A difference data fusion method and device based on weighted fractional calculus

Abstract

The application provides a difference data fusion method and device based on weighted fractional calculus, and the method comprises the following steps: obtaining to-be-fused detection data of a plurality of sensing devices at a plurality of sampling time points; determining a weighting factor, a device influence factor and an environment influence factor corresponding to the sensing devices according to the to-be-fused detection data; performing fractional order fusion processing on the to-be-fused detection data according to the device influence factor and the environment influence factor to obtain preliminary fusion data corresponding to each sensing device; and performing weighted fusion processing on the preliminary fusion data according to the weighting factor to obtain target fusion data corresponding to each sampling time point. The application obtains the final target fusion data by performing fractional order fusion and weighted fusion on the same type of data in discrete manufacturing, effectively improves the precision of the fused data, thereby realizing effective fusion of the detection data in the discrete manufacturing workshop and improving the accuracy of the machine tool operation data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a difference data fusion method and device based on weighted fractional calculus. BACKGROUND

[0002] The precision manufacturing workshop of aerospace structures mainly uses milling and turning processing, and processes metal and non-metal parts, sub-assemblies, large cabin sections and satellite bodies in various types of aerospace structures and mechanisms. Aerospace structures have complex product types, different shapes and structures, integration and large-scale, and difficult process characteristics. The development of aerospace products has the characteristics of multiple varieties, small batch, and rolling batch development, which belongs to typical discrete manufacturing. In the daily production and processing process, the accuracy of various machine tool operation data (spindle power, turret speed, spindle speed, etc.) is an important factor affecting product processing quality and processing efficiency.

[0003] Currently, enterprises and research institutions use Internet of Things technology to realize the interconnection of manufacturing resources in discrete manufacturing workshops to effectively control and manage personnel, materials, equipment and production processes, and to improve manufacturing efficiency, reduce manufacturing costs and rationalize production cycles. Although detection technology, network technology and computer technology have gradually matured, due to various factors, the detection data collected has differences, such as differences in data acquisition equipment performance, production equipment performance, working environment and transmission distance. In addition, there is diversity in production environment and uncertainty in manufacturing information in discrete manufacturing, resulting in numerical differences between similar detection data, which seriously affects the accuracy and real-time of system decision-making. SUMMARY

[0004] The technical objective of the present application is to provide a difference data fusion method and device based on weighted fractional calculus to solve the problem of numerical differences between similar detection data in current discrete manufacturing, which seriously affects the accuracy and real-time of system decision-making.

[0005] To solve the above technical problems, the present application provides a difference data fusion method based on weighted fractional calculus, comprising:

[0006] Obtain the to-be-fused detection data of a plurality of sensing devices at a plurality of sampling time points, each sensing device corresponding to a machine tool, and the sensing device being used to detect data of the same type;

[0007] Determine the corresponding weighting factor, device influence factor and environment influence factor according to the to-be-fused detection data;

[0008] perform fractional order fusion processing on the to-be-fused detection data according to the device influence factor and the environment influence factor, to obtain preliminary fusion data corresponding to each of the sensing devices;

[0009] perform weighted fusion processing on the preliminary fusion data according to the weighting factor, to obtain target fusion data corresponding to each of the sampling time points.

[0010] Specifically, the method described above, the method comprises:

[0011] obtaining original detection data of the sensing devices at the sampling time points;

[0012] performing filtering processing on the original detection data according to a preset filter, to obtain the to-be-fused detection data, wherein the preset filter is a filter subjected to fractional order integral processing.

[0013] Further, the method described above, the method comprises:

[0014] obtaining a first standard deviation and a first variance of the to-be-fused detection data of each of the sensing devices at the sampling time points, respectively, and determining the first standard deviation as the device influence factor corresponding to each of the sensing devices;

[0015] determining the weighting factor according to the first variance and a first preset formula;

[0016] obtaining a second standard deviation of the to-be-fused detection data of each of the sensing devices at the same sampling time point, and determining the first standard deviation as the environment influence factor corresponding to the sampling time point.

[0017] Specifically, the method described above, after determining the device influence factor and the environment influence factor, the method further comprises:

[0018] determining a step corresponding to the device influence factor and the environment influence factor respectively during fractional order fusion, according to a value range of the device influence factor and / or the environment influence factor.

[0019] Further, the method described above, the method comprises:

[0020] obtaining a relationship function between the device influence factor, the environment influence factor and the to-be-fused detection data;

[0021] According to the relationship function and the pre-acquired fractional order and step length, a difference data fusion model based on fractional order partial differential of the to-be-fused detection data is determined;

[0022] According to the difference data fusion model, fractional order fusion processing is performed on the to-be-fused detection data, and the preliminary fusion data is obtained.

[0023] Specifically, the method as described above, the relationship function between the device influence factor, the environment influence factor and the to-be-fused detection data comprises:

[0024] According to a first preset algorithm, a first function relationship between the to-be-fused detection data and the device influence factor and a second function relationship between the to-be-fused detection data and the environment influence factor are respectively acquired;

[0025] According to the first function relationship, the second function relationship and the influence degree of the device influence factor and the environment influence factor on the to-be-fused detection data, the relationship function is determined.

[0026] Another embodiment of the present application further provides a control device, comprising:

[0027] A first processing module is configured to acquire to-be-fused detection data of a plurality of sensing devices at a plurality of sampling time points, each of the sensing devices corresponds to a machine tool, and each of the sensing devices is configured to detect data of the same type;

[0028] A second processing module is configured to determine a corresponding weighting factor, a device influence factor and an environment influence factor according to the to-be-fused detection data;

[0029] A third processing module is configured to perform fractional order fusion processing on the to-be-fused detection data according to the device influence factor and the environment influence factor, and obtain preliminary fusion data corresponding to each of the sensing devices;

[0030] A fourth processing module is configured to perform weighted fusion processing on the preliminary fusion data according to the weighting factor, and obtain target fusion data corresponding to each of the sampling time points.

[0031] Still another embodiment of the present application further provides a terminal device, comprising a processor, a memory and a computer program stored in the memory and executable on the processor, and the computer program is executed by the processor to implement the steps of the difference data fusion method based on weighted fractional calculus as described above.

[0032] Still another embodiment of the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the weighted fractional order calculus based difference data fusion method.

[0033] Compared with the prior art, the weighted fractional order calculus based difference data fusion method and device provided by the embodiments of the present application have at least the following beneficial effects:

[0034] The present application obtains the final target fusion data by performing fractional order fusion and weighted fusion on the same type data in discrete manufacturing, effectively improves the precision of the fusion data, thereby achieving effective fusion of the detection data in the discrete manufacturing workshop and improving the accuracy of the machine tool operation data. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 Fig. 1 is one of the flowcharts of the weighted fractional order calculus based difference data fusion method of the present application;

[0036] Figure 2 Fig. 2 is another of the flowcharts of the weighted fractional order calculus based difference data fusion method of the present application;

[0037] Figure 3 Fig. 3 is still another of the flowcharts of the weighted fractional order calculus based difference data fusion method of the present application;

[0038] Figure 4 Fig. 4 is a comparison diagram of the algorithm processing results of the present application in an actual application;

[0039] Figure 5 Fig. 5 is a structural diagram of the control device of the present application. DETAILED DESCRIPTION

[0040] In order to make the technical problems, technical solutions and advantages of the present application clearer, specific embodiments will be described in detail below with reference to the accompanying drawings. In the following description, specific details such as specific configurations and components are provided only to help a comprehensive understanding of the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, in order to be clear and concise, the description of known functions and structures is omitted.

[0041] It should be understood that the reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment is included in at least one embodiment of the application. The appearances of the phrase "in one embodiment" or "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0042] In various embodiments of the present application, it should be understood that the size of the serial number of the following processes does not mean the order of execution, and the execution order of the processes should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0043] It should be understood that the term "and / or" herein is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after it.

[0044] In the embodiments provided in the present application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that the determination of B according to A does not mean that B is determined only according to A, but B can also be determined according to A and / or other information.

[0045] Referring to Figure 1 An embodiment of the present application provides a difference data fusion method based on weighted fractional calculus, comprising:

[0046] In step S101, a plurality of sensing devices are obtained at a plurality of sampling time points, each of the sensing devices corresponds to a machine tool, and the sensing devices are used to detect the same type of data.

[0047] In step S102, a corresponding weighting factor, a device influence factor and an environment influence factor are determined according to the to-be-fused detection data.

[0048] In step S103, the to-be-fused detection data is subjected to fractional order fusion processing according to the device influence factor and the environment influence factor, to obtain preliminary fusion data corresponding to each of the sensing devices.

[0049] In step S104, the preliminary fusion data is subjected to weighted fusion processing according to the weighting factor, to obtain target fusion data corresponding to each of the sampling time points.

[0050] In the embodiment, when data fusion is performed on the differential data, the to-be-fused detection data to be fused is acquired, which can be data corresponding to a plurality of sampling time points detected by a plurality of sensing devices, wherein the sensing devices are used to detect the same type of data, and each sensing device corresponds to a machine tool. Specifically, taking a lathe as an example, the sensing device is used to detect spindle power (%), lathe turntable speed (r / min) or spindle speed (r / min), and the length of the sampling time point is 1 minute.

[0051] Due to the differences between the sensing devices and the position differences of the machine tools corresponding to the sensing devices, the to-be-fused detection data obtained by each sensing device is at least affected by the sensing device itself and the environment (machine tool position, sampling time point) in which the sensing device is located. Therefore, after obtaining the to-be-fused detection data, the device influence factor and the environment influence factor corresponding to the to-be-fused detection data are determined, and at the same time, in order to improve the accuracy of the final fitting, the weighting factor (weight) corresponding to the sensing device is also determined.

[0052] After the above weighting factor, device influence factor and environment influence factor are acquired, data fusion can be performed. When the fusion is performed, first, the to-be-fused detection data is subjected to fractional order fusion processing according to the device influence factor and the environment influence factor, to obtain preliminary fusion data corresponding to each sensing device, so as to preliminarily improve the accuracy of the detection data. Further, the preliminary fusion data is subjected to weighted fusion processing according to the weighting factor, so as to integrate the local estimation of each sensing device in the overall sensing device network according to the weighting factor, to obtain the global estimation of the overall sensor system, and further improve the precision of the fusion data.

[0053] In the weighted fusion, the weighted fusion algorithm is used for fusion, and the weighted fusion algorithm is as follows:

[0054]

[0055] wherein Y(t) i is the target fusion data at the sampling time point t, W(t) i is the weighting factor corresponding to the sensing device i at the sampling time point t, U(t) i is the preliminary fusion data corresponding to the sensing device i at the sampling time point t, and n is the number of sensing devices.

[0056] In summary, the final target fusion data is obtained by fractional order fusion and weighted fusion on the same type of data in discrete manufacturing, which effectively improves the precision of the fusion data, thereby realizing effective fusion of the detection data in the discrete manufacturing workshop and improving the accuracy of the machine tool operation data.

[0057] It should be noted that when there are multiple types of sensing devices on the machine tool, after receiving the data uploaded by the sensing devices, the data will be classified according to the type of the sensing device, and the same type of data will be executed The technical solutions described herein.

[0058] Referring to Figure 2 , specifically, the method as described above, the obtaining of the to-be-fused detection data of the plurality of sensing devices at a plurality of sampling time points comprises:

[0059] Step S201, obtaining original detection data uploaded by a plurality of sensing devices at a plurality of sampling time points;

[0060] Step S202, filtering the original detection data according to a preset filter to obtain the to-be-fused detection data, wherein the preset filter is a filter processed by fractional order integration.

[0061] In this embodiment, when obtaining the to-be-fused data, the original detection data uploaded by each sensing device is first obtained, and the original detection data is filtered to remove noise in the signal, such as high-frequency noise. In this embodiment, when filtering the original detection data, a filter processed by fractional order integration is used for filtering. By performing fractional order calculus transformation on the filter and adjusting the order of fractional order calculus according to the noise condition, the function form can be changed differently, thereby increasing the flexibility and diversity of the filter and achieving better filtering effect. The filter using fractional order calculus transformation can adapt to different noise environments and provide better filtering effect. Moreover, fractional order calculus transformation also makes the design of the filter more targeted and more suitable for actual application environment.

[0062] It should be noted that the filter in this embodiment is preferably a Butterworth low-pass filter.

[0063] Further, as described above, the method comprises:

[0064] respectively obtaining a first standard deviation and a first variance of the to-be-fused detection data of each sensing device at a plurality of sampling time points, and determining the first standard deviation as the device influence factor corresponding to each sensing device;

[0065] determining the weighting factor according to the first variance and a first preset formula;

[0066] obtaining a second standard deviation of the to-be-fused detection data of each sensing device at the same sampling time point, and determining the first standard deviation as the environmental influence factor corresponding to the sampling time point.

[0067] In a specific embodiment of the present application, when determining the weighting factor, the device influence factor and the environment influence factor, the first standard deviation of the to-be-fused detection data of each sensing device at a plurality of sampling time points is obtained, and the first variance is obtained, wherein the first standard deviation corresponds to the to-be-fused data of the same sensing device at different sampling time points, that is, the first standard deviation is only related to the device performance of the sensing device, and thus the first standard deviation can be determined as the device influence factor corresponding to each sensing device. Further, the weighting factor is determined according to the first preset formula and the first variance obtained above, wherein the first preset formula is:

[0068]

[0069] wherein W i is the weighting factor corresponding to each sensing device, i is the sensing device number, n is the total number of sensing devices, is the first variance.

[0070] The second standard deviation of the to-be-fused detection data of each sensing device at the same sampling time point is obtained, and the first standard deviation is determined as the environment influence factor corresponding to the sampling time point; wherein the second standard deviation corresponds to the to-be-fused data of different sensing devices at the same sampling time point, that is, the second standard deviation is mainly related to the environment at different sampling time points, and thus the second standard deviation is determined as the environment influence factor corresponding to the sampling time point.

[0071] Specifically, the method described above, after determining the device influence factor and the environment influence factor, the method further comprises:

[0072] According to the value range of the device influence factor and / or the environment influence factor, the step length corresponding to the device influence factor and the environment influence factor respectively in fractional order fusion is determined.

[0073] According to the amplitude-frequency characteristics of the fractional calculus transform, it can be known that the required calculation step and the value range of the influencing factor and the frequency of the detection signal are associated. When the fractional order and the value range of the influencing factor are fixed values, the fused detection data value is slightly improved with the decrease of the step, and the decrease of the step means that the number of steps required in the fusion process increases, the time required in the data fusion process becomes longer, and the timeliness becomes poor. Therefore, in this embodiment, after the device influencing factor and the environment influencing factor are determined, in order to ensure the timeliness in the subsequent fractional order fusion processing, the step corresponding to the device influencing factor and the environment influencing factor respectively in the fractional order fusion is determined according to the value range of the device influencing factor and / or the environment influencing factor, wherein the value range of the device influencing factor is determined according to the maximum value and the minimum value of the obtained device influencing factor, and the value range of the environment influencing factor is determined according to the maximum value and the minimum value of the obtained environment influencing factor. The value range of the device influencing factor and / or the environment influencing factor can ensure the timeliness and accuracy of the fused data.

[0074] In a specific embodiment, the step is preferably the difference between the upper and lower limits of the value range divided by the preset number of steps, wherein the preset number of steps is preferably 10 steps, or a number of steps close to 10 steps, such as 8 steps, 9 steps, 11 steps, etc.

[0075] In another embodiment, after the device influencing factor and the environment influencing factor are obtained, the device influencing factor and the environment influencing factor can be output and displayed, and the user input step or preset number of steps of the device influencing factor and / or the environment influencing factor can be received, and then the step corresponding to the device influencing factor and the environment influencing factor is determined according to the input step or preset number of steps.

[0076] It should be noted that the steps of the device influencing factor and the environment influencing factor are the same.

[0077] It should be further noted that the position of the last non-zero digit of the step (such as the tenth, hundredth, thousandth, etc.) is greater than or equal to the position of the last non-zero digit of the upper limit or lower limit of the value, for example, the position of the last non-zero digit of the step is the hundredth, and the position of the last non-zero digit of the upper limit of the value is the hundredth or the thousandth.

[0078] Referring to Figure 3 Further, the method described above, the fractional order fusion processing of the to-be-fused detection data according to the device influencing factor and the environment influencing factor, to obtain the preliminary fused data corresponding to each sensing device, comprises:

[0079] Step S301, obtaining the relationship function of the device influencing factor and the environment influencing factor and the to-be-fused detection data;

[0080] In step S302, the relationship function is determined according to the relationship function and the pre-acquired fractional order and step length, and a difference data fusion model based on fractional order partial differentiation of the to-be-fused detection data is determined.

[0081] In step S303, the to-be-fused detection data is subjected to fractional order fusion processing according to the difference data fusion model, and the preliminary fusion data is obtained.

[0082] In the embodiment, when the to-be-fused detection data is subjected to fractional order fusion processing according to the device influence factor and the environment influence factor, the relationship function between the device influence factor and the environment influence factor and the to-be-fused detection data is preferentially acquired, wherein a first function relationship between the to-be-fused detection data and the device influence factor and a second function relationship between the to-be-fused detection data and the environment influence factor are respectively acquired according to a first preset algorithm (for example, a Weighted Least Squares (WLS) fitting algorithm).

[0083] S(x) = a0 + a1x + a2x 2 + … + a n x n

[0084] The second function relationship is:

[0085] S(y) = b0 + b1y + b2y 2 + … + b m y m

[0086] wherein x is the device influence factor, y is the environment influence factor, n is the number of sensing devices, m is the number of sampling time points, and a and b are fitting parameters.

[0087] The relationship function is determined according to the first function relationship, the second function relationship, and the influence degree of the device influence factor and the environment influence factor on the to-be-fused detection data. Specifically, in a specific embodiment, the influence degree of the device influence factor and the environment influence factor on the to-be-fused detection data is the same, and the relationship function S(x, y) = z1S(x) + z2S(y) is obtained.

[0088] wherein S(x, y) is the to-be-fused detection data, z1 is the influence degree of the device influence factor on the to-be-fused detection data, z2 is the influence degree of the environment influence factor on the to-be-fused detection data, and z1 + z2 = 1.

[0089] After obtaining the relationship function between the device influence factor, the environment influence factor and the to-be-fused detection data, a difference data fusion model based on fractional order partial differential of the to-be-fused detection data is determined according to the relationship function and the pre-acquired fractional order and step length, wherein the fractional order is a fixed value, and in a specific embodiment, the fractional order is a fixed value such as 0.5; the step length is a fixed value or is determined according to the value range of the device influence factor and / or the environment influence factor, for example, the step length related to the device influence is 10, and the step length related to the environment influence is 8.

[0090] In a specific embodiment, z1 = z2 = 0.5.

[0091] In a specific embodiment, the determined difference data fusion model is:

[0092]

[0093] wherein:

[0094]

[0095] wherein, v is the fractional order, x is the device influence factor, y is the environment influence factor, h1 is the step length corresponding to the device influence factor, h2 is the step length corresponding to the environment influence factor, i is the device number, j is the sampling time point number, n is the number of devices, m is the number of sampling time points, Γ is the gamma function, D is the fractional order function, λ is a preset parameter, and S0 is the value of the to-be-fused detection data.

[0096] After obtaining the difference data fusion model, the fractional order fusion processing of each to-be-fused detection data can be performed according to the difference data fusion model to obtain the preliminary fusion data, and the specific process of the processing is not described herein.

[0097] It should be noted that the theory of fractional calculus and its application direction has become the research object of experts and scholars in many fields, and the theory of fractional calculus has been proposed for more than a hundred years, and until now it has not yet formed a rigorous and specific definition. The commonly used definitions at present are Caputo definition, Grunwald-Letnikov (G-L) definition and Riemann-Liouville (R-L) three types, among which the G-L definition is widely used in the field of engineering manufacturing due to its simple operation process and strong timeliness, so the G-L definition of fractional calculus is applied to study the differential detection data fusion algorithm in the application, and the fractional derivative is extended to two-dimensional space to accurately describe the factors affecting the accuracy of the monitoring data, and the fractional partial differential expression of the detection data to be fused is obtained. Among them, the machine tool running data collected by the Internet of Things is mainly affected by two factors of sensor equipment performance and production environment, assuming that the two influencing factors independently affect the detection data value, so according to the bounded variation BV V (Ω) space function, the variational model of the detection data fusion algorithm based on the fractional partial differential theory under the Internet of Things can be obtained, and further according to the fractional partial differential expression of the detection data to be fused, the differential data fusion algorithm model based on the fractional partial differential theory under the G-L definition can be obtained, that is, the differential data fusion model.

[0098] In order to facilitate the understanding of those skilled in the art, the following is illustrated by an actual application, in which the data type is the spindle power of a vertical lathe, and the sampling time is 1 min.

[0099] 1) Data acquisition

[0100] The spindle power is processed, and after data conversion, the commonly used denoising filter Butterworth low-pass filter is used for processing the collected signal, which can effectively filter out high-frequency noise in the signal. And the Butterworth low-pass filter is processed by fractional integral, so that the filtering algorithm is improved accordingly. The vertical lathe spindle power sampling data results (detection data to be fused) after denoising processing are shown in Table 1, in which ①-⑧ are various types of machine tools or various sensor devices, and NO.1-NO.10 are sampling time points.

[0101] Table 1 Vertical lathe spindle power sampling data table (%)

[0102]

[0103]

[0104] Among them, we use the total standard deviation of all detection data as a standard to measure the discreteness of the detection value, and the larger the standard deviation, the stronger the discreteness of the detection data.

[0105] 2) Impact factor and weighting factor determination

[0106] According to the data of the same machine tool at different sampling time points, the standard deviation is calculated and taken as the equipment impact factor x, which measures the influence degree of equipment factors; and according to the data of different machine tools at the same sampling time point, the standard deviation is calculated and taken as the environmental impact factor y, which measures the influence degree of environmental factors. From Table 1, the total standard deviation of the denoising processed detection data to be fused is 0.1299, which is used for the effect comparison after the subsequent fusion algorithm processing.

[0107] The variance of each machine tool sensor is calculated from the data values of the same machine tool at different sampling time points in Table 1, and the first preset formula is combined to obtain the machine tool sensor weighting factor data table shown in Table 2.

[0108] Table 2 Weighting factor data table

[0109]

[0110] Among them, the machine tool sensor weighting factor obtained in Table 2 is used for the final weighted data fusion processing, representing the reliability of the data obtained by each machine tool sensor.

[0111] 3) Difference data fusion model parameter determination

[0112] Firstly, the amplitude-frequency characteristic of the fractional order integral transformation acting on the input signal in the foregoing shows that when the order v is in the interval 0 < v < 1, the detection signal strength in the high frequency band will increase with the increase of v; but with the increase of frequency, the difference of the enhanced value of the detection signal by the change of the fractional order decreases. This embodiment explores the application effect of the fractional order integral fusion algorithm in the machine tool difference operation data fusion processing when the intermediate value of the fractional order v = [0, 1] is 0.5.

[0113] Secondly, according to the amplitude-frequency characteristics of the fractional calculus transform, it can be known that the required calculation step and the value range of the influencing factors and the frequency of the detection signal are associated. When the fractional order and the value range of the influencing factors are fixed values, the fused detection data value is slightly improved with the decrease of the step, and the decrease of the step means that the number of steps required in the fusion process increases, the time required in the data fusion process becomes longer, and the timeliness becomes worse. Therefore, according to the value range of the influence factor x (characterizing the performance influence of the sensing device) [0.1325, 0.1421] and the value range of the influence factor y (characterizing the production environment) [0.0053, 0.0128] shown in Table 1, the fusion step h1 = h2 = 0.001 is taken by comprehensively considering the required time length of fusion and the fusion accuracy, and the calculation steps required in the fusion process corresponding to the device influence and the environmental influence are [0.1421-0.1325] / 0.001 = 10, [0.0128-0.0053] / 0.001 = 8, respectively.

[0114] According to the values of the influence factors x and y in Table 2 and the average value of the main shaft power, the following function relationship between the influence factors and the detection data to be fused S(x), S(y) can be obtained by the WLS fitting algorithm, and the mathematical relationship between the detection data to be fused S(x, y) and the influence factors x and y can be calculated.

[0115] For example:

[0116] S(x) = 374.488-8133.546x+59223.222x 2 -143684.489x 3

[0117] S(y) = 2.274-35.115y+7409.963y 2 -355157.178y 3

[0118] S(x, y) = 0.5S(x) + 0.5S(y)

[0119] The above determined parameters are substituted into the preset difference data fusion model, and the difference data fusion model corresponding to the row operation can be obtained.

[0120] 4) Fractional data fusion processing

[0121] The data in Table 1 is substituted into the difference data fusion model, and the fractional data fusion processing result shown in Table 3 can be obtained.

[0122] Table 3 Fractional data fusion processing result

[0123]

[0124]

[0125] From the fusion results shown in Table 3, we can see that the difference detection data (to be fused detection data) after the 0.5 order partial differential equation fusion processing, the data presents the following characteristics: 1) The data value is significantly increased. The average value of the data in Table 3 after the 0.5 order partial differential equation fusion processing is 13.484, which is 5.89 times of the average value before fusion 2.289, indicating that the detection signal intensity is significantly enhanced after the fractional order partial differential equation processing; 2) In order to facilitate the application effect of the algorithm, the fusion results of each sensor are divided by the amplification coefficient K = 5.89, and the detection data shown in Table 4 is obtained.

[0126] Table 4 Fractional order data fusion processing results after dividing by the amplification coefficient

[0127]

[0128] From Table 4, it can be seen that the total standard deviation of the preliminary fusion detection data after the fractional order data fusion processing is 0.0462, which is much smaller than the 0.1299 of the to-be-fused detection data, and the data accuracy is significantly improved, and the data dispersion is weakened.

[0129] The preliminary fusion detection data after the fractional order data fusion processing is further weighted and fused using the weighted factor data in Table 2, and the results are shown in Table 5.

[0130] Table 5 Data weighted fusion processing results

[0131]

[0132] From the results, it can be seen that the accuracy of the data is further improved after the weighted fusion processing, and the final standard deviation is 0.0299, which is smaller than 0.0462 in the fractional order fusion.

[0133] Through the analysis of the data shown in Table 5, it can be seen that after applying the weighted fractional order calculus fusion processing, the standard deviation between the detection values is 0.0299, which is much smaller than the 0.1299 before fusion and the 0.09 of the Kalman filter algorithm and the 0.11 of the average value method. According to the data shown in Table 5 and the comparison algorithm processing results, the distribution curves of various algorithm data fusion before and after fusion are drawn as shown in the figure. Figure 4 It can be seen that the detection data after fusion is randomly distributed around the measured true value, and the dispersion between the data is greatly reduced. It is shown that the weighted fractional order calculus algorithm proposed in this paper has high fusion processing accuracy for difference data. Referring to Figure 5 Another embodiment of the application also provides a control device, comprising:

[0134] The first processing module 501 is configured to acquire detection data to be fused of a plurality of sensing devices at a plurality of sampling time points, each of the sensing devices corresponding to a machine tool, and the sensing devices being configured to detect data of the same type;

[0135] The second processing module 502 is configured to determine a weighting factor, a device influence factor and an environment influence factor according to the detection data to be fused;

[0136] The third processing module 503 is configured to perform fractional order fusion processing on the detection data to be fused according to the device influence factor and the environment influence factor, to obtain preliminary fusion data corresponding to each of the sensing devices;

[0137] The fourth processing module 504 is configured to perform weighted fusion processing on the preliminary fusion data according to the weighting factor, to obtain target fusion data corresponding to each of the sampling time points.

[0138] Specifically, the device as described above, the first processing module comprises:

[0139] The first processing unit is configured to acquire original detection data of a plurality of sensing devices at a plurality of sampling time points;

[0140] The second processing unit is configured to perform filtering processing on the original detection data according to a preset filter to obtain the detection data to be fused, wherein the preset filter is a filter subjected to fractional order integral processing.

[0141] Further, the device as described above, the second processing module comprises:

[0142] The third processing unit is configured to acquire a first standard deviation and a first variance of the detection data to be fused of each of the sensing devices at a plurality of sampling time points, and determine the first standard deviation as the device influence factor corresponding to each of the sensing devices;

[0143] The fourth processing unit is configured to determine the weighting factor according to the first variance and a first preset formula;

[0144] The fifth processing unit is configured to acquire a second standard deviation of the detection data to be fused of each of the sensing devices at the same sampling time point, and determine the first standard deviation as the environment influence factor corresponding to the sampling time point.

[0145] Specifically, the device as described above, the second processing module further comprises:

[0146] The sixth processing unit is configured to determine a step corresponding to the device influence factor and the environment influence factor during fractional order fusion according to a value range of the device influence factor and / or the environment influence factor.

[0147] Further, the device as described above, the third processing module comprises:

[0148] A seventh processing unit is configured to acquire a relationship function between the device influence factor, the environment influence factor and the to-be-fused detection data;

[0149] An eighth processing unit is configured to determine a difference data fusion model based on fractional order partial differential of the to-be-fused detection data according to the relationship function and a pre-acquired fractional order and a step length;

[0150] A ninth processing unit is configured to perform fractional order fusion processing on the to-be-fused detection data according to the difference data fusion model to obtain the preliminary fusion data.

[0151] Specifically, the device as described above, the second processing unit comprises:

[0152] The first function relationship between the to-be-fused detection data and the device influence factor and the second function relationship between the to-be-fused detection data and the environment influence factor are respectively acquired according to a first preset algorithm;

[0153] The relationship function is determined according to the first function relationship, the second function relationship and the influence degree of the device influence factor and the environment influence factor on the to-be-fused detection data.

[0154] The device embodiment of the present application is corresponding to the device embodiment of the above-mentioned difference data fusion method based on weighted fractional order calculus, all the implementation means in the method embodiment are applicable to the system embodiment, and the same technical effects can also be achieved.

[0155] Still another embodiment of the present application further provides a terminal device, comprising a processor, a memory and a computer program stored in the memory and executable on the processor, when the computer program is executed by the processor, the steps of the difference data fusion method based on weighted fractional order calculus as described above are implemented.

[0156] Still another embodiment of the present application further provides a computer readable storage medium, the computer readable storage medium stores a computer program, when the computer program is executed by a processor, the steps of the difference data fusion method based on weighted fractional order calculus as described above are implemented. In addition, the reference numerals and / or letters can be repeated in different examples. Such repetition is for the purpose of simplification and clarity, and does not indicate the relationship between the various embodiments and / or settings discussed.

[0157] It is also need to point out that, in this article, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or sequence between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion.

[0158] The above is the preferred embodiment of the present application, it should be pointed out that, for those skilled in the art, without departing from the principles described in this application, can also make a number of improvements and refinements, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A weighted fractional calculus based dissimilarity data fusion method, characterized in that, The method comprises the following steps: obtaining detection data to be fused of a plurality of sensing devices at a plurality of sampling time points, each of the sensing devices corresponding to a machine tool, and the sensing devices being used to detect data of the same type; determining corresponding weighting factors, device influence factors and environment influence factors according to the detection data to be fused; performing fractional order fusion processing on the detection data to be fused according to the device influence factors and the environment influence factors to obtain preliminary fusion data corresponding to each of the sensing devices; performing weighted fusion processing on the preliminary fusion data according to the weighting factors to obtain target fusion data corresponding to each of the sampling time points; after the device influence factors and the environment influence factors are determined, the method further comprises the following steps: determining steps corresponding to the device influence factors and the environment influence factors respectively during fractional order fusion according to the value range of the device influence factors and / or the environment influence factors, the step being a value obtained by dividing the difference between the upper and lower limits of the value range by a preset number of steps.

2. The method of claim 1, wherein, The step of obtaining detection data to be fused of a plurality of sensing devices at a plurality of sampling time points comprises the following steps: obtaining original detection data of a plurality of the sensing devices at a plurality of the sampling time points; performing filtering processing on the original detection data according to a preset filter to obtain the detection data to be fused, wherein the preset filter is a filter that has been subjected to fractional order integral processing.

3. The method according to claim 1 or 2, characterized in that, The step of determining corresponding weighting factors, device influence factors and environment influence factors according to the detection data to be fused comprises the following steps: obtaining a first standard deviation and a first variance of the detection data to be fused of each of the sensing devices at a plurality of sampling time points respectively, and determining the first standard deviation as the device influence factor corresponding to each of the sensing devices; determining the weighting factors according to the first variance and a first preset formula; obtaining a second standard deviation of the detection data to be fused of each of the sensing devices at the same sampling time point, and determining the first standard deviation as the environment influence factor corresponding to the sampling time point.

4. The method of claim 1, wherein, The step of performing fractional order fusion processing on the detection data to be fused according to the device influence factors and the environment influence factors to obtain preliminary fusion data corresponding to each of the sensing devices comprises the following steps: obtaining a relationship function between the device influence factors, the environment influence factors and the detection data to be fused; determining a difference data fusion model based on fractional order partial differentiation of the detection data to be fused according to the relationship function and a fractional order and a step obtained in advance; performing fractional order fusion processing on the detection data to be fused according to the difference data fusion model to obtain the preliminary fusion data.

5. The method of claim 1, wherein, The step of obtaining a relationship function between the device influence factors, the environment influence factors and the detection data to be fused comprises the following steps: obtaining a first functional relationship between the detection data to be fused and the device influence factors and a second functional relationship between the detection data to be fused and the environment influence factors according to a first preset algorithm respectively; According to the first function relationship, the second function relationship, and the influence degree of the device influence factor and the environment influence factor on the to-be-fused detection data, the relationship function is determined.

6. A control device characterized by comprising: Comprise: The first processing module is used for acquiring to-be-fused detection data of a plurality of sensing devices at a plurality of sampling time points, each of the sensing devices corresponds to a machine tool, and the sensing device is used for detecting the same type of data; The second processing module is used for determining corresponding weighting factors, device influence factors and environment influence factors according to the to-be-fused detection data; The third processing module is used for performing fractional order fusion processing on the to-be-fused detection data according to the device influence factor and the environment influence factor, to obtain preliminary fusion data corresponding to each sensing device; The fourth processing module is used for performing weighted fusion processing on the preliminary fusion data according to the weighting factor, to obtain target fusion data corresponding to each sampling time point; The second processing module further comprises: The sixth processing unit is used for determining the step length corresponding to the device influence factor and the environment influence factor respectively during fractional order fusion according to the value range of the device influence factor and / or the environment influence factor, the step length being the difference between the upper and lower limits of the value range divided by the preset number of steps.

7. A terminal device, characterized by comprising: The computer program is stored on the computer readable storage medium and is executed by the processor to realize the steps of the weighted fractional calculus-based differential data fusion method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium and is executed by the processor to realize the steps of the weighted fractional calculus-based differential data fusion method according to any one of claims 1 to 5.

Images