Adaptive filter correction method, device and system for measurement data

Abstract

The application provides a kind of adaptive filtering correction method, device and system of measurement data, comprising: obtaining current measurement value;Determine measurement statistical characteristics based on historical measurement value and corresponding historical filter value, measurement statistical characteristics include measurement noise variance and measurement fluctuation factor;According to measurement statistical characteristics, determine target adjustment weight;Based on target adjustment weight, the weighted addition of current measurement value and last filter value obtains the filter value of current sample point.The method of the application realizes the tracking of measurement change trend by including measurement noise variance and measurement fluctuation factor in measurement statistical characteristics, and realizes the suppression of abnormal measurement value based on target adjustment weight dynamic filtering, thereby improving the adaptability and robustness of non-steady state measurement, while reducing the computational complexity.

Description

Technical Field

[0001] This invention relates to the field of physical quantity measurement technology, and more specifically, to an adaptive filtering correction method, apparatus, and system for measurement data. Background Technology

[0002] In the field of physical quantity measurement, the equipment itself typically generates background measurement noise due to thermal noise, temperature drift, quantization errors, etc. Electromagnetic interference and mechanical vibration from the surrounding environment also cause noise to exhibit time-varying characteristics. Furthermore, the signal strength and distance of the measured physical quantity can cause variations in the measurement signal-to-noise ratio. Due to unforeseen factors such as switching, poor contact, or sudden computational anomalies, measurement results may occasionally show significant deviations from the actual values. In engineering, the true value of a physical quantity often exhibits a non-linear trend over time. Therefore, in most scenarios, the statistical characteristics of the measured values, such as variance, are time-varying, exhibiting non-steady-state characteristics.

[0003] Most existing measurement filtering correction methods are designed to deal with specific measurement scenarios or data characteristics. When faced with complex and ever-changing engineering application environments, they suffer from poor filtering smoothness, filtering lag, poor adaptability to non-steady and nonlinear measurement data, poor robustness, and high computational complexity.

[0004] Specifically, existing technologies typically assume that noise is stationary and the overall data exhibits a linear trend, but they cannot combat occasional outliers that deviate from the actual values. For example, moving average and median filtering methods take the average or median of several consecutive samples, resulting in poor signal smoothness and a certain lag in the filtering results, making them unsuitable for real-time measurement tasks. Forgetting filtering assigns weights to previous filtering results; however, it suffers from phase lag, low sensitivity, and because the parameters are fixed, it cannot track nonlinear changes in the measured values ​​and cannot handle time-varying noise. Adaptive filtering methods, such as Least Mean Square (LMS), have high memory consumption, high computational complexity, and are prone to numerical divergence and convergence issues. Furthermore, applying machine learning techniques to learn the potential distribution relationship between noise and clean signals to achieve data-driven adaptive filtering improves the tracking ability and robustness to non-stationary conditions. However, manually labeling large amounts of high-quality training data is difficult, the model's generalization ability is limited, and the high computational and storage complexity of multi-layer neural networks makes them unsuitable for measurement applications with limited computing power.

[0005] Therefore, it is necessary to provide a new adaptive data filtering correction method to solve the problems of poor filtering smoothness, filtering lag, poor adaptability to unsteady and nonlinear measurements, poor robustness and high computational complexity. Summary of the Invention

[0006] The present invention aims to provide an adaptive filtering correction method, device and system for measurement data, so as to improve the smoothness of filtering, reduce the hysteresis of filtering, improve the adaptability to non-steady and nonlinear measurements, improve robustness and reduce computational complexity.

[0007] To achieve the above objectives, the present invention provides an adaptive filtering correction method for measurement data, comprising:

[0008] Step S1: Obtain measurement data as the measurement value of the current sample point;

[0009] Step S2: Determine the measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance and measurement fluctuation factor;

[0010] Step S3: Determine the target adjustment weights based on the measurement statistical characteristics;

[0011] Step S4: Adjust the weights based on the target, and add the measured value of the current sample point to the filtered value of the previous sample point to obtain the filtered value of the current sample point.

[0012] In step S2, if the number of historical measurements and historical filtered values ​​does not reach the effective statistical sample size, a preset static parameter is used as the measurement statistical characteristic; otherwise, the measurement noise variance is calculated based on multiple historical measurements and corresponding historical filtered values. and measuring volatility factor As a measurement statistical characteristic.

[0013] In step S2, a sliding window with the same size as the effective statistical sample size is used to extract historical measurement values ​​and corresponding historical filtered values; the measurement values ​​and filtered values ​​are cyclically stored in two storage spaces of the same size, and the effective statistical sample size and the size of the sliding window are both equal to the size of the storage space.

[0014] The measurement noise variance is obtained by calculating the variance of multiple historical measurements. The historical filter variance is obtained by calculating the variance of multiple historical filter values, and the measurement fluctuation factor is obtained by dividing the historical filter variance by a preset coefficient. .

[0015] Before step S3, there is also step S3': determine whether the measured value of the current sample point is the measured value of the first sample point. If so, use the measured value of the current sample point as the current filter value, skip the subsequent steps, and return to step S1.

[0016] Step S3 specifically includes:

[0017] Step S31: Based on the measured fluctuation factor of the current sample points Combined confidence variance with the previous sample point Adjusting weights with targets Obtain the corresponding measurement error variance and overall confidence variance ;

[0018] Step S32: Based on the overall confidence variance of the current sample points and measurement noise variance Obtain the target adjustment weights for the current sample points. .

[0019] In step S31, when the sample point index k=0 and 1, the measurement error variance of the current sample point is... Obtained through initialization;

[0020] In the sample point index When -1, the measurement error variance of the current sample point is... The value is updated to the measurement noise variance. ;

[0021] Otherwise, the measurement error variance of the current sample points The measurement error variance of the current sample point is updated based on the target-adjusted weights and overall confidence variance of the previous sample point. The update formula is:

[0022] ,

[0023] in, Adjust the weights of the target for the previous sample point. This represents the overall confidence variance of the previous sample point. It also stores the measurement error variance of the current sample point. , to be used in the calculation during the next filtering;

[0024] Overall confidence variance For measurement error variance Measurement fluctuation factor in measurement statistical characteristics The sum of;

[0025] In step S32, the target adjustment weight of the current sample point is... for:

[0026] ,

[0027] in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance.

[0028] In step S4, the filtered value of the current sample point for:

[0029] ,

[0030] in, The filtered value for the current sample point. Adjust the weights for the target. This is the filtered value of the previous sample point. This represents the measurement value of the current sample point.

[0031] On the other hand, the present invention provides an adaptive filtering correction device for measurement data, comprising:

[0032] The storage space is set to acquire measurement data, use it as the measurement value of the current sample point, and cyclically store the measurement value and the filtered value;

[0033] The statistical characteristic update unit is configured to determine the measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance and measurement fluctuation factor;

[0034] The parameter coupling unit is configured to determine the target adjustment weights based on measurement statistical characteristics; and

[0035] The weighted summation filtering unit is configured to adjust the weights based on the target, and to add the measured value of the current sample point to the filtered value of the previous sample point to obtain the filtered value of the current sample point, and then send it to the storage space.

[0036] On the other hand, the present invention provides an adaptive filtering correction system for measurement data, including a memory, a processor, and computer instructions stored in the memory, wherein the processor executes the computer instructions to implement the adaptive filtering correction method for measurement data described above.

[0037] The method of this invention determines the measurement statistical characteristics, including measurement noise variance and measurement fluctuation factor, by combining historical measurement values ​​and corresponding historical filtered values. Based on these statistical characteristics, a target adjustment weight is determined. Then, based on this target adjustment weight, the measurement value of the current sample point is weighted and summed with the filtered value of the previous sample point to obtain the current target filtered value. Thus, this invention tracks measurement trends by utilizing the measurement statistical characteristics, including measurement noise variance and measurement fluctuation factor, and suppresses abnormal measurement values ​​through dynamic filtering based on the target adjustment weight. This improves the adaptability and robustness to non-steady-state measurements, and also adapts to general nonlinear measurements, while reducing computational complexity. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating an adaptive filtering correction method for measurement data according to an embodiment of the present invention.

[0039] Figure 2 This is a schematic diagram of the architecture of an adaptive filtering correction device for measurement data according to an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the time-domain spectrum of the true value of a typical constant physical quantity and the corresponding discrete measurement data.

[0041] Figure 4 A schematic diagram of the time-domain spectrum of the measurement data after processing by the adaptive filtering and correction method of the present invention.

[0042] Figure 5 This is a comparison chart of the error in the measured value and the error in the filtered value obtained by the adaptive filtering correction method for the measured data of the present invention;

[0043] Figure 6 This is a schematic diagram of the time-domain spectrum of the true value of a typical nonlinear physical quantity and the corresponding discrete measurement data.

[0044] Figure 7 This is a schematic diagram of the time-domain spectrum of nonlinear measurement data after processing by the adaptive filtering correction method of the present invention.

[0045] Figure 8 This is a comparison chart of the measurement and filtering errors of nonlinear physical quantities obtained by the adaptive filtering correction method for measurement data of the present invention.

[0046] Figure 9 This is a schematic diagram of an adaptive filtering correction system for measurement data according to an embodiment of the present invention. Detailed Implementation

[0047] The various embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0048] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0049] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0050] It should be noted that in the embodiments of the present invention, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0051] In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0052] This invention provides an adaptive filtering correction method, apparatus, and system for measurement data, used to adaptively filter and correct measurement data so that the processed measurement results are closer to the physical true values. The measurement data refers to the measurement and / or estimation data of a target physical quantity, such as frequency offset, signal-to-noise ratio (SNR), temperature, acceleration, and voltage. The method of this invention is preferably used to correct the measurement and estimation data of SNR and frequency offset in wireless communication to better reflect the true value of the physical quantity; however, this invention is also applicable to the correction of measurement data in other fields, such as the discrete measurement data of physical quantities like temperature, acceleration, and voltage mentioned here. In communication, frequency offset and SNR are estimated and measured by communication hardware and software based on received data. Whether it is sensor measurement data or wireless communication frequency offset and SNR measurement data, these measurement data will exhibit the same characteristics such as discreteness, oscillation, fluctuation, and outliers. The fundamental reason is that these physical quantities are obtained by processing, converting, and calculating analog signals. These processes all require semiconductor devices, ADC sampling, etc., and are affected by the environment and internal hardware conditions.

[0053] This invention, by combining historical measurement values ​​and corresponding historical filtered values, determines measurement statistical characteristics such as measurement noise variance and measurement fluctuation factor. Based on these statistical characteristics, a target adjustment weight is determined. Then, based on this target adjustment weight, the measurement value of the current sample point is weighted and summed with the filtered value of the previous sample point to obtain the current target filtered value. Thus, this invention tracks the true trend of measurement data changes by utilizing measurement statistical characteristics such as measurement noise variance (nv) and measurement fluctuation factor (flu), and performs dynamic filtering based on the target adjustment weight to suppress abnormal measurement values. This improves the adaptability and robustness to non-steady-state measurements and also adapts to general nonlinear measurements, making the processed measurement values ​​closer to the physical true values ​​while reducing computational complexity.

[0054] First embodiment: An adaptive filtering correction method for measurement data

[0055] Figure 1 This is a flowchart of an adaptive filtering correction method for measurement data according to an embodiment of the present invention. Figure 1 As shown, an adaptive filtering correction method for measurement data according to the present invention includes:

[0056] Step S1: Obtain measurement data as the measurement value of the current sample point;

[0057] Measurement data refers to the measurement and / or estimation data of the target physical quantity, which can be frequency offset, signal-to-noise ratio, temperature, acceleration, voltage, etc. Measurement data is usually non-steady-state data.

[0058] Step S2: Determine the measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance. and measuring volatility factor ;

[0059] In cases where the number of historical measurements and filtered values ​​does not reach the effective statistical sample size, preset static parameters are used as the measurement statistical characteristics; otherwise, the measurement noise variance is calculated based on multiple historical measurements and corresponding historical filtered values. and measuring volatility factor As a measurement statistical characteristic.

[0060] In this embodiment, the size of the storage space , k is the sample point index starting from 0, when < In this case, preset static parameters are used as the measurement statistical characteristics. The preset static parameters include: measurement noise variance. Measuring volatility factor .

[0061] In this embodiment, a sliding window with a size consistent with the effective statistical sample size is used to extract historical measurement values ​​and corresponding historical filtered values.

[0062] Preferably, the measured value and filter value The measurements are stored cyclically in two identical storage spaces (i.e., measurement value storage spaces). Filter value storage space The effective statistical sample size and the sliding window size are both equal to the storage space size. Circular storage refers to a first-in, first-out (FIFO) storage mechanism. When the storage space is full, data is written in the order it was written, overwriting the previous data. In other words, when the corresponding storage space is updated, the measured value... Place them into storage space one by one in a loop. Measure the filtered value Place them into storage space one by one in a loop. If the storage space is full, the corresponding historical values ​​in the storage space will not be cleared. The data is overwritten starting from the first location in the corresponding storage space.

[0063] In this embodiment, the measurement noise variance is obtained by calculating the variance of multiple historical measurements. The historical filter variance is obtained by calculating the variance of multiple historical filter values. The measurement fluctuation factor is then obtained by dividing the historical filter variance by a preset coefficient (e.g., 10). In other words, the storage space for measured values. The array of measured values ​​inside is Filter value storage space The filter value array inside is ,when When -1, measure the noise variance. Measuring volatility factor Where k is the index of the sample point starting from 0. For the size of the storage space, The measurement noise variance for the current sample points , The measured fluctuation factor for the current sample point. To perform variance calculation on an array, An array of measured values, This is an array of filtered values.

[0064] Step S3: Determine the target adjustment weights based on the measurement statistical characteristics. ;

[0065] Before step S3, step S3' may be included: determining whether the measured value of the current sample point is the measured value of the first sample point (i.e., whether the sample point index k=0). If so, the measured value of the current sample point is used as the current filter value, and subsequent steps are skipped, returning to step S1.

[0066] In other words, the first measurement, sample point index Measure the filter value For measured values ,Right now = Then return to step S1 to begin the second measurement.

[0067] Step S3 specifically includes:

[0068] Step S31: Based on the measured fluctuation factor of the current sample points Combined confidence variance with the previous sample point Adjusting weights with targets Obtain the corresponding measurement error variance and overall confidence variance ;

[0069] When the sample point index k=0 and 1, the variance of the measurement error of the current sample point The initial value of the first measurement error variance was obtained through initialization. .

[0070] In the sample point index When -1, the measurement error variance of the current sample points is... The value is updated to the measurement noise variance. ,Right now .

[0071] Otherwise, the measurement error variance of the current sample points The measurement error variance of the current sample point is updated based on the target-adjusted weights and overall confidence variance of the previous sample point. The update formula is:

[0072] ,

[0073] in, Adjust the weights of the target for the previous sample point. This represents the overall confidence variance of the previous sample point. It also stores the measurement error variance of the current sample point. This will be used in the calculation during the next filtering process.

[0074] Overall confidence variance For measurement error variance Measurement fluctuation factor in measurement statistical characteristics The sum. That is, the overall confidence variance. for:

[0075] (1)

[0076] in, To synthesize the confidence level variance, To measure the variance of the error, To measure the volatility factor.

[0077] Step S32: Based on the overall confidence variance of the current sample points and measurement noise variance Obtain the target adjustment weights for the current sample points. .

[0078] The target adjustment weights for the current sample points for:

[0079] (2)

[0080] in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance.

[0081] In other words, the variance of measurement error Overall confidence level variance Target adjustment weight The recursive formula is:

[0082] ,

[0083] ,

[0084] ,

[0085] in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance, To measure the variance of the error, To measure the fluctuation factor. To measure the noise variance. Characterizes the degree of data deviation from the center; measures the volatility factor. Characterizing the degree of fluctuation in the true value of the data; measurement error variance Characterizes the magnitude of the estimation error.

[0086] Therefore, if the data noise is high, Increase, while the target adjusts the weight. Decreasing the weights reduces confidence in the measured values ​​because the target adjusts the weights. Reduce measurement error variance As the value increases, the uncertainty of the estimation error is preserved and propagated, making the filtering more conservative and reliant on historical estimates.

[0087] When the true value fluctuates greatly, the fluctuation factor is measured. The increase leads to a larger overall confidence variance. Increase the size, adjust the target weight. The larger the variance, the more trust is placed in the current measurement, and the more the measurement error variance is transmitted. The reduction in error leads to a rapid release of estimation error, enabling the filter to respond quickly to the latest changes rather than becoming more conservative.

[0088] Step S4: Adjust weights based on the target The current sample point's measured value is weighted and summed with the previous sample point's filtered value to obtain the current sample point's filtered value. .

[0089] In some embodiments, the filtered value of the current sample point for:

[0090] (3)

[0091] in, The filtered value for the current sample point. Adjust the weights for the target. This is the filtered value of the previous sample point. This represents the measurement value of the current sample point.

[0092] Therefore, this invention uses a sliding window to store a small amount of historical data to evaluate the statistical characteristics of measurements. It has low computational complexity, outputs the filtered correction results of the current sample point's measurement value in real time, and achieves rapid convergence.

[0093] This invention, by combining historical measurement values ​​and corresponding historical filtered values, determines the measurement statistical characteristics, including measurement noise variance and measurement fluctuation factor. Based on these statistical characteristics, a target adjustment weight is determined. Then, based on this target adjustment weight, the measurement value of the current sample point is weighted and summed with the filtered value of the previous sample point to obtain the current target filtered value. Thus, this invention tracks measurement trends by utilizing the measurement statistical characteristics, including measurement noise variance and measurement fluctuation factor, and suppresses abnormal measurement values ​​by applying a target adjustment weight adapted to these statistical characteristics. This improves the adaptability and robustness to unsteady-state measurements, and also makes it suitable for general nonlinear measurements, while reducing computational complexity.

[0094] Second embodiment: An adaptive filtering and correction device for measurement data

[0095] Figure 2 This is a schematic diagram of the architecture of an adaptive filtering correction device for measurement data according to another embodiment of the present invention. This adaptive filtering correction device is used to implement the adaptive filtering correction method described above.

[0096] like Figure 2 As shown, the adaptive filtering correction device for measurement data of the present invention includes a storage space 100, a statistical characteristic update unit 200, a parameter coupling unit 300, and a weighted summation filtering unit 400.

[0097] The storage space consists of two units: one for storing measured values. Filter value storage space Storage space 100 is configured to acquire measurement data, use it as the measurement value of the current sample point, and cyclically store the measurement values. and filter value Preferably, in this embodiment, the size of the storage space... .

[0098] In other words, when the corresponding storage space is updated, the measured value will be... Place them into storage space one by one in a loop. Measure the filtered value Place them into storage space one by one in a loop. If the storage space is full, the corresponding historical values ​​in the storage space will not be cleared. The data is overwritten starting from the first location in the corresponding storage space.

[0099] The statistical characteristic update unit 200 is configured to determine measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance. and measuring volatility factor ;

[0100] The parameter coupling unit 300 is configured to determine the target adjustment weights based on measurement statistical characteristics. ;

[0101] The weighted summation filter unit 400 is configured to adjust weights based on the target. The current sample point's measured value is weighted and summed with the previous sample point's filtered value to obtain the current sample point's filtered value. And send it to storage space 100.

[0102] In cases where the number of historical measurements and filtered values ​​does not reach the effective statistical sample size, preset static parameters are used as the measurement statistical characteristics; otherwise, the measurement noise variance is calculated based on multiple historical measurements and corresponding historical filtered values. and measuring volatility factor This is used to measure statistical characteristics. The effective statistical sample size and the sliding window size are both equal to the storage space size. .

[0103] In the statistical characteristic update unit 200, the measurement noise variance is obtained by calculating the variance of multiple historical measurements. The historical filter variance is obtained by calculating the variance of multiple historical filter values. The measurement fluctuation factor is then obtained by dividing the historical filter variance by a preset value (e.g., 10). In other words, the storage space for measured values. The array of measured values ​​inside is Filter value storage space The filter value array inside is ,when When -1, measure the noise variance. Measuring volatility factor Where k is the index of the sample point starting from 0. For the size of the storage space, The measurement noise variance for the current sample points , The measured fluctuation factor for the current sample point. To perform variance calculation on an array, An array of measured values, This is an array of filtered values.

[0104] The parameter coupling unit 300 is configured to: determine whether the measured value of the current sample point is the measured value of the first sample point (i.e., whether the sample point index k=0). If so, the measured value of the current sample point is used as the current filter value and sent to the storage space 100.

[0105] The parameter coupling unit 300 is configured as follows:

[0106] Based on the measured volatility factor of the current sample points Combined confidence variance with the previous sample point Adjusting weights with targets Obtain the corresponding measurement error variance and overall confidence variance ;

[0107] When the sample point index k=0 and 1, the variance of the measurement error of the current sample point The initial value of the first measurement error variance was obtained through initialization. .

[0108] In the sample point index When -1, the measurement error variance of the current sample points is... The value is updated to the measurement noise variance. ,Right now .

[0109] Otherwise, the measurement error variance of the current sample points The measurement error variance of the current sample point is updated based on the target-adjusted weights and overall confidence variance of the previous sample point. The update formula is:

[0110] ,

[0111] in, Adjust the weights of the target for the previous sample point. This represents the overall confidence variance of the previous sample point. It also stores the measurement error variance of the current sample point. This will be used in the calculation during the next filtering process.

[0112] Overall confidence variance For measurement error variance Measurement fluctuation factor in measurement statistical characteristics The sum. That is, the overall confidence variance. for:

[0113] (1)

[0114] in, To synthesize the confidence level variance, To measure the variance of the error, To measure the volatility factor.

[0115] Based on the overall confidence variance of the current sample points and measurement noise variance Obtain the target adjustment weights for the current sample points. .

[0116] The target adjustment weights for the current sample points for:

[0117] (2)

[0118] in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance.

[0119] In the weighted summation filtering unit 400, the filtered value of the current sample point for:

[0120] ,

[0121] in, The filtered value for the current sample point. Adjust the weights for the target. This is the filtered value of the previous sample point. This represents the measurement value of the current sample point.

[0122] Furthermore, the adaptive filtering correction device may also include a filtering control unit 500, which includes a storage order control unit and a filtering process control unit. The storage order control unit is configured to control the storage order of the storage space 100, and the filtering process control unit is configured to control the statistical characteristic update unit 200, the parameter coupling unit 300, and the weighted summation filtering unit 400 to execute sequentially.

[0123] Furthermore, the filtering process control unit is also configured to be invoked by the feature update unit 200 to initialize the filtering parameters when the number of historical measured values ​​and historical filtered values ​​does not reach the effective statistical sample size; and to initialize the measurement error variance of the first and second sample points. Perform initialization; initialize the filter value for the first sample point.

[0124] In this embodiment, the size of the storage space , k is the sample point index starting from 0, when < In this case, preset static parameters are used as the measurement statistical characteristics. The preset static parameters include: measurement noise variance. Measuring volatility factor .

[0125] When the sample point index k=0 and 1, the variance of the measurement error of the current sample point The initial value of the first measurement error variance was obtained through initialization. .

[0126] First measurement, Measure the filter value For measured values ,Right now = The weighted summation filtering unit directly buffers and outputs the result. .

[0127] Figure 3 This is a schematic diagram of the time-domain spectrum of a typical fixed physical quantity's true value and the corresponding discrete measurement data. Figure 4 This is a schematic diagram of the time-domain spectrum of the measurement data after processing by the adaptive filtering correction method of the present invention. Figure 5 This is a comparison chart of the error in the measured value and the error in the filtered value obtained by the adaptive filtering correction method for the measurement data of the present invention.

[0128] Combination Figures 3 to 5 The comparison of measurement and filtering correction effects of the fixed physical quantities shown in the figure demonstrates that the mean square error performance of the filtering correction using the adaptive filtering correction method of the measurement data of this invention is improved by 99.6% compared with the mean square error performance of the original measurement results.

[0129] Figure 6 This is a schematic diagram of the time-domain spectrum of the true value of a typical nonlinear physical quantity and the corresponding discrete measurement data. Figure 7 This is a schematic diagram of the time-domain spectrum of nonlinear measurement data after processing by the adaptive filtering correction method of the present invention. Figure 8 This is a comparison chart of measurement and filtering errors of nonlinear physical quantities obtained by the adaptive filtering correction method for measurement data of the present invention.

[0130] Combination Figures 6 to 8 The comparison of measurement and filtering correction effects of nonlinear physical quantities shows that the mean square error performance of the adaptive filtering correction method for measurement data of this invention is improved by 87.8% compared to the mean square error performance of the original measurement results. The measured discrete data all contain time-varying noise and sudden anomalies. This demonstrates that the adaptive filtering correction method for non-steady-state measurement data provided by this invention can effectively filter, correct, and smooth the original measurement data containing time-varying noise and sudden anomalies.

[0131] Third embodiment: An adaptive filtering and correction system for measurement data

[0132] Figure 9 This is a schematic diagram of an adaptive filtering correction system for measurement data according to an embodiment of the present invention. The system includes a memory 501, a processor 502, and computer instructions stored in the memory 501. The processor 502 executes the computer instructions to implement the adaptive filtering correction method for measurement data described above.

[0133] Optionally, a communication interface 503 and a bus 504 are provided between the memory 501 and the processor 502.

[0134] Processor 502 may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 502 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 502 may also be a combination of functions implementing computation, such as a combination of CPU0 and CPU1, a DSP, and a microprocessor.

[0135] The communication interface 503 includes a receiving unit and a transmitting unit, and is used to connect with other devices via a communication network. This communication network can be Ethernet, a wireless access network, a wireless local area network (WLAN), etc.

[0136] The memory 501 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0137] As one possible implementation, the memory 501 can exist independently of the processor 502. The memory 501 can be connected to the processor 502 via a bus 504 and is used to store instructions or program code. When the processor 502 calls the instructions or program code stored in the memory 501, it can implement the adaptive filtering correction method provided in this embodiment of the invention.

[0138] In another possible implementation, the memory 501 can also be integrated with the processor 502.

[0139] Bus 504 can be an extended industry standard architecture (EISA) bus, etc. Bus 504 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0140] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the service calling device can be divided into different functional modules to complete all or part of the functions described above.

[0141] This invention also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware. The program can be stored in the computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be any of the foregoing embodiments or memory. The computer-readable storage medium can also be an external storage device of the service invocation device, such as a pluggable hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the service invocation device. Further, the computer-readable storage medium can include both internal storage units of the service invocation device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the service invocation device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0142] This invention also provides a computer program product comprising a computer program that, when run on a computer, causes the computer to execute the adaptive filtering correction method provided in the above embodiments.

[0143] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An adaptive filtering correction method for measurement data, characterized in that, include: Step S1: Obtain measurement data as the measurement value of the current sample point; Step S2: Determine the measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance and measurement fluctuation factor; Step S3: Determine the target adjustment weights based on the measurement statistical characteristics; Step S4: Adjust the weights based on the target, and add the measured value of the current sample point to the filtered value of the previous sample point to obtain the filtered value of the current sample point; Step S3 specifically includes: Step S31: Based on the measured fluctuation factor of the current sample points Combined confidence variance with the previous sample point Adjusting weights with targets Obtain the corresponding measurement error variance and overall confidence variance ; Step S32: Based on the overall confidence variance of the current sample points and measurement noise variance Obtain the target adjustment weights for the current sample points. ; In step S31, when the sample point index k=0 and 1, the measurement error variance of the current sample point is... Obtained through initialization; In the sample point index When -1, the measurement error variance of the current sample point is... The value is updated to the measurement noise variance. ; Otherwise, the measurement error variance of the current sample points The measurement error variance of the current sample point is updated based on the target-adjusted weights and overall confidence variance of the previous sample point. The update formula is: ， in, Adjust the weights of the target for the previous sample point. Store the overall confidence variance of the previous sample point; store the measurement error variance of the current sample point. , to be used in the calculation during the next filtering; Overall confidence variance For measurement error variance Measurement fluctuation factor in measurement statistical characteristics The sum of; In step S32, the target adjustment weight of the current sample point is... for: ， in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance.

2. The adaptive filtering correction method according to claim 1, characterized in that, The measurement data are measurement and / or estimation data of a target physical quantity, which includes at least one of frequency offset, signal-to-noise ratio, temperature, acceleration, and voltage.

3. The adaptive filtering correction method according to claim 1, characterized in that, In step S2, if the number of historical measurements and historical filtered values ​​does not reach the effective statistical sample size, a preset static parameter is used as the measurement statistical characteristic; otherwise, the measurement noise variance is calculated based on multiple historical measurements and corresponding historical filtered values. and measuring volatility factor As a measurement statistical characteristic.

4. The adaptive filtering correction method according to claim 3, characterized in that, In step S2, a sliding window with the same size as the effective statistical sample size is used to extract historical measurement values ​​and corresponding historical filtered values; the measurement values ​​and filtered values ​​are respectively stored cyclically in two storage spaces of the same size, and the effective statistical sample size and the size of the sliding window are both equal to the size of the storage space; The measurement noise variance is obtained by calculating the variance of multiple historical measurements. The historical filter variance is obtained by calculating the variance of multiple historical filter values, and the measurement fluctuation factor is obtained by dividing the historical filter variance by a preset coefficient. .

5. The adaptive filtering correction method according to claim 1, characterized in that, Before step S3, there is also step S3': determine whether the measured value of the current sample point is the measured value of the first sample point. If so, use the measured value of the current sample point as the current filter value, skip the subsequent steps, and return to step S1.

6. The adaptive filtering correction method according to claim 1, characterized in that, In step S4, the filtered value of the current sample point for: ， in, The filtered value for the current sample point. Adjust the weights for the target. This is the filtered value of the previous sample point. This represents the measurement value of the current sample point.

7. An adaptive filtering and correction device for measurement data, characterized in that, include: The storage space is set to acquire measurement data, use it as the measurement value of the current sample point, and cyclically store the measurement value and the filtered value; The statistical characteristic update unit is configured to determine the measurement statistical characteristics based on historical measurement values ​​and corresponding historical filtered values, wherein the measurement statistical characteristics include measurement noise variance and measurement fluctuation factor; The parameter coupling unit is configured to determine the target adjustment weights based on the measurement statistical characteristics. and The weighted summation filtering unit is configured to adjust the weights based on the target, and to add the measured value of the current sample point to the filtered value of the previous sample point to obtain the filtered value of the current sample point, and then send it to the storage space. Among these, the target adjustment weights are determined based on the statistical characteristics of the measurement, specifically including: Step S31: Based on the measured fluctuation factor of the current sample points Combined confidence variance with the previous sample point Adjusting weights with targets Obtain the corresponding measurement error variance and overall confidence variance ; Step S32: Based on the overall confidence variance of the current sample points and measurement noise variance Obtain the target adjustment weights for the current sample points. ; In step S31, when the sample point index k=0 and 1, the measurement error variance of the current sample point is... Obtained through initialization; In the sample point index When -1, the measurement error variance of the current sample point is... The value is updated to the measurement noise variance. ; Otherwise, the measurement error variance of the current sample points The measurement error variance of the current sample point is updated based on the target-adjusted weights and overall confidence variance of the previous sample point. The update formula is: ， in, Adjust the weights of the target for the previous sample point. Store the overall confidence variance of the previous sample point; store the measurement error variance of the current sample point. , to be used in the calculation during the next filtering; Overall confidence variance For measurement error variance Measurement fluctuation factor in measurement statistical characteristics The sum of; In step S32, the target adjustment weight of the current sample point is... for: ， in, Adjust the weights for the target. To synthesize the confidence level variance, To measure the noise variance.

8. An adaptive filtering and correction system for measurement data, characterized in that, The method includes a memory, a processor, and computer instructions stored in the memory, wherein the processor executes the computer instructions to implement the adaptive filtering correction method for measurement data according to any one of claims 1-6.

Images