Work support system, work support device, work support method, and program

Abstract

A work support system (1) for supporting work performed by a plurality of workers in cooperation with each other comprises a skill level calculation means, a work allocation means, and a display means. The skill level calculation means obtains a skill level of a worker according to an evaluation value including an actual work time spent by the worker for the work. The work allocation means allocates each unit of work included in a series of work to the worker according to the skill level, and generates divided work procedure data representing content of the allocated unit of work. The display means displays the content of the divided work procedure data.

Description

Work support system, work support device, work support method, and program

[0001] The present disclosure relates to a work support system, a work support device, a work support method, and a program.

[0002] Techniques for supporting assembly work at a manufacturing site are known. For example, the assembly work device disclosed in Patent Document 1 supports work to prevent incorrect part selection by lighting an indicator lamp attached to each parts box on a workbench.

[0003] Japanese Patent Application Laid-Open No. 2006-167829

[0004] Among the workers in charge of assembly work, there are workers with different work capabilities, ranging from those with high work capabilities to those with relatively low work capabilities. Therefore, even with the same workload, it may be too much for some workers and, conversely, some workers may be in a situation where they have nothing to do. When the assembly work consists of a series of multiple work processes, the low throughput of some workers reduces the throughput of the entire work. For this reason, it is desirable to allocate an appropriate amount of work to each worker so as to increase the throughput of the entire work, but it is not easy. The technique of Patent Document 1 can prevent incorrect parts from being selected at the work site, but it cannot contribute to allocating an appropriate amount of work to the workers.

[0005] The present disclosure has been made in view of the above problems, and an object thereof is to provide a work support system, a work support device, a work support method, and a program that enable an appropriate amount of work to be allocated to the work personnel.

[0006] To achieve the above object, a work support system according to the present disclosure is a work support system that supports work performed by a plurality of workers in cooperation, and includes proficiency calculation means, work allocation means, and display means. The proficiency calculation means obtains the proficiency of a worker according to an evaluation value including the actual work time spent by the worker on the work. The work allocation means allocates each unit work included in a series of works to a worker according to the proficiency, and generates divided work要领 data representing the content of the allocated unit work. The display means displays the content of the divided work要领 data.

[0007] According to this disclosure, the skill level of the workers is determined, each unit task included in a series of tasks is assigned to the workers according to the determined skill level, and divided work procedure data representing the content of the assigned unit task is generated and displayed. Therefore, an appropriate amount of work can be assigned to the workers.

[0008] Figures showing the configuration of the assembly line and work support system according to the embodiment. Detailed view of the area around the workbench of the assembly system according to the embodiment. Block diagram showing the functional configuration of the linkage device shown in Figure 1. Block diagram showing the functional configuration of the work support device shown in Figure 1. Figures showing an example of the hardware configuration of the computer shown in Figure 3 and Figure 4. Flowchart of pre-work start processing according to the embodiment. Figure to explain an example of work assignment according to the embodiment. Flowchart of work support processing according to the embodiment. Figure to explain the monitoring process of the parts shelf according to the embodiment. Figure showing an example of work instructions according to the embodiment. Figure showing an example of the results of image inspection processing according to the embodiment.

[0009] (Embodiments) Hereinafter, the work support system, work support device, work support method, and program according to embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals.

[0010] The work support system and work support method according to the embodiments of this disclosure are systems that improve the overall efficiency of work by i) determining the skill level of the workers in charge of the work, ii) assigning work processes to each worker in accordance with the determined skill level in order to improve the overall efficiency of the work, and iii) presenting work procedure data showing the details of the assigned work processes to the workers.

[0011] Here, we will describe an embodiment using the example of supporting assembly work performed on assembly line 2 shown in Figure 1. On assembly line 2, a series of work processes constituting the assembly work are performed by workers P1 to Pn at n (n is a natural number of 2 or more) workbenches 111 to 11n to assemble the product. Hereinafter, the work performed in each work process will be called a unit work.

[0012] In this configuration, worker Pi (i=1 to n) performs the following steps: i) works at workbench 11i, ii) receives an intermediate product or finished product (hereinafter collectively referred to as "product") from workbench 11(i-1) which is responsible for the previous work process, iii) performs assembly work such as attaching parts to the received product, iv) performs further inspection, and v) passes the inspected product to the next workbench 11(i+1). Each worker P completes one product as a whole by performing these work processes. A slide rail 6 is used to transport the product. Hereafter, workbenches 111 to 11n will be collectively referred to as workbench 11 when not distinguished, and workers P1 to Pn will be collectively referred to as worker P when not distinguished.

[0013] The work support system 1 comprises a work support device 100 and coordinating devices 301 to 30n that operate in conjunction with the work support device 100. The work support device 100 and the coordinating devices 301 to 30n are connected via a network NW to enable communication. The work support device 100 and the coordinating devices 301 to 30n are each composed of computers. Hereinafter, when the coordinating devices 301 to 30n are not distinguished, they will be collectively referred to as the coordinating device 300. The network NW may be wired, wireless, WAN (Wide Area Network), LAN (Local Area Network), intranet, extranet, etc.

[0014] The work support device 100 acts as a master unit, or in other words, a server, for distributing and aggregating data. On the other hand, the linkage device 300 acts as a slave unit, or in other words, a client unit. The linkage device 30i is installed on the corresponding work desk 11i and, under the control of the work support device 100, provides support to efficiently execute the work process assigned to worker Pi. In the process of providing this support, work performance data is collected to estimate the skill level of worker Pi.

[0015] Figure 2 is a perspective view of an example of a workbench 11 equipped with a linkage device 300. Worker P performs assembly work while standing or sitting next to the workbench 11.

[0016] Each workbench 11, as part of the assembly line 2, is equipped with a workbench surface 5, a slide rail 6 for transporting products 10, and a parts shelf 7 on which parts boxes 9 are placed.

[0017] The work surface 5 is a table on which the worker P performs assembly and inspection work. The slide rail 6 is formed in a groove shape across the work surface 5 of multiple work tables 11 and transports the product 10 between the work tables 11. The width W of the slide rail 6 is adjustable to match the size of the product 10.

[0018] The parts shelf 7 is a shelf for arranging one or more parts boxes 9 that contain parts used in assembly work. The parts boxes 9 contain parts used for work at the workbench 11. The worker takes the necessary parts from the parts boxes 9 placed on the parts shelf 7 and performs the assembly work. The parts shelf 7 is movable in the front-to-back direction, and as the parts shelf 7 moves, the parts boxes 9 move and the width W of the slide rail 6 changes.

[0019] Furthermore, each workbench 11 is equipped with the following components as part of the interconnected device 300: lighting devices 3a, 3b, and 3c for illuminating the workbench surface 5 and the products 10; fluorescent marks 8a and 8b attached to the boundary between the slide rail 6 and the parts shelf 7; a computer 310 for executing work support processing, which will be described later; an imaging unit 331 for capturing images of the workbench surface 5 and the products 10; a reading unit 332 for reading code information; a power supply unit 333 for supplying power to the lighting devices 3a, 3b, and 3c; and a display unit 334 for displaying images.

[0020] The lighting devices 3a, 3b, and 3c are composed of, for example, LEDs (Light Emitting Diodes) and illuminate the workbench surface 5 and the product 10 during assembly and inspection work. The power supply unit 333 adjusts the brightness of the lighting devices 3a, 3b, and 3c to the JIS standard work assistance brightness during assembly work and to the inspection brightness during inspection work.

[0021] The fluorescent marks 8a and 8b are reflective plates of a specific color attached to the boundary between the slide rail 6 and the parts shelf 7, indicating the boundary between the slide rail 6 and the parts shelf 7. As the parts shelf 7 moves, the fluorescent marks 8a and 8b also move, continuously indicating the boundary between the slide rail 6 and the parts shelf 7.

[0022] The imaging unit 331 is positioned to capture images of the product 10, fluorescent marks 8a and 8b, parts shelves 7, etc., and captures these images. The imaging unit 331 is composed of, for example, an industrial camera with a resolution of 5 million pixels or more and a frame rate of 30 fps or more. The imaging unit 331 also has a variable focus lens 2 with a zoom function. Detailed images of the product 10 and the workbench surface 5 are acquired by such an imaging unit 331 and transmitted to a computer 310 for use in image inspection and identification of work position coordinates. The imaging unit 331 is an example of an imaging means according to this disclosure.

[0023] The reading unit 332 reads code information printed on work slips, products, etc. The code information consists of, for example, character codes, one-dimensional or two-dimensional barcodes. The reading unit 332 transmits the read code information to the computer 310. As a result, the computer 310 can obtain product data, worker data, lot number, manufacturing number, etc., read by the reading unit 332.

[0024] The power supply unit 333 supplies power to the lighting devices 3a, 3b, and 3c for illumination. The power supply unit 333 also dims the lighting devices 3a, 3b, and 3c by modulating the output voltage using PWM and adjusting its duty cycle, according to the control of the computer 310. The luminous brightness of the lighting devices 3a, 3b, and 3c can be switched, for example, to the work assistance brightness specified by JIS during assembly work and to the inspection brightness during inspection work. The luminous brightness of the lighting devices 3a, 3b, and 3c may be a preset brightness, or it may be configured so that the worker P can adjust it by operating a volume control or the like on the power supply unit 333.

[0025] The display unit 334 is installed on the workbench 11 with its display surface facing the worker P, and displays various images, videos, text, etc., such as instructions, messages, and inspection results for the worker P. The display unit 334 is an example of a display means according to this disclosure.

[0026] As shown in Figure 3, the computer 310 includes a communication unit 311 that sends and receives information to and from the work support device 100 via a network NW, an input / output (I / O) unit 313 that sends and receives data to and from the imaging unit 331, reading unit 332, power supply unit 333, and display unit 334, a processing unit 312 that executes work support processing described later, and a storage unit 320 that stores information.

[0027] The storage unit 320 includes a desk information storage unit 321 that stores desk information including the identification information of the workbench 11 itself (hereinafter referred to as the workbench ID) and the ID of the worker who works at the workbench 11 (hereinafter referred to as the worker ID); a processing information storage unit 322 that stores information obtained by the processing of the processing unit 312; a work procedure information storage unit 323 that stores work procedure information representing the content of the work process, including the type of work process performed at the workbench 11 and the type of parts used; an inspection master information storage unit 324 that stores information necessary for inspection, including a trained model for inferring whether the product 10 is a normal product; and a work log information storage unit 325 that stores work log information recording the content of the work and inspection actually performed at the workbench 11. The storage unit 320 also stores image processing software that identifies the positions of the worker's hands, parts boxes 9, etc. from images captured by the imaging unit 331 and calculates the coordinates of the identified positions.

[0028] The workbench information storage unit 321 stores the ID of the workbench 11 and the ID of the worker who works at the workbench 11. The processing information storage unit 322 stores the information obtained by the processing of the processing unit 312 of the workbench 11.

[0029] The work procedure information storage unit 323 selectively stores work procedure information related to the work processes performed at this workbench 11 from the overall work procedure information for the assembly work. Here, the work procedure information is a series of process information necessary for the assembly work. The work procedure information is a group of unit work procedure information, which is the work corresponding to one of the multiple work processes that make up the series of assembly work. The configuration of the unit work procedure information is arbitrary, but examples include the order of the work processes, whether the order of the work processes can be changed, the type of work process, work position coordinates, prohibited work coordinates, standard work time, part type, part name, part sample image, number of parts, notes and comments, work animation, information for unskilled workers, information showing how to deal with assembly defects, etc.

[0030] For example, if "Work Order" = 7, it indicates that this is the 7th task to be performed out of all work processes. If "Work Order Reversibility" = Yes, it indicates that the order is not limited to the order specified by the work order and can be changed. If "Work Order Reversibility" = No, it indicates that the work processes must be performed in the order specified by the work order. "Type of Work" indicates the type of work, such as inserting a part into a specified position or soldering. Work Position Coordinates refer to information such as the XY coordinates, width, and height of the position or area to which the work is performed, such as attaching a part. Work Prohibition Coordinates refer to information such as the XY coordinates, width, and height of the position or area where work is prohibited. Standard Work Time refers to the time required for a worker with standard skill to perform the work. For example, if Standard Work Time = 10 seconds, it means that it takes a worker with standard skill to take the specified part from the parts box 9, place it in the specified position, and perform the specified type of work in 10 seconds. The part type, part name, part sample image, and part quantity refer to the type of part used in the work, the name of the part, the image of the part, and the quantity of the part. The notes comment refers to points to be aware of when performing the work. The work animation refers to a video showing the work process. The information for unskilled workers refers to points that unskilled workers should be aware of. The information on how to deal with assembly defects refers to information on what to do when an assembly defect occurs, such as "call the work supervisor."

[0031] As described above, the work procedure information is, for example, a group of unit work procedure information. The work procedure information is represented, for example, as work process 1 (work order: 1, can work order be rearranged: yes, ...), work process 2 (work order: 2, can work order be rearranged: yes, ...), ... work process X (work order: X, can work order be rearranged: yes, ...). For example, work process 1 (work order: 1, can work order be rearranged: yes, ...) indicates the procedure information for the first unit work to be performed. The work procedure information storage unit 323 selectively stores sets of unit work procedure information for work processes to be performed on the workbench 11. This is because, in step S17 of the pre-work start processing described later, it receives and stores segmented work procedure data from the work support device 100 that includes only the work procedure data for the work processes to be performed on this workbench 11.

[0032] The inspection master information storage unit 324 stores a trained model for inferring whether the product 10 generated in each operation is a normal product, as well as feature information and coordinate information of the product 10. The trained model takes image data of the product 10 as input and outputs whether the product 10 is good or bad, the number of defective parts if the product 10 is judged to be defective, the deviation from a good product, and the defective parts in the image. The trained model is composed of, for example, a deep learning model such as Mobile Net that has been pre-trained on a large image dataset. During training, by passing an arbitrary set of normal images through this trained model, a feature vector is output for each local region of the intermediate layer—usually a convolutional layer. By calculating the mean vector and covariance matrix for each local region, the feature distribution of normal images is constructed. On the other hand, during inference, for example, the difference between the feature vector of the image of a defective product and the feature vector of the trained normal image, i.e., the deviation, is calculated and output.

[0033] Normal or defective product data of product 10 is input into a trained model, and defective areas in the image are identified using the weights of previously extracted features and the deviation range of normal products in each region. In addition to whether the product is good or bad, if it is found to be defective, the number of defective areas and the deviation from good products are output. The smaller the deviation from good products, the closer it is to good products, and the larger the value, the further it is from good products. Areas where the output deviation exceeds a pre-set threshold are output as defective areas. If the defective areas are connected, for example, they are treated as one defect, and if they are not connected, they are treated as separate defects, and the number of defects is also output. Within the image region of product 10, the work position coordinates of the unit work procedure data performed at the self-work desk 11 are the target of the judgment. A trained model is created for each work process and is generated by learning training data that associates image data of the product 10 manufactured in that work process with the quality of the product 10, the number of defective parts if the product 10 is judged to be defective, the deviation from a good product, and the defective parts in the image. The inspection master information storage unit 324 selectively stores the trained models for inspections performed on the workbench 11. This is because, in step S18 of the pre-work processing described later, it receives and stores the trained models for inspections performed on the workbench 11 that are transmitted by the work support device 100.

[0034] The work log information storage unit 325 collects work log information representing the content of the work performed at the workbench 11 while executing the work support processing described later, and stores it along with the time and worker ID. The work log information from each cooperating device 300 is aggregated in the work support device 100.

[0035] The processing unit 312 displays images, videos, guidance, and messages indicating work procedures on the display unit 334 based on segmented work procedure data, which is stored in the work procedure information storage unit 323 and indicates the work procedures for one or more work processes to be performed on the workbench 11. The processing unit 312 also identifies the position of the product 10 and performs image inspection based on the image data captured by the imaging unit 331. The processing unit 312 is an example of an image inspection means according to this disclosure.

[0036] Meanwhile, the work support device 100 aggregates information from the coordinating device 300, performs pre-work start processing as described later, and causes the coordinating device 300 to perform work support processing. As shown in Figure 4, the basic configuration of the work support device 100 is substantially the same as that of the computer 310 of each coordinating device 300.

[0037] However, the storage unit 12 differs in that it stores information that aggregates the information stored in each storage unit 320 of the multiple cooperating devices 300. Specifically, the storage unit 120 includes a desk information storage unit 121 that stores the ID of the work desk 11 and the ID of the worker working at that work desk 11 for all work desks 11, a processing information storage unit 122 that stores information obtained by the processing of the processing unit 112, a work procedure information storage unit 123 that stores work procedure information for all work processes, including the type of work and the type of parts for each work process that constitutes a series of operations, an inspection master information storage unit 124 that stores information necessary for inspection, including a trained model for inferring whether the product 10 is a normal product, for all work processes, and a work log information storage unit 125 that stores work log information for all work desks 11, recording the work process and the content of the inspection. The work log information stored in the work log information storage unit 125 is accumulated for a predetermined period and discarded in order from oldest to newest.

[0038] Furthermore, the work support device 100 includes a communication unit 111 that communicates with the coordinating device 300, an I / O unit 113 that exchanges data with an input unit 131, a display unit 132, etc., and a processing unit 112 that processes data. The processing unit 112 is an example of the proficiency calculation means and work assignment means according to this disclosure. The communication unit 111 is an example of the transmission means according to this disclosure.

[0039] Next, an example of the hardware configuration of the computer 110 of the work support device 100 and the computer 310 of the linkage device 300 having the above configuration will be described with reference to Figure 5. Since the configuration of computer 110 and the configuration of computer 310 are substantially the same, the computer 110 will be used as an example for explanation.

[0040] The computer 110 includes a processor 1001 that executes an operation program, a memory 1002 that serves as the main storage area of the processor 1001, an interface 1003 having a communication function, and a secondary storage device 1004 that stores the operation program and data for executing processing. The processor 1001, the memory 1002, the interface 1003, and the secondary storage device 1004 are connected to each other via a bus 1000.

[0041] The processor 1001 includes, for example, a CPU (Central Processing Unit). By the processor 1001 reading the operation program stored in the secondary storage device 1004 into the memory 1002 and executing it, each function of the work support device 100, specifically, the function of the processing unit 112, is realized.

[0042] 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. Also, the memory 1002 functions as a work memory when the processor 1001 executes the operation program.

[0043] The interface 1003 includes a communication device and an I / O port. The communication device includes a network board or the like and communicates with the cooperation device 300. The I / O port is an I / O (Input / Output) interface such as a serial port, a USB (Universal Serial Bus) port, or a network interface. The functions of the communication unit 111 and the I / O unit 113 of the work support device 100 are realized by the interface 1003.

[0044] The secondary storage device 1004 includes, for example, a flash memory, a HDD (Hard Disk Drive), and a SSD (Solid State Drive). The secondary storage device 1004 stores the operation program executed by the processor 1001, the desk information, the processing information, the work procedure information, the inspection master information, and the work log information.

[0045] Next, referring to FIG. 6, the pre-operation processing executed before the start of work will be described. The pre-operation processing is the processing executed by the work support device 100 before the start of the assembly work on the assembly line 2. Specifically, the pre-operation processing includes: i) a process of obtaining the proficiency levels of the workers on the assembly line 2; ii) based on the obtained proficiency levels, a process of allocating work processes to each workbench 11 so that the expected total assembly time of the assembly line 2 is minimized, a process of extracting the work要领 data of the work processes allocated to each workbench 11, and iii) a process of transmitting the allocated work processes, the divided work要领 data corresponding to the work processes, and other related information to each cooperation device 300.

[0046] As a premise, it is assumed that workers P1 to Pn are assigned to n workbenches 111 to 11n on the assembly line 2. The person in charge of the work, the leader of the workers, etc. input the product ID, lot number, worker IDs of all workers, etc. of the work target. The computer 310 transmits this information to the computer 110. Also, before the start of work, each worker Pi causes the reading unit 332 of the workbench 11i assigned to himself / herself to read, for example, the barcodes printed on the work ticket, the product 10, etc. and the barcodes printed on the worker's ID card. The reading unit 332 sends the product ID, worker ID, lot number, manufacturing number, etc. included in these barcodes to the computer 310. The computer 310 transmits the table information including these information and the workbench ID to the computer 110.

[0047] The processing unit 112 of the computer 110 of the work support device 100 acquires the product ID, worker ID, etc. from the cooperation device 300 via the communication unit 111 (step S11), and stores the acquired product ID, worker ID, etc. in the table information storage unit 121 of the storage unit 120 in association with each other.

[0048] Also, the processing unit 112 receives the table information including the worker ID and the workbench ID from each cooperation device 300 via the communication unit 111 (step S12). The processing unit 112 stores the received table information in the table information storage unit 121.

[0049] The processing unit 112 determines whether or not it has received console information from all of the cooperating devices 300 (step S13). If the processing unit 112 determines that it has not received console information from any of the cooperating devices 300 (step S13: No), it repeats the process in step S13.

[0050] If the processing unit 112 determines that it has received table information from all cooperating devices 300 (step S13: Yes), it calculates the proficiency level of each worker (step S14). Proficiency level is calculated using, for example, the following formula: Proficiency level = ((total actual work time - total standard work time) / number of tasks) + (number of mistakes in parts selection × maximum deviation distance from the correct parts box × weighting coefficient WE1) + (number of assembly mistakes × number of defective parts × maximum deviation deviation from good products × weighting coefficient WE2) Total actual work time is the total time actually spent by the worker on past tasks, which is collected by the work support processing described later and stored in the work log information storage unit 125. Total standard work time is the sum of the standard work times predetermined for each work process, i.e., each unit task. The standard work time for each task is stored in the work procedure information storage unit 123. The number of tasks is the total number of tasks performed by the worker and is recorded in the work log information storage unit 125.

[0051] Furthermore, the number of incorrect parts selections is the number of times the worker selected the wrong part. The number of incorrect parts selections is counted during the work support process and recorded in the work log information storage unit 125. The maximum deviation distance from the correct parts box is the maximum distance from the correct parts box when the wrong part is selected. The maximum deviation distance from the correct parts box is based on monitoring data from the parts shelf 7, is determined during the work support process, and recorded in the processing information storage unit 122.

[0052] Furthermore, the number of assembly errors refers to the number of mistakes that occurred during the assembly process and corresponds to the number of inspection defects. The number of assembly errors is calculated in the work support processing based on the inspection results and recorded in the work log information storage unit 125. The number of defective locations is the number of defective locations that occurred during the assembly process. The number of defective locations is determined by image inspection and recorded in the work log information storage unit 125. Note that the total actual work time, the number of parts selection errors, and the number of assembly errors are examples of evaluation values ​​related to this disclosure.

[0053] Furthermore, the maximum deviation from a good product represents the maximum deviation value indicating how much a defective product deviates from a good product. The value of the deviation from a good product is obtained by a trained model during the image inspection process and stored in the processing information storage unit 122. The processing unit 112 stores the maximum value of the deviation from a good product stored in the processing information storage unit 122 as the maximum deviation from a good product. In this way, this data is automatically collected by the work support device 100 and the linkage device 300, aggregated and stored in the computer 110. Skill level is calculated from this data. It is desirable that the skill level value expressed by the calculated result be small. The obtained skill level is stored in the processing information storage unit 122. Skill level may also be determined individually for each type of work and part. The weighting coefficients WE1 and WE2 are predetermined values, but they may be changed by production engineers, supervisors, etc.

[0054] The processing unit 112 sequentially assigns the work processes to be performed at each workbench 11 based on the skill level of each worker (step S15). Specifically, it assigns one or more work processes in order from work process 1 to the first workbench 111 and worker P1, assigns one or more work processes in order from the next work process to the second workbench 112 and worker P2, and so on, assigning the work processes.

[0055] For example, in the example in Figure 7, the first workbench 111 is assigned work process 1 and work process 2. The second workbench 112 is assigned work process 3 and ... Similarly, the nth workbench 11n is assigned ... and work process N. In other words, the first worker P1 is assigned work process 1 and work process 2, the second worker P2 is assigned work process 3 and ..., and the nth worker Pn is assigned ... and work process N. Next, we determine the work time required for each work process by the workbench 11 and worker P. The time tX required for the Xth work process is expressed by the following formula. tX = Standard work time for work process X × (1 + Skill level of worker P at that workbench 11 × Weighting coefficient WE3) The weighting coefficient WE3 is a predetermined value, but may be changed by the production engineer, supervisor, etc.

[0056] Next, the total expected time T for the work performed by worker P at each workbench 11 is calculated. For example, in the example in Figure 7, the first workbench 111 and worker P1 are assigned work process 1 and work process 2, so the expected work time t1 for work process 1 and the expected work time t2 for work process 2 are calculated, resulting in T1 = t1 + t2. Subsequently, the processing unit 112 calculates the expected work time required to complete a series of tasks for one product 10, that is, the expected total expected time = TS = T1 + T2 + ... Tn. The processing unit 112 changes the combination of processes assigned to each workbench 11, in other words, each worker P, - while maintaining the order of the work processes - and calculates the total expected time. That is, the processing unit 112 repeats the same process for multiple assignments.

[0057] As mentioned above, some work processes allow for changes in their order. Therefore, the processing unit 112 rearranges the order of the work processes that allow for changes in order and assigns them to each work table 11, and estimates the expected total work time TS. In the example in Figure 7, work processes 1 and 3 allow for changes in order. For this reason, the expected total work time TS is also calculated when, for example, work process 3 is assigned to the first work table 111 and work process 1 is assigned to the second work table 112.

[0058] The processing unit 112 identifies the smallest of the multiple expected total work time values ​​TS obtained in this way (step S16).

[0059] The processing unit 112 identifies the assignment to each work table 11 for each work process, when the minimum expected total work time TS is obtained, as the work assignment for the current work.

[0060] Next, the processing unit 112 extracts the unit work procedure for the assigned work process from the work procedure data for each work table 11 and generates divided work procedure data (step S17). For example, as shown in Figure 7, if two work processes, work process 1 and work process 2, are assigned to the first work table 111, divided work procedure data is created that includes the unit work procedure data for work process 1 and the unit work procedure data for work process 2.

[0061] The division work procedure data generated by the processing unit 112 is transmitted to each of the linking devices 300 (step S18). In this way, the work support device 100, acting as a server, performs optimal work assignment based on the skill level of each worker.

[0062] In this way, by optimizing the tasks assigned to each worker, the total working time can be reduced. Furthermore, by considering whether the order of tasks can be rearranged, assigning complex unit tasks to highly skilled workers and relatively simple unit tasks to less skilled workers, the expected total assembly time of the line is reduced, and overall efficiency is optimized.

[0063] The processing unit 112 transmits the data in the divided work procedure created in this manner to the linking device 300 of each workbench 11. Each linking device 300 stores the received divided work procedure data in the work procedure information storage unit 323 and displays the corresponding information on the display unit 334. This display includes a work preparation screen. The work preparation screen is a screen that instructs, for example, to change the width of the slide rail 6 according to the size of the product 10, to place the required number of parts boxes 9 on the parts shelf 7, and to put the required number of parts into the parts boxes 9. These screens are, for example, videos and are stored in the work procedure information storage unit 123 in video file format. The worker prepares for work according to the instructions displayed on the work preparation screen. The pre-work processing is completed.

[0064] Next, with reference to Figure 8, a work support process that assists the worker's work will be explained. The work support process includes a process of displaying work instructions and advice to the worker P on the display unit 334, and a process of performing image inspection based on the captured image of the product 10 after the work is completed. When each worker P has finished preparing for work, they indicate that they have finished preparing for work by, for example, pressing a key on the keyboard connected to the computer 310 to input the product ID, lot number, manufacturing number, etc. of the product 10, or by having the reading unit 332 read a two-dimensional code on which the product ID, lot number, manufacturing number, etc. of the product 10 is printed. After that, in the case of work table 111, the assembly work begins by receiving a new product 10, or in the case of subsequent work tables 11, by receiving a product 10 from the previous work table 11. Work log information linked to the lot number, manufacturing number, etc. is stored in the work log information storage unit 325.

[0065] When the processing unit 312 starts the work support process shown in Figure 8, it first determines whether or not there is an input instruction indicating that work preparation is complete (step S21). If the processing unit 312 determines that there is no input instruction indicating that work preparation is complete (step S21: No), it repeats the process in step S21.

[0066] If it is determined that an input instruction indicating that work preparation is complete has been received (Step S21: Yes), the processing unit 312 starts the internal timer to begin timing the work time and also starts continuous imaging on the imaging unit 331. A control command from the processing unit 312 to the power supply unit 333 turns on the lighting devices 3a, 3b, and 3c at the work assistance brightness. The imaging unit 331 acquires the position information of the fluorescent marks 8a and 8b by color search in the pixels. The processing unit 312 detects the position of the parts shelf 7 and defines a parts shelf boundary line m1 representing the boundary line of the parts shelf 7, as shown in Figure 9 (Step S22). Specifically, the processing unit 312 analyzes the image data sent from the imaging unit 331 and defines a parts shelf boundary line m1 that passes through the coordinates of the two points of the fluorescent marks 8a and 8b, as shown in Figure 9.

[0067] The processing unit 312 sets a group of monitoring ranges m2 aligned between fluorescent marks 8a and 8b and on the parts shelf boundary line m1, as shown in Figure 9, following the parts shelf boundary line m1 (step S23). The shape of each monitoring range m2 is not limited, but a rectangle as shown in Figure 9 is preferable. The size of each monitoring range m2 is not limited, but from the viewpoint of improving detection accuracy, it is preferable that the size in the X-axis direction be smaller than the size of the parts box 9 in the X direction. The processing unit 312 controls the imaging unit 331 to acquire and store the initial image information of each monitoring range m2. Thereafter, although not shown in Figure 8, the imaging unit 331 continuously monitors the pixel information of each monitoring range m2. If, during monitoring by the imaging unit 331, there is a significant change in the pixel information of one or more monitoring ranges m2 that exceeds a reference level, the monitoring range m2 is changed to a detection range m3, and it is determined that a worker's hand has entered that area. If the pixel information of the detection range m3 returns to the original initial image information, it is changed back from the detection range m3 to the monitoring range m2.

[0068] The following are specific examples of when there is a significant change in the pixel information of the aforementioned monitoring range m2: i) When a worker's hand enters the monitoring range m2, a significant change occurs in the pixel information of that monitoring range m2. In this case, the change in the color and brightness of the pixels due to the movement of the hand is detected. ii) When a part is taken out of the parts shelf 7 or moved to another location, etc., the worker's hand or arm crosses multiple monitoring ranges m2 in succession, causing significant changes in the pixel information of those monitoring ranges m2 in sequence. This allows the system to detect that the position of the part has changed. iii) Changes in lighting, the influence of external light, etc., can also cause changes in the pixel information of one or more monitoring ranges m2. For example, lighting may be temporarily blocked, or external light may become stronger or weaker. When the processing unit 312 detects these changes, it changes the monitoring range m2 to the detection range m3 and determines that a worker's hand has entered the range.

[0069] Here, we will explain specific examples of setting the monitoring range m2 and detection range m3. The detection range m3 is set using a specific position on the workbench 11 as the reference point. For example, as shown in Figure 9, the left front end of the workbench 11 or the parts shelf 7 is set as the reference point (0,0,0), and the relative position from the reference point is defined as each coordinate. Each monitoring range m2 and detection range m3 is set based on the relative position coordinates (X coordinate, Y coordinate, Z coordinate) from the reference point. For example, if a specific area of ​​the parts shelf 7 is to be used as the monitoring range m2 and detection range m3, the range of that area is set from (X1, Y1, Z1) to (X2, Y2, Z2). As a specific example, if the left front end of the parts shelf 7 is used as the reference point (0,0,0), and the entire parts shelf 7 is to be used as one monitoring range m2 and detection range m3, the range is set from (0,0,0) to (100,50,30). When worker P performs an action to remove a part within the monitoring range m2, the pixel information in the monitoring range m2 changes, the action is detected, and the monitoring range m2 is updated to the detection range m3.

[0070] The processing unit 312, in parallel with monitoring the group within the monitoring range m2, displays work instructions on the display unit 132, as illustrated in Figure 10, and starts measuring the work time using an internal timer (step S24). The processing unit 312 displays the work instructions for the first work process assigned to the workbench 11 from the divided work procedure data on the display unit 334. For example, in the example of Figure 7, the processing unit 312 of workbench 111 displays the work instructions for work process 1 on the display unit 334, and the processing unit 312 of workbench 112 displays the work instructions for work process 3 on the display unit 334. Note that the work instructions refer to both the display of the contents of the unit work procedure data and the playback of the work animation. The work animation is generated based on the unit work procedure data, and a part of the unit work procedure data is displayed as an animation. For example, the instruction location n14 shown in Figure 10 displays the work process being executed, in other words, the image of the parts used in the unit work, the part number, the hand used to process the parts, the number of insertions, etc. Instruction point n8 displays the remaining time for the work process using a time bar, with the bar's length and color changing as the remaining time decreases. Instruction point n10 displays comments, notes, etc., associated with the unit work using subtitles, dynamic highlighting, etc. Instruction point n11 is an overhead view of the product, in which the mounting position of parts indicated by semi-transparent rectangles, the mounting direction indicated by arrow icons, pre-registered part images, etc., and other coordinates that require attention indicated by semi-transparent rectangles, arbitrary icons, etc., are displayed in a blinking manner. Instruction point n12 can display enlarged images of the mounting positions, pre-registered work videos, etc.

[0071] Furthermore, the instruction location n1 shown in Figure 10 is a button to re-receive the division work procedure data, but it is not normally pressed. The instruction location n2 is a button to finish the assembly work. It is not normally pressed except at the end of the work. The instruction location n3 is a button to display the software log. The displayed content includes not only the assembly log but also the internal operation of the software, error information, etc. It is not normally pressed. The instruction location n4 is a button to display the monitoring process of the parts shelf 7 by the imaging unit 331. It is not normally pressed.

[0072] The instruction location n5 is a button for operating data related to the overall settings. The computer 310 operates on the work procedure data for each product, the inspection master data for each product, and the assembly log data for each worker. Normally, this is not operated by anyone other than production engineers. The instruction location n6 is an area that displays the product currently being worked on. The computer 310 is linked to the product ID read by the reading unit 332. The linkage device 300 is automatically set from the divided work procedure data received from the work support device 100. If there are multiple patterns for the product and manufacturing method, the patterns are also displayed.

[0073] Indicator area n7 is the area that displays the machine number of the workbench 11, i.e., the workbench ID. Indicator area n8 is the area that displays the standard work time for the current task and the actual elapsed time measured by the internal timer. If the standard work time is exceeded, that fact is displayed. Indicator area n9 is the area that displays the process number for the current task. The display is updated when the process changes. Indicator area n13 represents the processing target part for the previous work process. Indicator area n14 represents the processing target part for the currently performed process. Indicator area n15 represents the processing target part for the next work process. Note that the work animation is started by an operation program that is pre-stored in the memory unit 320.

[0074] After displaying the work instructions in step S24, the processing unit 312 determines whether or not there is an input instruction to complete the final work process among the work processes assigned to the workbench 11 (step S25). For example, if a specific key on the keyboard is pressed during the work of the final work process, the processing unit 312 determines that there is an input instruction to complete the final work process (step S25: Yes). Then, the process moves to the inspection process in step S31.

[0075] On the other hand, if it is determined that there is no input instruction to complete the final work process (step S25: No), the processing unit 312 determines whether the position of the hand overlaps with the parts shelf 7 (step S26). Specifically, the processing unit 312 determines whether there has been a significant change in the pixel information of the group in the monitoring range m2 from the initial image information that exceeds a certain standard. If there is no significant change, the processing unit 312 determines that the hand does not overlap with the parts shelf 7 (step S26: No), and repeats the process in step S26.

[0076] If the processing unit 312 determines that there has been a significant change in the pixel information, that is, that the hand has overlapped with the parts shelf 7 (step S26: Yes), it determines whether the hand's position is within the correct area (step S27). Here, the processing unit 312 determines the coordinate validity of the detection range m3. Specifically, in order to prevent the wrong parts from being picked up, the placement positions of the parts boxes 9 containing each part are predetermined. Also, the parts used in each work process are predetermined. Therefore, the processing unit 312 can identify which image information in the monitoring range m2 changes significantly above the standard in each work process. When the worker picks up a part, the processing unit 312 identifies the position of the hand from the position in the detection range m3 and determines whether the hand's position matches the coordinate range of the parts box 9 containing the parts used in the work process being processed. If the position of the parts box 9 containing the parts used in the work does not match the position of the worker's hand, it is determined that the part has been picked up incorrectly. In this way, by setting and verifying the correct position of the parts and the hand that picks them up, it becomes possible to confirm whether worker P is picking up the parts from the correct position and to respond immediately if an error occurs.

[0077] If the processing unit 312 determines that the hand position is not correct (step S27: No), it records a work error (step S28) as the worker having picked up the wrong part, and stores the recorded work error in the processing information storage unit 322. The details of the work error, for example, that the worker picked up a part from the wrong parts box 9 and that this was the first time the mistake occurred, are also stored in the work log information storage unit 325. The processing unit 312 also executes image processing software stored in the storage unit 320 to identify the coordinate position of the worker's hand, specifically the fingertips that are grasping the part, at the time of the error. The processing unit 312 calculates the distance between the coordinate position of the fingertips identified by the image processing software and the center point of the correct parts box 9. The processing unit 312 stores the calculated distance in the processing information storage unit 322. Furthermore, the processing unit 312 stores the maximum value among the distances between the coordinate position of the fingertip and the correct parts box 9, which are stored in the processing information storage unit 322, as the maximum deviation distance.

[0078] Next, the processing unit 312 displays an error message on the display unit 334 (step S29). In the case of incorrect part selection, the error message includes an instruction to select the correct part. The process returns to step S24, and the processing unit 312 again displays the work instructions for the current work process on the display unit 334. At this time, the internal timer for measuring work time may be reset or allowed to continue counting.

[0079] If the processing unit 312 determines that the hand position is in the correct area (step S27: Yes), it determines that the worker has finished the unit task of the current work process and is going to retrieve the part for the next work process, and calculates the work time (step S30). As a result, the elapsed time between work processes is calculated as the work time for that work process. The process returns to step S24, and the processing unit 312 displays the work indication for the next work process on the display unit 334.

[0080] If it is determined that there is an input instruction to complete the final work process (step S25: Yes), the processing unit 312 causes the imaging unit 331 to image the product 10 (step S31). The processing unit 312 uses image processing software to determine the position of the product 10 from the captured image. The processing unit 312 also sends a control command to the power supply unit 333 to temporarily switch the brightness of the lighting devices 3a, 3b, and 3c to inspection brightness. This allows the imaging unit 331 to capture the image area of ​​the product 10 more clearly. The processing unit 312 uses feature information, coordinate information, etc., from the captured image to locate and detect the product 10. This allows the accurate position of the product 10 to be determined and alignment performed.

[0081] The image captured by the imaging unit 331 includes background images such as human hands and the workbench 11, in addition to the product 10. Performing inspection on the entire image including these elements reduces the accuracy of the pass / fail judgment. The alignment detection process is performed to avoid such a decrease in accuracy by recognizing the coordinates where the object to be inspected exists. The alignment process uses template matching, acquires each edge pattern on the product 10, and recognizes the area with a high degree of agreement as the product 10. In this way, the imaging unit 331 performs alignment detection on the image area of ​​the product 10 based on the command of the processing unit 312. In this way, the processing unit 312 can identify the position of the product 10 from the captured image and perform accurate alignment.

[0082] The processing unit 312 performs image inspection by inputting image data of the portion corresponding to the product 10 identified by alignment detection in the image captured by the imaging unit 331 into a trained model stored in the inspection master information storage unit 324 (step S32). The trained model refers to, for example, a deep learning model such as MobileNet that has been pre-trained on a large image dataset. During training, an arbitrary set of normal images is passed through the neural network to output feature vectors for each local region of the intermediate layer. The feature distribution of normal images is constructed by calculating the mean vector and covariance matrix for each local region. On the other hand, during inference, for example, the difference between the feature vector of a defective product image and the feature vector of a trained normal image, i.e., the deviation, is calculated.

[0083] Normal or defective product data for product 10 is input into a trained model, and defective areas in the image are identified using the weights of previously extracted features and the deviation range of normal products in each region. In addition to whether the product is good or bad, if it is found to be defective, the number of defective areas and the deviation from good products are output. The smaller the deviation from good products, the closer it is to good products, and the larger the value, the further it is from good products. Areas where the output deviation exceeds a pre-set threshold are output as defective areas. If the defective areas are connected, for example, they are treated as one defect, and if they are not connected, they are treated as separate defects, and the number of defects is also output. Among the image regions of product 10, the work position coordinates of the unit work procedure data performed at the workbench 11 are the target of judgment. As described above, the unit work procedure data consists of work order, work type, work position coordinates, prohibited work position coordinates, standard work time, part type, part name, part sample image, number of parts, part mounting direction, note comments, work animation, information for unskilled workers, etc. Thus, the image inspection is performed based on image data of product 10, and the object of judgment is the set of work position coordinates from the unit work procedure data.

[0084] The processing unit 312 determines whether the inspection result is defective or not (step S33). If a defect is detected as a result of the image inspection, the processing unit 312 determines that the inspection result is defective (step S33: Yes) and identifies which work position coordinates caused the work error (step S35). Specifically, as shown in the indicated location n16 in Figure 11, the processing unit 312 identifies the defective area and displays the image with the defective area highlighted to inform the worker of the defect. Highlighting means editing the image to make the defective area easier to recognize, for example, by changing the color of the defective area, changing the brightness, making it blink, or surrounding it with a line of a conspicuous color. If the inspection result is defective, the processing unit 312 counts the number of assembly errors and saves the number of times the assembly error occurred (for example, the second time) in the work log information storage unit 325. The processing unit 312 also saves the number of defective areas and the deviation value from good products obtained by the trained model as a result of the image inspection in the processing information storage unit 322. The processing unit 312 stores the maximum value among the stored deviation values ​​from good products as the maximum deviation from good products in the processing information storage unit 322.

[0085] The processing unit 312 redisplays the work instructions for the work process determined to be defective on the display unit 334 (step S36) and prompts the worker to make corrections. The processing unit 312 determines whether there are any inputs that are at a level where correction is not possible (step S37). In step S36, along with redisplaying the work instructions, the processing unit 312 causes the display unit 334 to display a message allowing the worker P to choose whether the correction is possible or not. For example, if the correction is not possible, it displays a message indicating that the worker should press the "Y" key on the keyboard, and if the correction is not possible, it displays a message indicating that the worker should press the "N" key.

[0086] If the processing unit 312 determines that the correction work is at a level where it can be performed (step S37: No), it waits for input indicating that the correction work is complete (step S38). For example, if the operator presses a specific key on the keyboard, the processing unit 312 determines that the correction work is complete. If the processing unit 312 determines that there is no input indicating that the correction work is complete (step S38: No), it repeats the process in step S38.

[0087] If it is determined that the correction work has been completed (Step S38: Yes), the process returns to Step S31. Product imaging, image inspection, and other processes are performed again.

[0088] If the processing unit 312 determines that the correction work is not possible (step S37: Yes), it terminates the work support processing. In the case where it is determined that the correction work is not possible, a function to contact the work supervisor, production technology staff, etc., and an alarm device such as a call button and call light may be added, and a mechanism for calling production engineers may be added.

[0089] If the image inspection results in a good result, the processing unit 312 determines that the inspection result is not a defect (step S33: No) and transmits the work log information of the user's workstation to the work support device 100 (step S34). The work log information of the user's workstation is also stored in the work log information storage unit 325. The work log information of each workstation received by the work support device 100 is stored in the work log information storage unit 125, associated with the workstation ID, operator ID, etc. After that, the work support process ends.

[0090] As described above, the work support system 1 according to the embodiment stores the worker's work performance, determines the worker's skill level based on their past work performance, assigns each unit task to the worker according to the determined skill level, generates divided work procedure data representing the content of the assigned unit task, and displays its content. Therefore, an appropriate amount of work process can be assigned to the worker. In addition, the captured image of the product after work can be applied to a trained model for inferring whether the product is normal or not to determine the quality of the product.

[0091] (Modification) In the embodiment, the various information of the storage unit 120 was provided in the work support device 100, but it may also be provided outside the work support device 100. For example, the various information of the storage unit 120 may be provided on a cloud-type server outside the work support device 100. Similarly, with respect to the collaboration device 300, the storage unit 320 may be located outside the collaboration device 300. Furthermore, the configuration of the work support device 100 and the collaboration device 300 is arbitrary. For example, one collaboration device 300 may also function as the work support device 100. Also, the work support device 100 may have the functions of one or more collaboration devices 300.

[0092] In the above embodiment, normal product data or defective product data of product 10 is input to a neural network, and defective areas in the image are identified using the weights of previously extracted features and the deviation range of normal products in each region. However, inspection information from rule-based image processing may also be added. For example, pattern matching can be used as an example of rule-based image processing. If a case is assumed where a different mounting component is fitted in an area that is judged as NG, an image of the correct mounting component may be registered as a pattern image, and by pattern matching with the image of the product acquired by the imaging unit 331, it may be determined that the product is normal if the degree of agreement between the two images is above a standard value, and defective if the degree of agreement is below a standard value.

[0093] In the above embodiment, the formula for determining proficiency is merely an example and is not limited to it. For example, the number of mistakes in selecting parts and / or the number of assembly mistakes may also be evaluated by dividing by the number of tasks. In addition, in the above embodiment, the total actual work time, number of tasks, number of mistakes in selecting parts, maximum deviation distance from the correct parts box, number of assembly mistakes, number of defective parts, and maximum deviation from good products were used as actual data for determining proficiency, but it is not necessary to use all parameters. Other parameters based on actual data, such as years of experience, may also be added. Similarly, parameters such as age and gender may be added in addition to actual data. Furthermore, the training and placement of workers may be optimized based on the overall proficiency level, which is an indicator that shows the average level of workers' proficiency and its variability. The overall proficiency level of a worker is an indicator calculated based on the average, deviation, etc., of the individual proficiency calculation results for each worker. Specifically, it has the following meanings. In the sense of average, it means the average of repetitive assembly work and corresponds to the average value of the individual proficiency calculation results for each worker. This makes it possible to grasp the trend of the overall proficiency level of workers. Furthermore, deviation is an indicator that shows the variability in the skill level calculation results for each worker. A smaller deviation means that the worker's work quality is less varied.

[0094] Furthermore, in the display of the work preparation screen during the pre-work processing, instructions regarding the arrangement of the parts boxes 9 may be given, taking work efficiency into consideration. For parts boxes 9 that have a parts box 2D code m4, as shown in Figure 9, the imaging unit 331 may read the contents of the parts box 2D code m4, and after work preparation is complete, a determination of whether the arrangement is correct may be made based on that order. Specifically, the computer 310 has pre-stored information indicating the correct arrangement order of the parts boxes 9. This information is stored as part of the work procedure data for product 10. The imaging unit 331 reads the parts box 2D code m4 attached to the parts box 9. This 2D code contains identification information for the parts box 9. The processing unit 312 compares the information of the parts box 2D code m4 read by the imaging unit 331 with the pre-stored information on the correct arrangement order. Specifically, it determines whether the arrangement order is correct based on the identification information of each parts box. If the determination result is correct, it is displayed as "Correct arrangement", and if it is incorrect, it is displayed as "Incorrect arrangement". In this way, the processing unit 312 determines the correct arrangement order of the parts box 9.

[0095] Alternatively, the distance from the position of the parts box 9 to the coordinate position of the hand may be used. The computer 310 has a function to determine the coordinate position of the worker's hand and the distance from the position of the parts box 9 to the coordinate position of the hand from the image captured by the imaging unit 331. Specifically, the computer 310 has image processing software pre-stored. This software identifies the positions of the worker's hand and the parts box 9 from the captured image. Based on the identified positions, the coordinate position of the worker's hand (X coordinate, Y coordinate, Z coordinate) is determined. The processing unit 312 calculates the distance between the position of the parts box 9 and the coordinate position of the worker's hand. This enables the computer 310 to accurately monitor the movements of the worker's hand, the retrieval of parts, etc., and to provide appropriate work instructions.

[0096] The method of assigning the work processes shown in Figure 6 to each worker in the embodiment is just one example and may be modified. In Figure 6, in step S16, the work process assignment to each workbench 11 and worker P was selected from among multiple work process combinations such that the expected value TS of the total work time is minimized. This disclosure is not limited to this. It is not limited to the minimum, and any work process assignment that yields an expected value TS of the total work time that satisfies predetermined conditions may be adopted. Also, an example of accumulating each worker's work performance in the background during work is shown with reference to Figure 8, but this is just one example, and any steps may be taken as long as performance data can be collected while supporting the workers' work. In addition, although a display unit 334 is placed on each workbench 11, other examples of display means include projecting a screen onto multiple workbenches 11 with a single projection device, or displaying a screen for each worker P on a screen common to multiple workers.

[0097] Although preferred embodiments have been described in detail above, the invention is not limited to the embodiments described above, and various modifications and substitutions can be made to the embodiments described above without departing from the scope of the claims.

[0098] Furthermore, this disclosure allows for various embodiments and modifications without departing from its broad spirit and scope. The embodiments described above are for illustrative purposes only and do not limit the scope of this disclosure. That is, the scope of this disclosure is indicated by the claims, not by the embodiments. Various modifications made within the scope of the claims and the equivalent significance of the disclosure are considered to be within the scope of this disclosure.

[0099] This application is based on Japanese Patent Application No. 2024-170286, filed on 30 September 2024. The entire specification, claims, and drawings of Japanese Patent Application No. 2024-170286 are incorporated herein by reference.

[0100] 1 Work support system, 2 Lens, 3a, 3c, 3d Lighting device, 6 Slide rail, 7 Parts shelf, 8a, 8b Fluorescent mark, 9 Parts box, 10 Product, 11 Work desk, 100 Work support device, 110, 310 Computer, 111, 311 Communication unit, 112, 312 Processing unit, 113, 313 I / O unit, 120, 320 Storage unit, 121, 321 Desk information storage unit, 122, 322 Processing information storage unit, 123, 323 Work procedure information storage unit, 124, 324 Inspection master information storage unit, 125, 325 Work log information storage unit, 131 Input unit, 132, 334 Display unit, 300 Interoperability device, 331 Imaging unit, 332 Reading unit, 333 Power supply unit, 1000 Bus, 1001 Processor, 1002 memory, 1003 interface, 1004 secondary storage device, m1 parts shelf boundary, m2 monitoring range, m3 detection range, m4 parts box 2D code.

Claims

1. A work support system that assists in work performed collaboratively by multiple workers, comprising: a skill level calculation means for determining the skill level of the workers according to an evaluation value including the actual work time spent by the workers on the work; a work assignment means for assigning each unit task included in a series of tasks to the workers according to the skill level and generating divided work procedure data representing the content of the assigned unit tasks; and a display means for displaying the content of the divided work procedure data.

2. The work support system according to claim 1, wherein the work assignment means generates the divided work procedure data such that the expected value of the work time for the series of tasks is reduced.

3. A work support system according to claim 1 or 2, comprising: an imaging means for imaging a product obtained by the above work; and an image inspection means for applying the image of the product captured by the imaging means to a trained model for inferring whether the product is a normal product or not, to estimate whether the product is a normal product or a defective product, wherein the display means displays the contents of the division work procedure data again if the product is determined to be a defective product.

4. The work support system according to claim 3, wherein the trained model further estimates defective areas in the captured image, the image inspection means applies the captured image of the product to the trained model to estimate the defective areas, and the display means displays the image with the defective areas emphasized more than other parts.

5. A work support device comprising: a skill level calculation means for determining the skill level of each worker according to an evaluation value based on the worker's past work performance; a work assignment means for assigning each unit task included in a series of tasks to the worker according to the skill level and generating divided work procedure data representing the content of the assigned unit task; and a transmission means for transmitting the divided work procedure data to a cooperating device that operates in conjunction with the device.

6. A work support method that determines the skill level of a worker, assigns each unit task included in a series of tasks to the worker according to the skill level, generates divided work procedure data representing the content of the assigned unit task, and displays the content of the divided work procedure data.

7. A program that causes a computer to perform the following processes: determine the skill level of an operator; assign each unit task included in a series of tasks to the operator according to the skill level; generate divided work procedure data representing the content of the assigned unit task; and display the content of the divided work procedure data.

Images