An automatic calibration method for analog quantity modules

By calculating the gain coefficient and offset using the digital twin method and the loss reduction method, the analog quantity module is automatically calibrated, which solves the problem of insufficient accuracy in the entire range in the prior art, and achieves a calibration effect of higher quality and reliability.

CN118801885BActive Publication Date: 2024-11-29KUNMING UNIONSCIENCE TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411267348.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-11
Publication Date
2024-11-29
Estimated Expiration
2044-09-11

AI Technical Summary

Technical Problem

When calibrating the analog quantity module, the prior art can only perform well at a few specific points, and the accuracy cannot be guaranteed throughout the range, and it is prone to underfitting or overfitting problems caused by limited data volume.

Method used

An automatic calibration method is adopted, by selecting a known high accuracy analog module as the calibration source module, connecting it with the module to be calibrated, using the digital twin method and the loss drop method to calculate the gain coefficient and offset, and updating the module to be calibrated internal calibration parameters until the error is less than the preset threshold.

Benefits of technology

Ensure that the module to be calibrated not only performs well at the calibration point, but also achieves optimal performance over the entire range, improves the overall quality and reliability of the calibration, and avoids underfitting or overfitting.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118801885B_ABST
    Figure CN118801885B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of automatic calibration technology, and in particular, to an automatic calibration method for an analog module. The method comprises the following steps: using a digital twin method and combining a calibration source module to perform error calculation, using a loss descent method to calculate a gain coefficient K and an offset B according to the error, and sending the gain coefficient K and the offset B to a module to be calibrated, and the module to be calibrated updates an internal calibration parameter, and after multiple iterations, the error is less than a preset threshold value; the method ensures that the module to be calibrated not only performs well at the calibration point, but also approaches the optimal performance that can be achieved in theory as much as possible, thereby improving the overall quality and reliability of the calibration, while considering the errors in physical and virtual environments, ensuring that the performance of the module to be calibrated in actual applications and under ideal conditions can reach high standards, and calculating the errors in a virtual environment also increases the training data set, avoiding the problems of overfitting and underfitting.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of automatic calibration, and in particular to an automatic calibration method for an analog quantity module. Background Art

[0002] In modern industrial production, the improvement of automation and intelligence levels is crucial to improving production efficiency, product quality and safety. As a key component in the automation system, analog modules are responsible for converting physical quantities into electrical signals for processing and decision-making by the control system. The accuracy of these modules is directly related to the performance and reliability of the entire system. Over time, analog modules may shift due to environmental factors, equipment aging or mechanical wear, resulting in a decrease in the accuracy of the measurement results. Regular calibration is a necessary measure to ensure that the module continues to meet the accuracy standards. It helps to restore the original performance of the module and ensure the reliability and consistency of the measurement data. Traditional calibration methods often perform well only at a few specific points, but cannot guarantee accuracy over the entire range, and are prone to underfitting or overfitting problems caused by limited data. Therefore, an automatic calibration method for analog modules is designed. Summary of the invention

[0003] The purpose of the present invention is to provide an automatic calibration method for an analog module to solve the problem proposed in the above background technology that the accuracy is only good at a few specific points but cannot be guaranteed within the entire range, and is prone to underfitting or overfitting due to limited data volume.

[0004] To achieve the above object, the present invention provides an automatic calibration method for an analog module, comprising the following steps:

[0005] S1. Select an analog module with known high accuracy as a calibration source module, and the range of the calibration source module covers the range of the module to be calibrated;

[0006] S2. Use a signal line to connect the output terminal of the calibration source module to the input terminal of the module to be calibrated, and use an isolation amplifier to avoid signal interference;

[0007] S3. Use the digital twin method and combine it with the calibration source module to calculate the error. According to the error, use the loss reduction method to calculate the gain coefficient and offset , and the gain factor and offset The module to be calibrated updates its internal calibration parameters after multiple iterations until the error is less than a preset threshold.

[0008] S4. After the calibration is completed, the module to be calibrated is tested, and the accuracy of the calibration result is verified using data points that are not involved in the calibration. The calibration effect is verified by testing under different environmental conditions. At the same time, the host computer records the key data of the entire calibration process and generates a calibration report.

[0009] As a further improvement of the technical solution, the calibration source module has been rigorously calibrated for multiple times, and its output value is within an acceptable range with the theoretical value.

[0010] As a further improvement of the technical solution, the specific steps of S3 are as follows:

[0011] S31, controlling the calibration source through the host computer to generate a calibration signal, recording the first response value output by the module to be calibrated according to the calibration signal value; using the digital twin method to establish a virtual calibration source, simulating the same calibration signal value to output a second response value;

[0012] S32, calculating an error between a calibration signal value of the calibration source module and a first response value to obtain a first error, and calculating an error between a second response value and the first response value to obtain a second error, and evaluating the first error and the second error based on an error indicator;

[0013] S33, based on the first error and the second error, calculate the gain coefficient using the loss reduction method and offset ,Will Gain factor and offset Send to the module to be calibrated to update the internal calibration parameters;

[0014] S34, repeat S31 to S33 until the first error and the second error are both smaller than a preset threshold.

[0015] As a further improvement of the technical solution, in S32, the first error and the second error are evaluated based on the error index as follows:

[0016] The first error is evaluated based on the bias-weighted root mean square error as follows:

[0017] ;

[0018] ;

[0019] in, is the root mean square error weighted by the bias; is the number of first error samples; ; For the calibration signal value; For the First response value; Forecast bias;

[0020] The evaluation criteria for the second error are based on the normalized mixed error, as follows:

[0021] ;

[0022] in, is the normalized mixed error; is the number of error indicators; ; For the The weight of the error indicator; For the The value of the error indicator; For the The maximum value reached by the error indicator.

[0023] As a further improvement of the technical solution, in S33, based on the first error and the second error, the gain coefficient is calculated using the loss descent method. and offset The details are as follows:

[0024] Construct a least squares function:

[0025] ;

[0026] in, is the comprehensive error function; is the number of calibration points; ; is the error weight; For the The calibration signal value of each calibration point; is the gain coefficient; is the first response value; is the offset; is the second response value;

[0027] Use gradient descent to minimize the loss function:

[0028] ;

[0029] ;

[0030] in, is the current value of the gain coefficient; is the value of the gain coefficient in the previous iteration; is the learning rate; The current value of the offset; The value of the previous iteration for the offset.

[0031] As a further improvement of the technical solution, the gradient descent method with momentum is introduced in the minimization function using the gradient descent method, and after optimization, it is specifically as follows:

[0032] ;

[0033] ;

[0034] in, For the gain factor The momentum term of For the offset The momentum term of is the momentum factor; is the learning rate;

[0035] ;

[0036] ;

[0037] in, is the current value of the optimized gain coefficient; The current value of the optimized offset.

[0038] As a further improvement of the technical solution, before executing S3, an error threshold needs to be defined in advance for comparison with the first error and the second error, and a maximum number of iterations needs to be defined in advance to avoid an infinite loop.

[0039] As a further improvement of the technical solution, the specific steps of S4 are as follows:

[0040] S41, selecting low, medium and high data points that have not been used in the previous calibration process and that cover all parts of the measurement range, measuring these data points using the calibrated module to be calibrated, and recording the results, comparing the measurement results with the calibration signal value, calculating the error, and then quantifying the performance after calibration using the evaluation standard;

[0041] S42, testing the calibration effect under different temperature and humidity environments, and observing whether the calibration effect is consistent. If not, re-calibrate until the calibration effect under different environments is consistent with the calibration effect under ideal conditions;

[0042] S43. The host computer records all key data in the calibration process, and can identify data trends and abnormal values, and finally generate a calibration report.

[0043] As a further improvement of the present technical solution, the key data involved in S43 include the calibration signal value, the output value of the module to be calibrated, the calculated gain coefficient and offset and calibration results.

[0044] As a further improvement of the present technical solution, the calibration report generated in S43 includes the calibration date, time, information of the calibration source used, parameter changes before and after calibration, and an error curve.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. This automatic calibration method for analog modules not only considers the error between the physical calibration source module and the module to be calibrated, but also takes into account the error between the module and the ideal response value simulated by the digital twin. This ensures that the module to be calibrated not only performs well at the calibration point, but also approaches the theoretically optimal performance as much as possible, thereby improving the overall quality and reliability of the calibration. At the same time, the training data set is increased to avoid overfitting and underfitting problems.

[0047] 2. In this automatic calibration method for analog modules, the loss descent method combines the least squares method and the gradient descent method. It not only considers the actual calibration data, but also utilizes the simulation data of the digital twin technology, thereby improving the accuracy and robustness of the calibration. The introduction of the gradient descent method with momentum can not only accelerate the convergence process, but also improve the algorithm's ability to resist local minima, thereby more effectively optimizing the calibration parameters of the analog module. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flowchart of the overall method of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Example

[0050] See also Figure 1 As shown, a method for automatic calibration of an analog module is provided, comprising the following steps:

[0051] S1. Select an analog module with known high accuracy as the calibration source module. The range of the calibration source module covers the range of the module to be calibrated. The calibration source module has been strictly calibrated for many times, and its output value is within the acceptable range with the theoretical value.

[0052] S2. Use signal lines to connect the output terminals of the calibration source module to the input terminals of the module to be calibrated, and use an isolation amplifier to avoid signal interference; ensure that the connection is stable, avoid poor contact, and ensure electrical isolation between the two modules.

[0053] S3. Use the digital twin method and combine it with the calibration source module to calculate the error. According to the error, use the loss reduction method to calculate the gain coefficient. and offset , and the gain factor and offset The module to be calibrated updates its internal calibration parameters after multiple iterations until the error is less than a preset threshold.

[0054] The specific steps of S3 are as follows:

[0055] S31, controlling the calibration source through the host computer to generate a calibration signal, recording the first response value output by the module to be calibrated according to the calibration signal value; using the digital twin method to establish a virtual calibration source, simulating the same calibration signal value to output a second response value;

[0056] S32, calculating an error between a calibration signal value of the calibration source module and a first response value to obtain a first error, and calculating an error between a second response value and the first response value to obtain a second error, and evaluating the first error and the second error based on an error indicator;

[0057] The first error and the second error are evaluated based on the error index as follows:

[0058] The first error assessment is based on the bias-weighted root mean square error, which is an indicator for evaluating forecast accuracy. It aims to remove systematic biases in forecasts to make the assessment more fair and accurate, as follows:

[0059] ;

[0060] ;

[0061] in, is the root mean square error weighted by the bias; is the number of first error samples; ; For the calibration signal value; For the First response value; is the forecast deviation, which is the average deviation of all forecast values ​​relative to the actual values;

[0062] By subtracting the forecast bias from the forecast value , effectively eliminating systematic bias, making the evaluation more accurate and avoiding exaggeration or underestimation of errors caused by systematic bias. After removing the bias, Provides a fairer measure of model performance that reflects the true difference between the calibration signal value and the first response value, without being affected by systematic biases.

[0063] The evaluation standard of the second error is based on the normalized mixed error, which is a method of evaluating by integrating multiple error indicators. It ensures comparability at different scales or magnitudes by normalizing different error indicators and assigning different weights to each of them. The details are as follows:

[0064] ;

[0065] in, is the normalized mixed error; is the number of error indicators; ; For the The weight of the error indicator; For the The value of the error indicator; For the The maximum value reached by an error indicator depends on the specific error indicator definition and data range;

[0066] The specific steps are as follows: First, you need to calculate all The value of each error metric, such as root mean square error, mean absolute error, normalized root mean square error, etc. For each error metric, determine its maximum possible value. This usually depends on the range or characteristics of the data. For example, for normalized root mean square error, the maximum possible value is usually 1 (if the data range is used as a normalization factor); divide the value of each error metric by its maximum possible value to ensure that all metrics are on the same scale; multiply the normalized error metric values ​​by their corresponding weights , and then summing, we get The final value of .

[0067] Allowing to assign weights to different error metrics allows the evaluator to adjust the importance of the metric based on specific application needs. For example, if the accuracy of the virtual calibration source is particularly critical in a certain application, the second error can be given a higher weight. , a single evaluation value combining different error metrics can be obtained, which helps to quickly understand the overall performance of the calibration process and the consistency between the digital twin model and the actual module response.

[0068] Combining these two error assessment methods enables in-depth analysis and optimization of the calibration process. Focus on eliminating systematic deviations and ensuring the accuracy and reliability of calibration results. It provides a comprehensive framework that ensures that the performance of the calibration process at different levels and scales is fully considered through comprehensive evaluation of multiple indicators. This combined approach can help researchers understand the sources of errors in the calibration process in more detail, and then take targeted measures to improve the calibration effect.

[0069] S33, based on the first error and the second error, calculate the gain coefficient using the loss reduction method and offset , the gain factor and offset Send to the module to be calibrated to update the internal calibration parameters;

[0070] In S33, based on the first error and the second error, the gain coefficient is calculated using the loss reduction method. and offset The details are as follows:

[0071] Construct a least squares function:

[0072] ;

[0073] in, is the comprehensive error function; is the number of calibration points; ; is the error weight, which is a weight between 0 and 1, used to balance the importance of the first error and the second error. is close to 1, the algorithm will pay more attention to the first error; on the contrary, if If it is close to 0, more emphasis will be placed on the second error; For the The calibration signal value of each calibration point; is the gain coefficient; is the first response value; is the offset; is the second response value;

[0074] Gradient descent is an iterative optimization algorithm used to find the minimum value of the error function. In this example, the goal is to find and The optimal value of Minimum,

[0075] Use gradient descent to minimize the loss function:

[0076] ;

[0077] ;

[0078] in, is the current value of the gain coefficient; is the value of the gain coefficient in the previous iteration; is the learning rate, which is an important parameter in the gradient descent method and determines the step size of parameter update in each iteration. If it is too large, the algorithm may oscillate around the minimum value and fail to converge; if it is too small, the convergence speed will be very slow; The current value of the offset; The value of the previous iteration for the offset.

[0079] in, is the error function About parameters The partial derivative of , which represents the slope of the error function at the current parameter value, is obtained by subtracting a certain multiple of this slope (given by the learning rate Control), we move towards the goal of minimizing the error function; similarly, is the error function About parameters The partial derivative of

[0080] The least square method and gradient descent method are combined to form an effective automatic calibration algorithm for optimizing the analog module. and This approach not only considers the actual calibration data but also utilizes the simulated data from the digital twin technology, thus improving the accuracy and robustness of the calibration.

[0081] In order to more effectively combine the least squares method and the gradient descent method, the gradient descent method with momentum is introduced in the gradient descent method to minimize the function, which can accelerate convergence and avoid falling into the local minimum. After optimization, it is specifically as follows:

[0082] ;

[0083] ;

[0084] in, For the gain factor The momentum term of For the offset The momentum term of ; The two momentum terms accumulate the direction and magnitude information of the past gradients in the gradient descent method to accelerate the convergence process. As the iterations proceed, these momentum terms guide the direction of parameter updates and help the algorithm move toward the global minimum more effectively, rather than relying solely on the direction of the current gradient; is the momentum factor, usually a value between 0 and 1. Its function is to control the decay rate of the momentum term. Means more "memory" of past gradient directions, which helps the algorithm overcome local minima and flat areas, but may increase the risk of excessive oscillation; smaller This will lead to faster momentum decay, and the algorithm may be more easily dominated by the current gradient, but the convergence speed may be slower; is the learning rate, which controls the step size of each parameter update. This can make the algorithm converge faster, but may also skip the minimum region or cause instability; lower This will make the algorithm more stable, but converge more slowly. Choose the appropriate It is an important factor in balancing convergence speed and stability;

[0085] ;

[0086] ;

[0087] in, is the current value of the optimized gain coefficient; The current value of the optimized offset.

[0088] By introducing the gradient descent method with momentum, we can not only accelerate the convergence process, but also improve the algorithm's ability to resist local minima, thereby more effectively optimizing the calibration parameters of the analog module. and According to the current gradient and momentum factor Updated and then used to calculate new parameter values and , and iterate until convergence.

[0089] S34, repeat S31 to S33 until the first error and the second error are both smaller than a preset threshold.

[0090] Before executing S3, you need to define the error threshold in advance to compare with the first error and the second error. You also need to define the maximum number of iterations in advance to avoid infinite loops. The error threshold is the maximum allowable value you want the module output error to reach after calibration. The maximum number of iterations is to avoid infinite loops. Even if the error has not reached the threshold, the iteration is terminated. At this time, manual intervention is required to check whether there are other problems.

[0091] This step not only considers the error between the physical calibration source module and the module to be calibrated, but also takes into account the error with the ideal response value simulated by the digital twin. This ensures that the module to be calibrated not only performs well at the calibration point, but also approaches the theoretically optimal performance as close as possible, thereby improving the overall quality and reliability of the calibration. In short, this method considers the errors in both physical and virtual environments, ensuring that the performance of the module to be calibrated meets high standards in both actual applications and ideal conditions.

[0092] S4. After the calibration is completed, additional tests are performed on the module to be calibrated. The accuracy of the calibration results is verified using data points that are not involved in the calibration. Tests are also performed under different environmental conditions to verify the calibration effect. At the same time, the host computer records the key data of the entire calibration process and generates a calibration report.

[0093] The specific steps of S4 are as follows:

[0094] S41, selecting low, medium and high data points that have not been used in the previous calibration process and that cover all parts of the measurement range, measuring these data points using the calibrated module to be calibrated, and recording the results, comparing the measurement results with the known calibration signal value, calculating the error, and then quantifying the performance after calibration using the evaluation standard;

[0095] S42, testing the calibration effect under different temperature and humidity environments, and observing whether the calibration effect is consistent. If not, re-calibrate until the calibration effect under different environments is consistent with the calibration effect under ideal conditions;

[0096] S43. The host computer records all key data in the calibration process, and can identify data trends and abnormal values, and finally generate a calibration report.

[0097] Key data include calibration signal value, output value of the module to be calibrated, and calculated gain coefficient and offset And the calibration results; where the calibration signal value is a known accurate value used as a reference during the calibration process, usually from a standard or reference device, which provides an absolute benchmark for comparing the output value of the module to be calibrated to determine its accuracy. The calibration signal value helps identify deviations and adjust the module to the correct reading;

[0098] The output value of the module to be calibrated can be compared with the calibration signal value to identify deviations or errors. The recording of these values ​​at different calibration points helps to draw error curves and understand the linearity and accuracy of the module.

[0099] Gain factor The coefficient describes the slope of the module output as the input signal changes, the offset Describes the output value of the module when the input signal is zero. They are used to adjust the output of the module to make it closer to the calibration signal value. After calibration, these coefficients will be used to correct future measurements to improve accuracy;

[0100] The calibration results include error analysis before and after calibration, as well as the final measurement accuracy after the implementation of corrective measures. The calibration results indicate the effectiveness of the calibration process and prove whether the module reaches the expected accuracy level after calibration. In addition, the calibration results are also used to generate calibration certificates or reports to record the calibration status and history.

[0101] The calibration report includes the calibration date, time, calibration source information used, parameter changes before and after calibration, and error curves. The date and time provide the exact timestamp of the calibration activity, which is very important for tracking the calibration cycle and equipment usage history, helping to determine the scheduled time for the next calibration and ensuring that the equipment is always in a calibrated state;

[0102] The calibration source information used records the information of the standard instrument or calibration equipment used, including model, serial number, calibration certificate number and validity period, etc., to ensure the traceability and transparency of the calibration process, and verify the accuracy and effectiveness of the standard instrument itself;

[0103] The parameter changes before and after calibration show the difference between the original reading of the device in the uncalibrated state and the reading after calibration adjustment. This data helps to evaluate the calibration effect, confirm whether the device is within the allowable error range, and quantify the performance improvement brought by calibration;

[0104] The error curve depicts the deviation between the measured value and the standard value of the device over the entire measuring range. It provides a visual representation of the device performance and facilitates the analysis of the nonlinear error, repeatability and stability of the device. The error curve is particularly useful for understanding the accuracy of the device at different measurement points.

[0105] The above steps can ensure the effectiveness of calibration and provide valuable reference for subsequent maintenance and use. If the calibration results do not meet the requirements, it may be necessary to recheck the calibration procedure or look for potential problems in the hardware or software.

[0106] 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, and 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, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. An automatic calibration method for an analog module, characterized in that: The steps include: S1. Select an analog module with known high accuracy as a calibration source module, and the range of the calibration source module covers the range of the module to be calibrated; S2. Use a signal line to connect the output terminal of the calibration source module to the input terminal of the module to be calibrated, and use an isolation amplifier to avoid signal interference; S3. Use the digital twin method and combine it with the calibration source module to calculate the error. According to the error, use the loss reduction method to calculate the gain coefficient and offset , and the gain factor and offset The module to be calibrated updates its internal calibration parameters after multiple iterations until the error is less than a preset threshold. S4. After the calibration is completed, the module to be calibrated is tested, and the accuracy of the calibration result is verified using data points that are not involved in the calibration. The calibration effect is verified by testing under different environmental conditions. At the same time, the host computer records the key data of the entire calibration process and generates a calibration report. The specific steps of S3 are as follows: S31, controlling the calibration source through the host computer to generate a calibration signal, recording the first response value output by the module to be calibrated according to the calibration signal value; using the digital twin method to establish a virtual calibration source, simulating the same calibration signal value to output a second response value; S32, calculating an error between a calibration signal value of the calibration source module and a first response value to obtain a first error, and calculating an error between a second response value and the first response value to obtain a second error, and evaluating the first error and the second error based on an error indicator; S33, based on the first error and the second error, calculate the gain coefficient using the loss reduction method and offset , the gain factor and offset Send to the module to be calibrated to update the internal calibration parameters; S34, repeat S31 to S33 until the first error and the second error are both smaller than a preset threshold; In S33, based on the first error and the second error, the gain coefficient is calculated using the loss reduction method. and offset The details are as follows: Construct a least squares function: ; in, is the comprehensive error function; is the number of calibration points; ; is the error weight; For the The calibration signal value of each calibration point; is the gain coefficient; is the first response value; is the offset; is the second response value; Use gradient descent to minimize the loss function: ; ; in, is the current value of the gain coefficient; is the value of the gain coefficient in the previous iteration; is the learning rate; The current value of the offset; The value of the previous iteration for the offset.

2. The automatic calibration method for an analog module according to claim 1, characterized in that: The calibration source module has been rigorously calibrated for many times, and its output value is within an acceptable range with the theoretical value.

3. The automatic calibration method for an analog module according to claim 2, characterized in that: In S32, the first error and the second error are evaluated based on the error index as follows: The first error is evaluated based on the bias-weighted root mean square error as follows: ; ; in, is the root mean square error weighted by the bias; is the number of first error samples; ; For the calibration signal value; For the First response value; Forecast bias; The evaluation criteria for the second error are based on the normalized mixed error, as follows: ; in, is the normalized mixed error; is the number of error indicators; ; For the The weight of the error indicator; For the The value of the error indicator; For the The maximum value reached by the error indicator.

4. The automatic calibration method for an analog module according to claim 3, characterized in that: The gradient descent method with momentum is introduced in the minimization function using the gradient descent method, and the specific optimization is as follows: ; ; in, For the gain factor The momentum term of For the offset The momentum term of is the momentum factor; is the learning rate; ; ; in, is the current value of the optimized gain coefficient; The current value of the optimized offset.

5. The automatic calibration method for an analog module according to claim 4, characterized in that: Before executing S3, an error threshold needs to be defined in advance for comparison with the first error and the second error. A maximum number of iterations also needs to be defined in advance to avoid an infinite loop.

6. The automatic calibration method for an analog module according to claim 5, characterized in that: The specific steps of S4 are as follows: S41, selecting low, medium and high data points that have not been used in the previous calibration process and that cover all parts of the measurement range, measuring these data points using the calibrated module to be calibrated, and recording the results, comparing the measurement results with the calibration signal value, calculating the error, and then quantifying the performance after calibration using the evaluation standard; S42, testing the calibration effect under different temperature and humidity environments, and observing whether the calibration effect is consistent. If not, re-calibrate until the calibration effect under different environments is consistent with the calibration effect under ideal conditions; S43. The host computer records all key data in the calibration process, and can identify data trends and abnormal values, and finally generate a calibration report.

7. The automatic calibration method for an analog module according to claim 6, characterized in that: The key data involved in S43 include the calibration signal value, the output value of the module to be calibrated, the calculated gain coefficient and offset and calibration results.

8. The automatic calibration method for an analog module according to claim 7, characterized in that: The calibration report generated in S43 includes the calibration date, time, information of the calibration source used, parameter changes before and after calibration, and an error curve.

Citation Information

Patent Citations

  • MEMS accelerometer error compensation method based on LMS adaptive filtering and gradient descent

    CN109631889A

  • Automatic calibration method and system for analog input

    CN115865089A