Operation impact factor extraction device

Abstract

To provide a work influence factor extraction method for extracting a work influence factor that has influence on stabilization of work, and an apparatus for the method.SOLUTION: A work influence factor extraction method includes: acquiring data of multiple measurement items which are different from each other and related to mental conditions that first and second technicians in first and second groups have during a work, the groups being formed by classifying multiple technicians; determining the presence of significant difference between the mental conditions of the first and second technicians, by using Mahalanobis-Taguchi system of a unit space defined with the multiple measurement items, on the basis of the data of the first and second technicians; and extracting, when the presence of the significant difference is determined, a work influence factor from among the measurement items, using Mahalanobis distance of the Mahalanobis-Taguchi system, on the basis of the data of the first and second technicians.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The present invention relates to a work impact factor extraction device that extracts work impact factors that affect the stabilization of a predetermined work for handling a predetermined object. Make

Background Art

[0002] Skills are difficult to verbalize, document, and illustrate because they are actually acquired through repeated practice and experience. Therefore, skills are difficult to convey to others. As a technology related to such skills, for example, Patent Document 1 discloses a skill acquisition support system that detects differences between the actions of a user and those of a model.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, generally, for the stabilization of a certain predetermined quality, the stabilization of work is desired. However, as described above, since skills are difficult to convey to others, it is difficult to extract factors (work impact factors) that affect the stabilization of the work, and the extraction of the work impact factors is desired. In particular, when the work impact factor is a factor related to the mental state of the skilled worker who performs the work, since the mental state of the skilled worker is internal to the skilled worker, its extraction is more difficult.

[0005] The present invention has been made in view of the above circumstances, and an object thereof is to provide a work impact factor extraction device that can extract work impact factors that affect the stabilization of the work. Make

Means for Solving the Problems

[0006] ​​As a result of various studies, the present inventor has found that the above object can be achieved by the following present invention. That is, a method for extracting work influencing factors according to an aspect of the present invention includes a first skilled person in a first group and a second skilled person in a second group who are divided from a plurality of skilled persons who perform a predetermined work of handling a predetermined object, respectively. When each of them performs the predetermined work, a data acquisition step of acquiring data of a plurality of different measurement items related to the mental states of the first skilled person in the first group and the second skilled person in the second group, respectively; based on the data of the first skilled person and the data of the second skilled person acquired in the data acquisition step, by using the Mahalanobis distance system in the unit space defined by the plurality of measurement items, a significant difference determination step of determining whether there is a significant difference between the mental state of the first skilled person and the mental state of the second skilled person; when it is determined in the significant difference determination step that there is a significant difference, based on the data of the first skilled person and the data of the second skilled person acquired in the data acquisition step, by using the Mahalanobis distance of the Mahalanobis distance system, a factor extraction step of extracting a work influencing factor that affects the stabilization of the predetermined work from among the plurality of measurement items. Preferably, in the above method for extracting work influencing factors, the significant difference determination step includes a unit space generation step of generating unit space data based on one of the data of the first skilled person and the data of the second skilled person acquired by the data acquisition step in the unit space; a distance calculation step of obtaining the Mahalanobis distance with respect to the unit space data generated in the unit space generation step based on the other of the data of the first skilled person and the data of the second skilled person acquired by the data acquisition step; and a determination step of determining whether there is the significant difference based on the Mahalanobis distance obtained in the distance calculation step. Preferably, in the above method for extracting work influencing factors, the factor extraction step includes an SN ratio calculation step of obtaining the SN ratio of the maximization characteristic of the Mahalanobis distance by a two-level orthogonal array for the plurality of measurement items based on the data of the first skilled person and the data of the second skilled person acquired in the data acquisition step; and an extraction step of extracting the work influencing factor from among the plurality of measurement items based on the SN ratio obtained in the SN ratio calculation step.

[0007] Such a method for extracting work influence factors can extract work influence factors because it determines whether there is a significant difference between the data of the first skilled person and the data of the second skilled person, and when there is such a significant difference, it uses the data of the first skilled person and the data of the second skilled person with the significant difference.

[0008] In another aspect, in the above-mentioned method for extracting work influence factors, the plurality of measurement items are measurement items obtained from electrocardiogram data, electroencephalogram data, and eye gaze data.

[0009] Such a method for extracting work influence factors can preferably extract work influence factors related to the mental state of the skilled person because the plurality of measurement items are measurement items obtained from electrocardiogram data, electroencephalogram data, and eye gaze data.

[0010] In another aspect, in the above-mentioned method for extracting work influence factors, the significant difference determination step includes a unit space generation step of generating unit space data based on one of the data of the first skilled person and the data of the second skilled person obtained in the data acquisition step in the unit space, a distance calculation step of obtaining the Mahalanobis distance with respect to the unit space data generated in the unit space generation step based on the other of the data of the first skilled person and the data of the second skilled person obtained in the data acquisition step, and a determination step of determining whether there is the significant difference based on the Mahalanobis distance obtained in the distance calculation step. The one data is the data of the skilled person classified into the highest class among the plurality of classes when the plurality of skilled persons are classified into a plurality of classes based on the accuracy of the work and the work time from the start to the end of the work. The unit space generation step generates unit space data based on the data in the one data for which the Mahalanobis distance converges.

[0011] Skilled workers classified into the highest level can usually perform their work more steadily and reliably. Since the above-mentioned work influence factor extraction method generates unit space data using data of skilled workers classified into the highest level who can perform their work more steadily and reliably, more appropriate work influence factors can be extracted. Since the above-mentioned work influence factor extraction method generates unit space data based on the data of skilled workers classified into the highest level before the Mahalanobis distance converges, the unit space data serving as the evaluation criterion can be generated more appropriately.

[0012] In another aspect, in the above-mentioned work influence factor extraction method, the plurality of measurement items at least include β waves and δ waves of electroencephalogram.

[0013] Since such a work influence factor extraction method has the plurality of measurement items at least including β waves and δ waves of electroencephalogram, it can be determined whether β waves and δ waves can be work influence factors.

[0014] In another aspect, in the above-mentioned work influence factor extraction method, the predetermined work is work using tools.

[0015] Such a work influence factor extraction method can extract work influence factors in work using tools.

[0016] In another aspect, in the above-mentioned work influence factor extraction method, the predetermined work is work of grinding or polishing a mold.

[0017] Such a work influence factor extraction method can extract work influence factors in work of grinding or polishing a mold.

[0018] In another aspect, in the above-mentioned work influence factor extraction method, the mold is a mold for molding a member used in a vehicle, i.e., a mold for vehicle member molding.

[0019] Such a work influence factor extraction method can extract work influence factors in work of grinding or polishing a mold for vehicle member molding.

[0020] The working influence factor extraction device according to another aspect of the present invention includes a data acquisition unit that acquires data of a plurality of measurement items related to the respective mental states of a first skilled worker in a first group and a second skilled worker in a second group, who are separated from a plurality of skilled workers who perform a predetermined work of handling a predetermined object, when each of them performs the predetermined work; a significant difference determination unit that determines whether there is a significant difference between the mental state of the first skilled worker and the mental state of the second skilled worker by using the Mahalanobis distance system in the unit space defined by the plurality of measurement items, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit; and a factor extraction unit that extracts a working influence factor that affects the stabilization of the predetermined work from among the plurality of measurement items by using the Mahalanobis distance of the Mahalanobis distance system, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit, when it is determined by the significant difference determination unit that there is a significant difference.

[0021] Such a working influence factor extraction device can extract a working influence factor because it determines whether there is a significant difference between the data of the first skilled worker and the data of the second skilled worker, and when there is such a significant difference, it uses the data of the first skilled worker and the data of the second skilled worker having the significant difference.

Effect of the Invention

[0022] The Make working influence factor extraction device according to the present invention can extract a working influence factor.

Brief Description of the Drawings

[0023]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Mode for Carrying Out the Invention

[0024] Hereinafter, one or more embodiments of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to the disclosed embodiments. In each figure, components denoted by the same reference numerals indicate the same components, and the description thereof will be omitted as appropriate. In this specification, when referring generically, reference numerals without subscripts are used, and when referring to individual components, reference numerals with subscripts are used.

[0025] The method for extracting work influencing factors in the embodiment is a method for extracting factors (work influencing factors) that affect the stabilization of a predetermined work for handling a predetermined object. The method includes: a data acquisition step of acquiring data of a plurality of measurement items related to the respective mental states of the first skilled workers in the first group and the second skilled workers in the second group, where the first group of first skilled workers and the second group of second skilled workers, who are divided from a plurality of skilled workers performing the predetermined work, each perform the predetermined work; a significant difference determination step of determining whether there is a significant difference between the mental state of the first skilled workers and the mental state of the second skilled workers by using the Mahalanobis-Taguchi system in the unit space defined by the plurality of measurement items, based on the data of the first skilled workers and the data of the second skilled workers acquired in the data acquisition step; and a factor extraction step of extracting work influencing factors that affect the stabilization of the predetermined work from among the plurality of measurement items by using the Mahalanobis distance of the Mahalanobis-Taguchi system, based on the data of the first skilled workers and the data of the second skilled workers acquired in the data acquisition step, when it is determined in the significant difference determination step that there is a significant difference. Hereinafter, a work influence extraction device implementing such a work influencing factor extraction method will be described more specifically as an example.

[0026] FIG. 1 is a block diagram showing the configuration of a work influence factor extraction device according to an embodiment. FIG. 2 is a diagram showing, as an example, the measurement state of electrocardiogram data and the electrocardiogram data. FIG. 2A shows a state in which electrodes of an electrocardiograph are attached to a skilled person to measure electrocardiogram data, and FIG. 2B shows an example of the electrocardiogram data. The horizontal axis of FIG. 2B is the elapsed time (Time), and the vertical axis thereof is the electric potential. FIG. 3 is a diagram showing, as an example, the measurement state of electroencephalogram data (electroencephalic potential data) and the electroencephalogram data. FIG. 3A shows a state in which electrodes of an electroencephalograph are attached to a skilled person to measure electroencephalogram data, FIG. 3B shows an example of the electroencephalogram data, and FIG. 3C shows alpha waves, beta waves, theta waves, and delta waves obtained from the electroencephalogram data. The horizontal axis of FIG. 3B is the elapsed time (Time), and the vertical axis thereof is the electric potential. The horizontal axis of FIG. 3C is the elapsed time (Time), and the vertical axis thereof is the content rate. FIG. 4 is a diagram showing, as an example, an eye tracker for measuring gaze data.

[0027] In the embodiment, the work influence factor extraction device A includes, for example, as shown in FIG. 1, a data acquisition unit 1, a control processing unit 2, an input unit 3, an output unit 4, an interface unit (IF unit) 5, and a storage unit 6.

[0028]

[0029] The above-mentioned work can be any work as long as it is a work performed by physical movements. From the perspective of skill inheritance, it may include not only work such as manufacturing work of articles and transporting work of articles, but also arts, sports, etc. In the present embodiment, the work is preferably a work using a tool, and more specifically, it is a work of grinding or polishing a mold for molding a member used in a vehicle, which is a vehicle member molding die. In this case, the tool is, for example, a hand grinder in one example. Of course, it is not limited to this. For example, the work may be a welding work, and the tool may be a welding machine such as a torch, or, for example, the work may be a painting work, and the tool may be a painting tool such as a spray gun. Preferably, the work is a work related to the manufacture of a vehicle, and the tool is a tool used in the work related to the manufacture of the vehicle.

[0030] The grouping of a plurality of skilled workers into a plurality of groups is preferably carried out from the perspective of work stability in order to extract work influencing factors. For example, usually, the more work experience one has, the higher the work stability is considered to be. Therefore, the grouping may be carried out from the perspective of the amount of work experience. For example, the grouping may be carried out according to the working hours or the number of work pieces. Or, for example, usually, the higher the work accuracy (such as the dimensional accuracy of the work product, etc.), the higher the work stability is considered to be. Therefore, the grouping may be carried out from the perspective of work accuracy. In the present embodiment, the grouping is carried out by classifying the plurality of skilled workers into a plurality of classes based on the work accuracy and the working hours from the start to the end of the work. Such grouping is disclosed, for example, in Japanese Patent Laid-Open No. 2021-015225.

[0031] The plurality of different measurement items may be arbitrary as long as they are related to the mental state of a skilled person. For example, the plurality of measurement items are measurement items obtained from electrocardiogram data, electroencephalogram data, and gaze data. More specifically, the plurality of measurement items include the heart rate, the indicator of activity of the autonomic nerve, the indicator of activity of the parasympathetic nerve, and the indicator of activity of the sympathetic nerve obtained from electrocardiogram data. An electrocardiogram due to one heartbeat of a healthy person generally consists of a P wave, a QRS wave, and a T wave. The heart rate is the number of peaks R of the QRS wave observed in one minute. As the indicator of activity of the autonomic nerve, so-called RRSD, which is the coefficient of variation of the intervals between continuously adjacent peaks R, is used. As the indicator of activity of the parasympathetic nerve, so-called RMSSD, which is the square root of the average value of the squares of the differences between continuously adjacent peaks R, is used. As the indicator of activity of the sympathetic nerve, so-called LF / HF, which is the ratio of the power spectrum (LF) in the low frequency band (0.04 to 0.15 [Hz]) to the power spectrum (HF) in the high frequency band (0.15 to 0.4 [Hz]) when the heartbeats are frequency-analyzed, is used. The plurality of measurement items are the alpha wave (8 to 13 [Hz]), the beta wave (13 to [Hz] (above 13 Hz)), the theta wave (4 to 8 [Hz]), and the delta wave (~4 [Hz] (below 4 Hz)) obtained from electroencephalogram data. The alpha wave, the beta wave, the theta wave, and the delta wave are each represented by its content rate ([%]). The plurality of measurement items include the gaze data itself obtained from the gaze data. The gaze data is represented by the dispersion of the viewpoints.

[0032] The data acquisition unit 1 may be a measuring instrument that measures the measurement items. For example, the data acquisition unit 1 may be an electrocardiograph that measures electrocardiogram data, an electroencephalograph that measures brain waves, and a gaze meter that measures gaze data. In one example, for the electrocardiograph and the electroencephalograph, an organized measurement device intercross-413 manufactured by Intercross Co., Ltd. is used. As shown in FIG. 2A, two electrodes Pb1 and Pb2 for the electrocardiograph are attached to the chest around the heart of the skilled person Ob at intervals, and the electrocardiogram data shown in FIG. 2B is acquired. A headset BM having three electrodes for the electroencephalograph is attached to the head HD of the skilled person Ob as shown in FIG. 3A, and the electroencephalogram data shown in FIG. 3B is acquired. By performing a fast Fourier transform on the electroencephalogram data shown in FIG. 3B, the content ratio [%] of each wave is obtained, and thereby, the alpha wave, beta wave, theta wave, and delta wave shown in FIG. 3C are obtained. In one example, for the gaze meter, Tobii Pro Glass 2 (EM) manufactured by Tobii shown in FIG. 4 is used, and the variance is obtained from the coordinate values of the viewpoints (the fixation points of the skilled person Ob) obtained by converting the measured gaze data into an XYZ orthogonal coordinate system. In the measurement of these electrocardiogram data, electroencephalogram data, and gaze data, each data is measured by sampling at a predetermined sampling interval from the start to the end of the operation. In this embodiment, the average value during the operation time is finally used as the data of the measurement item.

[0033] Alternatively, for example, the data acquisition unit 1 may be an input unit that inputs each data of each measurement item obtained from each measurement result or each measurement result measured by such a measuring instrument. In this case, the data acquisition unit 1 may be used in combination with the input unit 3 described later.

[0034] Alternatively, for example, the data acquisition unit 1 may be an interface circuit for inputting and outputting data to and from an external device. In this case, the external device is a storage medium that stores data of each measurement item obtained from each measurement result or each measurement value. The storage medium is, for example, a USB (Universal Serial Bus) memory, an SD card (registered trademark), or the like. Alternatively, for example, the data acquisition unit 1 may be a drive device that reads data from a recording medium that records data of each measurement item obtained from each measurement result or each measurement value. In this case, the recording medium is, for example, a CD-ROM (Compact Disc Read Only Memory), a CD-R (Compact Disc Recordable), a DVD-ROM (Digital Versatile Disc Read Only Memory), a DVD-R (Digital Versatile Disc Recordable), or the like. Alternatively, for example, the data acquisition unit 1 may be a communication interface circuit that transmits and receives communication signals to and from an external device. In this case, the external device is a server device that is connected to the communication interface circuit via a network (such as a WAN (Wide Area Network, including a public communication network) or a LAN (Local Area Network)) and manages data of each measurement item obtained from each measurement result or each measurement value. When the data acquisition unit 1 is the above-described interface circuit or communication interface circuit, it may be shared with the IF unit 5 described later.

[0035] The input unit 3 is connected to the control processing unit 2, and is a device that inputs various commands such as a command for instructing the start of extraction of work influencing factors and various data necessary for operating the work influencing factor extraction device A such as the work name into the work influencing factor extraction device A. For example, it is a plurality of input switches, a keyboard, a mouse, etc. to which a predetermined function is assigned. The output unit 4 is connected to the control processing unit 2, and is a device that outputs commands, data input from the input unit 3, extracted work influencing factors, etc. according to the control of the control processing unit 2. For example, it is a display device such as a CRT display, a liquid crystal display, and an organic EL display, a printing device such as a printer, etc.

[0036] Note that a so-called touch panel may be configured from the input unit 3 and the output unit 4. When configuring this touch panel, the input unit 3 is a position input device that detects and inputs an operation position such as a resistive film method or a capacitance method, and the output unit 4 is a display device. In this touch panel, the position input device is provided on the display surface of the display device, candidates for one or more input contents that can be input to the display device are displayed, and when the user touches the display position where the input content to be input is displayed, the position is detected by the position input device, and the display content displayed at the detected position is input to the work influencing factor extraction device A as the user's operation input content. In such a touch panel, since the user can easily understand the input operation intuitively, a work influencing factor extraction device A that is easy for the user to handle is provided.

[0037] The IF unit 5 is a circuit that is connected to the control processing unit 2 and performs data input / output with external devices according to the control of the control processing unit 2. For example, it is an interface circuit of RS-232C using a serial communication method, an interface circuit using the Bluetooth (registered trademark) standard, an interface circuit for performing infrared communication such as the IrDA (Infrared Data Association) standard, and an interface circuit using the USB (Universal Serial Bus) standard. Further, the IF unit 5 is a circuit that communicates with external devices, and for example, it may be a data communication card or a communication interface circuit according to the IEEE802.11 standard or the like.

[0038] The memory unit 6 is a circuit connected to the control processing unit 2 and stores various predetermined programs and various predetermined data according to the control of the control processing unit 2. The various predetermined programs include, for example, a control processing program. The control processing program includes, for example, a control program for controlling each part 1, 3 to 6 of the work influence factor extraction device A, and based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1, by using the Mahalanobis distance system of the unit space defined by the plurality of measurement items, a significant difference determination program for determining whether there is a significant difference between the mental state of the first skilled worker and the mental state of the second skilled worker, and when it is determined that there is a significant difference in the significant difference determination program, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1, by using the Mahalanobis distance of the Mahalanobis distance system, a factor extraction program for extracting a work influence factor that affects the stabilization of the predetermined work from among the plurality of measurement items, etc. are included. The various predetermined data include, for example, data necessary for executing these programs. Such a memory unit 6 includes, for example, a ROM (Read Only Memory) which is a non-volatile memory element, an EEPROM (Electrically Erasable Programmable Read Only Memory) which is a rewritable non-volatile memory element, etc. And the memory unit 6 includes a RAM (Random Access Memory) etc. which serves as a working memory of the so-called control processing unit 2 for storing data etc. generated during the execution of the predetermined program. The memory unit 6 may be configured to include a hard disk device with a relatively large storage capacity.

[0039] The control processing unit 2 controls each of the parts 1, 3 to 6 of the work impact factor extraction device A according to the functions of the respective parts, and based on the data of the first skilled worker and the data of the second skilled worker, by using the Mahalanobis-Taguchi system, it is a circuit for extracting work impact factors that affect the stabilization of the predetermined work from among a plurality of different measurement items related to the mental state. The control processing unit 2 is configured to include, for example, a CPU (Central Processing Unit) and its peripheral circuits. The control processing unit 2 functionally includes a control unit 21, a significant difference determination unit 22, and a factor extraction unit 23 when the control processing program is executed.

[0040] The control unit 21 controls each of the parts 1, 3 to 6 of the work impact factor extraction device A according to the functions of the respective parts, and is in charge of the overall control of the work impact factor extraction device A. The control unit 21 stores each data of the plurality of measurement items acquired by the data acquisition unit 1 in the storage unit 6.

[0041] The significant difference determination unit 22 determines whether or not there is a significant difference between the mental state of the first skilled worker and the mental state of the second skilled worker by using the Mahalanobis-Taguchi system in the unit space defined by the plurality of measurement items based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1. The Mahalanobis-Taguchi system is a method of information processing for judging the approximation degree (dissociation degree) of target data based on the Mahalanobis distance considering the correlation between each item constituting the multi-dimension with respect to the data composed of multi-dimensional information with reference to the unit space. More specifically, the significant difference determination unit 22 includes a unit space generation unit 221, a distance calculation unit 222, and a determination unit 223.

[0042] The unit space generation unit 221 generates unit space data based on one of the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1 in the unit space of the Mahalanobis distance system. In the present embodiment, the one data is the data of the skilled worker classified into the highest class among the plurality of classes when classifying the plurality of skilled workers into a plurality of classes based on the accuracy of the work and the work time from the start to the end of the work, and the unit space generation unit 221 generates unit space data based on the data in which the Mahalanobis distance converges among the one data. Preferably, the unit space generation unit 221 generates unit space data based on the data in which the Mahalanobis distance converges near 1 among the one data. Converging near 1 means within a predetermined range according to the class interval (interval between classes) including 1. For example, when the class of the Mahalanobis distance in the frequency distribution of the Mahalanobis distance is 0.5 intervals, it converges within a range less than 2 (classes less than 2), and more specifically, it converges within a range of 1 to 1.5 (each class of 1 and 1.5), or, for example, when the class of the Mahalanobis distance in the frequency distribution of the Mahalanobis distance is 1 interval, it converges to 1 (class of 1).

[0043] The distance calculation unit 222 obtains the Mahalanobis distance with respect to the unit space data generated by the unit space generation unit 221 based on the other data of the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1.

[0044] The determination unit 223 determines whether or not there is a significant difference based on the Mahalanobis distance obtained by the distance calculation unit 222. More specifically, the determination unit 223 obtains the frequency distribution of the Mahalanobis distance of the unit space data and the Mahalanobis distance of the other data, and determines whether or not there is a significant difference based on whether or not there is a difference between the frequency distribution of the Mahalanobis distance of the unit space data and the frequency distribution of the Mahalanobis distance of the other data.

[0045] When it is determined by the determination unit 223 of the significant difference determination unit 22 that there is a significant difference, the factor extraction unit 23 extracts, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit, the working influence factors that affect the stabilization of the predetermined work from among the plurality of measurement items by using the Mahalanobis distance of the Mahalanobis system. More specifically, the factor extraction unit 23 functionally includes an SN ratio calculation unit 231 and an extraction unit 232.

[0046] The SN ratio calculation unit 231 obtains the SN ratio of the maximization characteristic of the Mahalanobis distance according to the two-level orthogonal array for the plurality of measurement items, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1. More specifically, the SN ratio calculation unit 231, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1, for each of the plurality of measurement items, obtains the SN ratio of the maximization characteristic of the Mahalanobis distance when using the measurement item (first level) and the SN ratio of the maximization characteristic of the Mahalanobis distance when not using the measurement item (second level) as the first SN ratio and the second SN ratio, and obtains the first SN ratio and the second SN ratio for each of the other data.

[0047] The extraction unit 232 extracts the working influence factors from among the plurality of measurement items based on the SN ratio obtained by the SN ratio calculation unit 231. More specifically, the extraction unit 232 counts, for each of the other data, the number of the other data in which the second SN ratio is smaller than the first SN ratio for each of the plurality of measurement items in the first SN ratio and the second SN ratio obtained for each of the other data, and extracts, as the working influence factors, a predetermined number from among the plurality of measurement items starting from the one with the larger counted number.

[0048] These control processing unit 2, input unit 3, output unit 4, IF unit 5, and storage unit 6 can be configured by, for example, a computer such as a desktop type, notebook type, or tablet type.

[0049] Next, the operation of this embodiment will be described. FIG. 5 is a flowchart showing the operation of the work factor extraction device.

[0050] When the power of the work influence factor extraction device A configured as described above is turned on, it executes initialization of necessary parts and starts its operation. In the control processing unit 2, a control unit 21, a significant difference determination unit 22, and a factor extraction unit 23 are functionally configured by executing its control processing program. In the significant difference determination unit 22, a unit space generation unit 221, a distance calculation unit 222, and a determination unit 223 are functionally configured. In the factor extraction unit 23, an SN ratio calculation unit 231 and an extraction unit 232 are functionally configured.

[0051] When an operator (user) inputs an instruction for work influence factor extraction from the input unit 3, in FIG. 5, the work influence factor extraction device A first acquires, by the control unit 21 of the control processing unit 2, each data of a plurality of measurement items for each of the first skilled workers in the first group and the second skilled workers in the second group by the data acquisition unit 1, and stores it in the storage unit 6 (S1).

[0052] Next, the work influence factor extraction device A generates unit space data by the unit space generation unit 221 in the significant difference determination unit 22 of the control processing unit 2 based on one of the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1 in the unit space of the Mahalanobis distance system in process S1 (S2). In this embodiment, the one data is the data of the first skilled worker classified into the highest class among the plurality of classes when classifying the plurality of skilled workers into a plurality of classes based on the accuracy of the work and the work time from the start to the end of the work. The unit space generation unit 221 selects, as the unit space data, the data for which the Mahalanobis distance converges among the data of the first skilled worker.

[0053] Next, the work influence factor extraction device A, by the distance calculation unit 222 in the significant difference determination unit 22 of the control processing unit 2, based on the data of the other party among the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1, that is, for the data of the second skilled worker, obtains the Mahalanobis distance with respect to the unit space data generated by the unit space generation unit 221 in process S2 (S3).

[0054] Next, the work influence factor extraction device A determines whether there is a significant difference based on the Mahalanobis distance obtained by the distance calculation unit 222 by the determination unit 223 in the significant difference determination unit 22 of the control processing unit 2 (S4). As a result of this determination, if there is no significant difference, the work influence factor extraction device A returns the process to process S1. That is, the process is restarted from the beginning with new data. On the other hand, as a result of the determination, if there is a significant difference, the work influence factor extraction device A, by the SN ratio calculation unit 231 in the factor extraction unit 23 of the control processing unit 2, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1, obtains the SN ratio of the maximization characteristic of the Mahalanobis distance according to the two-level orthogonal array for the plurality of measurement items (S5). More specifically, the SN ratio calculation unit 231, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit 1 in process S1, for each of the plurality of measurement items, obtains the SN ratio of the maximization characteristic of the Mahalanobis distance when using the measurement item and the SN ratio of the maximization characteristic of the Mahalanobis distance when not using the measurement item as the first SN ratio and the second SN ratio, and obtains the first SN ratio and the second SN ratio for each of the data of the other party, in this example, the data of the second skilled worker.

[0055] Subsequent to process S5, the work influence factor extraction device A extracts the work influence factor from among the plurality of measurement items based on the signal-to-noise ratio obtained by the signal-to-noise ratio calculation unit 231 by the extraction unit 232 in the factor extraction unit 23 of the control processing unit 2 (S6). More specifically, the extraction unit 232 counts, for each of the plurality of measurement items, the number of the other data in which the second signal-to-noise ratio is smaller than the first signal-to-noise ratio in the first signal-to-noise ratio and the second signal-to-noise ratio obtained for each of the other data, and extracts, as the work influence factor, a predetermined number from among the plurality of measurement items starting from the one with the larger counted number.

[0056] Subsequent to process S6, the work influence factor extraction device A outputs, by the control processing unit 2, the work influence factor extracted in process S6 to the output unit 4 and ends this process (S7). Incidentally, if necessary, the work influence factor extracted in process S6 may be output from the IF unit 5 to an external device.

[0057] Findings and verification regarding the ability of the work influence factor extraction device A to extract the work influence factor by such a configuration and operation will be described below.

[0058] FIG. 6 is a diagram for explaining a test piece used in the work. FIG. 6A is a perspective view schematically showing the test piece before the work, and FIG. 6B is a perspective view schematically showing the test piece after the work. FIG. 7 is a diagram showing the frequency distribution of the Mahalanobis distance in the unit space data and the measurement data. The horizontal axis of FIG. 7 is the Mahalanobis distance, and the vertical axis thereof is the frequency. FIG. 8 is, as an example, a two-level orthogonal array L 12It is a diagram showing the signal-to-noise ratio of the target characteristics according to []. Each horizontal axis in FIGS. 8A to 8D is a measurement item, and each vertical axis in these figures is the signal-to-noise ratio of the target characteristics. A to I are measurement items 1 to 9, respectively. For example, A is measurement item 1 (= heart rate), and B is measurement item 2 (= RRSD). Note that the illustration of each of the remaining 10 results excluding the above 4 results out of the 14 measurement data is omitted. FIG. 9 is a diagram showing the distribution of work influencing factors as an example. The horizontal axis of FIG. 9 is the Mahalanobis distance, and the vertical axis is the content rate of each electroencephalogram. FIG. 10 is a diagram for explaining the results when measures for skill improvement are implemented. FIG. 10A shows the case of beta waves, where the horizontal axis is the Mahalanobis distance and the vertical axis is the content rate of beta waves. FIG. 10B shows the case of delta waves, where the horizontal axis is the Mahalanobis distance and the vertical axis is the content rate of delta waves.

[0059] In this finding and verification, the above work is the above-mentioned die grinding or polishing work (grinder work) using a grinder as the above tool. More specifically, this grinder work includes two steps: a roughing process of roughly cutting the die and a finishing process of polishing the roughly cut surface to a smooth surface. However, in order to improve the cutting accuracy (dimensional accuracy and machining accuracy of the work target) by the grinder work, it is usually important to perform the roughing process work with high precision and high efficiency. For this reason, a test piece TP modeled after the die is prepared, and a skilled worker performs the roughing process work in the grinder work on the test piece TP, and the work time and data of each measurement item are acquired from the start to the end of this work. Then, the cutting accuracy of the test piece TP after the grinder work was measured with a three-dimensional shape measuring machine.

[0060] This test piece TP is, for example, as shown in Fig. 5A, a 100×100×35 [mm] SS material (rolled steel for general structure) in which a concave strip TPb extending in a strip shape with a width of 20 [mm] in one direction and a depth of 0.1 [mm] is formed by machining in the central part. For this test piece TP, the roughing process operation is an operation of uniformly grinding, for example, one of the convex parts TPa and TPc with a height of 0.1 mm formed on each side of the concave strip TPb by 0.1 mm using a handy-type grinder. After such roughing process operation, as shown in Fig. 5B, a grinding plane part TPd substantially flush with the bottom surface of the concave strip TPb is formed on the test piece TP.

[0061] The plurality of measurement items are the heart rate (measurement item 1, A), RRSD (measurement item 2, B), RMSSD (measurement item 3, C), LF / HF (measurement item 4, D), beta wave (measurement item 5, E), alpha wave (measurement item 6, F), theta wave (measurement item 7, G), delta wave (measurement item 8, H), and viewpoint dispersion (measurement item 9, I) described above.

[0062] The skilled workers are those belonging to the top class pre-classified by the method disclosed in Japanese Patent Application Laid-Open No. 2021-015225 and those belonging to the class next to the top class, based on the working time and cutting accuracy in such grinder work. In the method disclosed in Japanese Patent Application Laid-Open No. 2021-015225, skilled workers are classified into five classes (grades): beginners (Class A), intermediate-level workers (Class B), semi-advanced workers (Class C), advanced workers (Class D), and top-level workers (Class E). Here, the skilled workers belonging to the top class (top-level workers) are four, regarded as the first skilled workers in the first group, and the next class (advanced workers) are four, regarded as the second skilled workers in the second group. In grinder work, even for the same skilled worker, the working time and the accuracy of the work may vary by about 10% from day to day. Since it is considered that the top-level workers and advanced workers have fully mastered the operations of grinder work, it is considered that the above-mentioned variation of about 10% occurred because the potential of the skilled worker could not be exerted due to the mental state. Thus, here, the top-level workers and advanced workers are selected as the first and second skilled workers.

[0063] These four first skilled workers and four second skilled workers each performed grinder work on the test piece TP three times. As described above, during each work, the working time and the data of each measurement item were acquired, and the cutting accuracy of the test piece TP after the grinder work was measured. Each data of each measurement item was measured multiple times at a predetermined sampling interval, and the average value of each numerical value measured from the start to the end of the work was taken as the final data of each measurement item.

[0064] Twelve data (data sets consisting of each data of nine measurement items) were obtained by each of the three grinder operations performed by four first-level skilled workers. Among these twelve data, each of their working times and each cutting accuracy were compared with the criteria (threshold values) for classifying as the top-level workers, and ten data classified as the top-level workers' data were selected as unit space data in the unit space of the Mahalanobis distance system defined by the above nine measurement items. The remaining two data were classified as the data of the upper-level workers and were therefore treated as the data of the second-level skilled workers. Therefore, fourteen data consisting of twelve data obtained by each of the three grinder operations performed by four second-level skilled workers and these remaining two data were used as the measurement data. A part of the unit space data (Sample No. 1 to 5) is shown in Table 1, and a part of the measurement data (Sample No. 1 to 5) is shown in Table 2.

[0065]

Table 1

[0066]

Table 2

[0067] Then, for this unit space data, the Mahalanobis distance D in the fourteen measurement data was calculated by Equation 1 below, and the occurrence frequencies of these fourteen Mahalanobis distances were counted as frequencies for each distance. The ten unit space data were similarly calculated and counted. The resulting frequency distribution of the Mahalanobis distances is shown in Figure 7.

[0068]

Equation

[0069] In FIG. 7, the frequency represented by the black hatching is the counting result of the Mahalanobis distance of the unit space data, and the frequency represented by the slanted hatching is the counting result of the Mahalanobis distance of the measurement data. As can be seen from FIG. 7, the Mahalanobis distance of the unit space data converges near 1, and the inverse matrix R of the correlation matrix R -1There exists [a certain situation], and it was confirmed that it is appropriate as a unit space. The data obtained by the work of the top-level workers (the first-level skilled workers belonging to the highest rank) does not necessarily converge near 1. As described above, among the data obtained by the work of the top-level workers, it is necessary to select the data classified as the data of the top-level workers. In other words, it is necessary to generate unit space data based on the data where the Mahalanobis distance converges near 1. And as can be seen from FIG. 7, the Mahalanobis distances based on the respective data of the upper-level workers (the second-level skilled workers belonging to the rank next to the highest rank) are distributed above 2, and the Mahalanobis distances based on the respective data of the top-level workers (each data classified as the data of the top-level workers) and the Mahalanobis distances based on the respective data of the upper-level workers are clearly differently distributed, and a significant difference is recognized. Therefore, it can be understood that the skill difference between the top-level workers and the upper-level workers appears as the distribution difference of the Mahalanobis distance in the Mahalanobis Taguchi system defined by 9 measurement items. In other words, in the Mahalanobis Taguchi system defined by 9 measurement items, it is possible to determine whether it is a first-level skilled worker or a second-level skilled worker based on the Mahalanobis distance, and it is possible to determine the presence or absence of a significant difference based on the distribution difference of the Mahalanobis distance. One of the data for generating the unit space data (here, the data classified as the data of the top-level workers) has its Mahalanobis distance converging near 1. Therefore, only the Mahalanobis distance of the other data (here, the data of the upper-level workers) is obtained. If the obtained Mahalanobis distance of the other data does not converge near 1 (for example, distributed above 2 in the example shown in FIG. 7), it can be determined that there is a significant difference.

[0070] Compared with the upper-level workers, the top-level workers have shorter working hours, higher cutting accuracy, and more stable work. In other words, since the data obtained by the work of the upper-level workers contains components that destabilize the work, it is considered that by analyzing the data obtained by the work of the upper-level workers, work influencing factors that affect the stabilization of the work can be extracted. Here, the work influencing factors were extracted from the 9 measurement items by obtaining the larger-the-better characteristic signal-to-noise ratio of the Mahalanobis distance using a two-level orthogonal array for the 9 measurement items.

[0071] Since the influence of each measurement item on the Mahalanobis distance is analyzed, whether the measurement item is used in the calculation of the Mahalanobis distance (the calculation of the SN ratio of the larger-the-better characteristic), a two-level system is used. In order to efficiently generate the patterns of whether to use or not, an orthogonal array is used. Here, the two-level orthogonal array L shown in Table 3 below is used. 12 is used. "1" represents the first level of using the measurement item, and "2" represents the second level of not using the measurement item. At the first level "1", the data of the measurement item is directly used in the calculation, and at the second level "2", the data of the measurement item is replaced with 0 and used.

[0072]

Table 3

[0073] Since the true value of the data of the measurement item is unknown, the SN ratio η of the larger-the-better characteristic is used. This SN ratio η of the larger-the-better characteristic is obtained by the following formula (2). To the above 14 measurement data, sample numbers k from 1 to 14 are respectively assigned (k is an integer from 1 to 14), and D k is the Mahalanobis distance of the sample number k.

[0074]

Equation

[0075] Two-level orthogonal array L 12The signal-to-noise ratio η of the maximum likelihood characteristic of the Mahalanobis distance according to [reference] is shown in FIG. 8. For each of the nine measurement items, the signal-to-noise ratio η of the maximum likelihood characteristic of the Mahalanobis distance when using the measurement item and the signal-to-noise ratio η of the maximum likelihood characteristic of the Mahalanobis distance when not using the measurement item are obtained as the first signal-to-noise ratio η1 and the second signal-to-noise ratio η2, and these first signal-to-noise ratio η1 and second signal-to-noise ratio η2 are obtained for each of the 14 measurement data. FIGS. 8A to 8D show the first signal-to-noise ratio η1 and the second signal-to-noise ratio η2 in each measurement item obtained for 4 of the 14 measurement data. For example, FIG. 8A shows the first signal-to-noise ratio η1 and the second signal-to-noise ratio η2 of each measurement item in the measurement data of sample number 1 shown in Table 2, and FIG. 8B shows the first signal-to-noise ratio η1 and the second signal-to-noise ratio η2 of each measurement item in the measurement data of sample number 4 shown in Table 2.

[0076] Then, in the first signal-to-noise ratio η1 and the second signal-to-noise ratio η2 obtained for each of these 14 measurement data, for each of the nine measurement items, the number of measurement data in which the second signal-to-noise ratio η2 is smaller than the first signal-to-noise ratio η1 is counted, and from among the nine measurement items, a predetermined number is extracted as the work influence factor from the larger of the counted numbers. For example, in the case of the four measurement data shown in FIGS. 8A to 8D, the number in the heart rate (measurement item 1, A) is counted as 3, the number in the RRSD (measurement item 2, B) is counted as 1, and the number in the RMSSD (measurement item 3, C) is counted as 3. In the 14 measurement data, the top three measurement items in terms of the number, beta wave, alpha wave, and delta wave, are extracted as the work influence factor. Note that it is not limited to the top three, and it may be any number such as the top one, the top two, the top four, etc.

[0077] Figure 9 shows the distributions of beta waves, alpha waves, and delta waves with respect to the Mahalanobis distance D for unit space data and measurement data, respectively. From Figure 9, it can be seen that as the Mahalanobis distance D approaches 1, the beta wave increases and the delta wave decreases. Therefore, it is considered that the work can be stabilized by increasing the beta wave and decreasing the delta wave. Beta waves are generally said to increase in the so-called concentrated and awake states where the brain is activated. Note that delta waves generally do not appear much during wakefulness and are associated with slow-wave-sleep. The sum of each brain wave is 100%, and as described above, since beta waves are related to wakefulness and delta waves are related to slow-wave-sleep, there is a relationship between beta waves and delta waves such that as the beta wave increases, the delta wave decreases. Also, since work is considered to be a repetition of situation judgment and the reflection of the judgment result on the operation, it is considered that the top-level person who generated the data adopted as unit space data repeatedly makes situation judgments with high concentration. Therefore, the measures for work stabilization are, firstly, to incorporate actions that enhance concentration into skilled workers, or conversely, secondly, to eliminate environmental factors that inhibit concentration. Such actions that enhance concentration and, conversely, environmental factors that inhibit concentration often vary from person to person, and the specific actions and environmental factors are considered to depend on skilled workers.

[0078] While having the actions that enhance concentration for the senior workers, the above-mentioned grinder work was carried out 3 times, and during the work, beta waves and delta waves were acquired. The Mahalanobis distance D of the resulting beta waves is indicated by ○ in Figure 10A, and the Mahalanobis distance D of the resulting delta waves is indicated by ○ in Figure 10B. The working time and cutting accuracy are shown as No.N4, No.N5, and No.N6 in Table 4 below with their standard deviations. The Mahalanobis distances D of the beta waves and delta waves indicated by ● in Figure 10A and Figure 10B are the results obtained from the above-mentioned measurement data without taking the actions that enhance concentration, and the working time and cutting accuracy are shown as No.N1, No.N2, and No.N3 in Table 4 below with their standard deviations.

[0079]

Table 4

[0080] From FIGS. 10A and 10B, it was confirmed that by the action of enhancing concentration, the beta waves increased and the delta waves decreased in the grinder work. And from Table 4, it can be seen that the variation in cutting accuracy decreased and the work was stabilized. Therefore, from the nine measurement items, it was confirmed that the beta waves, alpha waves, and delta waves extracted as described above are work influencing factors. In other words, as described above, after determining that there is a significant difference between the mental states of the first skilled worker and the second skilled worker by using the Mahalanobis distance system of the unit space defined by a plurality of measurement items, by using the Mahalanobis distance of the Mahalanobis distance system, the work influencing factor can be extracted from among the plurality of measurement items.

[0081] As described above, the work influencing factor extraction device A in the embodiment and the work influencing factor extraction method implemented thereon determine whether there is a significant difference between the data of the first skilled worker and the data of the second skilled worker, and when there is a significant difference, use the data of the first skilled worker and the second skilled worker with the significant difference, so that the work influencing factor can be extracted.

[0082] Since the plurality of measurement items of the work influencing factor extraction device A and the work influencing factor extraction method are measurement items obtained from electrocardiogram data, electroencephalogram data, and eye gaze data, the work influencing factors related to the mental state of the skilled worker can be preferably extracted.

[0083] Skilled workers classified into the highest level (top - level workers) can usually perform work more reliably and stably. The above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method generate unit - space data using data of skilled workers classified into the highest level who can perform such stable work more reliably, so that more appropriate work - influencing factors can be extracted. The above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method generate unit - space data based on the data of skilled workers classified into the highest level before the Mahalanobis distance converges, so that the unit - space data serving as the evaluation criterion can be generated more appropriately.

[0084] Since the above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method include at least β - waves and δ - waves of electroencephalogram in a plurality of measurement items, it is possible to determine whether β - waves and δ - waves can be work - influencing factors. The above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method can extract work - influencing factors in work using tools. The above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method can extract work - influencing factors in the work of die grinding or polishing. The above - mentioned work - influencing factor extraction device A and work - influencing factor extraction method can extract work - influencing factors in the work of grinding or polishing a die for vehicle member molding.

[0085] In order to describe the present invention, the present invention has been appropriately and sufficiently described through embodiments with reference to the drawings above. However, it should be recognized that those skilled in the art can easily make changes and / or improvements to the above - mentioned embodiments. Therefore, as long as the changes or improvements made by those skilled in the art do not depart from the scope of the claims described in the claims, such changes or improvements are construed as being included in the scope of the rights of the claims.

Explanation of Signs

[0086] A Work - influencing factor extraction device 1 Data acquisition unit 2 Control processing unit 6 Storage unit 21 Control Unit 22 Significant Difference Determination Unit 23 Factor Extraction Unit 221 Unit Space Generation Unit 222 Distance Calculation Unit 223 Determination Unit 231 SN Ratio Calculation Unit 232 Extraction Unit

Claims

【Claim 1】 A data acquisition unit that acquires data of a plurality of measurement items related to the respective mental states of a first skilled worker in a first group and a second skilled worker in a second group, each of whom performs a predetermined operation of handling a predetermined object when performing the predetermined operation; A significant difference determination unit that determines whether there is a significant difference between the mental state of the first skilled worker and the mental state of the second skilled worker by using the Mahalanobis distance system in the unit space defined by the plurality of measurement items, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit; When it is determined by the significant difference determination unit that there is a significant difference, a factor extraction unit that extracts a work influence factor that affects the stabilization of the predetermined operation from among the plurality of measurement items by using the Mahalanobis distance of the Mahalanobis distance system, based on the data of the first skilled worker and the data of the second skilled worker acquired by the data acquisition unit. An operation influence factor extraction device.

Images