Stability testing method of six-dimensional force sensor based on data analysis

Through data analysis methods, the data anomalies and fluctuations of the six-dimensional force sensor are calculated, the abnormal period is screened, and the stability degree indicators are constructed, which solves the accuracy of the stability test of the six-dimensional force sensor in high humidity environments and improves the test results. accuracy.

CN118656759BActive Publication Date: 2024-11-05SHENZHEN AMPRON TECH CORP
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411124261.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-05
Estimated Expiration
2044-08-16

AI Technical Summary

Technical Problem

In environments with high humidity, rust internal parts of the six-dimensional force sensor and moisture insulating materials lead to a decrease in measurement accuracy and stability. The existing testing methods ignore the continuity characteristics of adjacent data in the time series, resulting in stability test results Inaccurate.

Method used

Using a data analysis method, by collecting the output data of the six-dimensional force sensor, calculating the data anomalies and fluctuations in each dimension, filtering the abnormal period, combining preset judgment thresholds, determining the data fluctuation period caused by sensor instability, and building stability degree indicator, considering the continuity characteristics of adjacent data.

Benefits of technology

It improves the accuracy of the stability test of the six-dimensional force sensor, can accurately detect abnormal fluctuations caused by sensor instability, distinguish data fluctuations caused by robotic arm impact and sensor instability, and provides a more accurate stability assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118656759B_ABST
    Figure CN118656759B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of six-dimensional force sensor stability testing, and specifically to a six-dimensional force sensor stability testing method based on data analysis, the method comprising: collecting the output data of the six-dimensional force sensor in each dimension; determining the data abnormality of each dimension and each period; screening the period of abnormal fluctuation in each dimension; screening the period of data fluctuation caused by sensor instability in each dimension, recorded as the second abnormal period; based on the number of second abnormal periods in all dimensions, the interval length between the second abnormal periods in each dimension, and the number of dimensions with the second abnormal period, determine the stability of the six-dimensional force sensor. The present application aims to improve the accuracy of the stability test results of the six-dimensional force sensor.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of six-dimensional force sensor stability testing, and in particular to a six-dimensional force sensor stability testing method based on data analysis. Background Art

[0002] The six-axis force sensor is a high-precision force sensor that can simultaneously measure forces and torques in three directions. The sensor is widely used in the industrial field and is usually installed at the end of a robotic arm to collect force and torque, and feed the data back to technicians for better control of the robotic arm.

[0003] In a dry environment, the six-axis force sensor can usually show excellent performance, accurately measuring and recording the force and torque of an object in six degrees of freedom. However, when the humidity increases, the internal parts of the sensor may rust and the insulating material may deteriorate due to moisture, which in turn affects the measurement accuracy and stability of the sensor. Therefore, when the working environment of the six-axis force sensor is a high humidity environment, in order to ensure the accuracy of the output data, the stability of the sensor needs to be tested regularly.

[0004] Data analysis methods are usually used to test the output data of the sensor. When performing anomaly detection on the output data, traditional anomaly detection algorithms often only consider the anomaly of a single data relative to the overall data, while ignoring the continuity characteristics of adjacent data in the time series, resulting in inaccurate stability test results of the six-dimensional force sensor. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a six-dimensional force sensor stability testing method based on data analysis to solve the existing problems.

[0006] The six-dimensional force sensor stability test method based on data analysis of this application adopts the following technical solutions:

[0007] An embodiment of the present application provides a six-dimensional force sensor stability testing method based on data analysis, the method comprising the following steps:

[0008] Collect the output data of the six-dimensional force sensor in each dimension;

[0009] Based on the differences between the predicted data of each period of each dimension and its adjacent previous period and the output data of each period of each dimension, as well as the fluctuation of the output data of each period of each dimension, the data anomaly of each period of each dimension is determined; based on the normal distribution of the data anomaly of each dimension in all periods, the period with abnormal fluctuations in each dimension is screened and recorded as the first abnormal period;

[0010] Based on the difference value changes of the output data of each first abnormal period of each dimension and its adjacent subsequent period, and the data fluctuation changes of each first abnormal period of each dimension, combined with the preset judgment threshold, the period of data fluctuation caused by sensor instability in each dimension is screened and recorded as the second abnormal period;

[0011] The stability of the six-dimensional force sensor is determined based on the number of second abnormal cycles in all dimensions, the interval length between the second abnormal cycles in each dimension, and the number of dimensions in which the second abnormal cycles occur.

[0012] Preferably, the method for determining the data abnormality of each dimension and each period includes:

[0013] Get the predicted value of each dimension and each period;

[0014] Based on the differences between the predicted values ​​of each period of each dimension and its adjacent previous period and the actual values ​​of each period of each dimension, as well as the maximum value difference of each period of each dimension, the data abnormality of each period of each dimension is determined; wherein the actual value is the output data.

[0015] Preferably, the method for obtaining the predicted value is: taking the period as the moving interval step, adopting the simple moving average method to obtain the moving average value of each period in each dimension, and recording it as the predicted value of each period in each dimension.

[0016] Preferably, the data abnormality of each dimension and each period is further determined by the expression: ; In the formula, is the data anomaly degree of the nth cycle under the i-th dimension; , They are the predicted values ​​of the nth period and the n-1th period under the i-th dimension respectively; is the average value of the actual value of the nth period in the i-th dimension; is the output data of the nth cycle in the i-th dimension; and are the maximum and minimum values ​​of all output data in the nth cycle under the i-th dimension respectively; is the first preset constant.

[0017] Preferably, the screening process for the periods in which abnormal fluctuations occur in each dimension is as follows: obtaining the values ​​of the data abnormality of each dimension in all periods that are outside the range of 2sigma, and recording the period in which the values ​​are located as the period in which abnormal fluctuations occur in each dimension.

[0018] Preferably, the method for determining the period of data fluctuation caused by sensor instability in each dimension includes:

[0019] Based on the difference value changes of the output data of each first abnormal period in each dimension and its adjacent subsequent period, and the data fluctuation changes of each first abnormal period in each dimension, determine the abnormal type discrimination coefficient of each first abnormal period in each dimension;

[0020] Based on the abnormal type discrimination coefficient and the preset judgment threshold, the periods of data fluctuation caused by sensor instability in each dimension are screened.

[0021] Preferably, the abnormal type discrimination coefficient of each first abnormal period in each dimension is further determined by the expression: ; In the formula, is the anomaly type discrimination coefficient of the mth first anomaly cycle in the ith dimension; is the mean of the differential sequence of output data in the next cycle adjacent to the mth first abnormal cycle in the i-th dimension; is the differential sequence of output data in the mth first abnormal period under the i-th dimension; is the variance function; is a sequence consisting of output data in the mth first abnormal period of the i-th dimension; is the information entropy function; is the normalization function; is the second preset constant.

[0022] Preferably, the screening process of the period of data fluctuation caused by sensor instability in each dimension is: when the abnormal type discrimination coefficient is greater than a preset judgment threshold, the period in which it is located is determined as the period of data fluctuation caused by sensor instability in its dimension.

[0023] Preferably, the method for determining the stability of the six-dimensional force sensor includes:

[0024] When the number of second abnormal cycles in all dimensions is 0 or 1, the six-dimensional force sensor is in a stable state;

[0025] Otherwise, the stability of the six-dimensional force sensor is determined based on the intervals between the second abnormal cycles in all dimensions and the number of dimensions in which the second abnormal cycles occur.

[0026] Preferably, the method for further determining the stability of the six-dimensional force sensor is:

[0027] Calculate the mean interval duration between all adjacent second abnormal cycles in each dimension;

[0028] The ratio of the average value of the interval durations in all dimensions to the number of dimensions in which the second abnormal cycle occurs is used as the stability of the six-dimensional force sensor.

[0029] This application has at least the following beneficial effects:

[0030] Based on the difference between the predicted data and the output data of the adjacent previous cycle, as well as the fluctuation of the output data, an abnormality index is constructed to detect the period with abnormal fluctuations in the data, solving the problem of missed detection of abnormal fluctuations; based on the difference of the output data of the adjacent next cycle with the abnormal fluctuation period, as well as the data fluctuation change of the abnormal fluctuation period, it is judged whether the abnormal fluctuation is the data trend fluctuation caused by sensor instability or the data trend fluctuation of the sensor caused by the impact of the robot arm, thereby improving the detection accuracy of abnormal fluctuations caused by sensor instability. Finally, according to the number of cycles, interval duration, and number of dimensions of data fluctuations caused by sensor instability in six dimensions, a stability index of the six-dimensional force sensor is constructed to judge the stability of the six-dimensional force sensor, fully considering the continuity characteristics of adjacent data, and improving the accuracy of the stability test results of the six-dimensional force sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 A flow chart of a six-dimensional force sensor stability testing method based on data analysis provided in one embodiment of the present application;

[0033] Figure 2 A flowchart for constructing an indicator of the stability of a six-dimensional force sensor provided in one embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the six-axis force sensor stability test method based on data analysis proposed in the present application, its specific implementation method, structure, features and its effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0035] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0036] The specific scheme of the six-dimensional force sensor stability testing method based on data analysis provided by the present application is described in detail below with reference to the accompanying drawings.

[0037] An embodiment of the present application provides a six-dimensional force sensor stability testing method based on data analysis.

[0038] Specifically, the following six-axis force sensor stability test method based on data analysis is provided, please refer to Figure 1 , the method comprises the following steps:

[0039] The first step is to collect the output data of the six-dimensional force sensor in each dimension.

[0040] In industrial manufacturing, the six-axis force sensor can monitor the force and torque of the robot arm and provide timely data feedback to the operator. When the humidity in the working environment is high, the parts of the six-axis force sensor may corrode and rust over time, which will affect the accuracy of its test data.

[0041] This example collects 2000 sets of six-dimensional output data from the six-dimensional force sensor installed at the end of the robot arm within three months. The six-dimensional output data includes forces in three directions and torques in three directions. These output data are arranged from left to right in the order of collection time to obtain the matrix E :

[0042] ;

[0043] The six rows in the above matrix represent the three forces and three moments on the x, y, and z axes output by the six-dimensional force sensor, respectively. Each column represents the six output data of the six-dimensional force sensor under the action of one force. s represents the number of forces acting on the six-dimensional force sensor within three months. In this embodiment, s is 2000.

[0044] At this point, the output data of the six-dimensional force sensor can be obtained through the above method.

[0045] The second step is to determine the data anomaly of each period of each dimension based on the differences between the predicted data of each period of each dimension and its adjacent previous period and the output data of each period of each dimension, as well as the fluctuation of the output data of each period of each dimension; based on the normal distribution of the data anomaly of each dimension in all periods, screen the period with abnormal fluctuations in each dimension and record it as the first abnormal period.

[0046] For the robotic arms in industrial manufacturing, they usually work according to pre-set actions, so the force and torque of the robotic arms should be periodic, and the length and start time of the cycle in each dimension should be consistent. In addition, the data of the six dimensions of the six-dimensional force sensor are independent of each other, that is, the data size of one dimension will not affect the data size of other dimensions.

[0047] When the performance of the six-dimensional force sensor is not affected by moisture, the overall trend of each dimension of data is stable; when affected by moisture, the force strain gauge becomes less sensitive due to rust or corrosion, and the resistance of the force strain gauge increases, so the data of one or more dimensions may tend to increase; in addition, moisture may also enter the sensor, causing a short circuit failure, resulting in a large error in the test data. In order to avoid serious deviations in the test data of the six-dimensional force sensor caused by moisture, it is necessary to monitor the stability of the sensor at the early stage when the sensor is affected by moisture.

[0048] The time interval from the start moment to the moment when the robot arm completes a set of actions and returns to the starting point is taken as a cycle. In this embodiment, every 10 forces are taken as a data cycle. The output data of the six-dimensional force sensor is used as the input of the simple moving average method, and the data cycle t is used as the moving interval step. The moving average value of each cycle is calculated in a step-by-step manner and recorded as the predicted value of each cycle. The predicted value can be used to predict future data based on the data when it is stable. The simple moving average method is a well-known technology and the specific process will not be repeated.

[0049] Ideally, if the output data of the sensor changes trend in the nth cycle, there will be a deviation between the predicted value and the actual value of the nth cycle, and it can be concluded that the data in this cycle is abnormal.

[0050] But in fact, the deviation between the predicted value and the true value in the nth period may be very small, which is not enough to distinguish it from the deviation in the stable state of the data; and the predicted value will only deviate from the true value in this period of trend change. After this period, the predicted value will be predicted in the direction of the trend change. At this time, the predicted value and the true value will be closer. It is impossible to judge whether the data is abnormal by the deviation between the predicted value and the true value.

[0051] When the data is abnormal in the nth period, not only will the predicted value of this period deviate from the true value, but the predicted value of the previous period will also deviate from the true value of this period. The maximum value of the true value in the nth period may be larger than the maximum value in the previous n-1 periods.

[0052] According to the above characteristics, a data anomaly index is constructed to determine whether there is an abnormal change in the data in the nth period. The data anomaly degree of the nth period under the i-th dimension is The expression is: ; In the formula, is the data anomaly degree of the nth cycle under the i-th dimension, ; , They are the predicted values ​​of the nth period and the n-1th period under the i-th dimension respectively; is the average value of the actual value of the nth period in the i-th dimension; is the output data of the nth cycle in the i-th dimension; and are the maximum and minimum values ​​of all output data in the nth cycle under the i-th dimension respectively; It is the first preset constant to prevent the denominator from being 0. In this embodiment, it is 0.01.

[0053] It should be understood that when When the value of is large, it means that there is a large difference between the predicted value and the true value of the nth period under the i-th dimension, indicating that the data is likely to fluctuate abnormally in this period; when When the value of is large, it means that there is a large difference between the predicted value of the n-1th period under the i-th dimension and the true value of the nth period, indicating that the data is likely to fluctuate abnormally in this period; The bigger, The smaller it is, the greater the data difference in the nth period under the ith dimension, which means that the data is likely to experience abnormal fluctuations within this period. The larger the value of , the greater the possibility of abnormal fluctuations in the nth period under the i-th dimension.

[0054] Calculate the data anomaly of all periods in each dimension. It should be noted that the data anomaly of all periods in each dimension should belong to the normal distribution. According to the 3σ (3sigma) criterion, if the data anomaly Not in range If the value is within , it is considered that the nth period under the ith dimension has abnormal fluctuation, and this period is recorded as the first abnormal period.

[0055] So far, the period of abnormal fluctuations in each dimension has been obtained through the above method.

[0056] The third step is to screen the period of data fluctuation caused by sensor instability in each dimension based on the difference value change of the output data of each first abnormal period in each dimension and its adjacent subsequent period, as well as the data fluctuation change of each first abnormal period in each dimension, combined with the preset judgment threshold, and record it as the second abnormal period.

[0057] The abnormal fluctuation periods obtained above are not necessarily due to the instability of the sensor itself, but may also be due to the impact of the robot arm causing abnormal data. Therefore, all the first abnormal periods obtained need to be further distinguished.

[0058] The damage to the internal parts of the sensor due to moisture is formed slowly and is irreversible, so the trend change is not very strong at the beginning, but it will continue to grow; in the case where the robotic arm is hit, the data will fluctuate greatly, and its value may increase or decrease, and will gradually return to a stable state after the fluctuation.

[0059] According to the above characteristics, an abnormality type discrimination coefficient index is constructed to determine whether the first abnormal period is caused by sensor instability or sensor data trend fluctuations caused by the impact of the robot arm. The abnormality type discrimination coefficient of the first abnormal period of the mth dimension is recorded as , the expression is: ; In the formula, is the anomaly type discrimination coefficient of the mth first anomaly cycle in the ith dimension; is the mean of the differential sequence of output data in the next cycle adjacent to the mth first abnormal cycle in the i-th dimension; is the differential sequence of output data in the mth first abnormal period under the i-th dimension; is the variance function; is a sequence consisting of output data in the mth first abnormal period of the i-th dimension; is the information entropy function; is the normalization function; It is the second preset constant to prevent the denominator from being 0. In this embodiment, it is 0.01.

[0060] Among them, the calculation of differential sequence, information entropy and variance are all well-known technologies and will not be repeated here. In other embodiments of the present application, the range of the sequence can also be calculated instead of calculating the information entropy function of the sequence. Both the range and information entropy can characterize the abnormal fluctuations of data in the sequence.

[0061] It should be understood that The larger the value is, the greater the data trend change is in the m+1th cycle, indicating that the abnormal fluctuation is more likely to be caused by sensor instability; The smaller the value is, the smaller the data fluctuation amplitude of the mth period in the i-th dimension is, which means that the abnormal fluctuation is more likely to be caused by sensor instability. The smaller it is, the smaller the data fluctuation amplitude in the mth cycle is, which means that the abnormal fluctuation in this cycle is more likely to be caused by sensor instability. The larger it is, the more likely the abnormal fluctuation of the mth first abnormal period in the i-th dimension is that the data trend fluctuation is caused by sensor instability.

[0062] Calculate the abnormal type discrimination coefficient of each first abnormal cycle. Take the preset judgment threshold s=0.6, when When , it means that the abnormal fluctuation of the first abnormal cycle of the mth dimension belongs to the data trend fluctuation caused by sensor instability; when When , it indicates that the abnormal fluctuation of the first abnormal period of the mth dimension belongs to the sensor data trend fluctuation caused by the impact of the robotic arm.

[0063] Therefore, the first abnormal period whose abnormal type discrimination coefficient is greater than the preset judgment threshold is recorded as the second abnormal period of the dimension in which it is located. The number of second abnormal periods in all dimensions is recorded as N.

[0064] The fourth step is to determine the stability of the six-dimensional force sensor based on the number of second abnormal cycles in all dimensions, the interval between the second abnormal cycles in each dimension, and the number of dimensions in which the second abnormal cycles occur.

[0065] To judge whether a six-dimensional force sensor is stable, we must first consider whether the data fluctuates and the number of dimensions in which the data fluctuates. Secondly, we must consider the time interval between the data fluctuations in two dimensions. If the interval is long, it means that the sensor is slowly affected by moisture and is relatively stable; if the interval is short, it means that the sensor is quickly affected by moisture and has poor stability.

[0066] If N=0, it means that the six-axis force sensor has no data fluctuation caused by moisture, indicating that the sensor is stable; if N=1, it means that the six-axis force sensor begins to be affected by moisture and data fluctuation occurs, but the sensor is still stable; if , indicating that the six-dimensional force sensor has experienced multiple data fluctuations, thus constructing The stability of the six-dimensional force sensor W is: ;W is the stability of the six-dimensional force sensor; N is the number of second abnormal cycles of the six-dimensional force sensor in all dimensions; L is the number of dimensions where the second abnormal cycle is located; , They are the starting times of the f-th and f-1-th second abnormal cycles in all dimensions respectively.

[0067] It should be understood that the smaller the value of N is, the fewer the number of data fluctuations are, indicating that the six-dimensional force sensor is more stable; The larger the value, the more slowly the six-axis force sensor is affected by moisture and the more stable its performance; the larger the value of W, the better the stability of the six-axis force sensor.

[0068] It should be noted that when the fth second abnormal period and the f-1th second abnormal period are not in the same dimension, they are not calculated. .

[0069] When testing the stability of a six-dimensional force sensor, first determine whether the data is affected by moisture based on the value of the second abnormal period number N in all dimensions. or , indicating that the six-dimensional force sensor is not greatly affected by moisture and is in a stable state; if , the stability of the six-dimensional force sensor when affected by moisture is judged according to the value of the stability degree W.

[0070] In one embodiment of the present application, a flowchart of constructing an index of the stability of a six-dimensional force sensor is shown in the attached figure. Figure 2 shown.

[0071] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0072] It should be noted that, unless otherwise specified and limited, terms such as "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such article or device. In the absence of further restrictions, an element defined by the sentence "including one..." does not exclude the existence of other identical elements in the article or device including the element. In addition, the term "and\or" used herein includes any and all combinations of one or more related listed items.

[0073] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not invented by the present application.

[0074] It should be understood that the present application is not limited to the exact construction that has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof.

Claims

1. Six-axis force sensor stability test method based on data analysis, It is characterized in that The method comprises the following steps: Collect the output data of the six-dimensional force sensor in each dimension; Based on the differences between the predicted data of each period of each dimension and its adjacent previous period and the output data of each period of each dimension, as well as the fluctuation of the output data of each period of each dimension, the data anomaly of each period of each dimension is determined; based on the normal distribution of the data anomaly of each dimension in all periods, the period with abnormal fluctuations in each dimension is screened and recorded as the first abnormal period; Based on the difference value changes of the output data of each first abnormal period of each dimension and its adjacent subsequent period, and the data fluctuation changes of each first abnormal period of each dimension, combined with the preset judgment threshold, the period of data fluctuation caused by sensor instability in each dimension is screened and recorded as the second abnormal period; Determine the stability of the six-dimensional force sensor based on the number of second abnormal cycles in all dimensions, the interval between the second abnormal cycles in each dimension, and the number of dimensions in which the second abnormal cycles occur; The method for determining the data abnormality of each dimension and each period includes: Get the predicted value of each dimension and each period; Based on the differences between the predicted values ​​of each period of each dimension and its adjacent previous period and the actual values ​​of each period of each dimension, and the maximum value difference of each period of each dimension, the data abnormality of each period of each dimension is determined; wherein the actual value is the output data; The expression for further determining the data abnormality of each dimension and each period is: ; In the formula, is the data anomaly degree of the nth cycle under the i-th dimension; , They are the predicted values ​​of the nth period and the n-1th period under the i-th dimension respectively; is the average value of the actual value of the nth period in the i-th dimension; is the output data of the nth cycle in the i-th dimension; and are the maximum and minimum values ​​of all output data in the nth cycle under the i-th dimension respectively; is the first preset constant; The method for determining the period of data fluctuation caused by sensor instability in each dimension includes: Based on the difference value changes of the output data of each first abnormal period in each dimension and its adjacent subsequent period, and the data fluctuation changes of each first abnormal period in each dimension, determine the abnormal type discrimination coefficient of each first abnormal period in each dimension; Based on the abnormal type discrimination coefficient and the preset judgment threshold, filter the period of data fluctuation caused by sensor instability in each dimension; The expression for further determining the abnormal type discrimination coefficient of each first abnormal period in each dimension is: ; In the formula, is the anomaly type discrimination coefficient of the mth first anomaly cycle in the ith dimension; is the mean of the differential sequence of output data in the next cycle adjacent to the mth first abnormal cycle in the i-th dimension; is the differential sequence of output data in the mth first abnormal period under the i-th dimension; is the variance function; is a sequence consisting of output data in the mth first abnormal period of the i-th dimension; is the information entropy function; is the normalization function; is the second preset constant.

2. The six-dimensional force sensor stability testing method based on data analysis as claimed in claim 1, It is characterized in that The method for obtaining the predicted value is: taking the period as the moving interval step, using the simple moving average method to obtain the moving average value of each period in each dimension, and recording it as the predicted value of each period in each dimension.

3. The six-dimensional force sensor stability testing method based on data analysis as claimed in claim 1, It is characterized in that The screening process of the periods in which abnormal fluctuations occur in each dimension is as follows: obtaining the values ​​of the data abnormality of each dimension in all periods that are outside the range of 2sigma, and recording the period in which the values ​​are located as the period in which abnormal fluctuations occur in each dimension.

4. The six-dimensional force sensor stability testing method based on data analysis as claimed in claim 1, It is characterized in that The screening process of the period of data fluctuation caused by sensor instability in each dimension is: when the abnormal type discrimination coefficient is greater than the preset judgment threshold, the period in which it is located is determined as the period of data fluctuation caused by sensor instability in its dimension.

5. The six-dimensional force sensor stability testing method based on data analysis as claimed in claim 1, It is characterized in that The method for determining the stability of the six-dimensional force sensor includes: When the number of second abnormal cycles in all dimensions is 0 or 1, the six-dimensional force sensor is in a stable state; Otherwise, the stability of the six-dimensional force sensor is determined based on the intervals between the second abnormal cycles in all dimensions and the number of dimensions in which the second abnormal cycles occur.

6. The six-dimensional force sensor stability testing method based on data analysis as claimed in claim 5, It is characterized in that The method for further determining the stability of the six-dimensional force sensor is as follows: Calculate the mean interval duration between all adjacent second abnormal cycles in each dimension; The ratio of the average value of the interval durations in all dimensions to the number of dimensions in which the second abnormal cycle occurs is used as the stability of the six-dimensional force sensor.

Citation Information

Patent Citations

  • Mechanical state monitoring method and system based on multiple heterogeneous sensors

    CN111551383A

  • Multivariable time sequence anomaly detection method and system based on graph neural network

    CN114881115A

  • Transformer operation abnormity monitoring method and system

    CN116955895A

  • Belt operation fault evaluation system based on multi-dimensional sensor data

    CN117272218A

  • Mechanical turntable state monitoring method based on sensor data

    CN118378125A