, and vice versa. The distribution of errors can be intuitively displayed by drawing histograms, box plots and other visualization methods of characteristic parameters. At the same time, combined with statistical hypothesis testing methods, such as t-test or F-test, the errors in different batches or under different conditions are compared and analyzed to determine the significant differences and changing trends of errors.

[0106] Specifically, in this embodiment, an improved swarm optimization algorithm model is constructed, the error analysis results are used as the input of the algorithm, the gear measurement parameters are used as optimization variables, and the improved swarm optimization algorithm model is used to dynamically adjust the measurement parameters, including the following steps:

[0107] Initialize the chicken flock, set the total number of chickens to 200, the maximum number of iterations to 200, and calculate the fitness of the chicken flock:

[0108]

[0109] Where e represents the gear error vector, which represents the set of gear errors found during the measurement process. J represents the weight matrix, which is used to measure the importance or weight of each error term. Δδ represents the adjusted measurement parameter vector, which represents the best measurement parameter combination found by the optimization algorithm. size(e) represents the number of elements in the gear error vector e, which usually refers to the number of different types of error terms considered during the measurement process. norm() represents the Euclidean norm function, which is used to calculate the length or modulus of a vector. Here, it represents the modulus of the residual vector after the error vector e is subtracted from the adjusted measurement parameter JΔδ. fit represents the fitness value, which is used to evaluate the optimization effect of the swarm optimization algorithm. This formula is used to calculate the fitness value fit of the swarm optimization algorithm, which measures the degree of improvement of the gear error e by the optimized measurement parameter Δδ. Specifically, first, the error vector e is subtracted from the weight matrix J and multiplied by the adjusted measurement parameter Δδ, then the Euclidean norm of the residual vector is calculated, and finally divided by the number of elements in the error vector size(e). The purpose of this is to make the fitness value fit comparable, so that the optimization effect can be effectively evaluated even at different error scales. Determine individual extreme values ​​and group extreme values, and classify the chicken flock in descending order of fitness to obtain roosters, hens and chicks, update the positions of the chicken flock members, and use an adaptive parameter adjustment mechanism to move the chicken flock in the search space; repeat the optimization iteration until the maximum number of iterations is reached, stop the iterative process, and output the optimized measurement parameters.

[0110] Specifically, in this embodiment, the improved chicken swarm optimization algorithm is used to dynamically adjust the measurement parameters of the gear measurement center for detection control according to the error analysis results, wherein the measurement parameters at least include the sampling frequency and the distribution of the measurement points, and the following steps are also included:

[0111] Firstly, a swarm optimization algorithm model suitable for gear measurement parameter optimization is constructed. On the basis of the traditional swarm algorithm, specific improvement strategies for gear measurement problems are introduced, such as adjusting individual update rules, adding local search mechanisms or adopting adaptive parameter settings, so as to improve the algorithm's search efficiency and convergence speed. The error analysis results are used as the input of the algorithm, and the measurement parameters such as sampling frequency and measurement point distribution are used as optimization variables. The fitness function is defined to evaluate the measurement effect under different measurement parameter combinations. The fitness function can be comprehensively designed based on multiple factors such as the degree of error reduction, measurement accuracy and stability. The algorithm initializes a group of chicken individuals, each of which represents a set of possible measurement parameter combinations. During the iteration process, the hen updates its position according to the guidance of the rooster and its own experience, and the chicks follow the hen to move, and continuously optimize the measurement parameters through competition and cooperation mechanisms. At the same time, random perturbations are introduced to avoid the algorithm from falling into local optimality. After a certain number of iterations, the algorithm converges to a set of optimal measurement parameter combinations, and realizes the dynamic adjustment of parameters such as sampling frequency and measurement point distribution to improve the accuracy and efficiency of gear measurement. Collect the measurement parameters adjusted by the flock optimization algorithm, convert the optimized measurement parameters into electrical signal format, and obtain the compensation detection signal: collect the optimized measurement parameters output by the flock optimization algorithm in real time, including but not limited to the precise sampling frequency value and the specific distribution location information of the measurement point. Use digital-to-analog conversion technology to convert these digital measurement parameters into corresponding electrical signals. During the conversion process, accurate quantization and encoding are performed according to the hardware requirements of the measurement system and the specifications of the electrical signal. The generated electrical signal is filtered, amplified, and processed to ensure that it has good stability and anti-interference ability. At the same time, the amplitude, frequency and other characteristics of the electrical signal are calibrated and verified to ensure that it accurately reflects the optimized measurement parameters. The final compensation detection signal is transmitted to the control system of the gear measurement center to guide subsequent measurement operations and achieve more accurate detection and compensation of gear errors.

[0112] Step 105: transmit the compensated detection signal to the signal control unit of the gear measurement center, fuse the compensated detection signal and the original detection signal through a difference filter compensation algorithm to generate a corrected detection signal, and control the measurement action of the gear measurement center through the signal control unit according to the corrected detection signal.

[0113] Specifically, in this embodiment, the compensation detection signal is transmitted to the signal control unit of the gear measurement center, and the signal control unit synchronizes and aligns the received compensation detection signal with the collected original detection signal; the compensation detection signal is converted into the compensation signal by the difference filtering compensation algorithm. and the original detection signal Fusion, calculation and compensation of detection signals and the original detection signal The difference:

[0114]

[0115] Compensation detection signal and the original detection signal The difference Perform filtering:

[0116]

[0117] Calculate the compensation value and based on the compensation value and the original detection signal Generate a modified detection signal :

[0118]

[0119]

[0120] In the formula, represents the compensation coefficient, Indicates the window size, represents the corrected detection signal, represents the difference after smoothing, is an index variable, indicating a specific position in the signal sequence, Indicates that in the summation process, it is used as an index variable The change of is used to traverse all positions in the window, Indicates at location difference; identify the equipment parameters that need to be adjusted according to the corrected detection signal, generate corresponding control instructions, and send the control instructions to the measuring equipment of the gear measurement center to control and execute the measurement task through the signal control unit, wherein the equipment parameters include at least the probe path and the scanning speed; monitor the measurement process in real time, collect measurement data and equipment status information, and generate feedback data to continuously iterate and optimize the signal control process based on the feedback data.

[0121] The beneficial effect lies in that the original detection signal of the gear measurement center is collected in real time through a variety of sensors, and the collected original detection signal is filtered by using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal; the filtered detection signal is amplified by an operational amplifier to obtain an amplified detection signal, the amplified detection signal is converted into an initial digital signal, and the initial digital signal is linearly fitted by a least squares algorithm to obtain a target digital signal; a random forest model is constructed, and a genetic algorithm is used to optimize the hyperparameters of the random forest model to obtain a parameter identification model, and various characteristic parameters of the gear are extracted through the parameter identification model based on the target digital signal, and the parameter identification result is output; the parameter error of the gear is analyzed based on the parameter identification result, and the improved chicken swarm optimization algorithm model is used to dynamically adjust the measurement parameters based on the error analysis result, and the compensation detection signal corresponding to the gear measurement center is generated according to the measurement parameters; the compensation detection signal is transmitted to the signal control unit of the gear measurement center, and the compensation detection signal and the original detection signal are fused by the difference filter compensation algorithm to generate a corrected detection signal, and the measurement action of the gear measurement center is controlled by the signal control unit according to the corrected detection signal. The following can be achieved: 1. The accuracy and stability of the detection signal in the gear measurement center are improved, the measurement error is reduced, and the accuracy of gear measurement is improved. 2. The response speed of the detection signal in the gear measurement center is accelerated, the measurement efficiency is improved, and the needs of rapid measurement of various types of gears in modern manufacturing can be met. 3. The use of advanced digital filtering and data analysis algorithms improves the anti-interference ability and adaptability of the system, and it can operate stably in complex working environments. 4. The detection signal control method and system have good versatility and scalability, and can be applied to gear measurement centers of different types and specifications.

[0122] See also Figure 2 , a schematic diagram of a second embodiment of a detection signal control method for a gear measurement center provided by an embodiment of the present invention, wherein the original detection signal of the gear measurement center is collected in real time by multiple sensors, and the collected original detection signal is filtered by using the Mallat fast wavelet transform algorithm, compression processing and OMP algorithm, and the filtered detection signal includes:

[0123] Step 1011: Perform a comprehensive scan of the running gears at the gear detection center through a set sensor array to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor;

[0124] Step 1013: Decompose the original detection signal in the wavelet domain using the Mallat fast wavelet transform algorithm to decompose the original detection signal into detail components at multiple scales, and perform threshold quantization processing according to the soft threshold to obtain wavelet coefficients:

[0125]

[0126]

[0127] In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points;

[0128] Step 1013: Select a Gaussian random matrix as the observation matrix, compress the wavelet coefficients obtained after decomposition to obtain an observation signal, reconstruct the compressed wavelet coefficients based on the observation signal using the OMP algorithm, and obtain the filtered detection signal by performing an inverse wavelet transform on the reconstructed wavelet coefficients.

[0129] Its beneficial effect is that the implementation of compression processing strategy greatly reduces the amount of data, improves the efficiency of subsequent data processing and transmission, and reduces the demand for storage and computing resources, which is particularly beneficial for large-scale deployment or remote monitoring scenarios. Compression processing optimizes the data management process without sacrificing the diagnostic value of the signal, making the system run more smoothly and respond faster. On this basis, the OMP (Orthogonal Matching Pursuit) algorithm is used to optimize the filtered signal at a deeper level. This algorithm performs well in signal denoising and feature extraction due to its high efficiency and accuracy. OMP separates the feature vector that best represents the gear state from the signal through sparse representation, effectively suppresses residual noise, improves signal quality, and ensures the reliability of subsequent analysis. This method realizes efficient and high-quality processing of the original detection signal of the gear measurement center through an integrated sensor network, advanced Mallat fast wavelet transform, efficient compression technology and precise OMP algorithm. The beneficial effects are reflected in the following aspects: Improve diagnostic accuracy: effectively extract gear fault features, reduce false alarms and missed alarms, and improve the accuracy and sensitivity of fault diagnosis. Enhance real-time performance: Compression processing shortens data processing time and speeds up the response speed from signal acquisition to fault warning. Optimize resource utilization: Reduce data storage and transmission costs, and adapt to resource-constrained environments or large-scale monitoring networks. Promote intelligent maintenance: Provide a high-quality data foundation for implementing predictive maintenance strategies, reduce unplanned downtime, extend equipment life, and improve production efficiency and economic benefits. Enhance system adaptability: Applicable to gear monitoring under various working conditions and environments, and improve the versatility and robustness of the system.

[0130] See also Figure 3 , a schematic diagram of a third embodiment of a detection signal control method for a gear measurement center provided by an embodiment of the present invention, constructing a random forest model, using a genetic algorithm to optimize the hyperparameters of the random forest model, obtaining a parameter identification model, extracting various characteristic parameters of the gear through the parameter identification model based on the target digital signal, and outputting the parameter identification results, including:

[0131] Step 1031, obtaining a target digital signal of a gear measurement center, and performing standardization processing and principal component analysis processing on the target digital signal in sequence to obtain sample data;

[0132] Step 1032, construct a random forest model, input sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20;

[0133] Step 1033: construct a fitness function according to the objective function and the penalty function, and calculate the fitness of each individual in the population according to the fitness function. The fitness function is expressed as:

[0134]

[0135] In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint;

[0136] Step 1034: select, crossover, and mutate the group based on the fitness of the individual, with the probability of crossover being 0.5 and the probability of mutation being 0.1, and obtain a new group as the population at the beginning of the next iteration;

[0137] Step 1035, repeat the iteration until the number of iterations t equals the maximum number of iterations, and take the individual with the largest fitness in the population as the optimal solution, assign the optimal solution to the random forest model, and obtain a parameter identification model;

[0138] Step 1036: input the target digital signal into the parameter identification model, predict the mapping relationship between the characteristics in the target digital signal and the gear parameters through the parameter identification model, extract the various characteristic parameters of the gear, and output the parameter identification result, wherein the parameter identification result at least includes the gear type, gear pitch, gear tooth shape, and gear tooth direction.

[0139] Its beneficial effect is that by constructing and optimizing the random forest model, the model hyperparameters are intelligently optimized with the efficient global search capability of the genetic algorithm, which enhances the prediction accuracy and generalization ability of the model. It not only overcomes the limitations and inefficiency of traditional manual parameter adjustment, but also ensures that the parameter identification model can more accurately adapt to complex and changeable data characteristics. Specifically, the genetic algorithm finds a set of optimal parameter combinations by iteratively optimizing the key hyperparameters in the random forest model, such as the number of trees, the maximum number of features, and the minimum number of samples required for node splitting. This strategy based on the principle of biological evolution effectively explores the hyperparameter space by simulating processes such as natural selection, inheritance, and mutation, avoids local optimal solutions, and ensures the optimization of model performance. The optimized random forest model is used to extract feature parameters of the target digital signal, realizing efficient identification and quantification of various key indicators of the gear, such as wear degree, defect type, and load state. The process not only improves the stability and reliability of feature extraction, but also significantly enhances the accuracy and sensitivity of gear health status assessment.

[0140] See also Figure 4 , a schematic diagram of a first embodiment of a detection signal control system for a gear measurement center provided by an embodiment of the present invention, the system includes the following modules:

[0141] 401. A signal filtering module, which is used to collect original detection signals of the gear measurement center in real time through a variety of sensors, and to filter the collected original detection signals using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain filtered detection signals;

[0142] Specifically, the signal filtering module in this embodiment includes a signal collection submodule, a wavelet decomposition submodule and a signal reconstruction submodule, wherein:

[0143] The signal collection submodule is used to comprehensively scan the running gears at the gear detection center through a set sensor array to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor;

[0144] The wavelet decomposition submodule is used to decompose the original detection signal in the wavelet domain using the Mallat fast wavelet transform algorithm, so as to decompose the original detection signal into detail components at multiple scales, and perform threshold quantization processing according to the soft threshold to obtain the wavelet coefficients:

[0145]

[0146]

[0147] In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points;

[0148] The signal reconstruction submodule is used to select a Gaussian random matrix as the observation matrix, compress the wavelet coefficients obtained after decomposition to obtain the observation signal, reconstruct the compressed wavelet coefficients based on the observation signal using the OMP algorithm, and obtain the filtered detection signal by performing an inverse wavelet transform on the reconstructed wavelet coefficients.

[0149] Specifically, the signal reconstruction submodule in this embodiment includes the following units:

[0150] The compression unit is used to select a Gaussian random matrix as an observation matrix, compress the wavelet coefficients obtained after decomposition, and obtain an observation signal;

[0151] A search unit is used to project the observed signal onto all atoms of the index dictionary, find the atomic number of the index dictionary with the largest projection coefficient, and obtain the low-resolution atomic number, wherein the index dictionary is a low-spectral-resolution Fourier dictionary;

[0152] A projection unit, used to determine the spectrum range where the observation signal is located according to the low-resolution atomic number, and project the observation signal to the corresponding atomic number in the fine dictionary according to the corresponding range, wherein the fine dictionary is a high-spectrum resolution Fourier dictionary;

[0153] The optimization unit is used to add the projected atomic numbers to the support set and enter the iterative process. In each iteration, the coefficients of the atoms in the support set are optimized using the least squares method. After each iteration, it is determined whether the residual between the reconstructed signal and the observed signal is less than a threshold. If so, the iteration ends and the reconstructed signal is obtained.

[0154] The conversion unit is used to convert the wavelet coefficients reconstructed by the OMP algorithm back to the time domain through the inverse Mallat transform to obtain the filtered detection signal.

[0155] 402. A signal conversion module, configured to amplify the filtered detection signal using an operational amplifier to obtain an amplified detection signal, convert the amplified detection signal into an initial digital signal, and perform linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal;

[0156] Specifically, the present embodiment also includes a signal filtering submodule for removing noise from the original detection signal, which may come from electromagnetic interference, the device itself or the surrounding environment. A low-pass filter, a high-pass filter or a bandpass / band-stop filter is used. After filtering, the quality of the signal is significantly improved, providing a purer basis for subsequent processing. An amplifier submodule is used to amplify the filtered signal. The operational amplifier performs well in signal amplification due to its high gain, low noise and high input impedance. Through circuit configuration (such as in-phase amplifier, inverting amplifier, etc.), the amplification factor can be accurately controlled to ensure that the amplified signal is neither distorted nor introduces additional noise, thereby obtaining a detection signal with sufficient amplitude to meet the dynamic range requirements of subsequent AD conversion. A digital signal conversion submodule is used to convert the amplified analog signal into a digital signal, which is completed by an analog-to-digital converter (ADC). Analog-to-digital conversion is the conversion of a continuous analog voltage or current into a discrete sequence of digital values. The least squares linear fitting submodule is used to perform linear fitting on the initial digital signal using the least squares algorithm. The least squares method finds the best fitting line by minimizing the sum of squared errors. The error refers to the vertical distance from each data point to the fitting line. The process can effectively eliminate the influence of random errors and extract the main trend of the signal, thereby obtaining a clearer target digital signal, which is extremely important for subsequent data analysis, pattern recognition or control strategy formulation. The extraction submodule is used to extract the original detection signal from the noise interference environment through the processing flow, and finally convert it into a highly pure and representative target digital signal after amplification, digitization, and optimization through mathematical models.

[0157] 403. A parameter identification module, which is used to construct a random forest model, optimize the hyperparameters of the random forest model using a genetic algorithm, obtain a parameter identification model, extract various characteristic parameters of the gear through the parameter identification model based on the target digital signal, and output the parameter identification result;

[0158] Specifically, the parameter identification module in this embodiment includes a principal component analysis submodule, an initialization submodule, a calculation submodule, a genetic processing submodule, a repeated iteration submodule and a prediction submodule, wherein: the principal component analysis submodule is used to obtain the target digital signal of the gear measurement center, and perform standardization processing and principal component analysis processing on the target digital signal in turn to obtain sample data; the initialization submodule is used to construct a random forest model, input the sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, and use these n individuals as the initial population, and the initial population size is set to 20; the calculation submodule is used to construct a fitness function according to the objective function and the penalty function, and calculate the fitness of each individual in the population according to the fitness function. The expression of the fitness function is:

[0159]

[0160] In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, The penalty factor represents the inequality constraint; the genetic processing submodule is used to select, crossover and mutate the group based on the fitness of the individual, with the probability of crossover being 0.5 and the probability of mutation being 0.1, and a new group is obtained as the population at the beginning of the next iteration; the repeated iteration submodule is used to repeat the iteration until the number of iterations t equals the maximum number of iterations, and the individual with the largest fitness in the population is taken as the optimal solution, and the optimal solution is assigned to the random forest model to obtain a parameter identification model; the prediction submodule is used to input the target digital signal into the parameter identification model, and predict the mapping relationship between the features in the target digital signal and the gear parameters through the parameter identification model, extract the various characteristic parameters of the gear, and output the parameter identification results, wherein the parameter identification results at least include the gear type, gear pitch, gear tooth shape, and gear tooth direction.

[0161] 404. A dynamic adjustment module, for analyzing the parameter error of the gear based on the parameter identification result, and dynamically adjusting the measurement parameters based on the error analysis result by using an improved flock optimization algorithm model, and generating a compensation detection signal corresponding to the gear measurement center according to the measurement parameters;

[0162] Specifically, the present embodiment also includes an extraction submodule, which is used to extract characteristic parameters related to the error from the parameter identification results, calculate the statistics of the characteristic parameters, analyze the error size and the degree of error fluctuation, and obtain the error analysis results, wherein the statistics at least include the mean, variance, and standard deviation; an optimization submodule, which is used to construct an improved swarm optimization algorithm model, use the error analysis results as the input of the algorithm, use the gear measurement parameters as the optimization variables, and use the improved swarm optimization algorithm model to dynamically adjust the measurement parameters, wherein the measurement parameters at least include the sampling frequency and the measurement point distribution; a collection submodule, which is used to collect the measurement parameters adjusted by the improved swarm optimization algorithm, convert the optimized measurement parameters into an electrical signal format, and obtain a compensated detection signal.

[0163] Specifically, the optimization submodule in this embodiment also includes the following units: an initialization unit, which is used to initialize the chicken flock, set the total number of chickens to 200, the maximum number of iterations to 200, and calculate the fitness of the chicken flock:

[0164]

[0165] Where e represents the gear error vector, J represents the error weight matrix, represents the adjusted measurement parameter vector, represents the number of elements of e, () represents the Euclidean norm function; the update unit is used to determine the individual extreme values ​​and the group extreme values, and classify the chicken flock in descending order of fitness to obtain roosters, hens and chicks, update the positions of the chicken flock members, and adopt an adaptive parameter adjustment mechanism to make the chicken flock move in the search space; the output unit is used to repeat the optimization iteration until the maximum number of iterations is reached, stop the iterative process, and output the optimized measurement parameters.

[0166] 405. A signal correction module is used to transmit the compensated detection signal to the signal control unit of the gear measurement center, fuse the compensated detection signal with the original detection signal through a difference filter compensation algorithm, generate a corrected detection signal, and control the measurement action of the gear measurement center through the signal control unit according to the corrected detection signal.

[0167] Specifically, the signal correction module in this embodiment includes: a signal transmission submodule, which is used to transmit the compensation detection signal to the signal control unit of the gear measurement center, and the signal control unit synchronizes and aligns the received compensation detection signal with the collected original detection signal; a signal fusion submodule, which is used to convert the compensation detection signal into a signal fusion signal by using a difference filtering compensation algorithm. and the original detection signal Fusion, calculation and compensation of detection signals and the original detection signal The difference:

[0168]

[0169] The filter processing submodule is used to compensate the detection signal and the original detection signal The difference Perform filtering:

[0170]

[0171] The signal generation submodule is used to calculate the compensation value and based on the compensation value and the original detection signal Generate a modified detection signal :

[0172]

[0173]

[0174] In the formula, represents the compensation coefficient, Indicates the window size, represents the corrected detection signal, represents the difference after smoothing, is an index variable, indicating a specific position in the signal sequence, Indicates that in the summation process, it is used as an index variable The change of is used to traverse all positions in the window, Indicates at location The difference between the two values ​​is obtained by the embodiment of the present invention; the instruction generation submodule is used to identify the equipment parameters that need to be adjusted according to the corrected detection signal, and generate corresponding control instructions, and send the control instructions to the measuring equipment of the gear measurement center to control and execute the measurement task through the signal control unit, wherein the equipment parameters include at least the probe path and the scanning speed; the iterative optimization monitoring submodule is used to monitor the measurement process in real time, collect measurement data and equipment status information, and generate feedback data, so as to continuously iterate and optimize the signal control process based on the feedback data.

[0175] Its beneficial effects are: 1. Comprehensive signal acquisition and efficient preprocessing: First, the system integrates multiple types of sensors to ensure real-time capture of the subtle dynamics of the gear measurement center from multiple dimensions, laying a solid foundation for subsequent analysis. The Mallat fast wavelet transform algorithm is applied to effectively filter out background noise and irrelevant interference to maintain signal purity. Subsequently, compression processing and OMP algorithm are implemented to further refine the signal, which not only reduces the data volume, but also enhances the recognition of key features in the signal, and improves data processing speed and storage efficiency. 2. Accurate signal amplification and digital conversion: Through the operational amplification processing of the filtered signal, even weak signal changes can be significantly amplified, ensuring the sensitivity of subsequent analysis. 3. Intelligent feature extraction and parameter identification: The constructed random forest model is combined with the hyperparameter optimization of the genetic algorithm to form a powerful parameter identification system. This model can efficiently and accurately extract the key characteristic parameters of the gear based on the target digital signal, which not only improves the accuracy of recognition, but also speeds up the speed of feature extraction, making the gear state assessment more scientific and comprehensive. 4. Dynamic error analysis and optimization adjustment: Based on the parameter identification results, the improved flock optimization algorithm model is adopted to realize the dynamic adjustment of the measurement parameters. The process is carried out in real time according to the error analysis, which significantly enhances the adaptability and self-optimization ability of the system and ensures the continuity and accuracy of the measurement results. 5. Compensation signal generation and fusion control: The compensation detection signal obtained through the above analysis specifically compensates for the deviation of the original detection signal, and the difference filter compensation algorithm is used to fuse the two sets of signals to generate a corrected detection signal. The correction signal is input into the signal control unit as a command to guide the measurement center to perform more accurate measurement actions, fundamentally improving the stability and accuracy of the measurement operation.

[0176] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A detection signal control method for a gear measurement center, It is characterized in that The detection signal control method of the gear measurement center comprises the following steps: The original detection signal of the gear measurement center is collected in real time by multiple sensors, and the collected original detection signal is filtered by using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal; Amplifying the filtered detection signal using an operational amplifier to obtain an amplified detection signal, converting the amplified detection signal into an initial digital signal, and performing linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal; Constructing a random forest model, optimizing the hyperparameters of the random forest model using a genetic algorithm to obtain a parameter identification model, extracting various characteristic parameters of the gear through the parameter identification model based on the target digital signal, and outputting the parameter identification result; Analyzing the parameter error of the gear based on the parameter identification result, and dynamically adjusting the measurement parameters based on the error analysis result by using an improved flock optimization algorithm model, and generating a compensation detection signal corresponding to the gear measurement center according to the measurement parameters; The compensated detection signal is transmitted to the signal control unit of the gear measurement center, and the compensated detection signal and the original detection signal are fused through a difference filtering compensation algorithm to generate a corrected detection signal. The measurement action of the gear measurement center is controlled by the signal control unit according to the corrected detection signal.

2. A detection signal control method for a gear measurement center according to claim 1, It is characterized in that The original detection signal of the gear measurement center is collected in real time by multiple sensors, and the collected original detection signal is filtered by using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain the filtered detection signal, including: By setting up a sensor array, a comprehensive scan is performed on the running gear at the gear detection center to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor; The Mallat fast wavelet transform algorithm is used to decompose the original detection signal in the wavelet domain to decompose the original detection signal into detail components at multiple scales, and threshold quantization processing is performed according to the soft threshold to obtain the wavelet coefficients: In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points; A Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain an observation signal. The compressed wavelet coefficients are reconstructed using the OMP algorithm based on the observation signal, and the reconstructed wavelet coefficients are inversely transformed to obtain a filtered detection signal.

3. The detection signal control method of a gear measurement center according to claim 2, It is characterized in that The Gaussian random matrix is ​​selected as the observation matrix, the wavelet coefficients obtained after decomposition are compressed to obtain an observation signal, the compressed wavelet coefficients are reconstructed based on the observation signal using the OMP algorithm, and the reconstructed wavelet coefficients are inversely transformed to obtain a filtered detection signal, including: The Gaussian random matrix is ​​selected as the observation matrix, and the wavelet coefficients obtained after decomposition are compressed to obtain the observation signal; Projecting the observed signal onto all atoms of an index dictionary, finding the atomic number of the index dictionary with the largest projection coefficient, and obtaining a low-resolution atomic number, wherein the index dictionary is a low-spectral-resolution Fourier dictionary; Determine the spectrum range where the observation signal is located according to the low-resolution atomic number, and project the observation signal to the corresponding atomic number in a fine dictionary according to the corresponding range, wherein the fine dictionary is a high-spectrum-resolution Fourier dictionary; The projected atomic numbers are added to the support set and the iterative process is started. In each iteration, the coefficients of the atoms in the support set are optimized using the least squares method. After each iteration, it is determined whether the residual between the reconstructed signal and the observed signal is less than a threshold. If so, the iteration ends and the reconstructed signal is obtained. The wavelet coefficients reconstructed by the OMP algorithm are converted back to the time domain through inverse Mallat transform to obtain the filtered detection signal.

4. The detection signal control method of a gear measurement center according to claim 1, It is characterized in that The random forest model is constructed, and the hyperparameters of the random forest model are optimized by a genetic algorithm to obtain a parameter identification model. The characteristic parameters of the gear are extracted through the parameter identification model based on the target digital signal, and the parameter identification results are output, including: Acquire a target digital signal of a gear measurement center, and perform standardization processing and principal component analysis processing on the target digital signal in sequence to obtain sample data; Construct a random forest model, input the sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20; The fitness function is constructed according to the objective function and the penalty function. The fitness of each individual in the population is calculated according to the fitness function. The expression of the fitness function is: In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint; The group is selected, crossed, and mutated based on the fitness of the individual. The probability of crossover is 0.5 and the probability of mutation is 0.

1. The new group is used as the population at the beginning of the next iteration. Repeat the iteration until the number of iterations t equals the maximum number of iterations, and take the individual with the largest fitness in the population as the optimal solution. Assign the optimal solution to the random forest model to obtain the parameter identification model; The target digital signal is input into the parameter identification model, and the mapping relationship between the characteristics in the target digital signal and the gear parameters is predicted by the parameter identification model, and various characteristic parameters of the gear are extracted, and the parameter identification result is output, wherein the parameter identification result at least includes the gear type, gear pitch, gear tooth shape, and gear tooth direction.

5. The detection signal control method of a gear measurement center according to claim 1, It is characterized in that The method of analyzing the parameter error of the gear based on the parameter identification result, dynamically adjusting the measurement parameters based on the error analysis result by using an improved flock optimization algorithm model, and generating a compensation detection signal corresponding to the gear measurement center according to the measurement parameters includes: Extracting characteristic parameters related to the error from the parameter identification results, calculating the statistics of the characteristic parameters, analyzing the error size and the error fluctuation degree, and obtaining the error analysis results, wherein the statistics at least include the mean, variance, and standard deviation; Constructing an improved swarm optimization algorithm model, taking the error analysis results as the input of the algorithm, taking the gear measurement parameters as the optimization variables, and dynamically adjusting the measurement parameters using the improved swarm optimization algorithm model, wherein the measurement parameters at least include sampling frequency and measurement point distribution; The measurement parameters adjusted by the improved flock optimization algorithm are collected, and the optimized measurement parameters are converted into an electrical signal format to obtain a compensation detection signal.

6. A detection signal control method for a gear measurement center according to claim 5, It is characterized in that The improved swarm optimization algorithm model is constructed, the error analysis result is used as the input of the algorithm, the gear measurement parameter is used as the optimization variable, and the improved swarm optimization algorithm model is used to dynamically adjust the measurement parameter, including: Initialize the chicken flock, set the total number of chickens to 200, the maximum number of iterations to 200, and calculate the fitness of the chicken flock: Where e represents the gear error vector, J represents the error weight matrix, represents the adjusted measurement parameter vector, represents the number of elements of e, () represents the Euclidean norm function; Determine individual extreme values ​​and group extreme values, and sort the chickens in descending order of fitness to obtain roosters, hens and chicks, update the positions of the chicken members, and use an adaptive parameter adjustment mechanism to move the chickens in the search space; Repeat the optimization iteration until the maximum number of iterations is reached, stop the iteration process, and output the optimized measurement parameters.

7. A detection signal control method for a gear measurement center according to claim 1, It is characterized in that The method of transmitting the compensated detection signal to a signal control unit of a gear measurement center, fusing the compensated detection signal with the original detection signal through a difference filter compensation algorithm to generate a corrected detection signal, and controlling the measurement action of the gear measurement center through the signal control unit according to the corrected detection signal, comprises: Transmitting the compensated detection signal to a signal control unit of a gear measurement center, and the signal control unit synchronizes and aligns the received compensated detection signal with the collected original detection signal; The compensation detection signal is converted into and the original detection signal Fusion, calculation of the compensated detection signal and the original detection signal The difference: The compensation detection signal and the original detection signal The difference Perform filtering: Calculate the compensation value, and based on the compensation value and the original detection signal Generate a modified detection signal : In the formula, represents the compensation coefficient, Indicates the window size, represents the corrected detection signal, represents the difference after smoothing, is an index variable, indicating a specific position in the signal sequence, Indicates that in the summation process, it is used as an index variable The change of is used to traverse all positions in the window, Indicates at location The difference between Identify the equipment parameters that need to be adjusted according to the corrected detection signal, generate corresponding control instructions, and send the control instructions to the measuring equipment of the gear measurement center to control and perform the measurement task through the signal control unit, wherein the equipment parameters at least include the probe path and the scanning speed; The measurement process is monitored in real time, the measurement data and device status information are collected, and feedback data is generated to continuously iterate and optimize the signal control process based on the feedback data.

8. A detection signal control system for a gear measurement center, It is characterized in that The detection signal control system of the gear measurement center includes a signal filtering module, a signal conversion module, a parameter identification module, a dynamic adjustment module and a signal correction module, wherein: A signal filtering module is used to collect the original detection signal of the gear measurement center in real time through a variety of sensors, and filter the collected original detection signal using Mallat fast wavelet transform algorithm, compression processing and OMP algorithm to obtain a filtered detection signal; A signal conversion module is used to amplify the filtered detection signal using an operational amplifier to obtain an amplified detection signal, convert the amplified detection signal into an initial digital signal, and perform linear fitting on the initial digital signal using a least squares algorithm to obtain a target digital signal; A parameter identification module is used to construct a random forest model, optimize the hyperparameters of the random forest model using a genetic algorithm to obtain a parameter identification model, extract various characteristic parameters of the gear based on the target digital signal through the parameter identification model, and output the parameter identification result; A dynamic adjustment module, used to analyze the parameter error of the gear based on the parameter identification result, and dynamically adjust the measurement parameters based on the error analysis result by using an improved chicken swarm optimization algorithm model, and generate a compensation detection signal corresponding to the gear measurement center according to the measurement parameters; A signal correction module is used to transmit the compensated detection signal to the signal control unit of the gear measurement center, fuse the compensated detection signal and the original detection signal through a difference filtering compensation algorithm to generate a corrected detection signal, and control the measurement action of the gear measurement center through the signal control unit according to the corrected detection signal.

9. A detection signal control system for a gear measurement center according to claim 8, It is characterized in that The signal filtering module includes a signal collection submodule, a wavelet decomposition submodule and a signal reconstruction submodule, wherein: A signal collection submodule is used to perform a comprehensive scan of the running gears at the gear detection center through a set sensor array to collect multi-dimensional original detection signals, wherein the sensor array includes at least a displacement sensor, a velocity sensor, and an acceleration sensor; The wavelet decomposition submodule is used to decompose the original detection signal in the wavelet domain by using the Mallat fast wavelet transform algorithm, so as to decompose the original detection signal into detail components at multiple scales, and perform threshold quantization processing according to the soft threshold to obtain the wavelet coefficients: In the formula, Represents the wavelet decomposition of the original detection signal The first layer in the layer detail coefficient group coefficients, represents the wavelet transform coefficient after threshold quantization, represents the threshold of wavelet transform, represents the standard deviation of the noise signal and n∈[1,N], where N represents the signal length, is a sign function, is the number of data points; The signal reconstruction submodule is used to select a Gaussian random matrix as the observation matrix, compress the wavelet coefficients obtained after decomposition to obtain an observation signal, reconstruct the compressed wavelet coefficients based on the observation signal using the OMP algorithm, and obtain the filtered detection signal by performing an inverse wavelet transform on the reconstructed wavelet coefficients.

10. A detection signal control system for a gear measurement center according to claim 8, It is characterized in that The parameter identification module includes a principal component analysis submodule, an initialization submodule, a calculation submodule, a genetic processing submodule, a repetition submodule and a prediction submodule, wherein: A principal component analysis submodule is used to obtain a target digital signal of a gear measurement center, and perform standardization processing and principal component analysis processing on the target digital signal in sequence to obtain sample data; The initialization submodule is used to construct a random forest model, input the sample data into the random forest model, initialize the population, set the maximum number of iterations of the genetic algorithm to 100 as the algorithm termination condition, initialize the iteration count to t=0, and randomly generate n individuals, use these n individuals as the initial population, and set the initial population size to 20; The calculation submodule is used to construct a fitness function based on the objective function and the penalty function, and calculate the fitness of each individual in the population according to the fitness function. The expression of the fitness function is: In the formula, represents the fitness function, represents the objective function, and represents the penalty function constructed by equality and inequality constraints, represents the penalty factor for the equality constraint, represents the penalty factor of the inequality constraint; The genetic processing submodule is used to select, crossover, and mutate the population based on the fitness of the individual. The probability of crossover is 0.5 and the probability of mutation is 0.

1. The new population is used as the population at the beginning of the next iteration. The repeated iteration submodule is used to repeat the iteration until the number of iterations t equals the maximum number of iterations, and the individual with the largest fitness in the population is taken as the optimal solution. The optimal solution is assigned to the random forest model to obtain a parameter identification model; A prediction submodule is used to input the target digital signal into the parameter identification model, predict the mapping relationship between the characteristics in the target digital signal and the gear parameters through the parameter identification model, extract the various characteristic parameters of the gear, and output the parameter identification result, wherein the parameter identification result at least includes the gear type, gear pitch, gear tooth shape, and gear tooth direction.

Images