Method and device for evaluating quality of original output matrix of force sensor, electronic equipment, storage medium and product

Abstract

The application provides a quality evaluation method and device of a raw output matrix of a force sensor, electronic equipment, a storage medium and a product; the method comprises the following steps: acquiring raw output data of the force sensor in all load directions; determining a raw output matrix based on the raw output data and corresponding load information; determining one or more evaluation indexes based on the raw output matrix; determining a comprehensive quality score based on the evaluation indexes; and determining quality evaluation information representing the usability of the raw output matrix based on the score interval in which the comprehensive quality score is located. The application does not only look at the final decoupling result, but also calculates a series of indexes for the raw output matrix obtained in the calibration process, calculates a comprehensive quality score according to the indexes, and determines the usability of the raw output matrix according to the score interval in which the comprehensive quality score is located, so that it can be determined whether the decoupling matrix can be shipped with the sensor, and the reliability and accuracy of the raw output matrix and the decoupling matrix are improved.

Description

Technical Field

[0001] This application relates to the field of sensor technology, and in particular to a method, apparatus, electronic device, storage medium, and product for evaluating the quality of the raw output matrix of a force sensor. Background Technology

[0002] Before leaving the factory, force sensors typically undergo static loading in multiple directions, collecting the output of each bridge circuit or acquisition channel to obtain the original output matrix, and then solving for the decoupling matrix based on the original output matrix. Related technologies typically evaluate the decoupling matrix by focusing on issues such as single-channel sensitivity, coupling error after calibration, linearity, repeatability, or decoupling algorithm error. However, even considering these issues, the decoupling results may still be unstable under actual installation, temperature drift, noise disturbance, or long-term use conditions. Summary of the Invention

[0003] In view of this, embodiments of this application provide a method, apparatus, electronic device, storage medium, and product for evaluating the quality of the raw output matrix of a force sensor, aiming to provide a method for evaluating the reliability of the raw output matrix of a sensor, which can improve the reliability and accuracy of the raw output matrix and the decoupling matrix.

[0004] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a method for evaluating the quality of the raw output matrix of a force sensor, the method comprising: Acquire raw output data of the force sensor under all load directions; The original output matrix is ​​determined based on the original output data and the corresponding load information; One or more evaluation indicators are determined based on the original output matrix; The overall quality score is determined based on the aforementioned evaluation indicators; Based on the score range of the comprehensive quality score, quality evaluation information representing the usability of the original output matrix is ​​determined.

[0005] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. Determining the comprehensive quality score based on the evaluation metrics includes: The comprehensive quality score is determined based on one or more of the following indicators: output strength, balance, output model similarity, inversion stability, weak independence, large number subtraction risk, and noise risk, along with their corresponding weighting coefficients. The method further includes: determining a matrix processing scheme based on the score range in which the comprehensive quality score is located.

[0006] In some embodiments, the evaluation metrics include an output intensity metric, and determining one or more evaluation metrics based on the original output matrix includes: The output intensity information for each load direction is determined based on the magnitude of each column vector of the original output matrix; The output intensity index is determined based on the output intensity information.

[0007] In some embodiments, the evaluation index further includes a balance index, and the step of determining one or more evaluation indices based on the original output matrix further includes: The balance index is determined based on the maximum and minimum output intensity information corresponding to the original output matrix.

[0008] In some embodiments, the evaluation metrics include an output pattern similarity metric, and determining one or more evaluation metrics based on the original output matrix includes: The output pattern similarity index is determined based on the degree of similarity between every two column vectors in the original output matrix.

[0009] In some embodiments, the evaluation metrics include at least one of a weak independence metric, an inverse stability metric, and a noise risk metric, and the determination of one or more evaluation metrics based on the original output matrix includes at least one of the following: The weak independence index is determined based on the minimum singular value of the original output matrix; The inversion stability index is determined based on the maximum and minimum singular values ​​of the original output matrix. The noise risk index is determined based on the reciprocal of the minimum singular value of the original output matrix.

[0010] In some embodiments, the evaluation metrics include a large number subtraction risk metric, and the determination of one or more evaluation metrics based on the original output matrix includes: Determine the decoupling matrix based on the original output matrix; The risk index for large number subtraction is determined based on the decoupling matrix, the original output data, or the column vectors of the original output matrix.

[0011] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. The method further includes: If one or more of the following indicators—output intensity, balance, output pattern similarity, inversion stability, weak independence, large number subtraction risk, and noise risk—are not within the corresponding safety threshold range, then processing action information is generated.

[0012] Secondly, embodiments of this application provide a quality evaluation device for the raw output matrix of a force sensor, the device comprising: The data acquisition module is used to acquire the raw output data of the force sensor under all load directions; The matrix construction module is used to determine the original output matrix based on the original output data and the corresponding load information; The indicator calculation module is used to determine one or more evaluation indicators based on the original output matrix. The decision-making module is used to determine the overall quality score based on the evaluation indicators; Based on the score range of the comprehensive quality score, quality evaluation information representing the usability of the original output matrix is ​​determined.

[0013] Thirdly, embodiments of this application provide an electronic device, the electronic device including a processor and a memory for storing a computer program capable of running on the processor, wherein, The processor is configured to execute the steps of the method as described in the first aspect when running a computer program.

[0014] Fourthly, embodiments of this application provide a storage medium storing a computer program, which, when executed by a control device, implements the steps of the method described in the first aspect.

[0015] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a controlled device, implements the steps of the method described in the first aspect.

[0016] The technical solution provided in this application includes the following methods: acquiring raw output data of a force sensor under all load directions; determining a raw output matrix based on the raw output data and corresponding load information; determining one or more evaluation indicators based on the raw output matrix; determining a comprehensive quality score based on the evaluation indicators; and determining quality evaluation information characterizing the usability of the raw output matrix based on the score range of the comprehensive quality score.

[0017] This application takes into account that, in practical engineering, even if the original output matrix of a sensor is mathematically invertible, it may still suffer from problems such as uneven output strength in different directions, similar output patterns in different load directions, ill-conditioned matrix, reliance on large signal cancellation for decoupling calculations, and amplification of noise and zero-point drift. Therefore, this application does not only consider the final decoupling result, but also performs a series of index calculations on the original output matrix obtained during the calibration process, calculates a comprehensive quality score based on each index, and determines the usability of the original output matrix based on the score range of the comprehensive quality score. This allows us to determine whether the decoupling matrix can be shipped with the sensor, improving the reliability and accuracy of both the original output matrix and the decoupling matrix. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for evaluating the quality of the raw output matrix of a force sensor provided in an embodiment of this application. Figure 2 This is a flowchart illustrating the quality evaluation method for the raw output matrix of the force sensor provided in the application example of this application; Figure 3 This is a schematic diagram of the structure of the quality evaluation device for the original output matrix of the force sensor provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0020] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0021] Before leaving the factory, force sensors typically undergo static loading in multiple directions, collecting the output of each bridge circuit or acquisition channel to obtain the original output matrix, and then solving for the decoupling matrix based on the original output matrix. Related technologies typically evaluate the decoupling matrix by focusing on issues such as single-channel sensitivity, coupling error after calibration, linearity, repeatability, or decoupling algorithm error. However, even considering these issues, the decoupling results may still be unstable under actual installation, temperature drift, noise disturbance, or long-term use conditions.

[0022] Therefore, the problem with the related technologies is that: evaluating sensor performance solely based on single-axis sensitivity or coupling error after calibration cannot reflect the stability of the original output matrix during inversion and decoupling; judging solely on whether the matrix is ​​invertible cannot identify situations that are close to ill-conditioned, have weak independent directions, have similar output modes, or have excessively large differences in output intensity in each direction under full-scale load.

[0023] This application takes into account that, in practical engineering, even if the original output matrix of a sensor is mathematically invertible, it may still suffer from problems such as uneven output strength in different directions, similar output patterns in different load directions, ill-conditioned matrix, reliance on large signal cancellation for decoupling calculations, and amplification of noise and zero-point drift. These problems can lead to the sensor appearing usable after calibration, but the decoupling results becoming unstable under actual installation, temperature drift, noise disturbance, or long-term use conditions.

[0024] Therefore, in various embodiments of this application, not only is the final decoupling result considered, but a series of indicators are calculated on the original output matrix obtained during the calibration process, and a comprehensive quality score is calculated based on each indicator. The usability of the original output matrix is ​​determined based on the score range of the comprehensive quality score, thereby determining whether the decoupling matrix can be shipped with the sensor, which improves the reliability and accuracy of the original output matrix and the decoupling matrix.

[0025] The method provided in this application can be applied to force / torque sensors such as resistance strain gauge, MEMS, capacitive, piezoresistive, silicon-based, and fiber optic sensors. The following description uses a six-dimensional force sensor as an example. A six-dimensional force sensor can simultaneously measure three-dimensional forces Fx, Fy, and Fz, as well as three-dimensional torques Mx, My, and Mz. It is widely used in robot force control, precision assembly, aerospace, human-computer interaction, medical devices, and micro-operating systems. Six-dimensional force sensors typically convert six-dimensional loads into multiple raw electrical signals through an elastic body structure, sensitive element arrangement, and a bridge circuit. The six-dimensional force / torque output is then calculated using a decoupling matrix. During calibration, the sensor can be loaded using weights, a standard force source, a torque loading arm, a six-dimensional loading stage, a robot loading device, or an automatic calibration device.

[0026] This application provides a method for evaluating the quality of the raw output matrix of a force sensor, such as... Figure 1 As shown, the method includes: Step 101: Obtain the raw output data of the force sensor under all load directions.

[0027] For a six-dimensional force sensor, there are a total of 6 load directions. During the calibration process, a preset load is applied to each load direction in sequence, and the original channel output (i.e., the original output data) of the bridge circuit or acquisition channel is obtained each time it is loaded. In this embodiment, the scheme is illustrated by taking the number of original output channels of the six-dimensional force sensor as 6. Formula (1) is the six-dimensional load vector, which represents the preset load value applied to a certain load direction each time.

[0028] F = [ F x F y Fz M x M y M z ] (1) For example, if only a lateral force is applied to the x-axis of the six-dimensional force sensor, the value of Fx is the corresponding preset load value, while the values ​​of Fy, Fz, Mx, My, and Mz are 0. It is understood that the preset load values ​​for each loading direction during calibration can be determined according to actual conditions, and this application embodiment does not limit this.

[0029] Formula (2) is the original channel output vector after the zero point is deducted after the six-dimensional force sensor is loaded, which is the original output data.

[0030] U = [ U 1 U 2 U 3 U 4 U 5 U 6] (2) Where Ui is the output of the i-th original output channel. The output content can be the bridge voltage in mV or V, the amplified voltage, the ADC digital value, or the original channel value after conversion by the acquisition circuit.

[0031] Step 102: Determine the original output matrix based on the original output data and the corresponding load information.

[0032] Based on the above example, the original output matrix can be obtained. Specifically, the original output matrix can be determined according to the ratio of the preset load value in each load direction to the full-scale load value (i.e., the full-scale value) in that direction. Formula (3) is the calculation formula for the full-scale normalized load corresponding to each load direction.

[0033] Fn = [ F x / F x,FS F y / F y,FS F z / F z,FS M x / M x,FS My / M y,FS M z / M z,FS ] (3) in, F x,FS This represents the full-scale load value in the Fx direction. F y,FS This represents the full-scale load value in the Fy direction. F z,FS This represents the full-scale load value in the Fz direction. M x,FS This represents the full-scale load value in the Mx direction. M y,FS This represents the full-scale load value in the My direction. M z,FS This represents the full-scale load value in the Mz direction. The purpose of normalization is to eliminate the influence of different ranges and units in different directions, so that the three force directions and the three moment directions can be compared in the same matrix.

[0034] Correspondingly, the original channel output vector corresponding to each direction can be adjusted according to the full-scale normalized load to obtain the full-scale normalized original channel output vector (or full-scale normalized original output data), and the full-scale normalized original channel output vectors of each column (corresponding to each load direction) can be combined into a full-scale normalized original output matrix (i.e., the original output matrix). For example, if a preset load value of 1 / 2 of the full scale is applied to the x-axis of the six-dimensional force sensor, then the full-scale normalized original channel output vector should be twice the corresponding original channel output vector.

[0035] Formula (4) is the relationship between the original channel output vector U after the zero point is removed from the six-dimensional force sensor and the original output matrix Sn.

[0036] U = SnFn + ε (4) Where Sn is the normalized original output matrix, Sn includes [Y1,Y2,Y3,Y4,Y5,Y6], and the j-th column output vector Yj represents the common output result of all original output channels when a full-scale load is applied in the j-th direction. ε represents the combined disturbances such as noise, zero-point drift, temperature drift, and assembly errors.

[0037] For example, when only the Fz load is applied (loads in other directions are 0), if F z / F z,FSY3 is the original channel output vector that corresponds to the full-scale normalized column of Fz in the original output matrix. The original channel output vector U corresponding to the 1 / 2 full-scale load in the Fz direction is Y3 / 2+ε.

[0038] Step 103: Determine one or more evaluation indicators based on the original output matrix; Here, at least one evaluation index can be calculated based on the original output matrix, including one or more of the following: output intensity index, balance index, output pattern similarity index, inversion stability index, weak independence index, large number subtraction risk index, and noise risk index.

[0039] Step 104: Determine the comprehensive quality score based on the evaluation indicators; Step 105: Determine the quality evaluation information representing the usability of the original output matrix based on the score range of the comprehensive quality score.

[0040] Understandably, the solution in this application performs a series of index calculations on the original output matrix obtained during the calibration process, calculates a comprehensive quality score based on each index, and determines the usability of the original output matrix based on the score range of the comprehensive quality score. This allows us to determine whether the decoupling matrix can be shipped with the sensor, thereby improving the reliability and accuracy of the original output matrix and the decoupling matrix.

[0041] In some embodiments, the evaluation metrics include an output intensity metric, and determining one or more evaluation metrics based on the original output matrix includes: The output intensity information for each load direction is determined based on the magnitude of each column vector of the original output matrix; The output intensity index is determined based on the output intensity information.

[0042] For example, formula (5) is the magnitude of the output vector. The calculation formula is used to calculate the output strength of six specified load directions under the same normalization conditions.

[0043] (5) Output vector magnitude This can be understood as: the total signal strength generated across all original output channels in a given loading direction. ij Let Yj be the i-th element of the j-th column output vector Yj. The number of elements in the output vector Yj is the same as the number of the original output channels.

[0044] The magnitude of each output vector Yj in the original output matrix (output intensity information) is calculated. Each output vector magnitude is then compared to a preset magnitude threshold. If at least one output vector magnitude is less than the preset threshold, the output in that load direction is considered weak, susceptible to noise, zero-point drift, and temperature drift, resulting in a higher output intensity index. The output intensity index value can be determined based on a mapping table between the magnitudes of each output vector and the output intensity index value. The more output vectors whose magnitudes are less than the preset threshold, the higher the output intensity index value. It can be understood that a smaller output intensity index value indicates sufficient output in all directions and a better output intensity index. This application does not limit the specific value of the preset magnitude threshold.

[0045] In some embodiments, the evaluation index further includes a balance index, and the step of determining one or more evaluation indices based on the original output matrix further includes: The balance index is determined based on the maximum and minimum output intensity information corresponding to the original output matrix.

[0046] For example, formula (6) is the formula for calculating the output balance in six load directions, R L The closer R is to 1, the more balanced the output in all directions; if R L If the value is too large, it indicates that the output scales in the six directions are unbalanced.

[0047] (6) Here, we can use R L A mapping table between R and the values ​​of the equilibrium index determines the values ​​of the equilibrium index. It can be understood that R... L The smaller the value (closer to 1), the more balanced the output scale in the six directions; the smaller the value of the balance index, the better. L The larger the value, the more unbalanced the output scale in the six directions, and the larger the value of the balance index.

[0048] In some embodiments, the evaluation metrics include an output pattern similarity metric, and determining one or more evaluation metrics based on the original output matrix includes: The output pattern similarity index is determined based on the degree of similarity between every two column vectors in the original output matrix.

[0049] Here, when a force is applied in a certain direction, each original output channel outputs various positive or negative values. The positive-negative relationship and magnitude distribution of the output content of each original output channel constitute the original output pattern. For example, when Fx is applied, channel 1 outputs positive values; channel 2 outputs negative values; channels 3 and 4 output smaller values; and other channels output close to zero. This positive-negative relationship and magnitude distribution constitute the original output pattern in the Fx direction.

[0050] For example, formula (7) is a formula for calculating the similarity of the original output modes corresponding to two load directions.

[0051] (7) in, and These are the output vectors corresponding to two different load directions in the original output matrix. The cosine correlation coefficients for different output vectors. The closer the absolute value is to 1, the more similar the output modes of the two load directions are, or only opposite in direction, making them easy to confuse each other during decoupling.

[0052] Furthermore, the output pattern similarity index can be determined according to formula (8), where, This represents the maximum similarity between different output vectors in the original output matrix.

[0053] (8) According to The mapping table between the output pattern similarity index and the output pattern similarity index value determines the output pattern similarity index value. It is understandable that... The smaller the value (less than 1), the less similar the output modes of the two load directions are, and the smaller the corresponding output mode similarity index value is. The larger the value (closer to 1), the more similar the output modes of the two load directions are, and the larger the corresponding output mode similarity index value is.

[0054] In some embodiments, the evaluation metrics include at least one of a weak independence metric, an inverse stability metric, and a noise risk metric, and the determination of one or more evaluation metrics based on the original output matrix includes at least one of the following: The weak independence index is determined based on the minimum singular value of the original output matrix; The inversion stability index is determined based on the maximum and minimum singular values ​​of the original output matrix. The noise risk index is determined based on the reciprocal of the minimum singular value of the original output matrix.

[0055] For example, formula (9) is the formula for calculating the i-th singular value of the original output matrix Sn.

[0056] (9) Formula (10) is used to evaluate the stability of matrix inversion. This represents the ratio of the maximum singular value to the minimum singular value of the original output matrix Sn.

[0057] (10) Formula (11) is the formula for calculating the decoupling matrix C. It can be understood that when the original output matrix Sn is invertible, the decoupling matrix C is the inverse matrix of Sn. Formula (12) is used to evaluate the maximum amplification risk of the original output noise after decoupling under the most unfavorable disturbance direction.

[0058] C = (11) (12) Here, according to The value of the inverse stability index is determined by the value of the index, which can be specifically determined based on the value of the index. The mapping table between the inverse stability index and the inverse stability index value is used to obtain the inverse stability index value. The smaller the value, the smaller the inverse stability index; An excessively large matrix indicates that it is close to ill-conditioned, making inversion unstable and resulting in a large inversion stability index.

[0059] You can also based on The value of the noise risk index is determined by the value of the [value], specifically based on [the value of the [value]]. The mapping table between noise risk index values ​​and noise risk index values ​​is used to obtain the noise risk index values. The larger the value, the easier it is for the original noise to be amplified after decoupling, and the higher the noise risk index value.

[0060] It can also be based on the minimum singular value The value of the weak independence index is determined by the value of the index, which can be specifically determined based on the value of the index. The mapping table between the weak independence index and the weak independence index value is used to obtain the weak independence index value. The smaller the value, the more likely there is weak independence; the larger the value of the weak independence index, the more likely there is weak independence.

[0061] In some embodiments, the evaluation metrics include a large number subtraction risk metric, and the determination of one or more evaluation metrics based on the original output matrix includes: Determine the decoupling matrix based on the original output matrix; The risk index for large number subtraction is determined based on the decoupling matrix, the original output data, or the column vectors of the original output matrix.

[0062] Based on the above example, the decoupling matrix C can be the inverse matrix of the original output matrix, as shown in formula (13), where the j-th row is used to calculate the six-dimensional load component in the j-th direction based on all original channel output vectors; then the formula for calculating the normalized load estimate in the j-th direction is shown in formula (14).

[0063] (13) (14) If multiple If the absolute value of a term is large but the positive and negative values ​​cancel each other out, the result is smaller. Therefore, the load estimate is more sensitive to noise, zero-point drift, and temperature drift. Formula (15) is defined as the formula for calculating the risk coefficient of the large number subtraction.

[0064] (15) in, To prevent extremely small positive numbers with a denominator of zero, Uk is the j-th row and k-th column of the decoupling matrix; preferably, Uk = Yj is taken when evaluating the j-th load direction.

[0065] According to The value of the subtraction of large numbers risk indicator can be determined based on the specific value. The mapping table between the large number subtraction risk index value and the large number subtraction risk index value is used to obtain the large number subtraction risk index value. The closer it is to 1, the less positive and negative offsetting there is, and the lower the risk of subtracting large numbers. The smaller the value of the risk indicator for subtracting large numbers is. The larger the value, the more likely the decoupling depends on subtracting a large number, indicating poor engineering stability. The larger the value of the risk index corresponding to subtracting a large number, the better.

[0066] It is understood that the specific mapping relationships in the above mapping relationship table can be determined according to the actual situation, and this application embodiment does not limit this.

[0067] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. The method further includes: If one or more of the following indicators—output intensity, balance, output pattern similarity, inversion stability, weak independence, large number subtraction risk, and noise risk—are not within the corresponding safety threshold range, then processing action information is generated.

[0068] In practical applications, the processed action information is used to instruct calibration personnel, calibration equipment, or R&D personnel on their operations.

[0069] For example, during or after sensor calibration, processing action information can be generated based on each indicator and its corresponding safety threshold range, enabling calibration personnel or equipment to adjust calibration strategies or recalibrate in a timely manner. Specifically: If a certain direction exists If the value is significantly smaller, the output intensity index will be larger. If the output intensity index exceeds the corresponding output intensity threshold, the processing action information will indicate to increase the strain of the sensor structure in that direction, adjust the patch position, increase the bridge gain, or reallocate the range, etc. In practical applications, it can also be directly based on... If the value exceeds the corresponding threshold, a weak output in a certain direction is determined. The processing action information indicates that the sensor structure strain in that direction should be increased, the patch position should be adjusted, the bridge gain should be increased, or the range should be reallocated.

[0070] If the balance index value is too large, exceeding the corresponding balance threshold, then an imbalance in the output scale across the six directions is identified. The motion information is then processed to adjust the force / torque direction sensitivity matching and optimize the gain and full-scale normalization settings. In practical applications, this can also be directly determined based on R... L If the value exceeds the corresponding threshold, it indicates that the output scale in the six directions is unbalanced. The motion information is processed to indicate the adjustment of the force / torque direction sensitivity matching, and the gain and full-scale normalization settings are optimized.

[0071] If the output pattern similarity index value is too high, exceeding the corresponding similarity threshold, it is determined that there are two similar output patterns in the load directions. The processing action information indicates adjustments to the elastomer structure, patch direction, bridge circuit combination, or mechanical decoupling structure. In practical applications, it can also be based on... If the value exceeds the corresponding threshold, it indicates that the output modes of the two load directions are similar. The processing action information indicates the adjustment of the elastomer structure, patch direction, bridge circuit combination, or mechanical decoupling structure.

[0072] If the inverse stability index value is too large, exceeding the corresponding inverse stability threshold, the processing action information indicates recalibration, optimization of weak directions, adjustment of bridging, or determination that the matrix is ​​unsuitable for stable decoupling. In practical applications, it can also be based on... If the corresponding threshold is exceeded, the matrix is ​​determined to be close to ill-conditioned and unstable in inversion. The processing action information indicates recalibration, optimization of weak directions, adjustment of bridging groups, or determination that the matrix is ​​not suitable for stable decoupling.

[0073] If the noise risk index value is too high, exceeding the corresponding noise risk threshold, the processing action information indicates measures to improve the acquisition signal-to-noise ratio, optimize the matrix, add filtering, and implement temperature compensation. In practical applications, this can also be done based on... If the corresponding threshold is exceeded, it is determined that the original noise is amplified after decoupling. The processing action information indicates that the acquisition signal-to-noise ratio should be improved, the matrix should be optimized, and filtering and temperature compensation should be added.

[0074] If the weak independence index value is too large, exceeding the corresponding weak independence threshold, the processing action information indicates an enhancement of the original output in the corresponding load direction or a redesign of the elastic structure. In practical applications, it can also be based on... If the load is less than the corresponding threshold, a weak independent direction is identified. The processing action information indicates whether to enhance the original output in the corresponding load direction or redesign the elastic structure.

[0075] If the value of the large number subtraction risk indicator is too large, exceeding the corresponding large number subtraction risk threshold, then the processing action information indicates optimization of the bridge path's positive and negative contribution allocation and reduction of zero-point drift and temperature drift sensitivity. In practical applications, it can also be based on... If the value is greater than the corresponding threshold (e.g., 1), it is determined that the positive and negative terms of decoupling dependency are offset, and the action information is processed to indicate the optimization of the positive and negative contribution allocation of the bridge path and to reduce the sensitivity of zero drift and temperature drift.

[0076] It is understandable that the aforementioned thresholds can be determined based on sensor range, acquisition circuit noise, A / D resolution, target accuracy, historical qualified sample statistics, and customer site allowable error.

[0077] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. Determining the comprehensive quality score based on the evaluation metrics includes: The comprehensive quality score is determined based on one or more of the following indicators: output strength, balance, output model similarity, inversion stability, weak independence, large number subtraction risk, and noise risk, along with their corresponding weighting coefficients. The method further includes: determining a matrix processing scheme based on the score range in which the comprehensive quality score is located.

[0078] For example, the formula for calculating the overall quality score Q is shown in formula (16).

[0079] (16) Specifically, the weights for the output strength index q1 and q2 are as follows: w1 for the balance index q2, w3 for the output model similarity index q3, w4 for the inverse stability index q4, w5 for the weak independence index q5, w6 for the large number subtraction risk index q6, and w7 for the noise risk index q7. It can be understood that the sum of the weights of all the indices involved in calculating the overall quality score Q is 1.

[0080] Furthermore, the quality evaluation information representing the usability of the original output matrix and the corresponding matrix processing scheme can be determined based on the score range of Q.

[0081] For example, if Q < T1, then the quality evaluation information is determined to be excellent, the matrix quality is good, the output in each direction is sufficient, the directions are easily distinguishable, and the decoupling is stable; the matrix processing scheme allows the calibration results and sensors to be released to the factory or finalized in the design.

[0082] If T1≤Q<T2, the quality evaluation information is determined to be good, indicating that the sensor as a whole meets the requirements, but some individual indicators are close to the warning value; matrix processing solution: it can be released, while the warning items are recorded, and temperature drift, noise or composite load verification is added.

[0083] If T2≤Q<T3, then the quality evaluation information is determined to be usable. The sensor can complete six-dimensional decoupling, but there is a risk of weakness in a certain direction or certain noise amplification. Matrix processing scheme: can be used in scenarios with relatively low accuracy requirements, or after adjusting the gain, filtering and effective range.

[0084] If T3≤Q<T4, the quality evaluation information is determined to be risky, with some key indicators exceeding the warning value. In actual use, noise amplification, temperature drift sensitivity, or direction confusion may occur. Matrix processing solution: Do not release the product to the factory for the time being, and re-check the calibration data, structure, patch position, bridging method, and acquisition gain.

[0085] If Q≥T4, the quality evaluation information is determined to be unusable, the original output matrix is ​​close to singular, cannot be stably inverted, or the noise is severely amplified; matrix processing solution: the currently calculated decoupling matrix must not be used, and the product needs to be reworked, re-bridged, redesigned, or the product is judged to be unqualified.

[0086] It is understandable that the quality of the original output matrix can also reflect defects in the structural design of the sensor product. That is, the quality evaluation information can reflect not only the quality of the calibrated original output matrix, but also the quality of the sensor product itself. In addition, the specific values ​​of T1-T4 can be determined according to the actual situation, and this application embodiment does not limit them.

[0087] It should be noted that the number of original output channels in this application embodiment is not limited to 6 channels, but can also be more than 6 channels of redundant bridge circuits, multiple half-bridge circuits, multiple ADCs or temperature compensation channels. In this case, the quality of sensor products can be evaluated by generalized inverse, least squares or dimension-reduced matrix.

[0088] The solution presented in this application can be used to evaluate the calibration of real sensors, i.e., the data source is real calibration data. In some embodiments, the data obtained by the solution presented in this application may also come from finite element simulation data, digital prototype data, literature calibration matrix conversion data, or installation status verification data.

[0089] The technical solution provided in this application includes the following methods: acquiring raw output data of a force sensor under all load directions; determining a raw output matrix based on the raw output data and corresponding load information; determining one or more evaluation indicators based on the raw output matrix; determining a comprehensive quality score based on the evaluation indicators; and determining quality evaluation information characterizing the usability of the raw output matrix based on the score range of the comprehensive quality score.

[0090] This application takes into account that, in practical engineering, even if the original output matrix of a sensor is mathematically invertible, it may still suffer from problems such as uneven output strength in different directions, similar output patterns in different load directions, ill-conditioned matrix, reliance on large signal cancellation for decoupling calculations, and amplification of noise and zero-point drift. Therefore, this application does not only consider the final decoupling result, but also performs a series of index calculations on the original output matrix obtained during the calibration process, calculates a comprehensive quality score based on each index, and determines the usability of the original output matrix based on the score range of the comprehensive quality score. This allows us to determine whether the decoupling matrix can be shipped with the sensor, improving the reliability and accuracy of both the original output matrix and the decoupling matrix.

[0091] The following section provides a more detailed description of this application with reference to application examples.

[0092] Figure 2 This is a flowchart illustrating the original output matrix quality assessment method provided in the application example of this application, including: Step 201: Obtain data; then proceed to step 202. Specifically, the raw output data of the six-dimensional force sensor under six load directions is obtained. The data sources may include finite element simulation, static calibration, literature matrix conversion, mass production calibration data, or installation verification data.

[0093] Step 202: Full-scale normalization; then proceed to step 203; Specifically, based on the rated full-scale range of Fx, Fy, Fz, Mx, My, and Mz, the load vector and the corresponding original channel output vector are uniformly normalized to eliminate differences in force and torque units and ranges.

[0094] Step 203: Construct the original output matrix; then proceed to step 204. Here, the original channel output vectors corresponding to the six unit full-scale normalized loads are used as column vectors to construct the normalized original output matrix Sn.

[0095] Step 204: Calculate the signal strength index; then proceed to step 205. Here, the magnitude of each column's output vector is calculated. and the ratio of the module length to R L It is used to determine whether the original output in each load direction is sufficient and balanced.

[0096] Step 205: Calculate the mode independence index; then proceed to step 206. Here, the cosine correlation coefficient between the output vectors in different load directions is calculated. and the maximum absolute off-diagonal correlation coefficient It is used to determine whether different load output modes are similar.

[0097] Step 206: Calculate the matrix stability index; then proceed to step 207. Here, we perform singular value decomposition on Sn to calculate the maximum singular value, minimum singular value, and condition number. and the most unfavorable noise amplification factor .

[0098] Step 207: Calculate the risk of subtracting large numbers; then proceed to step 208. Here, based on the inverse of the decoupling matrix C or Sn, we calculate whether there is a risk of positive and negative terms offsetting each load direction during the decoupling process.

[0099] Step 208, make a comprehensive judgment; then proceed to step 209; Here, based on the above indicators and preset thresholds, the output matrix decoupling stability level (quality evaluation information), risk type, and reason for non-compliance are determined.

[0100] Step 209, optimize feedback.

[0101] Specifically, based on the risk type, suggestions for structure, patch, bridge, gain, calibration, or product release (i.e., matrix processing scheme or processing action information) are generated, and the sensor design or calibration scheme is updated.

[0102] In this application example, a quality evaluation device for the raw output matrix of a force sensor is also provided. This includes: The data acquisition module is used to acquire six-dimensional loads and raw output data obtained from simulation, calibration, or mass production testing.

[0103] The normalization module is used to uniformly normalize the load and the original output matrix according to the full scale of each axis.

[0104] The matrix construction module is used to generate the normalized original output matrix Sn and the decoupling matrix C.

[0105] The indicator calculation module is used to calculate... R L , , , i , , , Indicators such as...

[0106] The risk diagnosis module is used to identify risks such as weak output, pattern similarity, matrix ill-conditioned, large number subtraction, and noise amplification based on thresholds or scoring rules.

[0107] The optimization suggestion module is used to output suggestions for structure, surface mount, bridging, gain, calibration, and product release.

[0108] The report generation module is used to generate matrix quality evaluation reports for R&D design review, calibration result verification, or mass production quality inspection.

[0109] Based on the above analysis, existing technologies struggle to distinguish between different sources of risk, such as weak output, similar patterns, excessively large condition numbers, large number subtraction, and noise amplification. Furthermore, they lack a closed-loop method to feed back matrix evaluation results to optimize the elastic body structure, strain gauge positions, bridge circuit combinations, signal conditioning gain, and calibration schemes. In mass production testing, there is a lack of an automatic method for judging the quality of the calibration matrix for each sensor, making it difficult to detect products with unstable matrix quality early. Therefore, a method for evaluating the stability of the original output matrix decoupling is needed, applicable to sensor structure design, finite element simulation, prototype calibration, bridge configuration comparison, and mass production testing. This method would allow for early assessment of whether the matrix is ​​suitable for stable software decoupling and for pinpointing specific risk sources.

[0110] The purpose of this application is to provide a method and system for evaluating and optimizing the decoupling stability of the original output matrix of a six-dimensional force sensor. By constructing a full-scale normalized original output matrix, the method calculates indicators such as output intensity, mode independence, matrix stability, cancellation risk, and noise amplification risk, automatically determines whether the matrix is ​​suitable for stable software decoupling, and outputs risk sources and optimization suggestions, thereby improving the reliability of six-dimensional force sensor structural design, bridging scheme selection, calibration matrix quality judgment, and mass production consistency control.

[0111] This application does not directly evaluate the final six-dimensional output error, but rather takes the original output vector formed by the six-dimensional sensor under a unit full-scale normalized load as the analysis object. The six load directions correspond to six original output vectors, which together constitute the full-scale normalized original output matrix Sn. By analyzing the column vector strength, correlation between column vectors, singular value distribution, condition number, cancellation characteristics of the decoupling matrix row vectors, and noise propagation characteristics of Sn, the matrix quality can be determined before decoupling calculations or after calibration.

[0112] Furthermore, this invention correlates the evaluation results with engineering optimization actions: if the output vector magnitude is too small, it suggests enhancing the structural sensitivity or bridge gain in the corresponding load direction; if the cosine correlation coefficient is too high, it suggests adjusting the elastomer structure, patch position, or bridging method; if the condition number or noise amplification factor is too high, it suggests redesigning the weak output direction, recalibrating, or determining that the matrix is ​​unqualified; if the risk of subtracting large numbers is high, it suggests optimizing the bridge combination and signal conditioning consistency.

[0113] In summary, the advantages of this application are: 1. It can determine in advance whether the original output matrix is ​​suitable for stable software decoupling after the structural design, bridge design, calibration and mass production testing stages are completed.

[0114] 2. It can distinguish different sources of risk, such as weak output, similar output patterns, ill-conditioned matrix, large number subtraction, and noise amplification.

[0115] 3. It can transform the evaluation results of mathematical matrices into specific engineering optimization actions, thereby improving the efficiency of sensor design iteration.

[0116] 4. It can reduce misjudgments caused by relying solely on coupling errors after calibration, and improve the noise immunity, drift immunity, and consistency of repeated calibration after installation.

[0117] 5. It can be used for calibration matrix quality screening in the mass production process to improve the consistency of six-dimensional force sensor products.

[0118] To implement the method of the embodiments of this application, the embodiments of this application also provide a quality evaluation device for the raw output matrix of a force sensor. This device corresponds to the above-described quality evaluation method, and each step in the embodiments of the above-described quality evaluation method is also fully applicable to the quality evaluation device embodiment for the raw output matrix of a force sensor.

[0119] like Figure 3 As shown, the quality evaluation device for the raw output matrix of the force sensor includes: a data acquisition module 301, a matrix construction module 302, an index calculation module 303, and a decision module 304. Data acquisition module 301 is used to acquire the raw output data of the force sensor under all load directions; Matrix construction module 302 is used to determine the original output matrix based on the original output data and the corresponding load information; The indicator calculation module 303 is used to determine one or more evaluation indicators based on the original output matrix; Decision module 304 is used to determine the comprehensive quality score based on the evaluation indicators; Based on the score range of the comprehensive quality score, quality evaluation information representing the usability of the original output matrix is ​​determined.

[0120] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. The decision module 304 is specifically used for: The comprehensive quality score is determined based on one or more of the following indicators: output strength, balance, output model similarity, inversion stability, weak independence, large number subtraction risk, and noise risk, along with their corresponding weighting coefficients. The method further includes: determining a matrix processing scheme based on the score range in which the comprehensive quality score is located.

[0121] In some embodiments, the evaluation index includes an output intensity index, and the index calculation module 303 is specifically used for: The output intensity information for each load direction is determined based on the magnitude of each column vector of the original output matrix; The output intensity index is determined based on the output intensity information.

[0122] In some embodiments, the evaluation index further includes a balance index, and the index calculation module 303 is specifically used for: The balance index is determined based on the maximum and minimum output intensity information corresponding to the original output matrix.

[0123] In some embodiments, the evaluation index includes an output pattern similarity index, and the index calculation module 303 is specifically used for: The output pattern similarity index is determined based on the degree of similarity between every two column vectors in the original output matrix.

[0124] In some embodiments, the evaluation index includes at least one of a weak independence index, an inverse stability index, and a noise risk index, and the index calculation module 303 is specifically used for at least one of the following: The weak independence index is determined based on the minimum singular value of the original output matrix; The inversion stability index is determined based on the maximum and minimum singular values ​​of the original output matrix. The noise risk index is determined based on the reciprocal of the minimum singular value of the original output matrix.

[0125] In some embodiments, the evaluation index includes a large number subtraction risk index, and the index calculation module 303 is specifically used for: Determine the decoupling matrix based on the original output matrix; The risk index for large number subtraction is determined based on the decoupling matrix, the original output data, or the column vectors of the original output matrix.

[0126] In some embodiments, the evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. The decision module 304 is further configured to: If one or more of the following indicators—output intensity, balance, output pattern similarity, inversion stability, weak independence, large number subtraction risk, and noise risk—are not within the corresponding safety threshold range, then processing action information is generated.

[0127] It should be noted that the quality evaluation device for the original output matrix of the force sensor provided in the above embodiments is only illustrated by the division of the above-described program modules. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. Furthermore, the quality evaluation device for the original output matrix of the force sensor provided in the above embodiments and the quality evaluation method embodiment for the original output matrix of the force sensor belong to the same concept, and their specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0128] Based on the hardware implementation of the above program modules, and in order to implement the method of the embodiments of this application, the embodiments of this application also provide an electronic device. Figure 4 The diagram shows only an exemplary structure of the electronic device, not the entire structure; implementation is possible as needed. Figure 4 The structure shown may be part or all of the structure.

[0129] like Figure 4 As shown, the electronic device 400 provided in this application embodiment includes at least one processor 401, a memory 402, and a user interface 403. The various components in the electronic device 400 are coupled together via a bus system 404. It can be understood that the bus system 404 is used to implement communication between these components. In addition to a data bus, the bus system 404 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 4 The general designated all buses as Bus System 404.

[0130] The user interface 403 may include a monitor, keyboard, mouse, trackball, click wheel, buttons, touchpad, or touch screen.

[0131] The memory 402 in this embodiment is used to store various types of data to support the operation of the electronic device 400. Examples of such data include any computer program used to operate on the electronic device 400.

[0132] The method for evaluating the quality of the raw output matrix of a force sensor disclosed in this application can be applied to, or implemented by, a processor 401. The processor 401 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the method for evaluating the quality of the raw output matrix of a force sensor can be completed by integrated logic circuits in the hardware of the processor 401 or by instructions in software form. The processor 401 can be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 401 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules can be located in a storage medium, specifically memory 402. The processor 401 reads information from memory 402 and, in conjunction with its hardware, completes the steps of the method for evaluating the quality of the raw output matrix of a force sensor provided in the embodiments of this application.

[0133] In an exemplary embodiment, the electronic device 400 may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned method.

[0134] It is understood that memory 402 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Sync Link Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memories described in the embodiments of this application are intended to include, but are not limited to, these and any other suitable types of memories.

[0135] In an exemplary embodiment, this application also provides a storage medium, namely a computer storage medium, specifically a computer-readable storage medium, such as a memory 402 that stores a computer program. The computer program can be executed by the processor 401 of the electronic device 400 to complete the steps described in the method of this application embodiment. The computer-readable storage medium can be a ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM, etc.

[0136] In an exemplary embodiment, this application also provides a computer program product, including a computer program that can be executed by a processor 401 of an electronic device 400 to perform the steps described in the method of this application embodiment.

[0137] It should be noted that terms such as "first" and "second" are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. In this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0138] It should be understood that the phrase "some embodiments" throughout the specification means that a particular feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "some embodiments" appearing throughout the specification does not necessarily refer to the same embodiment.

[0139] In the description of the embodiments of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" or "linked" should be interpreted broadly. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, or a connection that allows communication between the components; it can be a direct connection or an indirect connection through an intermediate medium; it can be the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0140] Furthermore, the technical solutions described in the embodiments of this application can be combined arbitrarily without conflict.

[0141] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating the quality of the raw output matrix of a force sensor, characterized in that, The method includes: Acquire raw output data of the force sensor under all load directions; The original output matrix is ​​determined based on the original output data and the corresponding load information; One or more evaluation indicators are determined based on the original output matrix; The overall quality score is determined based on the aforementioned evaluation indicators; Based on the score range of the comprehensive quality score, quality evaluation information representing the usability of the original output matrix is ​​determined; The process of determining the original output matrix based on the original output data and the corresponding load information includes: For each load direction, the full-scale normalized load value is determined based on the preset load value and the full-scale value. Based on the full-scale normalized load value and the corresponding original output data, determine the full-scale normalized original output data; The original output matrix is ​​determined based on the full-scale normalized original output data corresponding to each load direction. The evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. Determining the comprehensive quality score based on the evaluation metrics includes: The comprehensive quality score is determined based on one or more of the following indicators: output strength, balance, output model similarity, inversion stability, weak independence, large number subtraction risk, and noise risk, along with their corresponding weighting coefficients. The method further includes: determining a matrix processing scheme based on the score range in which the comprehensive quality score is located.

2. The method according to claim 1, characterized in that, The evaluation metrics include output intensity metrics, and the determination of one or more evaluation metrics based on the original output matrix includes: The output intensity information for each load direction is determined based on the magnitude of each column vector of the original output matrix; The output intensity index is determined based on the output intensity information.

3. The method according to claim 2, characterized in that, The evaluation indicators also include balance indicators, and the process of determining one or more evaluation indicators based on the original output matrix further includes: The balance index is determined based on the maximum and minimum output intensity information corresponding to the original output matrix.

4. The method according to claim 1, characterized in that, The evaluation metrics include output pattern similarity metrics, and the determination of one or more evaluation metrics based on the original output matrix includes: The output pattern similarity index is determined based on the degree of similarity between every two column vectors in the original output matrix.

5. The method according to claim 1, characterized in that, The evaluation metrics include at least one of the weak independence metric, the inverse stability metric, and the noise risk metric. The determination of one or more evaluation metrics based on the original output matrix includes at least one of the following: The weak independence index is determined based on the minimum singular value of the original output matrix; The inversion stability index is determined based on the maximum and minimum singular values ​​of the original output matrix. The noise risk index is determined based on the reciprocal of the minimum singular value of the original output matrix.

6. The method according to claim 1, characterized in that, The evaluation metrics include a large number subtraction risk metric, and the determination of one or more evaluation metrics based on the original output matrix includes: Determine the decoupling matrix based on the original output matrix; The risk index for large number subtraction is determined based on the decoupling matrix, the original output data, or the column vectors of the original output matrix.

7. The method according to claim 1, characterized in that, The method further includes: If one or more of the following indicators—output intensity, balance, output pattern similarity, inversion stability, weak independence, large number subtraction risk, and noise risk—are not within the corresponding safety threshold range, then processing action information is generated.

8. A quality evaluation device for the raw output matrix of a force sensor, characterized in that, The device includes: The data acquisition module is used to acquire the raw output data of the force sensor under all load directions; The matrix construction module is used to determine the original output matrix based on the original output data and the corresponding load information; The indicator calculation module is used to determine one or more evaluation indicators based on the original output matrix. The decision-making module is used to determine the overall quality score based on the evaluation indicators; Based on the score range of the comprehensive quality score, quality evaluation information representing the usability of the original output matrix is ​​determined; The matrix construction module is specifically used for: For each load direction, the full-scale normalized load value is determined based on the preset load value and the full-scale value. Based on the full-scale normalized load value and the corresponding original output data, determine the full-scale normalized original output data; The original output matrix is ​​determined based on the full-scale normalized original output data corresponding to each load direction. The evaluation metrics include one or more of the following: output intensity metric, balance metric, output pattern similarity metric, inversion stability metric, weak independence metric, large number subtraction risk metric, and noise risk metric. The decision module is specifically used for: The comprehensive quality score is determined based on one or more of the following indicators: output intensity, balance, output model similarity, inversion stability, weak independence, large number subtraction risk, and noise risk, along with their corresponding weighting coefficients; and... The matrix processing scheme is determined based on the score range of the comprehensive quality score.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, performs the steps of the method according to any one of claims 1 to 7.

10. A storage medium storing a computer program, characterized in that, When the computer program is executed by the control device, it implements the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the controlled device, it implements the steps of the method according to any one of claims 1 to 7.

Images