Learner support system, learner support method, and program

Abstract

To make it possible to select an expert suitable for learning of a learner from among a plurality of experts.SOLUTION: A work support information creation system 1 according to one aspect of the present invention includes a reference data determination unit 123 that selects an expert to be referred to by a worker from among a plurality of experts based on information on the similarity between motion data 130, which is information on the characteristics of the worker, and each of multiple pieces of expert data, which is information on the characteristics of each of the plurality of experts, and associates the expert with the worker, and a communication unit 121 that outputs presentation information 140 including correspondence information between the worker and the expert.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a learner support system, a learner support method, and a program. [Background technology]

[0002] There are known learner support systems for helping workers efficiently learn tasks. For example, Patent Document 1 discloses a work training system that includes a wearable display, a storage device that stores work data of model workers, and a processing device that generates work information to be displayed on an image display device. The wearable display described in Patent Document 1 includes an image display device that displays work information in front of the eyes of the learner, a gaze detection device that detects the gaze of the learner, and a work action capture device that captures the work actions of the learner.

[0003] Furthermore, Patent Document 1 describes that the processing device of the work training system performs a first-stage process of displaying the model work on an image display device using the work data of the model worker, a second-stage process of detecting the work of the learning worker using an eye gaze detection device and a work motion recording device and acquiring the work data of the learning worker, and a third-stage process of displaying the differences in the work data between the model worker and the learning worker on an image display device.The technology described in Patent Document 1 has the effect of making the differences in the work between the model worker and the learning worker clear for each learning worker, allowing the learning worker to easily grasp the tricks of the work.

[0004] Patent Document 2 discloses an educational support system that allows inexperienced workers engaged in field work to learn the skills of experienced workers. The educational support system includes a measuring device that detects physical quantity data of the movements and postures of a virtual participant performing a specific task, as well as data on the movement of the person's line of sight, and a display device that displays an image of the field of view in a virtual space.

[0005] Furthermore, Patent Document 2 describes that, when the virtual experiencer is an inexperienced worker, the processing device compares the physical quantity data and gaze movement data with teacher data to calculate a difference, and performs processing to determine the proficiency level corresponding to the difference based on a predetermined reference value. Furthermore, Patent Document 2 describes that the processing content of a highly skilled worker is compared with registered content, and if a different operation is performed, the reason is taken into consideration and collected as know-how and reflected in a knowledge database. According to the technology described in Patent Document 2, appropriate content according to the proficiency level is presented to the inspection worker, so that the inspection worker can absorb the inspection know-how of the highly skilled inspection worker and further improve the efficiency of the inspection work. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2018-180090 [Patent Document 2] Japanese Patent Application Publication No. 2019-57099 Summary of the Invention [Problem to be solved by the invention]

[0007] According to the technologies described in the above-mentioned patent documents, a worker (learner) can efficiently learn a task by referring to the content of the task performed by a highly skilled worker. However, there may be situations where there are multiple desirable ways of performing the task by a skilled worker, and if a desirable skilled worker is not appropriately selected in such a situation, the content presented by the system may not be appropriate for the worker. Neither Patent Document 1 nor Patent Document 2 describes selecting a skilled worker suitable for the worker from multiple skilled workers.

[0008] The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to enable a learner to select an expert suitable for learning from among a plurality of experts. [Means for solving the problem]

[0009] A learner support system according to one embodiment of the present invention includes a reference expert selection unit that selects an expert to be referenced by the learner from among a plurality of experts based on information on the similarity between learner data, which is information about the characteristics of the learner, and each of a plurality of expert data, which is information about the characteristics of each of a plurality of experts, and associates the expert with the learner; and an output unit that outputs correspondence information between the learner and the experts. [Effects of the Invention]

[0010] According to the present invention, it becomes possible to select an expert suitable for the learner's learning from among a plurality of experts. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram showing an example of the overall configuration of a work support information creation system according to an embodiment of the present invention; [Figure 2] 1 is a block diagram showing an example of the hardware configuration of a worker terminal and a work support information creation device according to an embodiment of the present invention. [Figure 3] 1 is a schematic diagram showing a state in which a worker wears a clothing-type sensor according to an embodiment of the present invention. FIG. [Figure 4] FIG. 4 is a diagram illustrating an overview of processing by a task feature extraction unit according to an embodiment of the present invention. [Figure 5] 1 is a graph plotting task feature amounts of a worker and task feature amounts of a plurality of experts according to an embodiment of the present invention. [Figure 6] FIG. 10 is a diagram illustrating an example of input and output information for a skill level estimation unit configured by a DNN according to one embodiment of the present invention. [Figure 7] FIG. 10 is a diagram showing an example of the configuration of a skill level advice correspondence table according to an embodiment of the present invention. [Figure 8] 10A and 10B are diagrams illustrating an example of presentation of presentation information generated by an information generation unit in an information presentation unit according to an embodiment of the present invention. [Figure 9]10 is a flowchart showing an example of a procedure for work support information creation processing by the work support information creation system according to one embodiment of the present invention. [Figure 10] 1 is a graph plotting task feature amounts of a worker and task feature amounts of a plurality of experts according to an embodiment of the present invention. [Figure 11] 10 is a diagram showing an example of display of presentation information on an information presentation unit based on the first modified example of the present invention. FIG. [Figure 12] 10 is a graph plotting task feature amounts of a worker and task feature amounts of a plurality of skilled workers acquired when performing a task in a first task process according to a third modified example of the present invention. [Figure 13] 10 is a graph plotting task feature amounts of a worker and task feature amounts of a plurality of skilled workers acquired when performing a task in a second task process according to a third modified example of the present invention. [Figure 14] FIG. 13 is a diagram showing an example of reference expert information history data stored in a database unit according to the fourth modification of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, examples of modes for carrying out the present invention (hereinafter referred to as "embodiments") will be described with reference to the accompanying drawings. However, the present invention should not be interpreted as being limited to the description of the embodiments shown below. Those skilled in the art will easily understand that the specific configuration can be changed within the scope of the idea or purpose of the present invention. In the configuration of the invention described below, the same parts or parts having similar functions are denoted by the same reference numerals in different drawings, and duplicate explanations may be omitted. Furthermore, when there are multiple elements having the same or similar functions, they may be described using the same reference numerals with different subscripts. However, when it is not necessary to distinguish between multiple elements, the subscripts may be omitted in the description.

[0013] The designations "first," "second," "third," etc. in this specification are used to identify components and do not necessarily limit the number, order, or content thereof. Furthermore, numbers used to identify components are used in different contexts, and numbers used in one context do not necessarily indicate the same configuration in another context. Furthermore, this does not prevent a component identified by a certain number from also serving the function of a component identified by another number.

[0014] In order to facilitate understanding of the invention, the position, size, shape, range, etc. of each component shown in the drawings etc. may not represent the actual position, size, shape, range, etc. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings etc. As used herein, elements referred to in the singular are intended to include the plural unless the context clearly indicates otherwise.

[0015] <Overall system configuration> Fig. 1 is a block diagram showing an example of the overall configuration of a work support information creation system 1 according to this embodiment. As shown in Fig. 1, the work support information creation system 1 (an example of a learner support system) includes a worker terminal 110 and a work support information creation device 120. The worker terminal 110 transmits motion data 130 of a worker performing a predetermined task to the work support information creation device 120. The worker is a learner who learns the task by referring to the motions of an expert.

[0016] The work support information creating device 120 creates presentation information 140 as work support content (an example of educational content) using the action data 130 received from the worker terminal 110, and transmits the created presentation information 140 to the worker terminal 110. The worker terminal 110 and the work support information creating device 120 can communicate with each other via a wired or wireless network (not shown).

[0017] [Worker terminal configuration] The worker terminal 110 includes a sensor 111, a communication unit 112, and an information presentation unit 113. The sensor 111 is a sensor that extracts characteristics of work performed by a worker, and is configured, for example, as a clothing-type sensor with an embedded IMU (Inertial Measurement Unit). An example configuration of the sensor 111 configured as a clothing-type sensor will be described with reference to FIG. 3 described later. The sensor 111 may be any device that can extract characteristics of work performed by a worker. The sensor 111 may be configured, for example, as a general camera that can photograph a worker, or a glasses-type device that can measure the worker's line of sight. The sensor 111 outputs a detected value to the communication unit 112 as operation data 130 (an example of worker data). In this embodiment, an example is given in which the sensor 111 detects characteristics of a worker when the worker is performing work of lifting a load.

[0018] The communication unit 112 transmits the worker's action data 130 to the work support information creation device 120 and receives presentation information 140 transmitted from the work support information creation device 120 .

[0019] The information presenting unit 113 is a device that displays the presentation information 140 transmitted from the work support information creating device 120, and is configured, for example, by a display device such as an LED (Liquid Crystal Display). Note that the information presenting unit 113 may also be configured by AR (Augmented Reality) glasses or the like worn by the worker.

[0020] [Configuration of the work support information creation device] The work support information creation device 120 includes a communication unit 121 , a work feature extraction unit 122 , a reference data determination unit 123 , a skill level estimation unit 124 , an information generation unit 125 , and a database unit 126 .

[0021] The communication unit 121 (an example of an output unit) receives the action data 130 transmitted from the worker terminal 110, and transmits (outputs) the presentation information 140 generated by the information generation unit 125 of the work support information creation device 120 to the worker terminal 110.

[0022] The task feature extraction unit 122 estimates and extracts feature quantities to express the task features desired by the worker based on the action data 130 from the worker terminal 110. The task features from which the task feature extraction unit 122 extracts feature quantities can be defined as appropriate by the system designer or the like depending on the purpose of the task evaluation, the type of sensor 111 used, and the like. One or more feature quantities are arbitrarily defined for multiple body parts of the worker Wu. For example, the feature quantities include the position, speed, and bending angle of the worker Wu's elbow, the bending angle of the waist, the direction of the head, and the angle at which the feet are spread.

[0023] The reference data determination unit 123 (an example of a reference expert selection unit) determines data to be referenced for estimating the skill level of a worker based on the feature amounts extracted by the task feature extraction unit 122. The proficiency level estimation unit 124 estimates the proficiency level of the worker by comparing the reference data determined by the reference data determination unit 123 with the feature amounts extracted by the task feature extraction unit 122. The task feature extraction unit 122, the reference data determination unit 123, and the proficiency level estimation unit 124 according to this embodiment can be configured using, for example, a neural network.

[0024] The information generation unit 125 generates work support content including advice for improving the work of the worker, based on the information on the worker's proficiency estimated by the proficiency estimation unit 124. Then, the information generation unit 125 outputs the work support content as presentation information 140.

[0025] The database unit 126 stores data referenced by the reference data determination unit 123, the presentation information 140 generated by the information generation unit 125, various data constituting the presentation information 140, etc. The data referenced by the reference data determination unit 123 includes, for example, information on task features extracted from each of a plurality of experts (an example of expert data, hereinafter also referred to as "task features"). The various data constituting the presentation information 140 includes a proficiency advice correspondence table T1 (see FIG. 7), etc. The proficiency advice correspondence table T1 is a table that stores correspondence information between each proficiency level and advice for improving the task. The proficiency advice correspondence table T1 will be described in detail with reference to FIG. 7 below.

[0026] The worker terminal 110 can be configured with a PC (Personal Computer) or the like, and the work support information creation device 120 can be configured with a general server or the like. Fig. 1 is a diagram showing functional blocks, and the programs executed by the worker terminal 110 and the work support information creation device 120, their functions, or means for realizing those functions may be called "functions," "means," "parts," "units," "modules," etc.

[0027] <Example of computer hardware configuration> Next, the hardware configuration of the device for realizing each function of the worker terminal 110 and the work support information creation device 120 that constitute the work support information creation system 1 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example of the hardware configuration of the worker terminal 110 and the work support information creation device 120. The calculator 50 shown in Fig. 2 is hardware used as a so-called computer.

[0028] The computer 50 includes a control unit 510, a nonvolatile storage 520, a display unit 530, an operation input unit 540, and a communication I / F (Interface) 550, which are all connected to a bus B.

[0029] The control unit 510 includes a CPU (Central Processing Unit) 511, a ROM (Read Only Memory) 512, and a RAM (Random Access Memory) 513.

[0030] The CPU 511 reads out program code of software that realizes each function according to this embodiment from the ROM 512, expands it in the RAM 513, and executes it. Variables, parameters, etc. that are generated during the calculation process are temporarily written to the RAM 513.

[0031] The control unit 510 may include a processing device such as an MPU (Micro-Processing Unit) instead of the CPU 511. Alternatively, the control unit 510 may use both a CPU and an MPU. The control unit 510 may also be configured with hardware such as an FPGA (Field Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit). When the functions of the task feature extraction unit 122, the reference data determination unit 123, and the proficiency level estimation unit 124 are realized by a neural network, the neural network may be implemented in an FPGA or the like.

[0032] The nonvolatile storage 520 may be, for example, a hard disk drive (HDD), a solid state drive (SSD), a flexible disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, or a nonvolatile memory card. In addition to the OS and various parameters, the nonvolatile storage 520 also stores programs for operating the computer 50. The programs may also be stored in the ROM 512.

[0033] The program is stored in the form of a computer-readable program code, and the CPU 511 sequentially executes operations in accordance with the program code. In other words, the ROM 512 or the non-volatile storage 520 is used as an example of a computer-readable non-transitory recording medium that stores a program to be executed by a computer.

[0034] The display unit 530 is, for example, a monitor configured with an LCD (Liquid Crystal Display) or the like, and displays the results of processing performed by the computer 50, etc. The operation input unit 540 is configured with, for example, a keyboard, a mouse, a touch sensor, etc., and generates an operation signal in response to an operation by the user and supplies it to the CPU 511. Display unit 530 and operation input unit 540 may be integrated into one unit as a touch panel.

[0035] The communication I / F 550 may be, for example, a network interface card (NIC), and may transmit and receive various data to and from external devices via a network or communication line.

[0036] Each component constituting the computer 50 shown in FIG. 2 may be configured within a single device, or any part of each component may be distributed and placed within other computers connected via a network not shown.

[0037] <Sensor configuration> Next, a configuration example of a clothing-type sensor 111 worn by an expert will be described. FIG. 3 is a schematic diagram showing a state in which a worker Wu wears the clothing-type sensor 111. As shown in FIG. 3, a plurality of IMUs 211, indicated by black rectangles, and a sensor hub 212 are embedded in the clothing-type sensor 111. Each of the plurality of IMUs 211 is connected to the sensor hub 212 directly or via another sensor 111, and transmits motion data 130 (sensor detection values) of the worker Wu acquired by the sensor to the sensor hub 212. The connection between the sensor 111 and the sensor hub 212 may be realized by either a wired or wireless connection.

[0038] The sensor hub 212 transmits the operation data 130 received from each sensor 111 to the worker terminal 110 (see FIG. 1). Then, the communication unit 112 of the worker terminal 110 transmits the operation data 130 received from the sensor hub 212 to the work support information creation device 120. The IMU 211 and the sensor hub 212 each operate using power supplied from a battery (not shown) or the like. Note that the placement position of the IMU 211 shown in FIG. 2 is an example, and the IMU 211 may be placed in a position different from the placement position shown in FIG. 2. For example, the IMU 211 may be placed only on the upper body clothing of the worker Wu or only on the lower body clothing, etc., depending on the content of the operation to be evaluated.

[0039] <Outline of processing by the task feature extraction unit> Next, an overview of the processing by the task feature extraction unit 122 of the task support information creation device 120 will be described with reference to Fig. 4. Fig. 4 is a diagram showing an overview of the processing by the task feature extraction unit 122. Fig. 4 shows an example in which the task feature extraction unit 122 is equipped with a lumbar state estimation unit 301. The lumbar state estimation unit 301 estimates and outputs the lumbar bending angle from among the task feature amounts of the worker Wu.

[0040] The left side of Fig. 4 shows how output values ​​from four IMUs 211 attached to the shoulders, chest, and waist of the worker Wu are input to the waist state estimation unit 301. The waist state estimation unit 301 estimates the rotation angle of the waist of the worker Wu in the forward / backward direction. If it is desired to have the task feature extraction unit 122 extract feature quantities other than the waist rotation angle, various feature quantity extraction units, such as an elbow rotation angle extraction unit and a head direction extraction unit, can be added to the task feature extraction unit 122.

[0041] The lumbar state estimation unit 301 acquires and outputs time-series data in which the estimated rotation angles of the waist are arranged in time series. The lumbar state estimation unit 301 can be configured with, for example, a deep neural network (DNN), and supervised learning or the like can be applied to the learning of the DNN.

[0042] The graph on the right side of FIG. 4 is a graph showing an example of time-series data of the waist rotation angle output from the waist state estimation unit 301. The vertical axis of the graph represents the waist rotation angle, and the horizontal axis represents time. Note that the feature quantities extracted by the task feature extraction unit 122 do not have to be time-series data, and may be composed of statistics such as the mean or variance of the time-series data. Furthermore, the feature quantities extracted by the task feature extraction unit 122 are not limited to human-interpretable feature quantities such as the waist bending angle. The feature quantities extracted by the task feature extraction unit 122 may be, for example, feature quantities obtained as a result of principal component analysis of sensor data obtained from the IMU 211 or as a result of direct analysis using a known statistical analysis method such as an autoencoder.

[0043] <Outline of processing by the reference data determination unit> Next, an outline of the processing by the reference data determination unit 123 of the work support information creation device 120 will be described. The reference data determination unit 123 compares the feature amounts extracted by the task feature extraction unit 122 with the task feature amounts of the worker Wu stored in advance in the database unit 126, thereby determining the skill data that the worker should refer to.

[0044] The reference data determination unit 123 first calculates the distance between the feature of the worker Wu extracted by the task feature extraction unit 122 and each of the feature of the multiple skilled workers Ws stored in the database unit 126. The distance between the feature of the worker Wu and each of the feature of the multiple skilled workers Ws is expressed, for example, by Euclidean distance. Note that the distance between the feature of the worker Wu and each of the feature of the multiple skilled workers Ws may be calculated using another distance function. Alternatively, the distance between the feature of the worker Wu and each of the feature of the multiple skilled workers Ws may be calculated using a distance function that measures the distance between features and is calculated using a known machine learning method. Furthermore, when time-series data is used as the task feature, the distance between the feature of the worker Wu and each of the feature of the multiple skilled workers Ws may be calculated using a known method and is the distance between features that takes time change into account. In the following description, a case where Euclidean distance is used as the distance between features will be exemplified. Furthermore, the type of feature amount to be compared between the worker Wu and the skilled worker Ws is not limited to one type, but may be multiple types.

[0045] FIG. 5 is a graph plotting task feature values ​​of a worker Wu and task feature values ​​of multiple skilled workers Ws. The vertical axis of the graph represents the average value of the time-series data of the elbow position of the worker Wu, and the horizontal axis represents the average value of the time-series data of the waist rotation angles of the worker Wu and the skilled workers Ws. In the example shown in FIG. 5, the feature value having the closest Euclidean distance (highest similarity) to the feature value of the worker Wu is the feature value of the skilled worker Ws2. Therefore, the reference data determination unit 123 determines that the skilled worker Ws that the worker Wu should refer to for the task is the skilled worker Ws2. Since task feature values ​​are considered to reflect the task method (features) of the skilled worker Ws, a skilled worker Ws with a close feature distance can be assumed to be a skilled worker who performs the task in a manner similar to that of the worker Wu. Therefore, by referring to a skilled worker Ws with a task method similar to that of the worker Wu, the worker Wu can improve his or her task skills more quickly.

[0046] <Outline of processing by the skill level estimation unit> Next, an overview of the processing performed by the proficiency estimation unit 124 of the work support information creation device 120 will be described. The proficiency estimation unit 124 estimates and determines the proficiency of the worker Wu based on the task feature amounts obtained by the task feature extraction unit 122 and the feature amounts of the expert to be referenced determined by the reference data determination unit 123. The proficiency estimation unit 124 can estimate the proficiency of the worker Wu, for example, based on information about the distance between each feature amount of the expert Wu calculated by the reference data determination unit 123 and the feature amount of the worker Wu. Alternatively, the proficiency estimation unit 124 can be configured with a DNN or the like that receives the feature amounts of the worker Wu and the feature amounts of the expert Ws as input and outputs the proficiency level (level) of the worker Wu. For example, supervised learning can be applied to the learning of the DNN.

[0047] FIG. 6 is a diagram showing an example of input / output information for the proficiency estimation unit 124 configured by a DNN. FIG. 6 shows how the feature amounts of the worker Wu and the feature amounts of the expert Ws2 are input to the proficiency estimation unit 124, and how the proficiency information of the worker Wu is output from the proficiency estimation unit 124. The proficiency levels shown in FIG. 6 are, from top to bottom, "beginner," "intermediate," "advanced," and "expert equivalent." Note that the types of proficiency levels of the worker Wu are not limited to the example shown in FIG. 6.

[0048] 6 shows an example in which the proficiency level estimation unit 124 estimates the proficiency level of the worker Wu based on information about one expert Ws determined by the reference data determination unit 123, but the present invention is not limited to this. The reference data determination unit 123 may output information about multiple experts Ws, and the proficiency level estimation unit 124 may estimate the proficiency level of the worker Wu based on information about the multiple experts Ws.

[0049] <Outline of processing by the information generation unit> Next, an outline of the processing by the information generation unit 125 of the work support information creation device 120 will be described. The information generation unit 125 receives as input the task feature amount of the worker Wu extracted by the task feature extraction unit 122 and the proficiency level estimated by the proficiency level estimation unit 124, searches the proficiency level advice correspondence table T1, and outputs information (advice) for improving the task of the worker Wu obtained as a result of the search.

[0050] Fig. 7 is a diagram showing an example of the configuration of a proficiency advice correspondence table T1. In the proficiency advice correspondence table T1 shown in Fig. 7, each level of proficiency is associated with various types of advice. In the example shown in Fig. 7, there are four types of proficiency levels: "beginner," "intermediate," "advanced," and "expert equivalent," and there are three types of advice: "advice regarding waist angle," "advice regarding elbow position," and "general advice."

[0051] In the top row of the proficiency level advice correspondence table T1 shown in Figure 7, the "advice regarding waist angle" is associated with the "beginner" level of proficiency, with the advice "Work carefully and lift the load close to your body. It's better to bend your waist more (more / less) as you work." Also, the "beginner" level is associated with the "advice regarding waist angle" advice "Work carefully and hold the load closer to your body to reduce strain on your arms. It will be easier to work if you keep your elbows higher / lower." Also, the "beginner" level is associated with the "advice regarding elbow position" advice "Be careful of accidents while working and be aware of safety" and "Squat down properly and lift the load by pulling it closer to your body."

[0052] There are advice options such as "It would be better to work with your waist bent more (more / less)", and the information generation unit 125 can select the more preferable advice based on the information on the work features of the worker Wu extracted by the work feature extraction unit 122.

[0053] Fig. 8 is a diagram showing an example of presentation information 140 generated by information generation unit 125, presented by information presentation unit 113 (see Fig. 1). Fig. 8 shows an example in which presentation information 140, which is advice for beginners, is displayed on the screen of information presentation unit 113. On the screen of information presentation unit 113 shown in Fig. 8, advice for beginners regarding waist angle and advice regarding elbow position are displayed from top to bottom under the heading "Tips for Improvement". In addition, to the left of the advice, an illustration of an ideal posture during work corresponding to the advice is shown.

[0054] Note that the presentation information 140 generated by the information generation unit 125 is not limited to the example shown in FIG. 8 . For example, the presentation information 140 may indicate differences between the feature amounts estimated by the reference data determination unit 123 and those of an expert. Furthermore, in the present embodiment, an example has been given in which the information generation unit 125 determines advice based on the proficiency-advice correspondence table T1, but the present invention is not limited to this. For example, the information generation unit 125 may be configured with a large-scale language model (LLM) or the like. When the information generation unit 125 is configured with a large-scale language model, the information generation unit 125 can receive as input the task feature amounts of the expert Ws to be referenced and the estimated proficiency of the worker Wu, and generate and output a message (advice) to be presented to the worker Wu based on the learning content.

[0055] <Work support information creation process> Next, a work support information creating method (one example of a learner support method) by the work support information creating system 1 according to this embodiment will be described. Fig. 9 is a flowchart showing an example of the procedure of work support information creating processing by the work support information creating system 1.

[0056] First, the communication unit 112 (see FIG. 1) of the worker terminal 110 transmits action data 130 including the features of the work of the worker Wu detected by the sensor 111 to the work support information creation device 120 (step S1). Next, the task feature extraction unit 122 of the work support information creation device 120 calculates the task feature amount of the worker Wu from the action data 130 received from the worker terminal 110 (step S2).

[0057] Next, the reference data determination unit 123 of the work support information creation device 120 selects, from the task feature quantities of the multiple experts Ws stored in the database unit 126, the task feature quantity that is closest (shortest distance) to the task feature quantity of the worker Wu calculated by the task feature extraction unit 122 (step S3). Next, the proficiency estimation unit 124 of the work support information creation device 120 estimates the proficiency of the worker Wu based on the feature quantities of the worker Wu extracted by the task feature extraction unit 122 and the feature quantities of the expert Ws selected by the reference data determination unit 123 (step S4).

[0058] Next, the information generation unit 125 of the work support information creation device 120 generates advice useful for improving the work by the worker Wu, based on the proficiency of the worker Wu estimated by the proficiency level estimation unit 124 and the task feature amount of the worker Wu calculated by the task feature extraction unit 122. Then, the information generation unit 125 transmits the generated advice as presentation information 140 to the worker terminal 110 via the communication unit 121 (step S5). Next, the information presentation unit 113 of the worker terminal 110 displays the presentation information 140 transmitted from the work support information creation device 120 (step S6). After the processing of step S6, the work support information creation process by the work support information creation system 1 ends.

[0059] In the work support information creation system 1 according to the embodiment described above, the reference data determination unit 123 selects a skilled worker Ws to be referenced by the worker Wu from among multiple skilled workers based on information on the similarity between the action data 130, which is information about the characteristics of the worker Wu, and each of the multiple skilled worker data, which is information about the characteristics of each of the multiple skilled workers Ws, and associates the selected skilled worker Ws with the worker Wu. Then, the communication unit 121 outputs presentation information 140 including information on the association between the worker Wu and the skilled worker Ws. Therefore, according to this embodiment, the skilled worker Ws that the learner worker Wu should aim to become is made clear, and the worker Wu can improve his or her own work based on the work of the skilled worker Ws to be referenced.

[0060] Furthermore, in the above-described embodiment, the proficiency estimation unit 124 estimates the work proficiency of the worker Wu by comparing the worker data with the expert data selected by the reference data determination unit 123. Then, the reference data determination unit 123 creates educational content to be presented to the worker Wu based on the information on the work proficiency of the worker Wu estimated by the proficiency estimation unit 124. Therefore, according to this embodiment, educational content according to the proficiency of the worker Wu is created and presented to the worker Wu, so that the worker Wu can efficiently improve his or her work by referring to educational content whose content matches his or her proficiency.

[0061] <Various modified examples> [Variation 1] In the above-described embodiment, an example has been given in which the reference data determination unit 123 selects only one feature of a plurality of experts Ws as the feature of the expert Ws to be referred to by the worker Wu, in accordance with the task feature of the worker Wu. However, a situation may be assumed in which it is difficult to uniquely determine the expert Ws to be referred to by the worker Wu. In Modification 1, the reference data determination unit 123 selects feature values ​​of the plurality of experts Ws. Then, the information generation unit 125 generates a plurality of pieces of advice based on the feature values ​​of the plurality of experts Ws selected by the reference data determination unit 123.

[0062] FIG. 10 is a graph plotting task feature amounts of a worker Wu and task feature amounts of multiple skilled workers Ws. The vertical axis of the graph represents the average value of the time-series data of the elbow position of the worker Wu, and the horizontal axis represents the average value of the time-series data of the waist rotation angles of the workers Wu and the skilled workers Ws. In the example shown in FIG. 10, the distance between the feature amounts of the worker Wu and the skilled workers Ws2 is approximately equal to the distance between the feature amounts of the worker Wu and the skilled workers Ws3. In such a case, it is difficult for the reference data determination unit 123 to determine which of the skilled workers Ws's feature amounts should be selected.

[0063] In the first modification, the reference data determination unit 123 selects both the expert Ws2 and the expert Ws3 as the experts Ws to be referred to by the worker Wu. Then, the information generation unit 125 includes, in the presented information 140, both the advice for improving the work based on the feature amount of the expert Ws2 and the advice for improving the work based on the feature amount of the expert Ws3.

[0064] FIG. 11 is a diagram showing an example of the display of presentation information 140 on the information presentation unit 113 based on the first modification. In FIG. 11, messages such as "Tips for improvement" and "Please refer to the following two pieces of advice, whichever you find easiest," are displayed. Below these messages, two pieces of advice generated based on the feature amounts of the two experts Ws are displayed. The advice on the left is "It would be better to work with your waist bent more deeply. If you lower your elbows, it will be easier to work." On the other hand, the advice on the right is "It would be better to work with your waist bent less deeply. If you raise your elbows, it will be easier to work." By viewing such a display, the worker Wu can refer to the advice that he finds easier to adopt and efficiently improve his work.

[0065] 11 shows an example in which the work instructions included in the multiple pieces of advice are different from each other, but the present invention is not limited to this. The work instructions included in the multiple pieces of advice may be partially common. For example, each piece of advice may include a common instruction such as "bend your waist deeply."

[0066] In the first modification, when there are multiple feature quantities of the expert Ws that are similar in distance to the feature quantities of the worker Wu, multiple pieces of advice are generated and presented based on the feature quantities of the multiple experts W. Therefore, according to the first modification, it becomes possible to more efficiently create work assistance information for the worker Wu without narrowing the scope of future learning for the worker Wu.

[0067] [Variation 2] In the above-described embodiment and modification 1, an example has been given in which the reference data determination unit 123 determines the feature quantities of the expert Ws to be referenced in accordance with the task feature quantities of the worker Wu. However, it is also conceivable that a more efficient creation of task support information can be implemented by determining a desirable expert Ws by taking into account not only the task feature quantities but also information such as the physical features and the learning process of the worker Wu.

[0068] In the second modification, the database unit 126 stores in advance data on the physical characteristics of the worker Wu and time-series data showing changes in the task feature quantities of the worker Wu. Similarly, for the skilled worker Ws, the database unit 126 stores the physical characteristics of the skilled worker Ws and time-series data showing changes in the task feature quantities over a period until the skilled worker Ws becomes proficient at the task. The physical characteristics are physical characteristics that are thought to have some effect on the task, such as height, weight, BMI (Body Mass Index), gender, arm length, and leg length.

[0069] In the second modification, the reference data determination unit 123 acquires time-series data of the physical characteristics and task feature amounts of each of the worker Wu, the skilled worker Ws2, and the skilled worker Ws3 from the database unit 126. Then, the reference data determination unit 123 uses the acquired time-series data of the physical characteristics and task feature amounts to determine the skilled worker Ws that the worker Wu should refer to. The reference data determination unit 123 can select the most similar feature amount of the skilled worker Ws based on the degree of similarity with the time-series data of any one or more types of physical characteristics or task feature amounts.

[0070] According to variant example 2, the skilled worker Ws that the worker Wu should refer to is selected taking into account information on physical characteristics that are expected to have some effect on the work, so that advice on work that is easier for the worker Wu to master can be provided to the worker Wu as presented information 140.

[0071] The reference data determination unit 123 according to the second modification may select each of the experts Ws based on the task feature amounts and the physical characteristics, and present information about each expert Ws to the worker Wu. If the expert Ws selected based on the task feature amounts and the expert Ws selected based on the physical characteristics are the same, the reference data determination unit 123 associates one expert Ws with the worker Wu. If the expert Ws selected based on the task feature amounts and the expert Ws selected based on the physical characteristics are different, the reference data determination unit 123 associates multiple experts Ws with the worker Wu. In this case, the reference data determination unit 123 may also add information indicating whether the expert Ws was selected based on the task feature amounts or the physical characteristics to the presented information 140. By performing such processing by the reference data determination unit 123, the worker Wu can select for himself an expert Ws who is easier to adopt the task.

[0072] [Variation 3] In the above-described embodiment and Modification 1, an example was given in which the reference data determination unit 123 determines the feature quantities of the expert Ws to be referenced in accordance with the task feature quantities of the worker Wu. However, when the task to be trained can be further decomposed into multiple task steps, a situation may arise in which it is not necessarily desirable to uniquely determine the expert Ws to be referenced. In Modification 3, the reference data determination unit 123 determines the expert Ws to be referenced for each of the multiple task steps. Then, the information generation unit 125 generates advice corresponding to each of the multiple task steps based on information on each feature quantity of the determined expert Ws.

[0073] Fig. 12 is a graph plotting the task feature of the worker Wu acquired when performing the task in the first task process and each task feature of the multiple skilled workers Ws. Fig. 13 is a graph plotting the task feature of the worker Wu acquired when performing the task in the second task process and each task feature of the multiple skilled workers Ws. In Figs. 12 and 13, the vertical axis of the graph represents the average value of the time series data of the elbow position of the worker Wu, and the horizontal axis represents the average value of the time series data of the waist rotation angle of the worker Wu and the skilled workers Ws.

[0074] In the first work process shown in FIG. 12, the feature quantity closest to the feature quantity of the worker Wu is the feature quantity of the expert Ws1. On the other hand, in the second work process shown in FIG. 13, the feature quantity closest to the feature quantity of the worker Wu is the feature quantity of the expert Ws3. Therefore, it is desirable for the worker Wu to refer to the actions of the expert Ws1 in the first work process, and to refer to the actions of the expert Ws3 in the second work process. Therefore, in Modification 3, the reference data determination unit 123 selects the expert Ws1 as the expert Ws to be referenced in the first work process, and selects the expert Ws3 as the expert Ws to be referenced in the second work process. Then, the information generation unit 125 generates advice corresponding to each of the first work process and the second work process based on information on each feature quantity of the determined expert Ws.

[0075] In Modification 3, an optimal expert Ws to be referred to is selected for each work process, and advice for improving the work corresponding to each work process is generated and provided based on information on the feature quantities of the selected expert Ws. Therefore, according to Modification 3, even if work needs to be performed in which the work content varies greatly for each work process, advice useful for improving the work in that work process is provided to the worker Wu. Therefore, Modification 3 makes it possible to create work support information more efficiently.

[0076] [Variation 4] In the above-described embodiment, an example has been given in which the reference data determination unit 123 determines one feature of the expert Ws to be referenced in accordance with the task feature of the worker Wu. However, it is also conceivable that the expert Ws to be referenced by the worker Wu may change as the worker Wu becomes more proficient at the task through daily repetition of the task. In Modification 4, the reference data determination unit 123 periodically reselects the expert Ws to be referenced. Information about the periodically re-determined expert Ws is then stored in the database unit 126 as history data. The information generation unit 125 regenerates advice for improving the task of the worker Ws based on the information about each feature of the expert Ws determined by the reference data determination unit 123.

[0077] Fig. 14 is a diagram showing an example of the reference expert information history data D1 stored in the database unit 126. The reference expert information history data D1 is history information of the determination of the expert Ws to be referred to by the worker Wu by the reference data determination unit 123. In the reference expert information history data D1 shown in Fig. 14, information on the expert Ws determined by the reference data determination unit 123 is recorded in association with information on the number of days.

[0078] For example, suppose that an expert Ws2 was initially associated with a worker Wu as a reference expert. However, as the worker Wu continues working while referring to presentation information created based on the features of the expert Ws, the expert Ws whose distance to the features of the worker Wu changes from expert Ws2 to expert Ws1. In the example shown in FIG. 14 , the expert Ws whose feature distance to the features of the worker Wu is closest is expert Ws2 until the 48th day, and changes to expert Ws1 from the 49th day onwards. In this case, the information presenter 113 generates advice based on the features of the expert Ws1 as the presentation information 140 from the 49th day onwards.

[0079] According to the fourth modification, the most suitable skilled worker Ws is reselected depending on the level of skill of the worker Wu in the task, so that task support information can be created more efficiently.

[0080] In the above-described embodiment and various modified examples, an example has been given in which the reference data determination unit 123 creates educational content as the presentation information 140, but the present invention is not limited to this. The reference data determination unit 123 may create, as the presentation information 140, correspondence information between the worker Wu and the skilled worker Ws that the worker Wu should refer to for the work. The created presentation information 140 is transmitted to the worker terminal 110 via the communication unit 121 and displayed on the information presentation unit 113 of the worker terminal 110, allowing the worker Wu to grasp information about the skilled worker Ws that the worker Wu should refer to for the work.

[0081] Furthermore, in the above-described embodiment and various modified examples, the task feature extraction unit 122 extracts task feature amounts of the worker Wu when performing a specific task such as lifting a load. However, the present invention is not limited to this. For example, the task feature extraction unit 122 may extract feature amounts for general tasks in industry, agriculture, etc., or for actions performed by a person moving their body for a specific purpose, such as dancing, gymnastics, or playing a musical instrument. Therefore, the learner support system of the present invention can be applied to, for example, a worker support system used in maintenance work training, an educational system used for practicing dance or yoga poses, and the like.

[0082] Furthermore, the above-described embodiments and variants provide detailed and specific descriptions of the configurations of the devices and systems in order to clearly explain the present invention, and are not necessarily limited to those having all of the configurations described.

[0083] 6, the control lines or information lines shown by solid lines are those considered necessary for explanation, and do not necessarily show all control lines or information lines in the product. In reality, it can be considered that almost all components are interconnected.

[0084] Furthermore, in this specification, processing steps describing chronological processing include not only processing that is performed chronologically in the order described, but also processing that is not necessarily performed chronologically but is performed in parallel or individually (for example, parallel processing or processing by objects). [Explanation of symbols]

[0085] 1...work support information creation system, 110...worker terminal, 111...sensor, 112...communication unit, 113...information presentation unit, 120...work support information creation device, 121...communication unit, 122...work feature extraction unit, 123...reference data determination unit, 124...proficiency level estimation unit, 125...information generation unit, 126...database unit, 130...motion data, 140...presentation information, 211...IMU, 212...sensor hub, 301...lower back state estimation unit

Claims

1. a reference expert selection unit that selects an expert to be referenced by the learner from among a plurality of experts based on information on the similarity between learner data, which is information on the characteristics of the learner, and each of a plurality of expert data, which is information on each of the characteristics of a plurality of experts, and associates the expert with the learner; an output unit that outputs correspondence information between the learner and the expert; Learner support system.

2. the learner data includes task features that are features of the learner's movements when the learner performs a predetermined task; The expert data includes task feature amounts that are feature amounts of the expert's movements when the expert performs a predetermined task. The learner support system according to claim 1 .

3. The system further includes a proficiency level estimation unit that estimates the proficiency level of the learner in the task by comparing the learner data with the expert data selected by the reference expert selection unit. The learner support system according to claim 2 .

4. The reference expert selection unit creates educational content to be presented to the learner based on the information on the learner's proficiency in the movement estimated by the proficiency estimation unit. The learner support system according to claim 3 .

5. a sensor for acquiring information about the learner's movements when the learner performs a predetermined task; a task feature extraction unit that extracts the task feature amount of the learner from the information acquired by the sensor, The learner support system according to claim 4.

6. the learner data includes information regarding physical characteristics of the learner; The expert data includes information about the physical characteristics of the expert. The learner support system according to claim 4.

7. The reference expert selection unit selects a plurality of experts to be referenced by the learner, and associates the selected plurality of experts with the learner. The learner support system according to claim 4.

8. The reference expert selection unit selects the expert corresponding to each of the task feature amount and the physical feature amount and associates the expert with the learner. The learner support system according to claim 6.

9. The reference expert selection unit associates the expert with the learner based on the task feature amount of the learner and the task feature amount of the expert for each of a plurality of steps constituting the task performed by the learner. The learner support system according to claim 4.

10. the learner data includes time-series data of the task feature amount of the learner who performed a predetermined task, The expert data includes time-series data of the task feature amounts in the process in which the expert mastered the predetermined task. The learner support system according to claim 4.

11. A learner support method using a learner support system including a reference expert selection unit and an output unit, a step in which the reference expert selection unit selects an expert to be referenced by the learner from among a plurality of experts based on information on the similarity between learner data, which is information on the characteristics of the learner, and each of a plurality of expert data, which is information on each of the characteristics of a plurality of experts, and associates the expert with the learner; and a step of outputting correspondence information between the learner and the expert by the output unit. Learner support methods.

12. A computer constituting a learner support system including a reference expert selection unit and an output unit, a step in which the reference expert selection unit selects an expert to be referenced by the learner from among a plurality of experts based on information on the similarity between learner data, which is information on the characteristics of the learner, and each of a plurality of expert data, which is information on each of the characteristics of a plurality of experts, and associates the expert with the learner; a procedure for the output unit to output correspondence information between the learner and the expert; program.

Images