Evaluation system and evaluation method

Abstract

Provided is an evaluation system capable of accurately evaluating a proficiency level. This evaluation system comprises: a sensor attached to a worker's head to detect the movement of the head; a sensor data creation unit that acquires detection values output from the sensor during a period in which the worker operates a device, and creates sensor data that associates the acquired detection values with an elapsed time from the start time of a target period to be evaluated; a movement data creation unit that extracts a movement using the sensor data, and creates movement data that associates the extracted movement with the elapsed time; and an evaluation unit that evaluates the worker's proficiency level using the movement data.

Description

Evaluation system and evaluation method

[0001] The present disclosure relates to an evaluation system and an evaluation method.

[0002] Machining processes using equipment such as machine tools to machine workpieces include processes performed by operators. Specific processes performed by operators include, for example, placing the workpiece on the machine tool and checking the machining status of the workpiece and changing the machining conditions. Therefore, the process time spent on a machining process and the quality of the machining vary depending on the operator's proficiency. Therefore, technologies for supporting training to improve operator proficiency and technologies for assessing operator proficiency have been proposed (e.g., Patent Document 1). The technology in Patent Document 1 supports training by using a record of the operator's operations on the equipment.

[0003] Japanese Patent Application Laid-Open No. 2001-228905

[0004] However, in the case of a worker with low proficiency, even if the operation performed by the worker is appropriate, the process time may be longer than that of an expert. Therefore, there is a need for a technology that can accurately evaluate proficiency.

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to one aspect of the present disclosure, an evaluation system is provided. The evaluation system includes: a sensor worn on a worker's head for detecting head movements; a sensor data creation unit that acquires detection values ​​output from the sensor during a period in which the worker operates a device and creates sensor data correlating the acquired detection values ​​with elapsed time from a start time of a target period to be evaluated; a movement data creation unit that extracts the movements using the sensor data and creates movement data correlating the extracted movements with the elapsed time; and an evaluation unit that uses the movement data to evaluate the worker's proficiency. According to this aspect, evaluation using data on the worker's head movements allows for accurate evaluation of proficiency. (2) In the evaluation system of the above aspect, the evaluation unit may evaluate the proficiency by comparing the movement data with predetermined reference data in which reference movements serving as a reference for the movements during the target period are arranged in chronological order. According to this aspect, proficiency can be evaluated by comparing the reference data with the movement data. (3) In the evaluation system of the above aspect, the reference data includes a plurality of the reference actions, the action data includes a plurality of the actions, and each of the plurality of reference actions included in the reference data is assigned a predetermined action type, the action data creation unit associates the action type with each of the plurality of actions included in the action data, and the evaluation unit performs at least one of a determination process of determining whether each of the plurality of reference actions included in the reference data can be associated with at least one of the actions included in the action data having the matching action type, a first assignment process of assigning a first action label to the reference action of the reference data that is determined to be unable to be associated in the determination process if it is determined that the action cannot be associated, and a second assignment process of assigning a second action label to the action of the plurality of actions in the action data that is not associated with any of the plurality of reference actions in the reference data in the determination process, and the proficiency level may be evaluated using at least one of an assignment status of the first action label and an assignment status of the second action label.According to this aspect, proficiency evaluation can be performed with high accuracy. An action assigned with a first action label indicates an action that was not performed by the worker despite being necessary. An action assigned with a second action label indicates an action that was performed by the worker but was unnecessary. Therefore, by evaluating proficiency using the assignment status of each of the first action label and the second action label, evaluation can be performed with high accuracy. (4) In the evaluation system of the above aspect, the reference data may be the action data created using the sensor data acquired during a period in which an expert, who is the worker with high proficiency, operates the device. According to this aspect, sensor data acquired during a period in which the worker with high proficiency operates the device can be used as the reference data. Therefore, data that reflects actual conditions can be used as the reference data. (5) In the evaluation system of the above aspect, the motion data includes a plurality of the motions, and the evaluation unit performs at least one of a first time evaluation process for assigning a first time label to each of the plurality of motions included in the motion data if the motion time from the start of the motion to the end of the motion is longer than a predetermined reference motion time, and a second time evaluation process for assigning a second time label to a target motion among the plurality of motions included in the motion data if the motion interval between a target motion and a subsequent motion of the target motion is longer than a predetermined reference motion interval, and evaluates the proficiency level using the assignment status of the first time label and the second time label. According to this aspect, the proficiency level can be evaluated accurately. A motion assigned with the first time label indicates that the motion itself takes a long time. A motion assigned with the second time label indicates that it takes a long time to start the next motion after performing the target motion. Therefore, by evaluating the proficiency level using at least one of the state of attachment of the first action label and the state of attachment of the second action label, the evaluation can be performed with high accuracy.(6) In the evaluation system of the above aspect, the device includes a reception unit that receives operations from the worker, and the evaluation system further includes an operation information acquisition unit that acquires operation information data that associates the elapsed time with operation information, which is information about the operations received by the reception unit, during the target period, and the elapsed time, the operation information data including a plurality of pieces of the operation information, and the evaluation unit may use the operation information data to divide the motion data into a plurality of sections using each piece of the operation information as a separator, and set each of the plurality of sections as a target section to perform evaluation. According to this aspect, evaluation can be performed for each section, thereby further improving the accuracy of the evaluation. (7) The evaluation system of the above aspect may further include at least one of a sound conversion unit that converts sound data into sound and a display unit that displays information using display data, a determination unit that extracts the motion using the sensor data and determines a next reference motion that the worker should perform by comparing the extracted motion with predetermined reference data in which reference motions serving as references for the motions during the target period are arranged in chronological order, and an output unit that performs at least one of a first output process that outputs the sound data that indicates the next reference motion determined by the determination unit to the sound conversion unit and a second output process that outputs the display data including the information that indicates the next reference motion to the display unit. This aspect allows training for novice workers. (8) The evaluation system of the above aspect may further include an ear-worn device worn on the worker's ear, the ear-worn device having a data acquisition unit that acquires sound data and a sound conversion unit that converts the acquired sound data into sound and emits the sound toward the ear, and the sensor may be built into the ear-worn device. According to this aspect, a sensor built into the ear-worn device can be used for evaluation.(9) The evaluation system of the above aspect may further include a microphone for acquiring voices uttered by the worker as voice data, a voice data creation unit for acquiring the voice data acquired by the microphone during the target period and creating voice time series data in which the acquired voice data is associated with the elapsed time, a call data creation unit for extracting calls of the worker using the voice time series data and creating call data in which the extracted calls are associated with the elapsed time, and an integrated data creation unit for creating integrated data in which the actions included in the action data and the calls included in the call data are arranged in chronological order using the elapsed time associated with each of them, wherein the reference data further includes a reference call that serves as a reference for the calls during the target period, and the reference action and the reference call are arranged in chronological order, and the evaluation unit may evaluate the proficiency by comparing the integrated data with the reference data. According to this aspect, the voices uttered by the worker can be included in the evaluation, allowing for more detailed evaluation. The present disclosure can be realized in various forms, and in addition to the above-mentioned evaluation system, it can also be realized in the form of, for example, an evaluation method, an evaluation program, etc.

[0007] 1 is an explanatory diagram showing a schematic configuration of an evaluation system. A block diagram showing the configuration of the evaluation system. A diagram explaining detection by a motion sensor. A flowchart showing the work procedure of a press-fitting process. A diagram explaining the position and orientation of a worker at the start of a target period. A diagram explaining detected values ​​of a motion sensor and operation data. A diagram comparing detected values ​​of the motion sensors of a beginner and an expert. A flowchart showing the procedure of evaluation processing performed by an evaluation unit. A diagram explaining the processing content of the evaluation processing. A flowchart showing the procedure of training processing. A block diagram of an information processing device according to a second embodiment. A diagram explaining the relationship between a target period and sections according to the second embodiment. A flowchart showing the procedure of multiple section evaluation processing according to the second embodiment. A block diagram of an information processing device according to a third embodiment.

[0008] A. First Embodiment: A1. Configuration of Evaluation System: Fig. 1 is an explanatory diagram showing the schematic configuration of an evaluation system 1. The evaluation system 1 shown in Fig. 1 is a system for evaluating the proficiency of an operator OP who operates a press-fitting device 60 as an apparatus. Fig. 1 shows an up-down direction, a front-rear direction, and a left-right direction that are orthogonal to one another. Each direction is a direction relative to the operator OP.

[0009] As shown in FIG. 1 , the evaluation system 1 includes an information processing device 10, an ear-worn device 80, and a mobile terminal device 90. In this embodiment, the information processing device 10 is disposed near the press-fitting device 60. The ear-worn device 80 is worn on the head of the worker OP. The ear-worn device 80 has a pair of housings 80a that are worn on each of the worker OP's ears. In this embodiment, the ear-worn device 80 is an earphone type, and each of the pair of housings 80a has a shape that can be inserted into an ear canal. The ear-worn device 80 is used to detect the movement of the worker OP's head. Note that the shape of the ear-worn device 80 is not limited to an earphone type, and may be a headphone type.

[0010] In this embodiment, the portable terminal device 90 is carried by the operator OP. Since the operator OP works near the press-fitting device 60, the portable terminal device 90 does not necessarily have to be carried by the operator OP, but may be placed near the press-fitting device 60. The portable terminal device 90 mediates communication between the ear-worn device 80 and the information processing device 10.

[0011] The press-fitting device 60 is a device that performs press-fitting on a workpiece W. The workpiece W includes a press-fitted part Wa and a press-fitting part Wb. During press-fitting, the press-fitting part Wb is press-fit into the press-fitted part Wa. The press-fitting device 60 has a receiving unit 61, a fixed unit 62, and a movable unit 63. The movable unit 63 is disposed above the fixed unit 62. The movable unit 63 is configured to be movable in the vertical direction. Specifically, the movable unit 63 moves closer to the fixed unit 62 by descending and moves away from the fixed unit 62 by ascending. The movable unit 63 moves by the driving force of a motor (not shown) provided in the press-fitting device 60. The movable unit 63 descends from above the workpiece W placed on the fixed unit 62 and presses against the workpiece W, thereby performing press-fitting. The receiving unit 61 has a touch panel 61a and an operation button 61b. The touch panel 61a has a display function for displaying information and a reception function for receiving operations. The operation buttons 61b are, for example, push buttons such as a power button.

[0012] FIG. 2 is a block diagram showing the configuration of the evaluation system 1. As shown in FIG. 2, the information processing device 10 is configured as a computer having a processor 20 and a storage unit 30. The processor 20 performs various processes by executing programs stored in the storage unit 30. The processor 20 has, as functional units, a sensor data creation unit 21, a motion data creation unit 22, an evaluation unit 23, a determination unit 24, and an output unit 25. Each functional unit is realized by executing a program stored in the storage unit 30. The storage unit 30 is realized by a memory such as a RAM or a ROM. The storage unit 30 stores sensor data 31, motion data 32, reference data 33, an evaluation program 34, and a training program 35.

[0013] The information processing device 10 further includes a communication unit 40. The communication unit 40 communicates with an external device such as a mobile terminal device 90 via a network such as a LAN (Local Area Network).

[0014] The ear-worn device 80 has a pair of housings 80a shown in Fig. 1 each incorporating a motion sensor 81 as a sensor shown in Fig. 2, a sound conversion unit 82, a microphone 83, and a communication unit 84. The motion sensor 81 and the microphone 83 may be incorporated in both of the pair of housings 80a, or may be incorporated in either one of the housings 80a.

[0015] The motion sensor 81 detects the movement of the head of the worker OP. Specifically, the motion sensor 81 is a general-purpose so-called six-axis motion sensor. Specifically, the motion sensor 81 includes a three-axis acceleration sensor and a three-axis angular velocity sensor. The motion sensor 81 detects the acceleration [m / s ２ ] and the angular velocity [deg / s] of each of the three axes.

[0016] FIG. 3 is a diagram illustrating detection by the motion sensor 81. As shown in FIG. 3, the three axes along which the motion sensor 81 detects acceleration and angular velocity are, specifically, the yaw axis, roll axis, and pitch axis, which are perpendicular to one another. The yaw axis is an axis parallel to the up-down direction relative to the worker OP. The roll axis is an axis parallel to the front-back direction relative to the worker OP. The pitch axis is an axis parallel to the left-right direction relative to the worker OP. With respect to the acceleration along the yaw axis, the motion sensor 81 outputs a positive value when the worker OP moves his / her head up, and a negative value when the worker OP moves his / her head down. With respect to the acceleration along the roll axis, the motion sensor 81 outputs a positive value when the worker OP moves his / her head forward, and a negative value when the worker OP moves his / her head back. With respect to the acceleration along the pitch axis, the motion sensor 81 outputs a positive value when the worker OP moves his / her head to the left, and a negative value when the worker OP moves his / her head to the right. The motion sensor 81 outputs a positive value for the angular velocity of the yaw axis when the worker OP turns his face to the right, and a negative value when the worker OP turns his face to the left. The motion sensor 81 outputs a positive value for the angular velocity of the roll axis when the worker OP tilts his head to the left, and a negative value when the worker OP tilts his head to the right. The motion sensor 81 outputs a positive value for the angular velocity of the pitch axis when the worker OP turns his face down, and a negative value when the worker OP turns his face up. Therefore, head movement can be detected from changes in acceleration and angular velocity along each of the three axes.

[0017] The sound conversion unit 82 shown in FIG. 2 converts sound data, which is an electrical signal, into sound and emits it toward the ear of the worker OP. The microphone 83 converts sound into sound data, which is an electrical signal. Specifically, the microphone 83 converts the voice uttered by the worker OP into voice data and acquires it. The microphone 83 outputs the acquired voice data to the communication unit 84. The communication unit 84 performs short-range communication with the portable terminal device 90. In this embodiment, the communication unit 84 and the portable terminal device 90 communicate in accordance with Bluetooth (registered trademark). The communication unit 84 has, as functional units, a data acquisition unit 85 and a data output unit 86. The data acquisition unit 85 receives voice data transmitted from an external device and outputs the received voice data to the sound conversion unit 82. The data output unit 86 transmits the voice data output from the microphone 83 to an external device. The communication unit 84 is realized by an electronic circuit or the like.

[0018] In this embodiment, the mobile terminal device 90 is a smartphone. The mobile terminal device 90 has a communication unit 91. The communication unit 91 of the mobile terminal device 90 performs short-range communication with the communication unit 84 of the ear-worn device 80, and also performs communication with the communication unit 40 via a network. Note that communication between the ear-worn device 80 and the mobile terminal device 90 is not limited to wireless communication, and may be performed via wired communication.

[0019] In this embodiment, when the ear-worn device 80 and the information processing device 10 communicate with each other, the communication is performed via the mobile terminal device 90. In another embodiment, the communication unit 40 of the information processing device 10 may have a function of performing short-range communication, and the communication unit 40 may perform short-range communication with the communication unit 84 of the ear-worn device 80 without the mobile terminal device 90. Alternatively, the communication unit 84 of the ear-worn device 80 may have a function of communicating via a network, and the communication unit 84 may communicate with the communication unit 40 via the network without the mobile terminal device 90.

[0020] In addition to the above configuration, the press-fitting device 60 also includes a processor 65, a storage unit 66, and a communication unit 68. The storage unit 66 has a program for creating operation information data 67 stored in advance. The processor 65 creates the operation information data 67 by executing the program stored in the storage unit 66. The operation information data 67 is information in which operation information, which is information about an operation accepted by the accepting unit 61, is associated with the time of acceptance. In this specification, "time" refers to time including a date. When the accepting unit 61 accepts an operation, the processor 65 creates the operation information data 67 by associating the accepted operation information with the time of acceptance. The processor 65 then stores the created operation information data 67 in the storage unit 66. Specific examples of operations accepted by the accepting unit 61 include an operation to change the state of a descent button for lowering the movable unit 63 from an off state to an on state and an operation to set the descent speed of the movable unit 63 by operating the touch panel 61a. The down button is included in the operation buttons 61b. The communication unit 64 communicates with the information processing device 10 via a network.

[0021] A2. Functions of the evaluation system: The evaluation system 1 has an evaluation function and a training function. The evaluation function is a function that evaluates the worker OP's proficiency in work after the worker OP performs work using the press-fitting device 60. The training function is a function that supports the worker OP in training for work using the press-fitting device 60.

[0022] A3. Overview of Evaluation Function: Evaluation by the evaluation system 1 is performed after the press-fitting process is performed by the worker OP. Figure 4 is a flowchart showing the procedure for the press-fitting process. As shown in Figure 1, the worker OP stands in a position where he can visually check the state of the movable part 63 and the workpiece W, and performs the work by operating the reception unit 61 without moving. In the press-fitting process, the movable part 63 is operated by the worker OP to slow down and descend to a position just before contacting the workpiece W, and then to increase the speed and descend further for press-fitting.

[0023] Before the start of the press-fitting process, the press-fitting device 60 is preset to a state in which the movable part 63 is positioned above and away from the fixed part 62. In step S10 of FIG. 4 , the operator OP places the workpiece W at a predetermined reference position of the fixed part 62. In step S12, the operator OP checks by looking at the touch panel 61a whether the descent speed of the movable part 63 is set to a predetermined first speed. In step S14, the operator OP checks by looking at the workpiece W whether it is positioned at the reference position. In step S16, the operator OP sets a descent switch, which is one of the operation buttons 61b of the reception unit 61, to an on state, thereby lowering the movable part 63 at the first speed.

[0024] In step S18, the operator OP checks the workpiece W and confirms that the movable part 63 has descended to the stop position. Here, the stop position is a position just before the movable part 63 comes into contact with the workpiece W, where the distance between the movable part 63 and the workpiece W is about several centimeters.

[0025] In step S20, the worker OP operates the reception unit 61 to set the lowering switch to the OFF state and change the speed of the movable part 63 to a second speed that is higher than the first speed. In step S22, the worker OP sets the lowering switch to the ON state and further lowers the movable part 63 to press-fit the press-fit part Wb into the press-fitted part Wa. The press-fitting device 60 is preset so that the application of pressure by the movable part 63 to the fixed part 62 and the workpiece W is stopped when a preset time has elapsed since the pressure received by the movable part 63 reached a preset pressure. The application of pressure by the movable part 63 is stopped, and the press-fitting process is completed.

[0026] The sensor data creation unit 21 shown in FIG. 2 performs a sensor data creation process in which it acquires detection values ​​output from the motion sensor 81 and creates sensor data 31 that associates the acquired detection values ​​with the elapsed time from the start time of a target period. Here, the target period is a period during which the operator OP performs the press-fitting process and is the subject of proficiency evaluation. Specifically, the detection values ​​are the angular velocities for each of the three axes and the accelerations for each of the three axes. The sensor data creation unit 21 then stores the created sensor data 31 in the storage unit 30. In this embodiment, the target period is the period from step S12 to step S16.

[0027] A4. Details of the evaluation function: Fig. 5 is a diagram illustrating the position and orientation of the operator OP at the start of the target period. When evaluation is performed using the evaluation system 1, the operator OP performs step S10, and then stands in a reference position where he or she can operate the reception unit 61 while facing the front of the press-fitting device 60, as shown in Fig. 5. When the operator OP stands in the reference position without tilting his or her head, the values ​​of each detection value of the motion sensor 81 are set to the reference value, i.e., zero.

[0028] The administrator of the evaluation system 1 starts the evaluation program 34 on the information processing device 10. The evaluation program 34 displays a GUI (Graphical User Interface) for inputting an evaluation start time, which is displayed on the display of the information processing device 10. The administrator inputs a start time into the GUI for inputting the evaluation start time, and instructs the worker OP to start work. The evaluation program 34 stores the start time input into the GUI in the storage unit 30.

[0029] When instructed to start work, the worker OP performs steps S12 to S16. After step S16, the administrator inputs an end time into a GUI for inputting an evaluation end time, which is displayed on the display of the information processing device 10 by the evaluation program 34. The evaluation program 34 stores the end time input into the GUI in the storage unit 30. The administrator selects an evaluation start button, which is a GUI displayed on the display of the information processing device 10 by the evaluation program 34. When the evaluation start button is selected, the evaluation program 34 requests the movement data creation unit 22 to create movement data 32. When requested, the movement data creation unit 22 performs a movement data creation step of extracting movements using the sensor data 31 and creating movement data 32 that associates the extracted movements with elapsed time.

[0030] 6 is a diagram illustrating the detection values ​​of the motion sensor 81 and the operation data 32. From the top to the bottom of Fig. 6, there are shown, in order, a graph showing the relationship between the yaw axis angle and elapsed time, a graph showing the relationship between the pitch axis angle and elapsed time, a graph showing the relationship between the yaw axis acceleration and elapsed time, and the operation data 32. The horizontal axes of the graph showing the relationship between the yaw axis angle and time, the graph showing the relationship between the pitch axis angle and time, and the graph showing the relationship between the yaw axis acceleration and time are aligned so that the elapsed time from the evaluation start time corresponds to each other.

[0031] The motion data creation unit 22 extracts motions using the sensor data 31. In this embodiment, the head motions performed by the worker OP are two types of motions: a motion of looking at the reception unit 61 and a motion of looking at the workpiece W. Therefore, the motion data creation unit 22 uses data on the angular velocity and angle of the yaw axis and data on the angular velocity and angle of the pitch axis to extract the motions. The motion data creation unit 22 obtains data on the yaw axis angle from the motion sensor 81 by converting the detected value of the angular velocity of the yaw axis into an angle. Similarly, the motion data creation unit 22 obtains data on the pitch axis angle from the motion sensor 81 by converting the detected value of the angular velocity of the pitch axis into an angle.

[0032] The motion data creation unit 22 extracts, from the data on the angular velocity of the yaw axis, the time when the absolute value of the angular velocity becomes greater than a predetermined first threshold and the time when the absolute value of the angular velocity becomes smaller than a predetermined second threshold. The first threshold is a threshold for extracting the time when the motion starts. The second threshold is a threshold for extracting the time when the motion stops. In this embodiment, the first threshold is a threshold for extracting the start of the motion of turning the face. The second threshold is a threshold for extracting the end of the motion of turning the face. The first threshold and the second threshold are values ​​obtained through experiments or the like.

[0033] In this embodiment, the motion data creation unit 22 determines that the absolute value of the angular velocity about the yaw axis is greater than the first threshold value at times ts1 and ts2 in Fig. 6. Also, the motion data creation unit 22 determines that the absolute value of the angular velocity about the yaw axis is smaller than the second threshold value at times t1 and t2.

[0034] Next, the action data creation unit 22 determines whether the object the worker OP is looking at is the touch panel 61a or the workpiece W at the time when the absolute value of the angular velocity of the yaw axis becomes smaller than the second threshold value. Specifically, the action data creation unit 22 identifies the object the worker OP is looking at from the angle of the yaw axis and the angle of the pitch axis at time t1. When the worker OP looks at the reception unit 61, he turns his face up and to the right, so a positive value is output for the angle of the yaw axis, and a negative value is output for the angle of the pitch axis. When the worker OP looks at the workpiece W, he turns his face down and to the left, so a negative value is output for the angle of the yaw axis, and a positive value is output for the angle of the pitch axis.

[0035] For example, at time t1, the yaw axis angle value is positive and the pitch axis angle value is negative. Therefore, the motion data creation unit 22 determines that the motion started at time ts1 is a "display check" motion of checking the reception unit 61. Furthermore, at time t2, the yaw axis angle value is negative and the pitch axis angle value is positive. Therefore, the motion data creation unit 22 determines that the motion started at time t2 is a "workpiece check" motion of checking the workpiece W. In this manner, the motion data creation unit 22 determines whether the extracted motion is a "display check" motion or a "workpiece check" motion. Furthermore, the motion data creation unit 22 determines the start time and duration of each motion. In this embodiment, the motion is a motion of looking at the reception unit 61 or the workpiece W. Therefore, the motion data creation unit 22 determines the time when the motion starts looking at the reception unit 61 or the workpiece W as the start time of the motion. The motion duration is then determined as the time until the next motion starts. Specifically, in "motion 1," time t1 is the start time, and the period from time t1 to time ts2 is the motion period.

[0036] As described above, the action data 32 is data in which actions are arranged in chronological order. Each action in the action data 32 is associated with an action type, a start time, and an action duration. In this embodiment, the action types are "display check" and "work check." Note that the start time may be indicated by the elapsed time from the evaluation start time, rather than the actual time.

[0037] The method by which the motion data creating unit 22 extracts motions is not limited to the above. For example, the motion data creating unit 22 may determine the period during which the angle is within a predetermined reference range as the motion period. The reference range corresponds to the angle at which the worker OP turns his / her face to look at the object. The reference range is, for example, a range of +10 degrees or more and +20 degrees or less, and a range of -10 degrees or more and -20 degrees or less.

[0038] When the operation data creation unit 22 finishes creating the operation data 32, it stores the created operation data 32 in the storage unit 30 and requests the evaluation unit 23 to perform evaluation processing. Upon receiving the request, the evaluation unit 23 performs the evaluation processing using the operation data 32. In this embodiment, the evaluation unit 23 evaluates the proficiency level by comparing the operation data 32 with pre-stored reference data 33. The reference data 33 is data similar to the operation data 32 and is predetermined data in which reference operations serving as a reference for operations during a target period are arranged in chronological order. In this embodiment, the reference data 33 is data created in the same manner as the operation data 32. That is, the reference data 33 is data created by the operation data creation unit 22 using sensor data 31 acquired during a period in which a skilled worker OP, who is a highly skilled worker, operates the press-fitting device 60. Hereinafter, the reference operation may be simply referred to as an "operation." The operation data creation unit 22 assigns a number to each operation in chronological order, starting from "1," for both the reference data 33 and the operation data 32. In the following description, for example, the operation numbered "1" will be referred to as "operation 1."

[0039] 7 is a diagram comparing the detection values ​​of the motion sensor 81 of a beginner, who is a worker OP with a low level of proficiency, with the detection values ​​of the motion sensor 81 of an expert. Fig. 7 is a diagram similar to Fig. 6. The time on the horizontal axis of each graph for the beginner and the time on the horizontal axis of each graph for the expert are aligned so that the elapsed time from the evaluation start time matches each other.

[0040] As shown in Figure 7, the target period for the expert is shorter than that for the beginner. This is because the expert does not perform any actions other than the actions for performing the processing steps specified in the press-fitting process shown in Figure 4. Specifically, the expert performs "action 1" corresponding to step S12, "action 2" corresponding to step S14, and "action 3" corresponding to step S16. Therefore, by comparing the action data 32 of the beginner with the action data 32 of the expert, it is possible to analyze the reason why the target period for the beginner is longer.

[0041] FIG. 8 is a flowchart showing the procedure of the evaluation process performed by the evaluation unit 23. The evaluation method is realized by the evaluation process. FIG. 9 is a diagram illustrating the evaluation process. The evaluation process will be described using the example shown in FIG. 9. In FIG. 9, the actions included in the reference data 33 and the action data 32 are associated with either "type A" or "type B," which are action types. The order of the actions included in the reference data 33 and the action data 32 is as shown in FIG. 9. For example, the reference data 33 indicates that "action 1," "action 2," and "action 3" were performed in this order.

[0042] In step S30 of FIG. 8 , the evaluation unit 23 sets the variable N, constant Ns, variable M, and constant Ms used in the evaluation process to initial values. Specifically, the evaluation unit 23 sets both the variable N and the variable M to "1." The evaluation unit 23 sets the variable Ns to the number of actions included in the reference data 33. The evaluation unit 23 sets the variable Ms to the number of actions included in the action data 32. In the case of FIG. 9 , the reference data 33 includes three actions, so the evaluation unit 23 sets the constant Ns to "3." In the case of FIG. 9 , the action data 32 includes three actions, so the evaluation unit 23 sets the constant Ms to "3."

[0043] In step S32 of the determination process in FIG. 8 , the evaluation unit 23 determines whether the action type of the Nth action in the reference data 33 matches the action type of the Mth action in the action data 32. If it is determined that the action type of the Nth action in the reference data 33 matches the action type of the Mth action in the action data 32, the evaluation unit 23 determines in step S36 whether the action period of the Mth action is longer than the action period of the Nth action in the reference action. If it is determined that the action period of the Mth action is longer than the action period of the Nth action in the reference action, the evaluation unit 23 assigns a “period length” label as a first time label to the Mth action in the action data 32 in step S40. Having completed the evaluation of the Mth action, the evaluation unit 23 increments the variable M in step S42.

[0044] If it is determined in step S32 that the duration of the Mth action is not longer than the duration of the Nth action in the reference action, the evaluation unit 23 assigns an "evaluated" label to the Mth action in the reference data 33 in step S38, and proceeds to step S42.

[0045] If it is determined that the action type of the Nth action in the reference data 33 does not match the action type of the Mth action in the action data 32, the evaluation unit 23 assigns a ``missing action'' label as a second action label to the Nth action in the reference data 33 in step S34 as a first assignment process.

[0046] After performing step S40 or step S42, the evaluation unit 23 has completed the evaluation using the Nth motion of the reference data 33, and therefore increments the variable N in step S44.

[0047] In step S46, the evaluation unit 23 determines whether the variable N is equal to or greater than the constant Ns. If it is determined in step S46 that the variable N is not equal to or greater than the constant Ns, there are still movements for which evaluation has not been completed for the reference data 33, so the evaluation unit 23 determines in step S48 whether the variable M is equal to or greater than the constant Ms. If it is determined in step S48 that the variable M is not equal to or greater than the constant Ms, there are still movements for which evaluation has not been completed for the movement data 32, so the evaluation unit 23 returns to step S32.

[0048] If it is determined in step S46 that the variable N is equal to or greater than the constant Ns, the evaluation of all the movements in the reference data 33 has been completed, and the evaluation unit 23 then determines in step S50 whether labels have been assigned to all the movements in the movement data 32. If it is determined in step S50 that labels have not been assigned to all the movements in the movement data 32, the evaluation unit 23 assigns an "unnecessary movement" label as a second movement label to the Mth and subsequent movements in the movement data 32 in step S52 as a second assignment process, and terminates this processing routine. If it is determined in step S50 that labels have been assigned to all the movements in the movement data 32, the evaluation unit 23 terminates this processing routine.

[0049] If it is determined in step S48 that the variable M is not equal to or greater than the constant Ms, then in step S54 the evaluation unit 23 determines whether labels have been assigned to all the actions in the action data 32, as in step S50. If it is determined in step S54 that labels have not been assigned to all the actions in the action data 32, then in step S56 the evaluation unit 23 assigns a "missing action" label as the first action label to actions from the Nth onward, and ends this processing routine. If it is determined in step S54 that labels have been assigned to all the actions in the action data 32, then the evaluation unit 23 ends this processing routine.

[0050] The evaluation process will be further described with reference to FIG. 9 . In step S32, the first evaluation process compares the first "action 1" in the reference data 33 with the first "action 1" in the action data 32. Since the types of "action 1" in the reference data 33 and "action 1" in the action data 32 are both "Type A," a "YES" determination is made in step S32. Next, in step S36, if the action duration of "action 1" in the action data 32 is longer than the action duration of "action 1" in the reference data 33, a "duration length" label is assigned to "action 1" in the action data 32 in step S40. On the other hand, if the action duration of "action 1" in the action data 32 is not longer than the action duration of "action 1" in the reference data 33 in step S36, a "evaluated" label is assigned to "action 1" in the action data 32 in step S38.

[0051] In step S42, the variable M is updated to "2." In step S44, the variable N is updated to "2." In the case of FIG. 9, since the constants Ns and Ms are both "3," if the variable N is "2" and the variable M is "2," the determinations in steps S46 and S48 are both "NO."

[0052] In the second iteration of step S32 after starting the evaluation process, "movement 2," the second movement in the reference data 33, is compared with "movement 2," the second movement in the movement data 32. Since the type of "movement 2" in the reference data 33 is "type B," while the type of "movement 2" in the movement data 32 is "type A," the result in step S32 is "NO." Therefore, in step S34, a "missing movement" label is assigned to "movement 2" in the reference data 33. In step S44, the variable N is updated to "3."

[0053] In step S32 for the third time after starting the evaluation process, "action 3," which is the third action in the reference data 33, is compared with "action 2," which is the second action in the action data 32. Since the type of "action 3" in the reference data 33 and the type of "action 2" in the action data 32 are both "Type A," the evaluation process is started and either the "duration length" label or the "evaluated" label is assigned to "action 2" in the action data 32, as in the initial processing flow.

[0054] In step S42, the variable M is updated to "3." In step S44, the variable N is updated to "4." Therefore, in step S46, the answer is determined to be "YES." In the next step S50, the answer is determined to be "NO" because no label has been assigned to "action 3" in the action data 32. In the next step S52, the "unnecessary action" label is assigned to "action 3" in the action data 32.

[0055] As described above, the reference data 33 includes all the actions necessary for a process, and therefore, by comparing the reference actions included in the reference data 33 with the actions included in the action data 32 in chronological order, it is possible to evaluate whether there are any missing actions or unnecessary actions in the action data 32. Furthermore, when the same action type is performed in the reference data 33 and the action data 32, it is possible to evaluate whether the action is taking a long time by comparing the action period of the action data 32 with the action time of the reference data 33.

[0056] In the case of the motion data 32 shown in FIG. 7 , the motion types of the beginner ("motion 1") through "motion 3") match those of the expert ("motion 1") through "motion 3". However, "motion 4" and "motion 5" in the motion data 32 are not included in the reference data 33. Therefore, when an evaluation process is performed on the motion data 32 shown in FIG. 7 , "motion 4" and "motion 5" are assigned an "unnecessary motion" label. This makes it possible to identify that the long target period for the beginner is due to the performance of unnecessary motions. Furthermore, since the motion durations of "motion 1" and "motion 2" are longer than those of the reference data 33, when an evaluation process is performed, "motion 1" and "motion 2" are assigned a "duration length" label. This indicates that the beginner takes a long time to perform the motions.

[0057] When the evaluation unit 23 finishes the evaluation process, it passes the evaluation data, in which the action data 32 is associated with the reference data 33 and a label is assigned to each action, to the evaluation program 34, as shown in Fig. 9. The evaluation program 34 displays the evaluation data that has been passed on to the display of the information processing device 10. Note that the method for outputting the evaluation data is not limited to outputting it as image data, and it may also be, for example, text data in which actions are associated with labels.

[0058] Furthermore, the evaluation unit 23 performs an evaluation step of evaluating the proficiency of the worker OP using evaluation data indicating the status of the assignment of the "omission action" label and the status of the assignment of the "unnecessary action" label. The evaluation unit 23 performs an evaluation such that the greater the number of "omission action" labels or "unnecessary action" labels, the lower the proficiency. In detail, the evaluation unit 23 may deduct points for actions to which either the "omission action" label or the "unnecessary action" label is assigned, and calculate the total value for all actions as the proficiency. Furthermore, the evaluation unit 23 may total the number of "omission action" labels and the number of "unnecessary action" labels, and determine a rank as the proficiency using a correspondence table in which the number of labels is associated with a rank.

[0059] A5. Other Embodiments of the Evaluation Process: In the evaluation process described above, the evaluation unit 23 evaluates the proficiency level by comparing the reference data 33 with the action data 32. As another embodiment, the evaluation unit 23 may perform a first time evaluation process in which, for each of a plurality of actions included in the action data 32, a first time label is attached to the action if the action time from the start to the end of the action is longer than a predetermined reference action time. This will be specifically described with reference to FIG. 6. The evaluation unit 23 attaches a first time label to "action 1" if the action period "TD1" of "action 1" shown in FIG. 6 is longer than a predetermined reference action time. The evaluation unit 23 similarly performs the first time evaluation process for the other actions. Note that the reference action time may be set to a common time for "action 1" to "action 5," or may be set to different times.

[0060] Furthermore, the evaluation unit 23 may perform a second time evaluation process in which a second time label is assigned to a target action among the multiple actions included in the action data 32 if the action interval between the target action and the action following the target action is longer than a predetermined reference action interval. This will be specifically described with reference to FIG. 6 . The evaluation unit 23 assigns the second time label to "action 1" if the time from the end time ts2 of "action 1" (the target action shown in FIG. 6 ) to the start time t2 of "action 2" is longer than a predetermined reference action interval. The evaluation unit 23 similarly performs the second time evaluation process on other action intervals, for example, the interval between "action 1" and "action 2." Note that the reference action interval may be set to a common interval or may be different intervals for multiple action intervals. By assigning the second time label, it is possible to determine that an action to which the second time label is assigned takes a long time before the next action is performed.

[0061] The evaluation unit 23 performs an evaluation step of evaluating the proficiency of the worker OP using evaluation data indicating the attachment status of the first time labels and the attachment status of the second time labels, similar to the evaluation step described above. That is, the evaluation unit 23 performs an evaluation such that the proficiency level decreases as the number of first time labels or second time labels increases.

[0062] If the worker OP has a low level of proficiency, the steps of the process may not proceed smoothly, and the operation time may be significantly long, or it may take a long time to perform the next operation. In such cases, the evaluation can be made more easily by performing the first time evaluation process or the second time evaluation process. Note that the evaluation unit 23 may perform only one of the first time evaluation process and the second time evaluation process.

[0063] A6. Details of the Training Function: In training using the evaluation system 1, instructions for the next process to be performed are output to the worker OP by voice from the sound conversion unit 82. In the case of the evaluation function described above, the sensor data creation unit 21 creates the action data 32 after the worker OP has completed the press-fitting process. In contrast, in the training function, the determination unit 24 sequentially creates sensor data 31 while the worker OP is working. The determination unit 24, like the sensor data creation unit 21, extracts actions using the sensor data 31. Then, the determination unit 24 compares the extracted actions with the reference data 33 to determine the next action that the worker OP should perform next. Note that the reference data 33 used in the training function includes not only the actions during the target period but also all actions performed in the process.

[0064] 10 is a flowchart showing the procedure of the training process. When starting training, the operator OP starts the training program 35 in the information processing device 10. The training program 35 requests the output unit 25 to perform step S60.

[0065] In step S60, the output unit 25 instructs the first operation of the reference data 33. Specifically, the output unit 25 transmits audio data as sound data of a message instructing the next operation to the ear-worn device 80. As a result, the sound conversion unit 82 of the ear-worn device 80 outputs the audio instruction. The worker OP hears the audio instruction and performs the first operation.

[0066] In step S62, the determination unit 24 determines whether the worker OP has started to move. Specifically, when the determination unit 24 determines that the angular velocity of the yaw axis has become greater than the first threshold value, the determination unit 24 determines that the worker OP has started to move.

[0067] If it is determined in step S62 that the operation has not started, the determination unit 24 repeats step S62 at predetermined time intervals until it determines that the operation has started. The predetermined time is, for example, several milliseconds.

[0068] If it is determined in step S62 that a movement has been started, the determination unit 24 identifies the type of movement in step S64. Specifically, similar to the movement data creation unit 22, the determination unit 24 identifies the type of movement using the angular velocity and angle of the yaw axis and the angular velocity and angle of the pitch axis.

[0069] In step S66, the determination unit 24 determines whether the identified action matches the action in the reference data 33. Specifically, the determination unit 24 compares the actions in the reference data 33 starting from the first action. If the identified action type matches the action type of the action to be compared in the reference data 33, the determination unit 24 determines that the identified action matches the action in the reference data 33. On the other hand, if the identified action type does not match the action type of the action to be compared in the reference data 33, the determination unit 24 determines that the identified action does not match the action in the reference data 33. If it is determined that the identified action does not match the action in the reference data 33, the determination unit 24 requests the output unit 25 to perform step S68. In step S68, the output unit 25 outputs audio data instructing the reference action in the reference data 33 to the sound conversion unit 82. Specifically, the output unit 25 transmits audio data of a message instructing the reference action to the ear-worn device 80. As a result, the audio instruction is output from the sound conversion unit 82 of the ear-worn device 80. The worker OP hears the audio instruction and performs the standard action. After the output unit 25 performs step S68, the determination unit 24 returns the process to step S62.

[0070] If it is determined in step S66 that the movement matches the reference data 33, the decision unit 24 determines whether the process has ended in step S70 as the first output process. If it is determined in step S70 that the process has not ended, the decision unit 24 requests the output unit 25 to perform step S72. The output unit 25 instructs the next movement of the reference data 33 in step S72 as the first output process. Specifically, the output unit 25 transmits audio data of a message instructing the next movement of the reference to the ear-worn device 80. This causes the sound conversion unit 82 of the ear-worn device 80 to output the instruction audio. After the output unit 25 performs step S72, the decision unit 24 returns the process to step S62. If it is determined in step S70 that the process has ended, the decision unit 24 ends this processing routine. When the decision unit 24 ends the training process, the evaluation program 34 displays a message indicating the end of training on the display of the information processing device 10.

[0071] A7. Other Embodiments of the Instruction Method: In step S72, the output unit 25 outputs voice data instructing the next operation to the sound conversion device. In another embodiment, the evaluation system 1 may include a display unit that displays information, and the output unit 25 may perform a second output process in which display data including information instructing the next operation is output to the display unit. Specifically, in the second output process, the output unit 25 outputs display data including text information, such as "Please press the switch next." to the display unit. This allows the worker OP to perform the next step by looking at the text displayed on the display unit. The second output process may be performed instead of the first output process, or both the second output process and the first output process may be performed. When the display of the information processing device 10 is located in a position visible to the worker OP, the display of the information processing device 10 may be used as the display unit. In another embodiment, a display separate from the information processing device 10 may be used as the display unit.

[0072] According to the first embodiment described above, the evaluation system 1 includes a sensor data creation unit 21, a motion data creation unit 22, and an evaluation unit 23. The sensor data creation unit 21 creates sensor data 31 that associates detection values ​​output from a motion sensor 81 attached to the head of the worker OP with elapsed time. The motion data creation unit 22 extracts head movements using the sensor data 31 and creates motion data 32 that associates the extracted movements with elapsed time. The evaluation unit 23 can accurately evaluate the proficiency level by using the data on the head movements of the worker OP for evaluation. Even when the process time is short, there is a possibility that a necessary movement has been omitted. Therefore, by using the data on the head movements of the worker OP for evaluation, it is possible to evaluate whether a necessary movement has been omitted, thereby enabling accurate evaluation.

[0073] Furthermore, by using a motion sensor 81 worn by the worker OP instead of motion capture using captured video to extract the movements of the worker OP, it is possible to provide an evaluation system 1 with a high degree of freedom in installation. Specifically, one method of acquiring the movements of the worker OP as analyzable electronic data is to install a camera in a position where the movements of the worker OP can be imaged and acquire the video. However, in this case, space is required for installing the camera. In this regard, by using the motion sensor 81 worn by the worker OP, it is not necessary to secure space for installing the camera. Therefore, the evaluation system 1 can be installed even in places where it is difficult to secure sufficient space, and therefore it is possible to provide an evaluation system 1 with a high degree of freedom in installation.

[0074] Furthermore, in the evaluation process, the evaluation unit 23 evaluates the proficiency level by comparing the reference data 33 with the action data 32. This allows the evaluation unit 23 to evaluate the proficiency level by comparing the reference data 33 with the action data 32. In the evaluation process, the evaluation unit 23 performs step S32 as a determination process, step S34 as a first assignment process, and step S50 as a second assignment process. In step S34, if it is determined that a reference action included in the reference data 33 is not included in the action data 32, a "missing action" label is assigned to the action determined not to be included. An action assigned the "missing action" label indicates that the action was not performed by the worker OP despite being a necessary action. In step S50, if it is determined that an action included in the action data 32 is not included in the reference data 33, a "unnecessary action" label is assigned to the action determined not to be included. An action assigned the "unnecessary action" label indicates that the action was not performed by the worker OP. Therefore, by assigning the "missing action" label and the "unnecessary action" label, the evaluation can be performed more accurately.

[0075] Furthermore, the reference data 33 is operation data 32 created using sensor data 31 acquired while the skilled worker is operating the press-fitting device 60. This allows the operation data 32 of the skilled worker to be used as the reference data 33. Since the reference data 33 is data acquired when the skilled worker actually performs the work, data that reflects the actual situation can be used as the reference data 33.

[0076] Furthermore, according to another embodiment of the evaluation process described above, if the action time of an action included in the action data is longer than a reference action time, a first time label is assigned. Furthermore, if the action interval between a target action included in the action data and the action following the target action is longer than a reference action interval, a second time label is assigned. Therefore, an action assigned with the first time label indicates that it takes a long time to perform the action. A action assigned with the second time label indicates that it takes a long time to perform the next action. Therefore, by assigning the first time label and the second time label, evaluation can be performed with high accuracy.

[0077] The evaluation system 1 also includes a determination unit 24 and an output unit 25. The determination unit 24 extracts an action using the sensor data 31 and compares the extracted action with the reference data 33 to determine a next reference action that the worker OP should perform next. The output unit 25 outputs voice data instructing the next action to the sound conversion unit 82. Therefore, the evaluation system 1 can be used to provide training to beginners.

[0078] Furthermore, the motion sensor 81 is built into the ear-worn device 80. Therefore, the motion sensor 81 built into the ear-worn device 80 can be used for evaluation.

[0079] B. Second Embodiment: FIG. 11 is a block diagram of an information processing device 10 according to a second embodiment. FIG. 12 is a diagram illustrating the relationship between a target period and a section according to this embodiment. In the second embodiment, in addition to the detection value of the motion sensor 81, information on the operation received by the press-fitting device 60 is used for evaluation. The same components and processing steps as those in the first embodiment are denoted by the same reference numerals, and detailed explanations will be omitted as appropriate.

[0080] 11 , in the present embodiment, the information processing device 10 has, in addition to the configuration of the first embodiment, an operation information acquisition unit 26 and an integrated data creation unit 27 as functional units of the processor 20. The storage unit 30 stores sensor data 31, action data 32, reference data 33, an evaluation program 34, operation information data 67, and integrated data 36.

[0081] In the first embodiment, the processes from step S12 to step S16 of the press-fitting process shown in FIG. 4 are the subject of evaluation. In contrast, in the present embodiment, all processes from step S12 to step S22 of the press-fitting process are the subject of evaluation. In the processes from step S12 to step S22, in addition to the operation on the reception unit 61 in step S16, the operator OP also operates the reception unit 61 in steps S20 and S22. Therefore, in the multiple-section evaluation process, the evaluation unit 23 divides the action data 32 into multiple sections using operations as separators and performs evaluation processing for each section. This allows for accurate correspondence between the actions of the reference data 33 and the actions of the action data 32, thereby improving the accuracy of the evaluation.

[0082] The operation information acquisition unit 26 shown in Fig. 11 acquires operation information data 67 from the press-fitting device 60 and associates the time associated with the operation information with the elapsed time from the start time of the target period. As shown in Fig. 12, the operation information data 67 is data in which the elapsed time from the evaluation start time is associated with the operation information. The operation information data 67 shown in Fig. 12 indicates that at time to1, an operation to set the descent switch corresponding to step S18 to the ON state was performed, at time to2, an operation to change the descent speed corresponding to step S20 was performed, and at time to3, an operation to set the descent switch corresponding to step S22 to the ON state was performed.

[0083] 11 creates integrated data 36 in which the sensor data 31 and the operation information data 67 are associated with each other so that the time axes of the data correspond to the elapsed time from the evaluation start time. As shown in Fig. 12, the integrated data 36 can associate the head movement of the worker OP indicated by the sensor data 31 with the operation performed by the worker OP.

[0084] 13 is a flowchart showing the procedure of the multiple section evaluation process performed by the evaluation unit 23. As in the first embodiment, when requested by the evaluation program 34, in step S80 shown in FIG. 13, the evaluation unit 23 uses the operation information data 67 to divide the action data 32 into multiple sections, using each piece of operation information as a separator. In the case shown in FIG. 12, step S18 is performed at time to1, step S20 is performed at time to2, and step S22 is performed at time to3. Therefore, the evaluation unit 23 divides the evaluation period into three sections: "Section 1" from the evaluation start time to time to1, "Section 2" from time to1 to time to2, and "Section 3" from time to2 to time to3.

[0085] In step S82, the evaluation unit 23 sets the first section, "Section 1," as the evaluation target. In step S84, the evaluation unit 23 performs the evaluation process shown in FIG. 8 as a sub-processing routine for the section to be evaluated. In step S86, the evaluation unit 23 determines whether the evaluation has been completed for all sections. If it is determined in step S86 that the evaluation has not been completed for all sections, the evaluation unit 23 sets the next section as the evaluation target in step S88 and returns the process to step S84. On the other hand, if it is determined in step S86 that the evaluation has been completed for all sections, the evaluation unit 23 ends this processing routine. Upon completing the multiple section evaluation process, the evaluation unit 23 passes the evaluation data to the evaluation program 34. Furthermore, the evaluation unit 23 performs an evaluation step of evaluating the proficiency of the worker OP using evaluation data indicating the status of the assignment of the "missing action" label and the "unnecessary action" label.

[0086] According to the second embodiment described above, the evaluation system 1 includes an operation information acquisition unit 26 that acquires operation information data 67, and an evaluation unit 23. The evaluation unit 23 uses the operation information data 67 to divide the movement data 32 into a plurality of sections using the operation information as a delimiter, and performs evaluation by setting each of the plurality of sections as a target section. When the movement data 32 includes a plurality of movements associated with the operation of the reception unit 61, the sections are divided using the operations as delimiters, and for each section, the movements in the reference data 33 are associated with the movements in the movement data 32, thereby improving the accuracy of the evaluation.

[0087] C. Third Embodiment: FIG. 14 is a block diagram of an information processing device 10 according to a third embodiment. In this embodiment, in addition to the head movement of the worker OP, the voice uttered by the worker OP is set as a subject for evaluating proficiency. In more detail, in addition to the operation of the reception unit 61, the process may include, as a reference process, a so-called pointing and confirmation call by the worker OP, such as "Speed, good." Therefore, in this embodiment, the voice uttered by the worker OP is also included as a subject for evaluation. The same configurations and processing steps as those in the above embodiments are designated by the same reference numerals, and detailed descriptions thereof will be omitted as appropriate.

[0088] In this embodiment, the information processing device 10 further includes, as functional units of the processor 20, a voice data creation unit 28 and a call data creation unit 29. The storage unit 30 stores sensor data 31, motion data 32, reference data 33, an evaluation program 34, integrated data 36, ​​voice time series data 37, and call data 38.

[0089] The voice data creation unit 28 acquires voice data output from the microphone 83 shown in FIG. 2 and creates voice time series data 37 that associates the acquired voice data with the elapsed time from the start time of the target period. The call data creation unit 29 uses the voice time series data 37 to extract calls from the worker OP and creates call data 38 that associates the extracted calls with the elapsed time. Specifically, the call data creation unit 29 determines that a voice was uttered by the worker OP at the time when the sound pressure level [dB] indicated by the voice time series data 37 exceeded a predetermined reference sound pressure level [dB], and extracts the voice as a call. The call data creation unit 29 may also extract the text information as a "call" using a program that converts general-purpose voice data into text information indicated by the voice.

[0090] The integrated data creation unit 27 creates integrated data 36 in which the actions included in the action data 32 and the calls included in the call data 38 are arranged in chronological order using the elapsed times associated with each. The reference data 33 according to this embodiment includes not only the reference actions but also reference calls that serve as the basis for calls during the target period. The reference actions and reference calls included in the reference data 33 are arranged in chronological order according to the predetermined standard procedure of the process.

[0091] The evaluation unit 23 handles calls included in the integrated data 36 in the same way as actions included in the sensor data 31, and performs evaluation processing in the same way as in the first embodiment. That is, if a reference call included in the reference data 33 is not included in the integrated data 36, ​​the evaluation unit 23 assigns a label of "missing call" to the reference call included in the reference data 33. On the other hand, if a call included in the integrated data 36 is not included in the reference data 33, the evaluation unit 23 assigns a label of "unnecessary call" to the call included in the integrated data 36.

[0092] The evaluation unit 23 may evaluate the action data 32 and the call data 38 using separate processing routines. Specifically, the evaluation unit 23 performs the same evaluation process as in the first embodiment, and when evaluating the call, may evaluate the proficiency by comparing the call data 38 with reference call data that includes a reference call, separately from the reference data 33 that includes a reference action.

[0093] According to the third embodiment described above, the evaluation system 1 includes a voice data creation unit 28, a call data creation unit 29, and an integrated data creation unit 27. The call data creation unit 29 extracts calls from the worker OP using the voice time series data 37 created by the voice data creation unit 28, and creates call data 38 that associates the extracted calls with elapsed time. The integrated data creation unit 27 creates integrated data 36 that chronologically arranges the actions included in the action data 32 and the calls included in the call data 38 using the elapsed time associated with each. The evaluation unit 23 evaluates the proficiency level by comparing the integrated data 36 with the reference data 33. Therefore, voices uttered by the worker OP can be included in the evaluation, further improving the accuracy of the evaluation.

[0094] D. Other Embodiments: (D1) In the first embodiment, the information processing device 10 is located near the press-fitting device 60. In another embodiment, the information processing device 10 may be located at a location remote from the press-fitting device 60. Exchange of information between the information processing device 10, the press-fitting device 60, and the mobile terminal device 90 is performed via a network. Therefore, even if the information processing device 10 is located at a remote location away from the factory where the press-fitting device 60 is located, data can be exchanged by communication via a network including, for example, the Internet. Therefore, proficiency evaluation by the evaluation system 1 can be realized.

[0095] (D2) In the evaluation step of the first embodiment, the evaluation unit 23 evaluates the proficiency level using the "missing action" label and the "unnecessary action" label in the evaluation data. The evaluation unit 23 may further evaluate the proficiency level using the "evaluated" label and the "duration" label. Specifically, points may be added for actions assigned with the "evaluated" label, and points may be subtracted for actions assigned with any of the "duration" label, the "missing action" label, and the "unnecessary action" label, and the total value for all actions may be calculated as the proficiency level. The evaluation unit 23 can visualize the proficiency level by calculating the proficiency level as a value.

[0096] (D3) In the third embodiment, in addition to the detection value of the motion sensor 81, the voice data emitted by the worker OP is evaluated. Furthermore, the detection value of a motion sensor for detecting the arm movements of the worker OP may also be added to the evaluation targets. Specifically, a motion sensor similar to the motion sensor 81 is built into a wristband-type device worn on the wrist of the worker OP. The motion data creation unit 22 then extracts the wrist movements using the detection value of the motion sensor built into the wristband-type device. According to this embodiment, the arm movements of the worker OP can also be added to the evaluation targets, thereby further improving the accuracy of the proficiency evaluation.

[0097] (D4) In the third embodiment described above, in addition to the detection value of the motion sensor 81, the voice data emitted by the worker OP is evaluated. Furthermore, the detection value of a sensor built into the power assist suit worn by the worker OP may also be added to the evaluation targets. Specifically, the power assist suit is a device that assists the movements of the worker OP, for example, the movement of the waist. The sensor built into the power assist suit detects the angle of the waist of the worker OP. Therefore, according to this embodiment, the waist movement of the worker OP can also be added to the evaluation targets, allowing for a more detailed evaluation of the proficiency level.

[0098] (D5) In the first embodiment, the press-fitting device 60 was used as an example of the device. However, the device is not limited to the press-fitting device 60. The evaluation system 1 can be applied to devices that require operation by an operator OP during a process. Furthermore, the evaluation system 1 is more suitable for evaluating operators OP who operate machine tools, for example, whose process time and processing quality vary depending on the operator OP's proficiency. In the first embodiment, the actions that change the direction of the face, such as "checking the display" and "checking the workpiece," were used as examples of actions, but this is not limited to these. For example, actions that change the head position, such as the operator OP crouching, standing up, or walking, can be included as evaluation targets. Actions that change the head position can be extracted using the detection value of the acceleration sensor of the motion sensor 81. The start time and duration of an action may be determined based on the type of action. That is, for example, in the case of walking, the walking action itself may be extracted as an action, or the action of placing the workpiece W after walking may be extracted.

[0099] (D6) The evaluation system 1 may further include a proficiency evaluation unit that evaluates changes in proficiency. Specifically, the evaluation system 1 further includes a sensor that detects the pulse rate of the worker OP and a sensor that detects temperature and humidity. The evaluation unit 23 continuously evaluates the proficiency of the same worker OP, for example, every day for a week. For the same worker OP, the proficiency is unlikely to change significantly in a short period of time. Therefore, a significant change in proficiency may indicate that the worker OP's physical condition is deteriorating or that the temperature and humidity are not appropriate. Therefore, the proficiency evaluation unit calculates the amount of change in proficiency, and if the amount of change is greater than a predetermined standard amount of change, determines whether the pulse, temperature, or humidity is outside the standard range. If the proficiency evaluation unit determines that the pulse, temperature, or humidity is outside the standard range, it sends an alarm, for example, to an information processing device managed by the worker OP's supervisor. This allows the supervisor to know that the worker OP is abnormal. Specifically, the change in proficiency may be the change in a numerical value that quantifies the proficiency, as in (D2) above, or the change in the length of the target period of the action data 32, or the change in the action time of the action in the action data 32.

[0100] (D7) In the first embodiment, both step S34 and step S52 are performed in the evaluation process, but only one of these steps may be performed. By performing at least step S34, it is possible to determine that the movement data 32 does not contain a necessary movement. Similarly, by performing at least step S52, it is possible to determine that the movement data 32 contains an unnecessary movement.

[0101] The present disclosure is not limited to the above-described embodiments and can be realized in various configurations without departing from the spirit thereof. For example, the technical features of the embodiments corresponding to the technical features in each aspect described in the Summary of the Invention section can be appropriately replaced or combined to solve some or all of the above-described problems or achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be appropriately deleted.

[0102] 1...Evaluation system, 10...Information processing device, 20...Processor, 21...Sensor data creation unit, 22...Movement data creation unit, 23...Evaluation unit, 24...Decision unit, 25...Output unit, 26...Operation information acquisition unit, 27...Integrated data creation unit, 28...Speech data creation unit, 29...Call data creation unit, 30...Memory unit, 31...Sensor data, 32...Movement data, 33...Reference data, 34...Evaluation program, 35...Training program, 36...Integrated data, 37...Speech time series data, 38...Call data, 4 0...communication unit, 60...press-fitting device, 61...reception unit, 61a...touch panel, 61b...operation button, 62...fixed unit, 63...movable unit, 65...processor, 66...storage unit, 67...operation information data, 68...communication unit, 80...ear-worn device, 80a...housing, 81...motion sensor, 82...sound conversion unit, 83...microphone, 84...communication unit, 85...data acquisition unit, 86...data output unit, 90...portable terminal device, 91...communication unit, OP...operator, W...workpiece, Wa...press-fitted part, Wb...press-fitted part

Claims

1. An evaluation system comprising: a sensor attached to the head of a worker for detecting head movements; a sensor data creation unit that acquires detection values ​​output from the sensor during a period in which the worker is operating equipment and creates sensor data that correlates the acquired detection values ​​with the elapsed time from a start time of a target period to be evaluated; an action data creation unit that uses the sensor data to extract the movements and creates action data that correlates the extracted movements with the elapsed time; and an evaluation unit that uses the action data to evaluate the proficiency of the worker.

2. An evaluation system as described in claim 1, wherein the evaluation unit evaluates the proficiency by comparing the behavior data with predetermined reference data in which reference actions that serve as the basis for the actions during the target period are arranged in chronological order.

3. An evaluation system as described in claim 2, wherein the reference data includes a plurality of the reference actions, the action data includes a plurality of the actions, and each of the plurality of reference actions included in the reference data is assigned a predetermined action type, the action data creation unit associates the action type with each of the plurality of actions included in the action data, and the evaluation unit performs at least one of the following: a judgment process for determining whether each of the plurality of reference actions included in the reference data can be associated with at least one of the plurality of actions included in the action data having the same action type; a first assignment process for assigning a first action label to the reference action of the reference data that is determined not to be associated with the reference action in the reference data when it is determined in the judgment process that the action cannot be associated; and a second assignment process for assigning a second action label to the action of the plurality of actions in the action data that is not associated with any of the plurality of standard actions in the reference data in the judgment process, and evaluates the proficiency using at least one of the assignment status of the first action label and the assignment status of the second action label.

4. An evaluation system as described in claim 2 or 3, wherein the reference data is the operation data created using the sensor data acquired during a period in which the skilled worker, who is a highly skilled worker, is operating the device.

5. An evaluation system as described in claim 1, wherein the action data includes a plurality of the actions, and the evaluation unit performs at least one of the following: a first time evaluation process for attaching a first time label to each of the plurality of actions included in the action data if the action time from the start of the action to the end of the action is longer than a predetermined reference action time; and a second time evaluation process for attaching a second time label to a target action among the plurality of actions included in the action data if the action interval between a target action and the next action of the target action is longer than a predetermined reference action interval, and evaluates the proficiency using at least one of the attachment status of the first time label and the attachment status of the second time label.

6. An evaluation system as described in claim 1, wherein the device includes a reception unit that receives operations from the worker, and the evaluation system further includes an operation information acquisition unit that acquires operation information data that correlates the operation information received by the reception unit during the target period with the elapsed time, the operation information data including a plurality of pieces of operation information, and the evaluation unit uses the operation information data to divide the motion data into a plurality of sections using each piece of operation information of the plurality of pieces of operation information as a separator, and sets each of the plurality of sections as a target section to perform evaluation.

7. An evaluation system as described in claim 1, further comprising at least one of a sound conversion unit which converts sound data into sound and a display unit which displays information using display data; a determination unit which extracts the action using the sensor data and determines a next reference action which is the reference action to be performed next by the worker by comparing the extracted action with predetermined reference data in which reference actions which are standards for the actions during the target period are arranged in chronological order; and an output unit which performs at least one of a first output process which outputs the sound data which gives an instruction for the next reference action determined by the determination unit to the sound conversion unit, and a second output process which outputs the display data including the information which gives an instruction for the next reference action to the display unit.

8. An evaluation system as described in claim 1, further comprising an ear-worn device to be worn on the worker's ear, the ear-worn device having a data acquisition unit that acquires sound data and a sound conversion unit that converts the acquired sound data into sound and emits it toward the ear, and the sensor is built into the ear-worn device.

9. An evaluation system as described in claim 2, further comprising: a microphone for acquiring voice uttered by the worker as voice data; a voice data creation unit for acquiring the voice data acquired by the microphone during the target period and creating voice time series data in which the acquired voice data is associated with the elapsed time; a call data creation unit for extracting calls of the worker using the voice time series data and creating call data in which the extracted calls are associated with the elapsed time; and an integrated data creation unit for creating integrated data in which the actions included in the action data and the calls included in the call data are arranged in chronological order using the elapsed time associated with each of them, wherein the standard data further includes a reference call that serves as a standard for the calls during the target period, and the reference action and the reference call are arranged in chronological order, and the evaluation unit evaluates the proficiency by comparing the integrated data with the standard data.

10. An evaluation method comprising: a sensor data creation process for acquiring detection values ​​output from a sensor worn on the head of a worker operating an apparatus for detecting head movements, and creating sensor data that correlates the acquired detection values ​​with the elapsed time from the start time of a target period to be evaluated; a motion data creation process for extracting the movements using the sensor data, and creating motion data that correlates the extracted movements with the elapsed time; and an evaluation process for evaluating the proficiency of the worker using the motion data.

Images