Work improvement support apparatus, work improvement support system, and work improvement support method, and program

Abstract

To support determining a preferable task to be selected for improvement.SOLUTION: In a work improvement support system 1, a work improvement support apparatus includes: a calculation unit which calculates, based on a video obtained by imaging a task and a work video file including data indicating a distance between a worker and a target object, a representative value of work time for each task, a representative value of distance for each task, and the number of same-distance tasks which is the number of tasks for which the representative values of the distances are the same, and selects, as task candidates for improvement, the tasks corresponding to the largest number of same-distance tasks; an improvement target task specifying unit which specifies a task having the largest representative value of work time, as a task to be improved, out of the task candidates for improvement; and an output unit which outputs information representing the task to be improved. The calculation unit preferentially selects, from among tasks corresponding to the largest number of same-distance tasks, tasks not included in a task management table that specifies the contents of tasks, as task candidates for improvement. The improvement target task specifying unit specifies a task having the largest representative value of work time, as a task to be improved.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an activity improvement support device, an activity improvement support system, an activity improvement support method, and a program. [Background technology]

[0002] There are known techniques for providing support for improving the productivity of production processes. For example, Patent Document 1 discloses a method for inspecting parts in inspection processes with low loads by deleting overlapping parts to be inspected from inspection targets in order of inspection process load, starting from the inspection process with the highest load, from the viewpoint of improving productivity. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2003-258499 A Summary of the Invention [Problem to be solved by the invention]

[0004] In a production site, similar tasks may be customarily performed in multiple processes in an overlapping manner. For example, tasks not listed in a work management table that lists tasks to be performed in each process at the production site may be customarily performed. In this case, it is required to identify tasks not listed in the work management table as targets for improvement. However, the technology of Patent Document 1 focuses on inspection contents and parts to be inspected that are specified in advance, and considers deleting overlapping parts to be inspected from the inspection targets, but does not consider unspecified inspection contents and parts to be inspected. For this reason, support is required to more easily determine which tasks, including tasks not specified in advance, should be targets for improvement.

[0005] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a work improvement support device, a work improvement support system, a work improvement support method, and a program that support easily determining which tasks should be targeted for improvement. [Means for solving the problem]

[0006] In order to achieve the above object, the task improvement support device according to the present disclosure is a task improvement support device that supports the improvement of a task for producing a product, and includes a calculation means for calculating a representative value of task time and a representative value of distance for each task, as well as the number of tasks with the same representative distance, which is the number of tasks with the same representative distance, based on a task video file including a video of the tasks included in each process of producing a product and data indicating the distance between the worker performing the task and the target item being worked on, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; The system includes an improvement target work identification means for identifying an improvement target work as an industry, and an output means for outputting information representing the improvement target work identified by the improvement target work identification means, and the calculation means selects, from the works counted as the maximum number of same-distance works, works not listed in the work management table in accordance with data showing the contents of predetermined works, as candidates for improvement in priority over works listed in the work management table, and when there are candidates for improvement selected with priority, the improvement target work identification means identifies, from the candidates for improvement selected with priority, the work with the largest representative value for work time as the work to be improved. Effect of the Invention

[0007] According to the present disclosure, the calculation means selects, from among the tasks counted as the largest number of tasks over the same distance, tasks not listed in the task management table, as candidate tasks to be improved, in preference to tasks listed in the task management table, based on the task management table in which data indicating the contents of predetermined tasks is listed, and identifies, from among the candidate tasks to be improved selected with priority, the task with the largest representative value of task time as the task to be improved. This makes it possible to more easily support the determination of which tasks are desirable to be targeted for improvement, including tasks not specified in advance. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram showing a configuration example of an operation improvement support system according to a first embodiment of the present disclosure. [Diagram 2] A diagram to explain the distance between the worker's face and the target item [Diagram 3] FIG. 1 is a diagram showing an example of a QC process chart used in the improvement target task identification process executed by the task improvement support system according to the first embodiment of the present disclosure. [Figure 4] FIG. 1 is a diagram showing an example of a QC process chart used in the improvement target task identification process executed by the task improvement support system according to the first embodiment of the present disclosure. [Diagram 5] FIG. 1 is a diagram showing an example of a QC process chart used in the improvement target task identification process executed by the task improvement support system according to the first embodiment of the present disclosure. [Figure 6] FIG. 1 is a diagram showing an example of a QC process chart used in the improvement target task identification process executed by the task improvement support system according to the first embodiment of the present disclosure. [Figure 7] A diagram showing the average work time for each process and the distance between the worker's face and the product. [Figure 8] A flowchart of the process of identifying an improvement target task executed by the task improvement support system shown in FIG. [Figure 9] A flowchart of a specific process executed by the work improvement support system shown in FIG. [Figure 10] FIG. 1 is a diagram illustrating an example of a hardware configuration of an operation improvement support device according to a first embodiment of the present disclosure. [Figure 11] A block diagram showing a configuration example of an operation improvement support system according to a second embodiment. [Figure 12] FIG. 1 shows an overview of a neural network according to a second embodiment. [Figure 13] Flowchart of learning process according to the second embodiment [Figure 14] Flowchart of inference processing according to the second embodiment DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] (Embodiment 1) Hereinafter, an activity improvement support device, an activity improvement support system, an activity improvement support method, and a program according to embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same reference numerals are used to denote the same or corresponding parts.

[0010] The work improvement support system 1 according to the first embodiment takes videos of all work in a production line and identifies work that should be improved as a priority, in other words, overlapping work, thereby supporting work improvement that takes into consideration the elimination of overlapping work. Note that work that needs to be improved as a priority over other work will be referred to as the work to be improved hereinafter. Here, work refers to the work carried out in each process of the production line, and broadly includes, for example, transportation, conveyance, processing, machining, assembly, inspection, packaging, removal, etc.

[0011] As shown in FIG. 1, the work improvement support system 1 comprises a photography device 100 that photographs work at a production site, a work improvement support device 200 that identifies work to be improved based on the average work time for each work and the average distance between the worker and the work item, and a display device 300 that displays information on the identified work to be improved.

[0012] In the following explanation, it is assumed that the manufacturing line is made up of four processes, with each process in a different work area. The four processes are a parts arrival work process, an assembly work process, a pre-shipment inspection work process, and a packaging work process. One worker is in charge of each process, so a total of four workers perform all tasks included in the processes they are in charge of. For example, if worker A is in charge of the parts arrival work process, which is made up of three tasks, worker A will perform all three tasks.

[0013] One imaging device 100 is placed at each process. In this embodiment, the imaging devices 100 are attached to the sides of the eyes of four workers, for example, at positions on their work caps, helmets, etc., where they can capture the distance from the product. The imaging devices 100 include wearable cameras and cameras equipped in smart glasses.

[0014] The image capturing device 100 is an image capturing device that captures images of work including the work object, tools, measuring instruments, etc., which are the subject of the work, and is, for example, a video camera. In this embodiment, a wearable video camera is attached to the side of the eye of each worker at each of the four process work stations. Therefore, in this embodiment, there are a total of four image capturing devices 100.

[0015] The image capturing device 100 captures images of the manufacturing line at all times while it is in operation, and generates live view images. The image capturing device 100 stores the external shape of the work target in an internal memory together with the identification number of the work target. The image capturing device 100 identifies the work target based on the live view images and pre-registered images, for example, by using an image recognition technique such as a pattern matching method.

[0016] Furthermore, software capable of estimating the distance to a detected object is built into the image capturing device 100. The image capturing device 100 recognizes a work target Tp and estimates a distance L to the recognized work target Tp. Any method can be used to estimate the distance L, but for example, the distance L can be obtained from the size of the image of the target item Tp in the captured image. An example will be described with reference to FIG. 2. The upper part of FIG. 2 illustrates the positional relationship between the worker HM and the target item Tp during three tasks: measuring dimensions with a scale, visually inspecting the appearance, and inspecting with an inspection machine. The bold frames in the lower part of FIG. 2 show images during each of the three tasks. The distance L between the imaging device 100 and the target item Tp changes depending on the difference in the task, and the size of the image of the target item Tp in the image changes according to the change in distance L. The imaging device 100 estimates the distance L based on the actual size of the target item Tp and the proportion of the target item Tp in the image indicated by the bold line.

[0017] In the following description, in the case of visual appearance inspection, the distance L between the face of the worker HM and the work target Tp is estimated to be 40 cm (centimeters). The distance L is estimated in units of 10 cm. In addition, in the case of measuring dimensions with a scale, the worker HM will bend forward more than in the case of visual appearance inspection, so the proportion of the image occupied by the size of the target Tp in the image will be larger, and the distance L is estimated to be 10 cm, which is smaller than the distance in the case of visual appearance inspection. In addition, in the case of inspection work in which an inspection machine IM is placed between the worker HM and the work target Tp, the distance L is larger than in the case of visual appearance inspection, and the proportion of the image occupied by the size of the target Tp in the thick frame will be smaller. The distance L during inspection with the inspection machine is estimated to be 100 cm. In addition, in the image in the thick frame shown in Figure 2, only the work target Tp is shown, and the scale, tools including the inspection machine, measuring instruments, etc. are omitted. When the work in each process moves to the next work, the value of the distance L changes significantly. The distance L is estimated at a predetermined time interval.

[0018] Before starting work, the worker HM in charge of each process registers the identification number of the worker HM in the image capturing device 100 worn by the worker HM. The worker HM also registers an image of the work target Tp, the identification number of the work target Tp, and the identification number of the work process he or she is in charge of in the image capturing device 100. The image capturing device 100 associates the captured video with the time measured by a built-in clock and the distance L estimated at each time, and accumulates the video in its own memory as a work video file. The image capturing device 100 adds its own identification number, the identification number of the worker HM, the identification number of the work target Tp, and the identification number of the process to each work video file.

[0019] As shown in Fig. 1, the photographing device 100 is communicatively connected to the work improvement support device 200. When the work improvement support device 200 is started, the photographing device 100 transmits the work video file stored in the memory to the work improvement support device.

[0020] The work improvement support device 200 includes a calculation processing unit 210 that executes processing to acquire or generate various data, and a storage unit 220 that stores data necessary for the processing performed by the calculation processing unit 210.

[0021] The arithmetic processing unit 210 includes, for example, a CPU (Central Processing Unit). The arithmetic processing unit 210 executes a program stored in the storage unit 220 to function as a work video acquisition unit 211 that stores a video file acquired from the image capture device 100 in the storage unit 220, a calculation unit 212 that calculates values ​​necessary to identify an improvement target work, including the average values ​​of work time and distance L, and the number of works for which the average value of distance L is the same, an improvement target work identification unit 213 that identifies the improvement target work, and an output unit 214 that outputs the improvement target work. Hereinafter, the average value of work time will be referred to as the average work time, and the average value of distance L will be referred to as the average distance. The number of works for which the average value of distance L is the same is an example of the "number of works with the same distance" according to the present disclosure.

[0022] The storage unit 220 includes a non-volatile semiconductor memory including, for example, a flash memory and an EPROM (Erasable Programmable Read Only Memory). The storage unit 220 may also include a non-volatile memory including a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, and a DVD (Digital Versatile Disc).

[0023] The storage unit 220 has a moving image file storage unit 221 that stores a work moving image file, a reference image file storage unit 222 that stores an image that serves as a reference when calculating the work time of each work in each process, a calculated value storage unit 223 that stores a calculated value calculated by the calculation unit 212, an improvement target work storage unit 224 that stores an improvement target work, and a work management data storage unit 225 that stores a QC process chart (Quality Control Chart) created at the time of work design, an updated QC process chart, etc. The storage unit 220 also stores a program executed by the calculation processing unit 210.

[0024] The moving image file storage unit 221 stores a work moving image file including a moving image of a work. Note that a work moving image acquisition unit 211 described later acquires a work moving image file from the imaging device 100, and saves the acquired work moving image file in the moving image file storage unit 221.

[0025] The reference image file storage unit 222 stores images of work tools, work machines, work equipment, etc., which are used as a reference when calculating the work time of each task in the captured video acquired from the image capture device 100, for each process. Note that work tools, work machines, work equipment, work background, etc. are collectively referred to simply as tools. For example, when the work is performed visually in the appearance inspection of the target item Tp, an image without tools, measuring instruments, etc. is stored in the reference image file storage unit 222 together with the identification number of the appearance inspection. In addition, in the case of the measurement work of the dimensions of the target item Tp, an image of a scale used to measure the dimensions is stored in the reference image file storage unit 222 as a reference image together with the identification number of the dimension measurement work. In addition, in the case of the performance inspection work of the target item Tp, an image of a performance inspection machine is stored in the reference image file storage unit 222 as a reference image together with the identification number of the performance inspection. These reference images are merely examples, and there may be multiple reference images. Hereinafter, the "reference image" will also be referred to as the "reference image".

[0026] The calculated value storage unit 223 stores the calculated values ​​calculated by the calculation unit 212. The calculated values ​​include the average task time, the average distance, the number of tasks for which the average distance shows the same value, etc. Note that the average task time, the average distance, and the number of tasks for which the average distance shows the same value are examples of characteristic values ​​according to the present disclosure. The improvement target task storage unit 224 stores the improvement target tasks identified by the improvement target task identifying unit 213 .

[0027] The work management data storage unit 225 stores data indicating the contents of the designed work, such as the work name, work procedure, and tools used for the work. For example, the work management data storage unit 225 stores the QC process charts shown in Figs. 3 to 6. Here, the QC process chart means a table that describes the control points, control methods, and the like at each stage of the process from the arrival of parts to the shipment as a finished product. Note that Fig. 3 shows the QC process chart for the "work at the time of arrival of parts" process. Fig. 4 shows the QC process chart for the "assembly work" process. Fig. 5 shows the QC process chart for the "pre-shipment inspection work" process. Fig. 6 shows the QC process chart for the "packing work" process. In addition, in the QC process chart, the tools, measuring instruments, inspection machines, and the like used in each work are assigned work identification numbers. In addition, for work that does not use tools, measuring instruments, inspection machines, and the like, such as visual inspection work, the visual inspection work identification number is assigned to the work name in the QC process chart. Also, the identification numbers of parts, products, etc. are assigned to the target items in the QC process chart as the identification numbers of the target items Tp. Note that in Figs. 3 to 6, the QC process chart is in a form that can be easily recognized by a person for ease of understanding, but the format of the table is arbitrary. The QC process chart is an example of a work management chart according to the present disclosure.

[0028] The various data in the storage unit 220 of the work improvement support device 200 may be stored in a cloud-type server external to the work improvement support device 200. The files saved in the video file storage unit 221 may be compressed before being saved. The storage capacity of the video file storage unit 221 may be managed by automatically deleting old files after a period determined for each work site.

[0029] The work video acquisition unit 211 of the calculation processing unit 210 acquires a work video file from the imaging device 100 and stores it in the video file storage unit 221 of the storage unit 220. For example, in the case of a manufacturing line for train air conditioners, the work video stored in the video file storage unit 221 is a video of each work in each process of the manufacturing process for train air conditioners. Note that the target manufacturing process is not limited to the manufacturing process for train air conditioners, and may be a manufacturing process for any other type of electrical appliance.

[0030] The calculation unit 212 calculates the work time of each task for each process based on the video data acquired from the video file storage unit 221 of the storage unit 220. In this embodiment, as shown in FIG. 7, the target model is a vehicle air conditioner. In addition, the analysis period for performing the work improvement support is determined in advance. In this embodiment, the analysis period is from October 1 to December 31, 2023, as shown in FIG. 7. The calculation unit 212 plays the work video and identifies the tools used at each point in time using the reference image stored in the reference image file storage unit 222 and an image recognition technique such as pattern matching. The calculation unit 212 identifies the task from the identification number associated with the identified tool. For example, if a scale for measuring dimensions is shown in the image, it is identified that a dimension measurement task is being performed at that point in time from the identification number associated with the scale. The calculation unit 212 calculates the boundary point of consecutive tasks based on the change in the tool shown in the video. For example, the calculation unit 212 determines the boundary point between a visual appearance inspection task and a dimension measurement task using a scale as the time when an image of a scale appears on the image.

[0031] The calculation unit 212 calculates the task time from the difference between the time at the boundary between the previous task and the time at the boundary between the next task. The calculation unit 212 also adds up the calculated task times for each task. The calculation unit 212 calculates the average task time by dividing the value obtained by adding 1 to the number of additions N (N+1). The average task time calculated by the calculation unit 212 is accumulated and stored in the calculated value storage unit 223 together with the task identification number. In FIG. 7, the average task time is calculated in units of one minute.

[0032] The calculation unit 212 also calculates an average value of the distance L of each task for each process based on the task video file acquired from the video file storage unit 221 of the storage unit 220. Specifically, the calculation unit 212 adds data indicating the distance L between the time of the boundary point of the previous task and the time of the boundary point of the subsequent task. The calculation unit 212 calculates a value obtained by dividing the value obtained by the addition by the number of additions N plus 1 (N+1) as the average distance. The average distance calculated by the calculation unit 212 is accumulated and stored in the calculation unit storage unit 223 together with the task identification number. In this way, the calculation unit 212 forms an analysis table in which the type of task, the average task time, and the distance between the worker and the target item, here the product, are associated with each other for each process, as exemplified in FIG. 7. Note that in FIG. 7, the worker HM wears the imaging device 100 on his face, so the average value of the distance L between the face and the product is shown. Also, the average task time is in minutes, and the average value of the distance L is in 10 cm units. The calculation unit 212 accumulates and stores the formed analysis table in the calculated value storage unit 223 .

[0033] Furthermore, the calculation unit 212 calculates the number of tasks having the same average distance value among the multiple tasks on the created analysis table. In the case of Fig. 7, the calculation unit 212 calculates, for example, the number of tasks having an average distance of 40 cm to be 3, and the number of tasks having an average distance of 100 cm to be 2. The number of tasks calculated by the calculation unit 212 is accumulated and stored in the calculated value storage unit 223 together with the task identification number. The calculation unit 212 is an example of a calculation means according to the present disclosure.

[0034] The improvement target work identifying unit 213 identifies an improvement target work for which improvement should be given priority over other works from among the series of works, based on the average work time and number of works calculated by the calculation unit 212 and the QC process chart. Details will be described later. The improvement target work identifying unit 213 is an example of an improvement target work identifying means according to the present disclosure.

[0035] The output unit 214 outputs the identified improvement target work and a QC process chart corresponding to the improvement target work to the display device 300. Note that the output unit 214 is an example of an output means according to the present disclosure.

[0036] The display device 300 displays the information output by the arithmetic processing unit 210. The display device 300 includes, for example, a liquid crystal display.

[0037] Next, the process of identifying an improvement target task for identifying an improvement target task will be described with reference to FIG. 8. As a premise for the process of identifying an improvement target task, first, as described above, the image capturing device 100 captures images at all times while the production line is in operation. The image capturing device 100 identifies an image in the live view image that matches an image of a target item Tp registered in advance, for example, by using an image recognition technique such as pattern matching, and adjusts the focus position with the maximization of the contrast of the identified target item Tp as an index. The image capturing device 100 also estimates the distance L at predetermined time intervals using software built into itself.

[0038] The imaging device 100 stores the captured video in its memory as a work video file in association with the time measured by a built-in clock and the distance L estimated at each time. The imaging device 100 adds its own identification number, the identification number of the worker HM, the identification number of the work target item Tp, and the identification number of the process to each work video file.

[0039] When a person in charge of work on a production line, such as a manager or an operator, starts up the work improvement support device 200, the improvement target work identification process shown in FIG. 8 is executed.

[0040] When the process starts, the work video acquisition unit 211 of the work improvement support device 200 acquires a work video file from the photography device 100 and additionally stores it in the video file storage unit 221 (step S201). As a result, the identification number of the photography device 100, the identification number of the worker HM, the identification number of the target item Tp, the identification number of the process, the work video recorded together with the time measured by the built-in clock, and the distance L recorded together with the time measured by the built-in clock are additionally stored as a work video file in the video file storage unit 221.

[0041] The calculation unit 212 determines whether or not the present time is within a predetermined analysis period based on the time measured by a clock built into the task improvement support device 200 (step S202). As described above, in this embodiment, the analysis period is from October 1 to December 31, 2023, so if it is determined that the present time is within the analysis period (step S202: Yes), the improvement target task identification process ends.

[0042] If it is determined that the current analysis period is not in progress (step S202: No), the calculation unit 212 calculates the average work time during the analysis period (step S203). Specifically, the calculation unit 212 plays the work video stored in the video file storage unit 221. Then, based on the reference image stored in the reference image file storage unit 222, the calculation unit 221 identifies the work being performed at that time point using, for example, an image recognition technique such as a pattern matching method, and further determines the time of the boundary point between the previous work and the current work and the time of the boundary point between the next work. As described above, the calculation unit 212 calculates, for example, the boundary point between the work of inspecting the appearance of the target product Tp and the work of measuring the dimensions of the target product Tp using a scale as the time point when the scale appears in the image. The calculation unit 212 calculates the work time for each work by calculating the difference between the time of the boundary point with the next work and the time of the boundary point with the previous work. The calculated operation time is accumulated and stored in the calculated value storage unit 223 together with the identification number of the operation and the identification number of the target item Tp. For example, in the case of a part dimension measurement operation using a scale, the calculation unit 212 stores the calculated operation time in the calculated value storage unit 223 together with the identification number of the "dimension measurement operation" corresponding to the identification number associated with the scale and the identification number of the part. In the case of a part visual inspection operation, since no specific tool is used, the calculation unit 212 stores the calculated operation time in the calculated value storage unit 223 together with the predetermined identification number of the "visual inspection operation" and the identification number of the part. The predetermined identification number of the "visual inspection operation" is stored in the operation management data storage unit 225.

[0043] Furthermore, the calculation unit 212 adds up the task times accumulated and stored in the calculation value storage unit 223 for each task. The calculation unit 212 then divides the value obtained by adding 1 to the number of additions N (N+1), and obtains the value obtained by division as the average task time. The obtained average task time is accumulated and stored in the calculation value storage unit 223 together with the task identification number and the identification number of the target item Tp. For example, in the case of a task of measuring the dimensions of a part using a scale, the calculation unit 212 stores the obtained average task time in the calculation value storage unit 223 together with the identification number of the "dimension measurement task" which corresponds to the identification number associated with the scale, and the identification number of the part.

[0044] The calculation unit 212 calculates the average value of the distance L between the worker HM and the target item Tp, in other words, the average distance (step S204). Specifically, the calculation unit 212 adds up the data of the distance L between the boundary point with the previous task and the boundary point with the subsequent task for each task. Then, the calculation unit 212 divides the value obtained by adding 1 to the number of additions N (N+1), and calculates the value obtained by the division as the average distance. The calculation unit 212 accumulates and stores the calculated average distance in the calculated value storage unit 223 together with the identification number of the task and the identification number of the target item Tp. For example, in the case of a task of measuring the dimensions of a part using a scale, the calculation unit 212 stores the calculated average distance in the calculated value storage unit 223 together with the identification number of the "dimension measurement task" corresponding to the identification number associated with the scale and the identification number of the part.

[0045] The calculation unit 212 calculates the number of tasks with the same average distance (step S205). For example, in the case of FIG. 7, there are three tasks with an average distance of 40 cm, namely, "visual inspection of incoming parts", "visual inspection of parts", and "visual inspection of product", so the calculation unit 212 calculates the number of tasks as 3. Also, there are two tasks with an average distance of 100 cm, namely, "performance inspection of incoming parts" and "transportation of product to packing workplace", so the calculation unit 212 calculates the number of tasks as 2. The number of tasks for other average distance values ​​is 1, so the calculation unit 212 calculates the number of tasks as 1. The calculated number of tasks is stored in the calculation value storage unit 223 for each average distance together with the task identification number.

[0046] The calculation unit 212 determines whether the calculated number of tasks includes two or more tasks (step S206). In the example of Fig. 7, the calculated number of tasks includes two or more tasks (step S206: Yes), so the process proceeds to the identification process (step S207).

[0047] If it is determined that the number of tasks found is not two or more (step S206: No), there are no overlapping tasks to be improved, and the improvement target task identification process ends.

[0048] If it is determined that there are two or more tasks among the calculated number of tasks (step S206: Yes), the process proceeds to the identification process of step S207. Here, the identification process refers to a process of identifying an improvement target task from among the tasks corresponding to the average distance, which is the maximum number of tasks with the same average distance as the target. The specific process of step S207 is shown in FIG. 9.

[0049] When the process proceeds to the identification process in step S207, the calculation unit 212 selects the average distance with the maximum number of tasks (step S207a). In the case of Fig. 7, the maximum number of tasks is 3, and the average distance is 40 cm. Therefore, the calculation unit 212 selects the average distance of 40 cm.

[0050] The calculation unit 212 selects one average distance with the largest number of tasks and the task that indicates that average distance (step S207b). In the case of FIG. 7, the average distance corresponding to the largest number of tasks, 3, is one, 40 cm. Therefore, the calculation unit 212 selects the tasks "visual inspection of incoming parts", "visual inspection of parts", and "visual inspection of product" that correspond to the average distance of 40 cm and the tasks that indicate an average distance of 40 cm. Note that in the case of FIG. 7, the tasks "visual inspection of incoming parts", "visual inspection of parts", and "visual inspection of product" are examples of "candidate tasks to be improved" according to the present disclosure.

[0051] The calculation unit 212 determines whether or not there is any work not listed in the QC process chart among the works corresponding to the average distance selected in step S207b (step S207c). Specifically, in the case of FIG. 7, the calculation unit 212 determines whether or not there is any work not listed in the QC process charts shown in FIG. 3 to FIG. 6 among the works corresponding to an average distance of 40 cm, specifically, the work "visual inspection of incoming parts", the work "visual inspection of parts", and the work "visual inspection of products". The QC process charts for each process are stored in the work management data storage unit 225.

[0052] The calculation unit 212 judges whether the identification number indicating the work stored in the calculation value storage unit 223 together with the calculated average distance and average work time and the identification number of the target item Tp match the identification numbers assigned to the tools, measuring instruments, inspection machines, etc. of each work in the QC process schedule, or the identification numbers assigned to the work names in the QC process schedule and the identification numbers assigned to the target items in the QC process schedule. For example, as shown in FIG. 3, the QC process schedule for the "work at the time of arrival of parts" process includes the work "visual inspection of incoming parts". Also, as shown in FIG. 4, the QC process schedule for the "assembly work" process does not include the work "visual inspection of parts". Also, as shown in FIG. 5, the QC process schedule for the "pre-shipment inspection work" process includes the work "visual inspection of products". In this case, the calculation unit 212 determines that the identification number indicating the "visual inspection work" stored in the calculation value storage unit 223 together with the obtained average work time, average distance, etc. of the "visual inspection of parts" work is not included in the identification numbers assigned to the tools, measuring instruments, inspection machines, etc. used for each work in the QC process schedule for the "assembly work" process, or the identification numbers assigned to the work names in the QC process schedule for the "assembly work" process. Therefore, in step S207c, the calculation unit 212 determines that there is a work not listed in the QC process schedule (step S207c: Yes).

[0053] When it is determined that there is an operation not described in the QC process chart (step S207c: Yes), the calculation unit 212 selects the operation not described (step S207d). In the case of Fig. 7, the calculation unit 212 selects the "visual inspection of parts" operation that corresponds to the operation not described.

[0054] The calculation unit 212 determines whether or not there is any work among the selected works that is the same as another work (step S207f). Through this process, the selected works have the same average distance, but it is further determined whether or not they are the same work in the QC process chart. For example, as shown in FIG. 7, the work "performance inspection of incoming parts" and the work "transporting products to the packing workplace", which correspond to a work with an average distance of 100 cm, are different works even though they have the same average distance. For this reason, the process of step S207f is a process that prevents different works, even if they have the same average distance, from being excluded as duplicate works.

[0055] If it is determined that the selected tasks include tasks that are the same as other tasks (step S207f: Yes), the calculation unit 212 selects the selected tasks that are the same as other tasks (step S207g). In the case of FIG. 7, the calculation unit 212 selects the "visual inspection of parts" task. Specifically, in the case of FIG. 7, the calculation unit 212 determines whether the selected "visual inspection of parts" task is the same as another "visual inspection of incoming parts" task or a "visual inspection of product" task, whose average distance corresponds to 40 cm. For example, the calculation unit 212 determines whether the identification number of the part corresponding to the target product Tp and the identification number of the visual inspection task, which are stored in the calculation value storage unit 223 together with the average work time, average distance, etc. of the obtained "visual inspection of parts" task, are present in other tasks. In the QC process schedule, the identification number of the target item Tp indicating the identification number of the part and the identification number of the visual inspection associated with the work name are in the "visual inspection of incoming parts" work in the part arrival time work process, so the calculation unit 212 determines that the selected "visual inspection of parts" work is the same as the other work (step S207f: Yes). Note that, when the selected work and the other work are both works described in the QC process schedule, the determination in step S207f may be made, for example, depending on whether or not a predetermined item in the QC process schedule is the same in the selected work and the other work.

[0056] Calculation unit 212 determines whether or not multiple operations are selected in step S207g (step S207i). If it is determined that multiple operations are not selected in step S207g (step S207i: No), the process proceeds to step S207k. In the case of FIG. 7, the operation selected in step S207g is one operation, "visual inspection of components," so the process proceeds to step S207k.

[0057] If it is determined that a plurality of tasks have been selected (step S207i: Yes), the calculation unit 212 selects the task with the longest average task time from among the tasks selected in step S207g (step S207j). Then, the process proceeds to step S207k.

[0058] The improvement target work identifying unit 213 identifies the work selected in step S207g or the work selected in step S207j as the work to be improved (step S207k). In the case of FIG. 7, the improvement target work identifying unit 213 identifies the "visual inspection of parts" work selected in step S207g as the work to be improved. As a result, of the overlapping works between the "visual inspection of incoming parts" work and the "visual inspection of parts" work, the "visual inspection of parts" work, which is not listed in the QC process chart and may be considered as a customary work, is identified as the work to be improved. The identified work to be improved is stored in the improvement target work memory unit 224.

[0059] The calculation unit 212 determines whether there is another average distance with the maximum number of tasks (step S207m). In the case of Fig. 7, the average distance with the maximum number of tasks, 3, is only 40 cm. Therefore, it is determined that there is no other average distance with the maximum number of tasks (step S207m: No), and the process proceeds to step S208.

[0060] If it is determined that the average distance with the maximum number of operations is elsewhere (step S207m: Yes), the process returns to step S207b.

[0061] In the example of FIG. 7, there is a task not listed in the QC process schedule. However, when it is determined that there is no task not listed in the QC process schedule (step S207c: No), the calculation unit 212 selects a task listed in the QC process schedule (step S207e). As described above, the calculation unit 212 determines whether or not the task is listed in the QC process schedule based on whether or not the identification number indicating the task stored in the calculation value storage unit 223 along with the calculated average distance and average task time matches the identification number assigned to the tool, measuring instrument, inspection machine, etc. used for each task in the QC process schedule, or the identification number assigned to the task name in the QC process schedule. The calculation unit 212 determines whether or not there is a task that is the same as another task among the tasks listed in the QC process schedule selected in step S207e (step S207f).

[0062] 7, the task "visual inspection of parts" selected in step S207d and not listed in the QC process chart is the same as another task "visual inspection of incoming parts" (step S207f: Yes). However, if it is determined that none of the selected tasks is the same as another task (step S207f: No), the process proceeds to step S207h.

[0063] The calculation unit 212 determines whether the selected work is a work not listed in the QC process chart (step S207h). If it is determined that the selected work is a work not listed in the QC process chart (step S207h: Yes), the calculation unit 212 selects a work listed in the QC process chart (step S207e). Then, the processing from step S207f onwards is performed. As a result, if there is no work equivalent to overlapping work among the works not listed in the QC process chart, but there is work equivalent to overlapping work among the works listed in the QC process chart, it can be identified as a work to be improved.

[0064] If it is determined that the selected work is not a work not listed in the QC process chart, in other words, if it is determined that the selected work is a work listed in the QC process chart (step S207h: No), this corresponds to a case where there is no overlapping work, and processing proceeds to step S207m.

[0065] If it is determined that there is no other average distance with the maximum number of operations (step S207m: No), the process proceeds to step S208.

[0066] The output unit 214 outputs information including the improvement target tasks stored in the improvement target task storage unit 224 and the QC process chart stored in the task management data storage unit 225 to the display device 300 (step S208). This can support task improvement, including consideration of whether or not improvement target tasks that correspond to overlapping tasks can be eliminated.

[0067] Next, an example of the hardware configuration of the activity improvement support device 200 will be described with reference to Fig. 10. The activity improvement support device 200 in Fig. 10 is realized by a computer such as a personal computer or a microcontroller.

[0068] The work improvement support device 200 comprises a processor 1001 that executes an operation program which is a program for the operation of the work improvement support device 200, a memory 1002 that serves as the main storage area, an interface 1003 that realizes the communication function of the work improvement support device 200, and a secondary storage device 1004 that stores the operation program for executing processing, all of which are connected to each other via a bus 1000.

[0069] The processor 1001 is, for example, a CPU (Central Processing Unit). The processor 1001 loads an operation program stored in a secondary storage device 1004 into a memory 1002 and executes the program, thereby implementing each function of the work improvement support device 200.

[0070] The memory 1002 is a main storage device constituted by, for example, a RAM (Random Access Memory). The memory 1002 stores the operation program read by the processor 1001 from the secondary storage device 1004. The memory 1002 also functions as a work memory when the processor 1001 executes the operation program.

[0071] The interface 1003 is an I / O (Input / Output) interface such as a serial port, a USB (Universal Serial Bus) port, a network interface, etc. The interface 1003 realizes the communication function of the work improvement assistance device 200.

[0072] The secondary storage device 1004 is, for example, a flash memory, a hard disk drive (HDD), or a solid state drive (SSD). The secondary storage device 1004 stores the operation program executed by the processor 1001, the task video file acquired by the task acquisition unit 211, the reference image file serving as a reference for identifying the task, the calculated value obtained by the calculation unit 212, the improvement target task identified by the improvement target task identification unit 213, and data for managing the tasks including the QC process chart.

[0073] As described above, in the work improvement support system 1 according to the first embodiment, the calculation unit 212 calculates the average value of each work time for each process, the average value of the distance L from the face of the worker for each work to the target product Tp, and the number of works for which the average value of the distance L is the same, based on the video captured by the imaging device 100. The calculation unit 212 selects the work corresponding to the largest average distance among the calculated number of works, and selects candidates for the work to be improved from the selected works based on the QC process chart and the average work time, etc. The work to be improved is identified by the improvement target work identification unit 213. The output unit 214 outputs information including the work to be improved stored in the improvement target work memory unit 224 and the QC process chart stored in the work management data memory unit 225 to the display device 300. The display device 300 displays the information output by the output unit 214. Therefore, the people involved in the work, including the workers and the person in charge, can be supported in taking improvement measures, including considering whether or not to eliminate the work to be improved that corresponds to overlapping work.

[0074] In addition, the process of identifying tasks to be improved judges whether there are tasks not listed in the QC process chart, and if there are tasks not listed, the unlisted tasks are selected as candidates for tasks to be improved in preference to listed tasks. This makes it easier to identify tasks that are customarily performed even if they are not listed in the QC process chart as targets for improvement.

[0075] (Modification of the first embodiment) In the first embodiment, the image capturing device 100 is a wearable video camera, but the present disclosure is not limited to this. The image capturing device 100 may be, for example, a video camera installed above a workbench on which a work target Tp is placed.

[0076] Furthermore, the work time for each task in each process may be calculated from the video captured by the imaging device 100 using any analysis software.

[0077] In the first embodiment, in the process of identifying the work to be improved, even if the work has the same average distance, it is determined whether or not the same work as the selected work is in the QC process schedule (step S207f), but the present disclosure is not limited to this. The process of identifying the work to be improved may be performed on the premise that the work having the same average value of the distance L between the face of the worker and the target item is the same work. Specifically, after the process of step S207d or step S207e in FIG. 9, the process may proceed to step S207i. In other words, the processes of steps S207f, S207g, and S207h may not be performed. This makes it easier to identify the work to be improved based on the average work time, the average distance, the number of work having the same average distance, and whether or not it is written in the QC process schedule.

[0078] In the first embodiment, the calculation unit 212 calculates the work time of each task for each process based on the work video, but the present disclosure is not limited to this. The calculation unit 212 may calculate the work time of each task for each process based on the time when the value of the distance L between the face of the worker and the target product changes beyond a predetermined threshold. In the case of FIG. 7, for example, the predetermined threshold is set to ±5 cm. The calculation unit 212 may set the time T1 when the value indicating the data of the distance L included in the work video file changes from 40 cm to 10 cm as the end time of the "visual inspection of the incoming part" task and the start time of the "dimension measurement of the incoming part" task. In addition, the calculation unit 212 may set the time T2 when the value indicating the data of the distance L changes from 10 cm to 100 cm as the end time of the "dimension measurement of the incoming part" task and the start time of the "performance inspection of the incoming part" task. The calculation unit 212 may set the value obtained by calculating the difference between the time of the time T2 and the time of the time T1 as the work time of the "dimension measurement of the incoming part" task.

[0079] In the first embodiment, the QC process chart shown in Figs. 3 to 6 is stored in the work management data storage unit 225 as data showing the contents of the designed work, such as the work name, work procedure, tools used in the work, measuring instruments, etc., but the present disclosure is not limited to this. For example, depending on the production site, there are cases where the QC process chart is not prepared, making it difficult to compare the work standard with the measured work. In such a case, the necessity of overlapping work may be confirmed using the work procedure manual, work instruction manual, etc. created within the process. Also, only the work procedure may be shown, or improvement viewpoints may be added.

[0080] In embodiment 1, the distance L was the distance from the face of the worker HM to the work object Tp, but it is not limited to the face of the worker HM and may be, for example, the distance from the chest, waist, etc. of the worker HM to the work object Tp.

[0081] Furthermore, although the information indicating the work to be improved and the procedure of the work within the process is displayed on the display device, it may be displayed by audio, video, etc.

[0082] In addition, if there is a clear reason why additional work or additional inspections have occurred in a certain process due to specific circumstances such as the introduction of new employees or trainees, that period may be excluded from the evaluation.

[0083] In addition, the calculation unit 212 calculated the average value of the working time for each task in each process and the average value of the distance L between the worker's face and the target item, but the values ​​are not limited to the average value, and the median, mode, etc. may be calculated as representative values.

[0084] Additionally, the target product may be any type of product that can be the target of workplace improvements on a manufacturing line.

[0085] (Embodiment 2) A work improvement support device 200 according to embodiment 2 will be described with reference to Fig. 11 to Fig. 14. Embodiment 2 differs from embodiment 1 in that the work improvement support device 200 presents work to be improved using AI (Artificial Intelligence) based on the average work time of each work in each process, an improvement target evaluation value indicating an evaluation value for selecting work to be improved, and a trained model. The following describes embodiment 2, focusing on the differences from embodiment 1.

[0086] The improvement target evaluation value corresponds to an index value for evaluating the work, including the degree of overlap of the work and the degree of necessity of the work. When the improvement target evaluation value is large, there is a high possibility of overlapping work. The calculation unit 212 calculates the improvement target evaluation value based on the number of work with the same average distance and a value predetermined depending on the presence or absence of a QC process table. Specifically, the calculation unit 212 multiplies the number of work with the same average distance by a value predetermined depending on the presence or absence of a QC process table, and calculates the value obtained by multiplication as the improvement target evaluation value. In addition, the value predetermined depending on the presence or absence of a QC process table is set to 1 for work that is described in the QC process table, and is set to 3 for work that is not described in the QC process table. This makes it easier to select work that is customarily performed despite not being described in the QC process table as the improvement target work.

[0087] 11, the task improvement support device 200 according to this embodiment includes an improvement presenting unit 216 instead of the improvement target task identification unit 213 of the task improvement support device 200 according to embodiment 1. The improvement presenting unit 216 presents tasks to be improved based on the average task time, the improvement target evaluation value, and the trained model.

[0088] The memory unit 220 further includes a trained model memory unit 226.

[0089] The trained model storage unit 226 stores trained models. The trained models are generated by, for example, a "supervised learning method" that uses a known machine learning library and trains a set of data with a correct answer label attached to training data. Note that the trained models may be trained and generated by other learning methods such as an "unsupervised learning method," a "reinforcement learning method," or a "semi-supervised learning method."

[0090] The learning algorithm may be a known learning algorithm such as a neural network. The neural network has an input layer composed of a plurality of nodes to which different input parameters are input, an intermediate layer to which signals output from each node of the input layer are input, and an output layer to which signals output from the intermediate layer are input and which outputs output parameters. The intermediate layer of the neural network may be composed of one or more layers.

[0091] The neural network will be outlined with reference to FIG. As shown in the figure, the neural network is composed of an input layer, a hidden layer, and an output layer, each of which contains multiple neurons. Here, there is one hidden layer, but the number of hidden layers can be any number. The number of neurons in the output layer, k, is the total number of tasks in all processes to be evaluated, where n = 2·k, and m is an integer of 3 or more. For example, if the average task time of the jth task j is 2j-1 The improvement target evaluation value of task j is fed to the 2j-th neuron X 2j j = 1 to k. The neurons Z1 to Zk is assigned to the tasks 1 to k to be evaluated, and the task corresponding to the fired neuron Z becomes a candidate for improvement.

[0092] In this configuration, the neurons X1 to X n are the input values ​​to each neuron Y1 to Y m Output to.

[0093] Each neuron in the hidden layer, Y1 to Y m weights the input values ​​W 11 ~W nm The sum is multiplied by and summed, and the sum is applied to the output function to output the result. For example, the i-th neuron in the hidden layer, Y i is the jth neuron X in the input layer. j Input from I j Weight W ji Multiply by ΣW ji I j Then, apply this to the output function φ to obtain φ(ΣW ji I j -θ) to each neuron Z1 to Z k The output is as follows, where i = 1 to m and j = 1 to n.

[0094] Neurons Z1 to Z in the output layer k Also, each neuron Y1 to Y m Each input value from 11 ~V mk The sum is multiplied by the above and summed to obtain a total value, and the total value is applied to an output function to output the obtained value. Neurons Z1 to Z in the output layer k The output from is the weights W 11 ~W nm and the weight V 11 ~V mk depends on the value of . As shown in FIG. 12, when 11 tasks are included, n=22 and k=11. The average task time of task 1 is input to neuron X1 in the input layer, the improvement target evaluation value of task 1 is input to neuron X2, the average task time of task 2 is input to neuron X3, the improvement target evaluation value of task 2 is input to neuron X4, etc. In addition, neurons Z1 to Z 11 When one of these neurons fires, the task corresponding to the fired neuron becomes a candidate for improvement.

[0095] The neurons in the input layer are X1 to X n is supplied with the average task time, which is normalized to the value 1 for the longest average task time among the k tasks, and the k improvement target evaluation values, which is normalized to the value 1 for the largest improvement target evaluation value among the tasks.

[0096] The learning data is created based on a combination of the average work time and improvement target evaluation value for each task by process, which are obtained in advance by an expert or experienced person, by analyzing each product, and the task to be improved. Based on the learning data, a learning process is performed for the task to be improved. The weights W obtained by the learning process are 11 ~W nm , V 11 ~V mk The structure of the neural network including the values ​​n, k, etc. is stored as a trained model in the trained model storage unit 226.

[0097] The improvement presentation unit 216 presents the work to be improved by using the trained model.

[0098] As shown in FIG. 11, the improvement presentation unit 216 includes a learning unit 216a that generates a trained model, and an inference unit 216b that infers the work to be improved using the trained model.

[0099] The learning process by the learning unit 216a will be described below. The learning process shown in Fig. 13 is started when an operation unit (not shown) receives an operation by a user to instruct the execution of the learning process.

[0100] As described above, the learning unit 216a receives training data including a plurality of combinations of the average task time and the improvement target evaluation value of k tasks designed in advance, and the tasks to be improved (step S301). The training data is input to the learning unit 216a via an input unit (not shown).

[0101] The learning unit 216a repeatedly uses the teacher data to learn the relationship between the combination of the average work time and the improvement target evaluation value of the k tasks and the improvement target tasks (step S302), and calculates the weight W 11 ~W nm , V 11 ~V mk The learning unit 216a stores the trained model in the trained model storage unit 226 (step S303), and ends the learning process.

[0102] The inference unit 216b uses the trained model to execute the inference process shown in Fig. 14. The inference process is executed at a timing preset by a user. The timing for executing the inference process may be set to, for example, every week or every month.

[0103] The inference unit 216b applies, as input parameters, normalized values ​​of the average work time and the improvement target evaluation value of the k tasks to the trained model (step S401).

[0104] The inference unit 216b uses the trained model to execute inference based on the input parameters (step S402), outputs data indicating the work to be improved as output parameters (step S403), and ends the inference process.

[0105] The improvement presenter 216 acquires the output parameters output from the inference unit 216b as improvement target task data. Upon acquiring the improvement target task data, the improvement presenter 216 stores the improvement target task data in the improvement target task memory unit 224 shown in FIG.

[0106] When the improvement presentation unit 216 acquires the improvement target work data, it presents the improvement target work to the worker based on the acquired improvement target work data. Specifically, the improvement presentation unit 216 outputs the improvement target work data to the display device 300, and the improvement target work data is displayed on the display device 300, thereby presenting it to the worker.

[0107] The improvement presentation unit 216 may use an application such as a spreadsheet software based on the improvement target work data to display the improvement target work on the display device 300. The output unit 214 may also display the QC process chart stored in the work management data storage unit 225 together with the improvement target work data on the display device 300. The QC process chart can be used as a study material for improvement measures, including the elimination of duplicate work.

[0108] As described above, according to the task improvement support device 200 of the second embodiment, the improvement presentation unit 216 presents the task to be improved to the worker using a trained model trained with data including the average task time, the improvement target evaluation value, and the task to be improved. The worker can carry out efficient improvement activities by implementing improvements to the task to be improved presented by the improvement presentation unit 216. As a result, the occurrence of tasks with low task efficiency, including overlapping tasks, is reduced, and the burden of task improvement support is lightened. Therefore, the efficiency of task improvement support can be improved.

[0109] In the second embodiment, the learning target of the trained model used by the improvement presentation unit 216 may be a plurality of types of products. In addition, the learning unit 216a may learn about actual work performed in one area, or may learn about actual work performed independently in a plurality of different areas.

[0110] In addition, a product to be learned by the trained model may be added or removed from the trained model in the middle of the process. Furthermore, any trained model may be applied to a trained model of another model, and the other model may be retrained based on the trained model. In addition, the training data may be, for example, a combination of the average work time and the improvement target evaluation value as input, and the output may be the work that maximizes the product of the two.

[0111] (Modification of the second embodiment) In the second embodiment, the learning unit 216a uses a neural network as a learning algorithm. However, the learning unit 216a may use other known methods, such as genetic programming, functional logic programming, or a support vector machine, as a learning algorithm.

[0112] In the second embodiment, the learning unit 216a and the inference unit 216b are provided in the task improvement support device 200, but the learning unit 216a and the inference unit 216b may be provided in a device different from the task improvement support device 200. The improvement presentation unit 216 may acquire output parameters output from the inference unit 216b provided in a device different from the task improvement support device 200 as task data to be improved.

[0113] A trained model resulting from a learning process performed elsewhere may be stored in the trained model storage unit 226, and the learning unit 216a may be configured not to be included in the improvement presentation unit 216.

[0114] In the second embodiment, the learning unit 216a is provided in the work improvement support device 200. However, once a trained model is created, the learning unit 216a may be removed.

[0115] In addition, the improvement presenting unit 216 may use data that combines the average work time, the improvement target evaluation value, and the work to be improved not only for the same model but also for similar models. That is, the improvement presenting unit 216 can also use data that combines the average work time, the improvement target evaluation value, and the work to be improved for similar models as learning data.

[0116] In the second embodiment, the work to be improved is presented based on the average work time and the improvement target evaluation value. However, the present invention is not limited to this, and the most effective improvement viewpoint may be presented based on the average work time and the improvement target evaluation value. The work to be improved and the most effective improvement viewpoint may be presented together.

[0117] In this case, for example, the combination of the average work time, the evaluation value of the improvement target, the work to be improved, and the most effective improvement viewpoint is prepared as learning data from the analysis of past performance data. Also, neurons to which each improvement viewpoint is assigned are placed in the output stage of the neural network shown in FIG.

[0118] In the inference stage, a pair of the average task time and the evaluation value for the improvement target is input to the input stage, the task to be improved is identified from the firing state of the neuron to which the task is assigned in the output stage, and the most effective improvement perspective is identified from the firing state of the neuron to which the improvement perspective is assigned.

[0119] In the above embodiment, an example was shown in which the work improvement assistance device 200 is configured by a computer equipped with a processor and a memory, but the device configuration is arbitrary. For example, all or part of the processing executed by the processor may be executed by an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or the like.

[0120] Although the preferred embodiments have been described in detail above, the present invention is not limited to the above-described embodiments, and various modifications and substitutions can be made to the above-described embodiments without departing from the scope of the claims.

[0121] Various aspects of the present disclosure are summarized below as appendices.

[0122] (Appendix 1) A work improvement support device that supports improvement of a work for producing a product, a calculation means for calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of the tasks included in each process of producing the product and data indicating the distance between the worker of the task and the target product that is the subject of the task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; an improvement target task identification means for identifying a task having the largest representative value of the task time as the improvement target task from among the improvement target task candidates; an output means for outputting information representing the improvement target work identified by the improvement target work identifying means; Equipped with the calculation means, based on a work management table in which data indicating the contents of predetermined work is recorded, selects the work not recorded in the work management table as a candidate for the work to be improved from among the works counted as the maximum number of same-distance works, giving priority to the work recorded in the work management table; the improvement target task identification means, when there are improvement target task candidates selected with priority, identifies, from among the improvement target task candidates selected with priority, a task having the largest representative value of the task time as the improvement target task; Work improvement support device. (Appendix 2) The calculation means calculates the number of tasks having the same average value of the distance as the number of tasks having the same distance, the improvement target task identification means identifies, from among the improvement target task candidates, the task having the largest average task time as the improvement target task; 2. A work improvement support device according to claim 1. (Appendix 3) The output means outputs the work management table together with information representing the work to be improved. 3. A work improvement support device as described in appendix 2. (Appendix 4) the calculation means calculates a value obtained by multiplying the number of tasks of the same distance by a value that is determined in advance depending on whether the task is recorded in the task management table, as an improvement target evaluation value corresponding to an index value for evaluating the task; the improvement target task identification means presents the improvement target task based on the average task time and the improvement target evaluation value, using a combination of the average task time and the improvement target evaluation value, and a trained model that has trained on the improvement target task; 4. A work improvement support device according to claim 2 or 3. (Appendix 5) The improvement target task identification means A learning unit that generates a trained model based on training data including the combination of the average task time and the improvement target evaluation value and the improvement target task; an inference unit that infers the work to be improved using the trained model based on the combination; having 5. A work improvement support device according to claim 4. (Appendix 6) A work improvement support device according to any one of appendices 1 to 5; An imaging device for imaging the work; a display device that displays the information output by the output means; Equipped with Work improvement support system. (Appendix 7) calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of tasks included in each process of manufacturing a product and data indicating the distance between a worker performing each task and a target item that is the subject of each task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; identifying a task having the largest representative value of the task time as the task to be improved from among the candidate tasks to be improved; Work improvement support method. (Appendix 8) On the computer, calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of tasks included in each process of manufacturing a product and data indicating the distance between a worker performing each task and a target item that is the subject of each task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; identifying a task having the largest representative value of the task time as the task to be improved from among the candidate tasks to be improved; A program that executes a process. [Explanation of symbols]

[0123] 1 Work improvement support system, 100 imaging device, 200 work improvement support device, 210 calculation processing unit, 211 work video acquisition unit, 212 calculation unit, 213 improvement target work identification unit, 214 output unit, 216 improvement presentation unit, 216a learning unit, 216b inference unit, 220 memory unit, 221 video file memory unit, 222 reference image file memory unit, 223 calculated value memory unit, 224 improvement target work memory unit, 225 work management data memory unit, 226 trained model memory unit, 300 display device, 1000 bus, 1001 processor, 1002 memory, 1003 interface, 1004 secondary storage device, HM worker, Tp target item.

Claims

1. A work improvement support device that supports improvement of a work for producing a product, a calculation means for calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of the tasks included in each process of producing the product and data indicating the distance between the worker of the task and the target product that is the subject of the task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; an improvement target task identification means for identifying a task having the largest representative value of the task time as the improvement target task from among the improvement target task candidates; an output means for outputting information representing the improvement target work identified by the improvement target work identifying means; Equipped with the calculation means, based on a work management table in which data indicating the contents of predetermined work is recorded, selects the work not recorded in the work management table as the candidate work to be improved from among the works counted as the maximum number of same-distance works, giving priority to the work recorded in the work management table; the improvement target task identification means, when there are improvement target task candidates selected with priority, identifies, from among the improvement target task candidates selected with priority, a task having the largest representative value of the task time as the improvement target task; Work improvement support device.

2. The calculation means calculates the number of tasks having the same average value of the distance as the number of tasks having the same distance, the improvement target task identification means identifies, from among the improvement target task candidates, the task having the largest average task time as the improvement target task; The work improvement support device according to claim 1.

3. The output means outputs the work management table together with information representing the work to be improved. The work improvement support device according to claim 2.

4. the calculation means calculates a value obtained by multiplying the number of tasks of the same distance by a value that is determined in advance depending on whether the task is recorded in the task management table, as an improvement target evaluation value corresponding to an index value for evaluating the task; the improvement target task identification means presents the improvement target task based on the average task time and the improvement target evaluation value, using a combination of the average task time and the improvement target evaluation value, and a trained model that has trained on the improvement target task; 4. The work improvement support device according to claim 2 or 3.

5. The improvement target task identification means A learning unit that generates a trained model based on training data including the combination of the average task time and the improvement target evaluation value and the improvement target task; an inference unit that infers the work to be improved using the trained model based on the combination; having The work improvement support device according to claim 4.

6. The work improvement support device according to claim 1 ; An imaging device for imaging the work; a display device that displays the information output by the output means; Equipped with Work improvement support system.

7. calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of tasks included in each process of manufacturing a product and data indicating the distance between a worker performing each task and a target item that is the subject of each task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; identifying a task having the largest representative value of the task time as the task to be improved from among the candidate tasks to be improved; Work improvement support method.

8. On the computer, calculating a representative value of the task time for each task, a representative value of the distance for each task, and a number of tasks with the same distance, which is the number of tasks with the same representative value of the distance, based on a task video file including a video of tasks included in each process of manufacturing a product and data indicating the distance between a worker performing each task and a target item that is the subject of each task, and selecting the task with the largest number of tasks with the same distance as a candidate task to be improved; identifying a task having the largest representative value of the task time as the task to be improved from among the candidate tasks to be improved; A program that executes a process.

Images