JP2020052465A - Process estimation device and process estimation method - Google Patents

Abstract

To estimate accurately a worker's execution process even in a situation where it is difficult to estimate only by the worker's work area and motion feature value.SOLUTION: A process estimation device 1 acquires two or more pieces of information out of first information related to the operation of a device, second information related to a work performed by a worker on the device, and third information related to the work performed by the worker other than the second information, calculates a feature value related to the device from one or both of the first information and the second information, calculates a feature value related to the worker from one or both of the second information and the third information, and estimates a work process of the worker on the basis of the calculated feature value related to the device and the calculated feature value of the worker. In an estimation, a value of an objective variable calculated by a function using the time of an estimation target as an explanatory variable is calculated as a feature value at the time.SELECTED DRAWING: Figure 1

Description

  The present disclosure relates to a process estimating apparatus and a process estimating method for estimating a process performed by an operator.

  Generally, when manufacturing a product, operations of a plurality of types of processes are performed. For example, in order to create a printed material, it is necessary to perform a plurality of processes such as printing, bookbinding, cutting, and inspection. In the process of manufacturing a product, it is widely used to record who worked in what process, when, where, and to analyze the work efficiency of each process and use it to manage the operator's budget. . If the number of types of processes required for manufacturing or the amount of work increases, the method of recording work results by manual input by workers increases the number of targets that need to be input. Load increases.

  As a technique for creating a work record without imposing a load on an operator, there is, for example, Patent Document 1. In Patent Literature 1, an operation type is specified with high accuracy based on a movement feature amount of a worker and a feature amount stored in advance with respect to a field candidate where the worker is located. Further, based on the specified work type, a field under work is specified from the field candidates.

  On a farm to which the technology described in Patent Literature 1 is applied, an operator performs work related to a field where the worker is located, and when moving to another field, performs work related to that field. Therefore, the type of work can be specified with high accuracy by using the feature amount related to the field where the worker is working.

JP 2010-161991 A

  At a printing site, a plurality of work steps configured by similar operations may be performed in the same work area (corresponding to a field in Patent Document 1). In such a site, it is difficult for the worker to correctly specify the process being performed by the method using the position and the movement feature amount of the worker described in Patent Document 1. In particular, in a work environment in which data relating to the movement of the worker cannot be acquired with sufficient accuracy, it is difficult to correctly specify the work process being performed unless estimation is performed using other information.

  Patent Literature 1 describes a means for specifying the work type of a worker to be estimated, assuming that the work type of the worker to be estimated is the same as the work type of another worker working nearby. ing. However, at a printing site, a plurality of workers may be working in different processes near the same work area. In such a printing site, it is inappropriate to apply the method of using information of another worker described in Patent Document 1.

  The present disclosure has been made in view of such a problem, and an object of the present disclosure is to use a relationship between an operation time (or time period) of an apparatus and a work execution time (or time period) by an operator. Another object of the present invention is to provide a technique capable of accurately estimating an execution process of an operator.

  One embodiment of the present disclosure relates to a process estimation device. The process estimating apparatus includes two pieces of first information related to the operation of the apparatus, second information related to work performed by the worker on the apparatus, and third information related to work performed by the worker other than the second information. Acquiring means for acquiring one or more pieces of information; device characteristic amount calculating means for calculating a characteristic amount relating to the apparatus from one or both of the first information and the second information; and second information and third information From any one or both of the above, an operator feature amount calculation unit that calculates a feature amount regarding the worker, a feature amount calculated by the device feature amount calculation unit, and a feature amount calculated by the worker feature amount calculation unit, Estimating means for estimating the work process of the worker based on the The device feature value calculating means calculates a value of the target variable calculated by a function using the time of the estimation target as an explanatory variable as a feature value at the time. The function is a linear or non-linear function determined based on one or both of the first information and the second information.

  According to the present disclosure, it is possible to accurately estimate an execution process of an operator even in a situation where it is difficult to estimate only by a work area and a movement feature amount of the operator.

FIG. 1 is a diagram illustrating an example of a configuration of a process estimation system according to a first embodiment. The figure which shows an example of the hardware constitutions of a process estimation apparatus. The figure which shows an example of the table structure holding the content of work definition information. The figure which shows an example of the table structure which holds the content of an order and a plan. FIG. 2 is a diagram illustrating an example of an apparatus layout at a printing site, an operable area of the apparatus, and a work area. FIG. 2 is a diagram illustrating functions of a process estimation device according to the first embodiment. FIG. 9 is a diagram illustrating a first example of a table configuration that holds data relating to a feature amount. FIG. 4 is a diagram for explaining a method of creating a feature amount related to a position and an orientation of an operator. FIG. 9 is a diagram illustrating a second example of a table configuration that holds data relating to a feature amount. The figure which shows the concept of a reference time and a reference time parameter. FIG. 9 is a diagram illustrating an example of a table configuration that stores a distribution pattern used when calculating a feature amount regarding the device. The figure which shows the processing flow of an operator terminal and an apparatus at the time of acquiring the information regarding an operator and an apparatus. FIG. 6 is a diagram illustrating a processing flow of an apparatus feature amount creation unit and an operator feature amount creation unit according to the first embodiment. FIG. 6 is a diagram illustrating a processing flow of a correction unit that corrects an estimation result in the first embodiment. The figure which shows the function of the process estimation apparatus in 2nd Embodiment. FIG. 9 is a diagram illustrating the concept of a table configuration that holds data relating to a feature amount according to the second embodiment. FIG. 13 is a diagram illustrating an example of a table configuration that stores a distribution pattern used when calculating a feature amount regarding an apparatus according to the second embodiment. FIG. 13 is a diagram showing a processing flow relating to creation of a feature amount related to an apparatus in the second embodiment. FIG. 13 is a diagram showing a processing flow relating to creation of a feature amount related to an apparatus in the second embodiment. FIG. 13 is a diagram showing a processing flow relating to creation of a feature amount related to an apparatus in the second embodiment.

  Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Note that the configurations shown in the following embodiments are merely examples, and the present disclosure is not limited to the illustrated configurations. The same or equivalent components, members, and processes shown in each drawing are denoted by the same reference numerals, and the repeated description will be omitted as appropriate. In each drawing, some of the members that are not important for the description are omitted.

  At the printing site, the time between the time (or time zone) when the apparatus operates, the time (or time zone) at which the operator worked on the apparatus, and the time (or time zone) at which the operator performs other work. Often have significant relationships. Regardless of whether or not the area where the operating device exists, the area where the worker performs the work on the device, and the area where the worker performs other work are the same, Relationships can exist.

  As a first example, for example, in a certain printing site, an operator moves the printed matter output to the output tray to another place every five minutes after the printing press starts printing. There are many. As a second example, at a certain printing site, the operator performs inspection work of the cut printed material during a time period during which the automatic cutting machine is operating and a time period until one hour after the operation is stopped. Often implemented. As a third example, in a certain printing site, automatic cutting and bookbinding are often performed in parallel, and a time period from 5 seconds to 30 seconds after the automatic cutting machine starts operating, It is highly probable that the worker is binding the printed material to be cut next.

  That is, irrespective of whether or not the device operates in the area where the worker is working, information indicating when the operation of the device has started (finished) or scheduled to be started (scheduled). By taking this into account, it is possible to accurately estimate the process in which the worker is working.

<First embodiment>
The proposed process estimating apparatus and process estimating method are constructed at a site where products are produced by humans and devices. Hereinafter, a description will be given by taking a printing factory as an example, but the present invention is not limited to this, and a process estimation device can be constructed at another production site.

  FIG. 1 shows a configuration diagram of a process estimation system according to the first embodiment. The process estimating device 1 collects data obtained from the printing plant, such as operation logs of devices operating in the printing plant, information on the positions of workers, information on plans, and the like, to determine the work process being performed at the printing site. It is a device for estimating. The printing press 2 is a device for printing on paper. The bookbinding machine 3 is a device for binding printed paper. The cutting machine 4 is a device for cutting a sheet. The portable information terminal 5 refers to a terminal such as a mobile phone, a PHS, and a notebook PC that can be carried by an operator and can communicate with a wireless LAN. The portable information terminal 5 has a sensor for measuring acceleration, angular acceleration, geomagnetism, and the like inside. Further, the portable information terminal 5 can transmit the measured sensor information together with the time information to the process estimating device 1 and the position estimating device 7 via the wireless access point 6. The position estimating device 7 is a device for estimating the position of the worker holding the portable information terminal 5 from the sensor information received from the portable information terminal 5. The work definition information DB 8 is a DB (Data Base) that stores information used when creating a work plan for workers and apparatuses at the printing site. The order DB 9 is a DB that manages orders received by the printing factory. The plan making device 10 is a device that makes use of the order DB 9 and the work definition information DB 8 to create a plan to be followed by the worker and the device. The devices 1, 2, 3, 4, 7, 10, DBs 8 and 9, and access point 6 are connected by a local area network 11.

  FIG. 2 shows a hardware configuration diagram of the process estimation device 1. The input device 12 is a mouse, a digitizer, or the like, and is used to input an instruction or a selection from a user of the process estimation device 1. An administrator who manages the work site, an operator who operates the process estimation device 1, an operator who actually works at the site, and the like can be users of the process estimation device 1. The output device 13 is an LED (Light Emitting Diode) for displaying the state of the apparatus or the processing content, a liquid crystal panel, or the like. The CPU (Central Processing Unit) 14 reads the program stored in the secondary storage device 16 into the primary storage device 15. Further, the CPU 14 interprets and executes the read program to execute various controls and calculations in the apparatus and display a UI (User Interface). The primary storage device 15 is mainly a memory such as a RAM (Random Access Memory). The secondary storage device 16 stores a process estimation program necessary for operating the process estimation device 1. The secondary storage device 16 is, for example, a storage medium such as a hard disk or a flash memory, but it goes without saying that the present disclosure does not depend on a specific storage medium. Generally, the capacity of the primary storage device 15 is smaller than the capacity of the secondary storage device 16, and programs and data that cannot be stored in the primary storage device 15 are stored in the secondary storage device 16. Further, data that must be stored for a long time is also stored in the secondary storage device 16. The communication IF (Interface) 17 is connected to the local area network 11 and is a communication interface for communicating with various devices, DBs, access points, and portable information terminals 2 to 9 in the factory. Reference numeral 18 denotes a bus for communicatively connecting the components 12 to 17.

  FIG. 3 shows an example of a storage format of work definition information managed by the work definition information DB8. The work definition information is information necessary for manufacturing each product. For example, information such as a process required for manufacturing, the number of unit processes and unit processing time thereof, and work resources (operators, devices) in charge of the process are included. The work definition information also includes layout information of the arrangement of the devices, a work area in which work on each device is performed, and area information for performing operations on each device.

  FIG. 3A shows a product table 3100. A column 3101 stores information of a product ID for uniquely identifying each product. A column 3102 stores a product type indicating a product as a product name. A column 3103 stores information on manufacturing accuracy as an attribute item. The manufacturing accuracy is information on the accuracy of manufacturing when manufacturing a product. In the example of FIG. 3A, three levels of manufacturing accuracy of high, medium, and low are shown, but the present invention is not limited to this. For example, it may be expressed by inputting a numerical value representing accuracy. A column 3104 stores information on whether or not the order is an express order as an attribute item. The express order or not is information indicating whether or not the order between the order receiving date and the deadline date is short. In the example of FIG. 3A, when 1 is input, it indicates that it is an express order. When 0 is entered, it is not an express order. In the case where the manufacturing procedure of the product is the same regardless of whether the order is an express order or not, a value such as “1, 0” may be stored in the column 3104 and shared. In the example of FIG. 3A, whether or not an express order is represented by a binary value of 0 or 1 is not limited thereto. For example, the number of days from the order receiving date to the deadline or a range of the number of days may be indicated as the attribute value. Although two pieces of information are shown here as attribute items, the present invention is not limited to them, and attributes may be added as needed.

  FIG. 3B shows a worker table 3200 for managing information on workers who work at the printing site. A column 3201 stores information of a worker ID for uniquely identifying each worker. A column 3202 stores information on the names of the respective workers, and is displayed on a screen that displays plan information.

  FIG. 3C shows a device table 3300 for managing information on devices operating at the printing site. A column 3301 stores information of a device ID that uniquely identifies each device. Column 3302 stores information on the name of each device, which is displayed on a screen displaying plan information. A column 3303 stores information on where each device is located at the printing site. A column 3304 stores information of an area to which the worker can move when performing the work on each device. A column 3305 stores information of an area where an operator can perform an arbitrary operation on each device. The details of the information stored in columns 3303, 3304, and 3305 will be described later with reference to FIG.

  FIG. 3D shows a work definition table 3400 for specifying on which work definition each product is manufactured. The column 3401 stores the above-described product ID. A column 3402 stores a work definition ID 3502 included in a process table 3500 shown in FIG. In the process table 3500, the same work definition ID is assigned to the processes included in the same work definition information. That is, the work definition table 3400 is a table for linking each product and its manufacturing procedure, and can have a many-to-many relationship. That is, there may be a plurality of manufacturing procedures for manufacturing a certain product, and different products may be manufactured by the same manufacturing procedure.

  FIG. 3E shows a process table 3500 for managing main processes necessary for manufacturing a product. It should be noted that sub-processes to be performed along with the main process are managed in a sub-process table 3600 shown in FIG. A column 3501 stores information of a process ID for uniquely identifying each process. In a column 3502, a work definition ID indicating which work definition is associated with the process is stored. In a column 3503, a type of a process indicating what kind of work is stored as a process name. A column 3504 stores information on a worker who performs a process. Note that the ID stored in the column 3504 refers to the worker ID in the worker table 3200 described above. If there is a plurality of IDs stored in the column 3504 for one process ID, it means that the process can be performed by any worker as long as the worker corresponds to the stored worker ID. A column 3505 stores information indicating what device is used to execute the process. The information stored in the column 3505 refers to the device ID in the device table 3300 described above. In addition, “−” is described in the column 3504 of the process that does not require an operator's operation, and “−” is described in the column 3505 of the process that does not require an apparatus. For example, the printing process in the process ID = 1 is performed by operating the apparatus without an operator. In addition, the inspection process of process ID = 3 is performed by visual confirmation of an operator without using the apparatus. On the other hand, in the cutting process with the process ID = 2, the value is stored in both the column 3504 and the column 3505 because the operator manually performs the cutting using the manual cutting machine.

  A column 3506 stores the number of unit processes for each process, and a column 3507 stores a unit process time that defines how much work time is required for each unit process. As a specific target of the number of unit processes, for example, the number of pages is assumed. A column 3508 stores information on the standard deviation as information on the variation regarding the unit processing time. The information in the columns 3506 to 3508 is used when creating a work plan and in processing for correcting the result of the process estimation described later. For example, when creating a work plan, it is necessary to calculate how many man-hours each main process requires. For that purpose, first, the total number of pages required for manufacturing is acquired from the number of pages of ordered products and the number of ordered copies, which will be described later. Next, the total number of processes is obtained by dividing the total number of pages by the number of unit processes. Finally, the man-hour required for each main process is calculated by adding the unit processing time to the total number of processes. A column 3509 stores information on constraint conditions between processes. In FIG. 3E, for example, the cutting of the process ID of 2 can be performed only on the printed intermediate product having the process ID of 1. More specifically, when the printing with the process ID of 1 is performed and the intermediate products corresponding to the number of unit processes defined in the column 3506 are collected, the cutting with the process ID of 2 can be performed. In this case, printing with a process ID of 1 and cutting with a process ID of 2 may be performed in parallel.

  FIG. 3F defines a sub-process table 3600 for managing sub-processes to be performed when the process defined in the process table 3500 is performed. The sub-process is an auxiliary work process necessary for performing the main process described in the process table 3500. For example, in order to execute a printing process with the process ID = 1 in the process table 3500, a pre-process such as inputting settings of a printing press and setting sheets necessary for printing is required. Further, post-processing such as taking out the printed paper from the tray is required. These pre-processing and post-processing are managed in the sub-process table 3600 as sub-processes. A column 3601 stores information of a sub-process ID for uniquely identifying each sub-process. Column 3602 stores a process ID indicating which main process is a sub-process associated with the main process. The process ID stored in the column 3602 refers to the process ID in the process table 3500 described above. In a column 3603, information for specifying whether the process is a pre-process or a post-process is stored as a type. Specifically, the value of Pre is stored in the case of pre-processing, and the value of Post is stored in the case of post-processing.

  Column 3604 stores information on the worker who performs the sub-process. The ID stored in the column 3604 refers to the worker ID in the worker table 3200 described above. If there is a plurality of IDs stored in the column 3604 for one sub-process ID, it means that the process can be performed by any worker as long as the worker corresponds to the stored worker ID. Column 3605 stores information indicating what device is used to execute the sub-process. Note that the information stored in the column 3605 refers to the device ID in the device table 3300 described above. It should be noted that “−” is described in the column 3604 of the sub-process that does not require the operation by the operator, and “−” is described in the column 3605 of the sub-process that does not require the apparatus. A column 3606 stores the number of unit processes of each sub-process, and a column 3607 stores a unit processing time that defines how much work time is required for each unit process. As a specific target of the number of unit processes, for example, the number of pages is assumed. A column 3608 stores information on the standard deviation as the information on the variation regarding the unit processing time. The information in the columns 3606 to 3608 is used when creating a work plan and in the process of correcting the result of the process estimation described later, similarly to the columns 3505 to 3508.

  The data format is not limited as long as it has information equivalent to the tables 3100 to 3600. For example, the data may be converted into a data format defined by XML or the like and stored.

  FIG. 4 shows an example of an order received from a customer and a storage format of a manufacturing plan created in response to the order. Information on the order is stored in the order DB 9. Information on the plan is created by the plan creation device 10, and the created plan information is transmitted to the process estimation device 1.

  FIG. 4A shows an order table 4100 for managing information on orders received from customers. Column 4101 stores an order ID for uniquely identifying each order. Column 4102 stores the type of ordered product as a product name. A column 4103 stores information on the number of pages per product of the ordered product. A column 4104 stores the number of copies, which is information on the number of orders received. For example, when the product to be produced is a flyer composed of one page, the number of copies is the number of leaflets, and when the product to be produced is a booklet, the number of copies is the number of books to be shipped as a booklet. A column 4105 stores information on manufacturing accuracy required when manufacturing a product. Column 4106 stores information as to whether the order is an express order. A column 4107 stores date information of the order received. Column 4108 stores information on the delivery date of each order.

  FIG. 4B shows a worker plan table 4200 for managing information on a plan when each worker performs each main process or each sub-process. Column 4201 stores a worker plan ID for uniquely identifying each plan. Column 4202 stores a worker ID for specifying a worker who performs the planned work. Note that the information stored in the column 4202 refers to the worker table 3200 in FIG. When a plurality of worker IDs are described in the column 4202 for one worker plan, this indicates that the plan is likely to be executed by a plurality of workers. The work may be performed by only one worker, or a plurality of workers may perform the work jointly. Column 4203 stores order ID information for specifying which order is a work plan for processing. Note that the information stored in the column 4203 refers to the order table 4100 in FIG. A column 4204 stores a task definition ID for specifying the task definition information based on which the task is planned. Note that the information stored in the column 4204 refers to the process table 3500 in FIG. A column 4205 stores a process ID for specifying which process is to be performed. Note that the information stored in the column 4205 refers to the process table 3500 in FIG. A column 4206 stores a sub-process ID for specifying which sub-process is to be performed. Note that the information stored in the column 4206 refers to the sub-process table 3600 in FIG.

  As shown in FIG. 4 as an example, each work plan is a plan relating to either a main process or a sub-process. Therefore, if an operation related to the main process is planned, “−” is stored in the sub-process ID in column 4206, and if an operation related to the sub-process is planned, “−” is stored in the process ID in column 4205. You. Column 4208 stores information on the date and time at which work is to be started. Column 4209 stores information on the date and time at which the work is to be completed. Column 4207 shows what percentage of the time zone between columns 4208 and 4209 is planned as the net working time of each planned process. For example, in the plan with the work plan ID = 1, the sub-process ID = 1 is scheduled from 9:00:00 to 9:01 on 2017/12/11. That is, it means that the operation of the sub-process ID = 1 is performed during 1 minute from 9:00 to 9:01. In addition, since the implementation ratio is 50%, the actual time required for the work is 30 seconds, which is 50% of the one minute. Note that the actual time required for the work is calculated by multiplying the number of unit processes of the process stored in the process table and the sub process table, the unit processing time, and the number of pages and the number of copies stored in the order table. . Also, when a plurality of worker IDs are stored in the column 4202, there is a possibility that the work is performed by a plurality of persons. In that case, the total time obtained by adding the time required for each worker to perform the work corresponds to the actual time required for the work.

  FIG. 4C shows an apparatus plan table 4300 that manages information on plans when each apparatus performs each main step or each sub-step. A column 4301 stores an apparatus plan ID for uniquely identifying each plan. Column 4302 stores a device ID for specifying a device that performs the planned work. Note that the information stored in the column 4302 refers to the device table 3300 in FIG. Column 4303 stores order ID information for specifying which order is a work plan for processing. Note that the information stored in the column 4303 refers to the order table 4100 in FIG. A column 4304 stores a task definition ID for specifying which task definition information is used as a basis for planning. Note that the information stored in the column 4304 refers to the process table 3500 in FIG. A column 4305 stores a process ID for specifying which process is to be performed. Note that the information stored in the column 4305 refers to the process table 3500 in FIG. A column 4306 stores a sub-process ID for specifying which sub-process is to be performed. Note that the information stored in the column 4306 refers to the sub-process table 3600 in FIG.

  As shown in FIG. 4 as an example, each work plan is a plan relating to either a main process or a sub-process. Therefore, when an operation related to the main process is planned, “−” is stored in the sub-process ID of column 4306, and when an operation related to the sub-process is planned, “−” is stored in the process ID in column 4305. You. Column 4308 stores date and time information of the scheduled start of the work. Column 4309 stores information on the date and time when the work is to be completed. Column 4307 shows what percentage of the time zone between columns 4308 and 4309 is planned as the net working time of each planned process. The specific meaning is the same as the execution ratio of the column 4207 of the worker table 4200.

  The worker planning table 4200 and the device planning table 4300 are created by the plan creating device 10. Note that the format of the data does not matter as long as it has information equivalent to the tables 4100, 4200, and 4300. For example, the data may be converted into a data format defined by XML or the like and stored.

  FIG. 5 shows an apparatus layout, an operable area (an area where an operator can perform some operation on the apparatus) and a work area at an assumed printing site. FIG. 5A shows a floor map of the entire printing site. At the printing site, a plurality of devices, work tables (desks 1 to 5), and places (shelf 1, shelf 2) for storing materials and the like are arranged. Further, there are output places (output trays 1 to 3) for intermediate products output from the apparatus and temporary storage places for temporarily storing completed products before shipping. Information on the location of each device is stored in the device table 3300 and is managed as information on three or more coordinates. Specifically, it is assumed that the coordinate points stored in column 3303 are located in an area created by connecting the stored coordinate points with a line in the order of the stored coordinate points. By increasing the number of coordinate points to be stored, the position of each device can be accurately represented. The expression format of the position of the device is not limited to the above example. For example, the position of the device may be managed by one coordinate, or the center coordinates of the device may be stored in the device table 3300.

  FIGS. 5B and 5C show the details of a part of the floor in FIG. 5A. In FIG. 5B, a range in which an operator can move when performing a main process or a sub-process related to bookbinding using the bookbinding machine in the drawing is indicated by thin hatching. This thin shaded area is defined as a work area for the main process or the sub process. In the example of FIG. 5B, it is suggested that the worker may use the desk 2 or the desk 3 to prepare an object to be put into the bookbinding machine when performing the main process or the sub-process related to bookbinding. . Also, in FIG. 5B, the range in which the operator can perform some work on the bookbinding machine in the figure is indicated by dark hatching. This dark shaded area is defined as an operable area of the apparatus. Examples of operations that can be performed in the dark shaded area include changing the setting of the bookbinding machine, setting an intermediate product to be bound in the bookbinding machine, and taking out the bound intermediate product from the bookbinding machine. In FIG. 5C, the work area and the operable area of the cutting machine 2 are shown by light and dark hatching, respectively. Information on the work area and the operable area of each device is stored in columns 3304 and 3305 of the device table 3300. The expression format of the work area and the operable area is information of three or more coordinates, and the area defined thereby is the same as the arrangement position of the above-described device.

  FIG. 6 shows a functional block diagram for explaining main functions of the process estimating apparatus 1 according to the present embodiment. Each block shown here can be realized in terms of hardware by elements such as a computer CPU and other mechanical devices, and can be realized in terms of software by a computer program or the like. The functional blocks are drawn. Therefore, it will be understood by those skilled in the art referred to in this specification that these functional blocks can be realized in various forms by a combination of hardware and software.

  The process estimation device 1 includes a communication unit 101, an acquisition unit 102, an operator feature amount creation unit 103 (operator feature amount calculation unit), an apparatus feature amount creation unit 104 (device feature amount calculation unit), and a feature amount. It includes a data holding unit 105, a correct answer data adding unit 106, an estimation model creation unit 107, a process estimation unit 108, a correction unit 109, and an estimation result storage unit 110.

  The communication unit 101 realizes a function for another function of the process estimation device 1 to perform data communication with a device, a DB, or an access point in a factory connected via the local area network 11.

  The acquisition unit 102 acquires information on the movement of the worker, operation information of the device, plan information, work definition information, and order information via the communication unit 101. The information on the movement of the worker includes measurement results of various sensors measured by the portable information terminal 5 held by the worker, and position information of the portable information terminal 5 estimated by the position estimation device 7. Since the portable information terminal is held by the worker who works in the printing factory, the position information of the worker is acquired. The operation information of the apparatus includes an apparatus log, an apparatus operation sound, and the like obtained from apparatuses such as the printing press 2, the bookbinding machine 3, and the cutting machine 4 operating in the factory. A specific method of transmitting data from the portable information terminal 5 to the acquisition unit 102 and a method of transmitting data from the devices 2 to 4 to the acquisition unit 102 will be described later with reference to FIG. The plan acquired by the acquisition unit 102 is plan information created by the plan creation device 10, and more specifically includes a worker plan table 4200 and a device plan table 4300. The work definition information acquired by the acquisition unit 102 includes the product table 3100, the worker table 3200, the device table 3300, the work definition table 3400, the process table 3500, and the sub process table 3600 shown in FIG. The order information acquired by the acquisition unit 102 includes an order table 4100. The work definition information and the order information are acquired from the work definition information DB 8 and the order DB 9, respectively.

  The worker feature value creation unit 103 and the device feature value creation unit 104 receive the information acquired by the acquisition unit 102 and create a feature value used in machine learning or deep learning at each time. The device feature value creation unit 104 calculates, as the feature value at the time, a value of the target variable calculated by a function using the time of the estimation target as an explanatory variable. This function is a linear or non-linear function determined based on one or both of information relating to the operation of the apparatus and information (operation log) relating to work performed by the operator on the apparatus. The outline of the feature amounts created by the worker feature amount creation unit 103 and the device feature amount creation unit 104 and the creation procedure will be described later with reference to FIGS.

  The feature amount data holding unit 105 holds the feature amount data created by the worker feature amount creation unit 103 and the device feature amount creation unit 104.

  The correct answer data assigning unit 106 determines the type of process (printing, bookbinding, inspection, etc.) actually performed at the printing site at each time on the feature amount for each time held by the feature amount data holding unit 105 and corrects the machine learning. Assigned as data. The correct answer data is added only to the feature amounts that can grasp the processes performed at the printing site at each time. The feature quantity to which the correct answer data is added is used as learning data when the estimation model creation unit 107 described later creates an estimator (hereinafter also referred to as an estimation model). The feature amount at the time to which no correct answer data is given is a target to be estimated by the estimator created by the estimation model creating unit 107.

  The estimation model creation unit 107 creates an estimator for estimating the process using the feature amount to which the correct answer data assigning unit 106 has assigned the correct answer data as learning data.

  The process estimating unit 108 uses the estimator created by the estimation model creating unit 107 to execute the feature amount at any time to which the correct data held by the feature amount data holding unit 105 is not added, at that time. Estimate the type of work process that was performed. Further, the process estimating unit 108 transmits the estimation result to the correction unit 109 and the estimation result storage unit 110 described below.

  The correction unit 109 specifies a portion having a high possibility of being erroneously estimated from the results estimated by the process estimation unit 108 and corrects the result to a correct result. At the time of correction, the information acquired by the acquiring unit 102 and the past estimation result are used. The specific procedure of the correction will be described later with reference to FIG.

  The estimation result storage unit 110 stores the estimation result estimated by the process estimation unit 108. Further, the estimation result corrected by the correction unit 109 is stored.

  FIG. 7 shows a feature amount table 7000 stored in the feature amount data storage unit 105. Each row of the feature amount table 7000 holds data on the printing factory at each time. Note that the feature amount table 7000 is created for each worker to be estimated. The feature amount table 7000 is used for learning the above estimator. The feature amount table 7000 is also input data used when the created estimator estimates the type of process performed by the worker to be estimated at an arbitrary time. A column 7001 stores time information for specifying which time data corresponds to each row of the feature amount table 7001. The information stored in the columns 7002 and 7003 expresses where the portable information terminal 5 was located in the printing factory as information on a plane, and indicates the x coordinate and the y coordinate in the printing factory, respectively. The information in the columns 7002 and 7003 is created by the worker feature amount creating unit 103 based on the location information of the portable information terminal 5 at each time estimated by the location estimation device 10. The information stored in column 7004 indicates in which direction the portable information terminal 5 is facing as information on a plane, and is indicated by a value (unit: degrees) between 1 and 360. The information stored in the column 7004 is created by the worker feature amount creating unit 103 based on the information on the orientation at each time from the portable information terminal 5 obtained by the acquiring unit 102. In addition, although the feature quantity relating to the position and the orientation is provided in the feature quantity table 7000 as information on a plane, that is, two-dimensional information, it may be provided as three-dimensional information, ie, three-dimensional information. . Columns 7005, 7006, and 7007 store feature quantities including information on the printing machine 2, the bookbinding machine 3, and the cutting machine 4, respectively. Specifically, a value calculated based on a predetermined calculation method based on operation information of each device or information that a worker has performed an arbitrary operation on the device is stored as a feature amount. ing. Note that, although not shown in FIG. 7, the feature amounts related to other apparatuses operating in the printing factory are also included in the feature amount table 7000 (column 7008). The feature values of each device stored in columns 7005 to 7008 are created by the device feature creating unit 104 using the information acquired by the acquiring unit 102. A specific method of calculating the feature value of each device will be described later with reference to FIGS. The column 7009 stores the work process actually performed by the estimation target at each time as correct answer data in machine learning. It should be noted that the assignment of the correct answer data is explicitly performed by the user of the process estimation device 1 via the input device 12, and is performed by the function of the correct answer data assigning unit 106.

  Although FIG. 7 illustrates the feature amount table 7000 used in the present embodiment, the feature amount table that can be used in the present disclosure is not limited to the example illustrated in FIG. A feature amount related to the posture and movement of the worker and a feature amount related to the planning information of the worker and the device may be included. Further, the information used when creating the feature amounts in FIG. 7 may be used based on different feature amount designs.

  As an example, another method of creating a feature amount from the position and orientation of the worker will be described with reference to FIGS. According to the methods described in FIGS. 8 and 9, it is possible to create a feature amount in which the degree of abstraction of information on the position and orientation of the worker is increased. FIG. 8A defines a route that can be moved by an operator at a printing site by dividing the route into a plurality of areas. FIG. 8B is a conceptual diagram showing a situation where a worker is facing the device when there is a worker in the operable area of the device. When the method described with reference to FIGS. 8 and 9 is applied, it is assumed that an operable area is defined for desks, shelves, output trays, and the like other than the devices arranged at the printing site.

  FIG. 9A shows a feature amount table 7000 in which the degree of abstraction of the position and orientation information of the worker is increased. In the feature amount table 7000 shown in FIG. 9A and the feature amount table shown in FIG. 7, the feature amounts (columns 7005 to 7008) regarding the device are the same. The information included in the feature amounts in the columns 7002 to 7004 in FIG. 7 has the feature amounts in the column 7010 in FIG. A method of calculating the feature amount in column 7010 will be described. First, at each time in the column 7001, it is determined whether or not the operator is in the operable area of the device, desk, shelf, or output tray located at the printing site (YES or NO). If the operator is in the operable area, it is determined whether the worker is facing the device, desk, shelf, or output tray at that time (YES or NO). If YES in both steps, categorical information (first worker summary information) that can identify the device, desk, shelf, and output tray is stored in column 7010. In both steps, if either is NO, categorical information (which can identify which path the worker is located in and defined in which direction the path is located in FIG. 8A) The second worker summary information is stored in the column 7010. The direction of the passage may be determined by identifying the flow line created by the movement of the worker and determining the direction of the flow line.

  FIG. 9B shows a feature amount table 7000 in which time-series context information is added to the position and orientation feature amounts. In the feature quantity table 7000 shown in FIG. 9B and the feature quantity table 7000 shown in FIG. 9A, the feature quantities stored in the columns 7010 and 7005 to 7008 are the same. The feature amounts in the columns 7011 and 7012 have information for grasping how the feature amounts in the column 7010 have changed over time. The feature amount in the column 7013 has information for grasping how long the value of the feature amount in the column 7010 has been continuously generated. A specific creation procedure of the feature amounts in the columns 7011 to 7013 will be described. First, the value (NOW VALUE) of the column 7010 at the creation target time of the feature amount is specified. Next, the value (BEFORE VALUE, immediately preceding worker summary information) of the latest column 7010 that occurred before the creation target time different from the specified value (NOW VALUE) and the latest time (BEFORE TIME) are specified. In addition, a value (AFTER VALUE, immediately after worker summary information) of the latest column 7010 generated after the creation target time different from the value (NOW VALUE) and its latest time (AFTER TIME) are specified. In the columns 7011 and 7012 at the creation target time, a value (BEFORE VALUE) and a value (AFTER VALUE) are stored, respectively. A column 7013 at the creation target time stores a value (summary information duration) for identifying the length of the time zone of (BEFORE TIME) to (AFTER TIME).

  The outline of the method of calculating the value of the feature amount for each device stored in columns 7005 to 7008 of the feature amount table 7000 will be described with reference to FIGS. 10 and 11. A more detailed calculation method will be described later with reference to FIG.

  FIG. 10 is a diagram showing the concept of a reference time and a reference time parameter used when creating a feature quantity relating to the device. A row 1001 in FIG. 10 illustrates the operation status of the cutting machine 1 in chronological order. A box 1011 in a row 1001 indicates a time zone during which the cutting machine 1 is operating. A row 1002 indicates a reference time required when determining a distribution created in rows 1003 to 1008 described below. In FIG. 10, as the reference time, a time when the cutting machine 1 starts operation (Start), a time when the operation ends (End), and a time (Center) which is the center of the time period from the operation start to the operation end are taken as an example. Has defined. In actual operation, another time may be defined as the reference time. Rows 1003 to 1008 each show a concept in which an arbitrary distribution is created based on the reference time and the reference time parameter determined in FIG. Since FIG. 10 is intended to explain the concept of the reference time and the reference time parameter, the same normal distribution is uniformly described as an arbitrary distribution, but the type of distribution actually applied is shown in FIG. And is determined by the method described later.

  In FIG. 10, the center position on the time axis of the normal distribution shown in rows 1003 to 1008 is determined based on the reference time and the reference time parameter set respectively. The reference time indicates the initial time at the center of the normal distribution, and the reference time parameter indicates how much the center of the normal distribution is moved from the reference time. When the reference time parameter has a positive value, the normal distribution is translated in the future from the reference time. When the reference time parameter is a negative value, the normal distribution is translated from the reference time to the past. When the value of the reference time parameter is 0, the normal distribution is not moved, and its center remains coincident with the reference time. In row 1003, the reference time is Center and the reference time parameter is 0. Therefore, the center of the distribution of the row 1003 is the time at the center of the box 1011. In row 1004, the reference time is Center and the reference time parameter is D. Therefore, the center of the distribution of the row 1004 is the time when D has elapsed from the time of the center of the box 1011. In row 1005, the reference time is Start and the reference time parameter is 0. Therefore, the center of the distribution of the row 1005 is the time when the cutting machine 1 starts operating. In row 1006, the reference time is Start and the reference time parameter is -2D. Therefore, the center of the distribution of the row 1006 is a time 2D in the past from the time when the cutting machine 1 starts operating. In row 1007, the reference time is End and the reference time parameter is 0. Therefore, the center of the distribution of the row 1007 is the time when the cutting machine 1 ends the operation. In row 1008, the reference time is End and the reference time parameter is 2D. Therefore, the center of the distribution of the row 1008 is the time when 2D has elapsed from the time when the cutting machine 1 ends the operation.

  FIG. 11A illustrates a distribution pattern definition table 1100 that stores distribution pattern information used when calculating a feature amount related to an apparatus. Note that the distribution pattern definition table 1100 is stored in the secondary storage device 16 as a setting file to which the process estimation device 1 refers. The arbitrary distribution in FIG. 10 is created based on the distribution pattern definition table 1100. The values of the feature values 7005 to 7008 at each time are distribution values at each time. A column 1101 stores a pattern ID for uniquely identifying a distribution pattern. Column 1102 stores the device ID. The device ID refers to the device ID in the column 3301 of the device table 3300. A column 1103 stores information serving as a base of a distribution finally created (hereinafter also referred to as a distribution base). Specifically, information indicating an outline of the distribution is stored. For example, information for specifying a basic property of the distribution such as a normal distribution, a Poisson distribution, and an exponential distribution is stored. Information specifying detailed properties of the distribution, for example, how much the distribution of the type shown in column 1103 has a variation, etc., is stored in column 1104. A column 1104 stores various parameters (hereinafter also referred to as distribution parameters) necessary for actually creating a distribution using the distribution base defined in the column 1103. For example, when the distribution base of the column 1103 is a normal distribution, the column 1104 stores the degree of distribution variation, information corresponding to the distribution enlargement / reduction ratio, and the like. The distribution base and the distribution parameters stored in the columns 1102 and 1103 are applied to the operation of the device corresponding to the device ID in the column 1102. Column 1105 stores the reference time of the distribution described with reference to FIG. The column 1106 stores the reference time parameter of the distribution described with reference to FIG. The reference time and the reference time parameter stored in the columns 1104 and 1105 are applied to the distribution determined by the distribution base and the distribution parameters.

  The distribution based on the distribution base (b) and its creation procedure in this embodiment, and the distribution based on the distribution base (c) and its creation procedure are shown in FIGS. This will be described with reference to FIG.

  In the present embodiment, the distribution base (b) is a normal distribution. The box illustrated in FIG. 11B indicates that the device with the device ID of 12 operates during the time period from 13:00 to 17:00. In this situation, a procedure for creating a distribution using a distribution base and a distribution parameter will be described. First, with respect to the time period of the box (13:00 to 17:00), the center of the time period (15:00) is the center of the distribution, and a normal distribution with a standard deviation of 5 is created. After that, scaling processing (enlargement / reduction processing) is performed on the entire distribution so that the maximum value of the distribution is 0.8. Information that the standard deviation of the distribution is 5 and the maximum value of the distribution is 0.8 is obtained from the column 1104. From column 1105, since the reference time of this distribution is Center and the reference time parameter is 0, the process of parallel translation of the distribution is not performed. When other values are stored in the columns 1105 and 1106, the distribution is translated in accordance with the procedure described with reference to FIG.

  In the present embodiment, the distribution base (c) is a distribution in which the first half is monotonically increasing, the middle is constant, and the second half is monotonically decreasing. The box illustrated in FIG. 11C indicates that the device with the device ID of 13 has performed from 13:00 to 17:00. In the distribution shown in FIG. 11C, the size of the distribution is uniformly 0.7 during the time period (13:00 to 17:00) in which the device with the device ID 13 is operating (column 1104). reference). In the time period before 13:00, the size of the distribution gradually decreases from 0.7 as the distance from 13:00 to the past increases. In the time zone after 17:00, the size of the distribution gradually decreases from 0.7 as the distance from 17:00 to the future increases. A specific method of determining the distribution of the part before 13:00 and the part after 17:00 (hereinafter, also referred to as “the tail of the distribution” and “the tail part”) will be described. As described in the column 1104 of the distribution pattern definition table 1100, the standard deviation of the base with the pattern ID of 12 is 3. Therefore, first, a normal distribution having an average of 0 on the t-axis and a standard deviation of 3 is created, and scaling processing (enlargement / reduction processing) is performed so that the maximum value of the created normal distribution becomes 0.7 in the apparatus operation time zone. Apply. The t-axis 0 or less of the distribution after the scaling processing is adopted as the distribution of the portion before 13:00 in FIG. At this time, the position having the maximum value after the scaling processing is aligned with the position at 13:00 in FIG. 11C. Similarly, the t-axis 0 or more of the distribution after the scaling processing is adopted as the distribution of the portion after 17:00 in FIG. 11C. At this time, the position having the maximum value after the scaling is aligned with the position at 17:00 in FIG. 11C. From column 1105, the reference time of this distribution is Center, and from column 1106, the reference time parameter is 0, so that the parallel translation process is not performed. When other values are stored in the columns 1105 and 1106, the distribution is translated in accordance with the procedure described with reference to FIG.

  Note that a distribution other than those described above may be used as the base distribution. For example, a Poisson distribution, an exponential distribution, a distribution defined by a non-parametric method, or another generally defined distribution may be used. Furthermore, the base of the distribution of the distribution pattern (c) may have a shape corresponding to the length of time during which the device has been operated. Further, information indicating whether each time is a time earlier than the operation of the apparatus or a time in the future may be included in the distribution. As an example, for a distribution created from each device operating time zone by the above-described method, for a time earlier than the operating time zone, a value obtained by multiplying the distribution value by -1 is used as a feature value. May be used. Although the distribution pattern is defined for each device ID in the present embodiment, the distribution pattern may be defined using other attributes.

  According to the method for calculating the characteristic amount of the device described above, the value of the characteristic amount of the device at each time includes information indicating how far past or future time has elapsed since the time the device was operating. Can be.

  In FIG. 10, the distribution is created from the time zone 1011 during which the device is operating. As described above, the distribution may be created only from the time zone in which it is confirmed that the operation has actually been performed, or the distribution may be created from the time zone scheduled to operate. In particular, when estimating a work process in real time, instead of performing estimation processing in a state where feature amount data is sufficiently accumulated, it is particularly effective to create a distribution from a scheduled time zone.

  A description will be given of a case where a plurality of operation time zones of the same device exist in the row 1001 in FIG. First, a distribution is created for each operation time zone according to the above-described procedure. In that case, at each time, there are as many values as the number of distributions representing the magnitude of the feature amount regarding the device. In the present embodiment, the total of the values from the respective distributions at each time is defined as the value of the feature value at that time. It should be noted that the value of the feature value of the device at each time may be defined by another method. For example, the value of the feature amount for the device at each time may be calculated from only the distribution created from the time zone operated in the time zone closest to that time. Alternatively, the value of the feature value for the device at each time may be calculated from only the distribution created from the closest time zone among the time zones in which the operation started earlier than that time. Further, a method of handling a case where a plurality of operation time zones of the same device exist in the row 1001 may be separately defined for each device.

  In the distribution pattern definition table 1100, how to handle feature amounts when a plurality of distribution patterns are defined for the same device ID will be described. In the present embodiment, an independent feature amount is created from each defined pattern. In the example of FIG. 11A, pattern ID = 11 and pattern ID = 12 are defined for the device with the device ID = 13. In this case, the feature amount table 7000 is provided with two columns (columns) for storing the feature amounts related to the device with the device ID = 12. In each column, a feature amount calculated based on a distribution created from pattern ID = 11 and a feature amount calculated based on a distribution created from pattern ID = 12 are stored. When a plurality of distribution patterns are defined for the same device, the distributions created from the respective distribution patterns may be combined, and the value of the combined distribution may be used as the feature value of the device. In that case, even if a plurality of distribution patterns are defined for the device, only one feature amount column for the device is provided in the feature amount table 7000. Further, which of the two types of feature amount creation procedures described above is applied may be defined for each device.

  In the description with reference to FIGS. 10 and 11, the object for which the distribution is to be created is the device operation. The first object is a device operation that operates regardless of whether a worker is in a work area or an operable area for the device. The second object is that the worker performs some kind of work on the device in the operable area related to the device. The device operation when the printing press actually performs the printing operation is an example of the first target. On the other hand, in order for the printing press to perform the actual printing work, the operator must set paper on the printing press and enter the settings of the printing press (printing conditions, printing speed, etc.). This is an example. In the case where the definitions of the distribution targets are strictly distinguished as described above, a distribution pattern is defined for each of the first target and the second target in FIG.

  The process estimating unit 108 according to the present disclosure estimates a type of a process at each time using a learned estimation model in machine learning. The creation of an estimation model in machine learning using the feature amount table 7000 described above, the accuracy of the created estimation model, and the actual estimation using the estimation model will be described. The first purpose of the process estimation unit 108 according to the present disclosure is to perform estimation at a timing when a certain time has elapsed from the estimation target time. The certain time is a time period in which data necessary for estimation can be sufficiently collected. Note that the content described in the present disclosure can be applied to estimation in real time.

  First, a procedure for creating an estimation model will be described. In order to create an estimation model, feature amount data in which features included in various data collected from a printing site are quantified is created for each time. Further, with respect to the created feature amount data for each time, the type of the process that the estimation target person was actually performing at that time is associated as the correct answer data. In the example of FIG. 7, the feature amount data is created at one second intervals. An estimation model can be created by applying feature amount data with correct data of a certain degree or more as learning data to an arbitrary machine learning algorithm. Since a method for generating the estimation model is known and is not a main requirement of the present disclosure, a detailed description will be omitted. In the following, the contents of the present disclosure will be described assuming that a random forest is used as a machine learning algorithm. However, the contents of the present disclosure can be implemented using other machine learning algorithms. In the random forest, the estimation result and the likelihood indicating the likelihood at the time of estimation can be calculated for each estimation result candidate. In the present embodiment, the feature amounts 7002 to 7004 use, as learning data, the calculation data of the position where the portable information terminal 5 was actually located at each time and the direction in which the portable information terminal 5 was actually facing. In addition, the feature amounts regarding the devices in columns 7005 to 7009 are created by the procedure described with reference to FIGS. 10 and 11.

  Next, the accuracy of the created estimation model will be described. The feature data with the correct answer data that is not used at the time of learning is input as test data to the created estimation model, and is actually estimated. The estimation accuracy of the estimation model can be obtained from the estimation result estimated by the estimation model and the correct answer data linked to the input feature amount data. There are several ways to express the estimation accuracy. For example, there is a recall that expresses the estimation accuracy by indicating how much the correct answer was correctly determined as the XX process from the correct answer in the XX process. Also, among the results estimated by the estimation model that the process is the XX process, there is a precision that expresses the estimation accuracy by indicating how many truly XX processes exist.

  Next, a process for actually estimating using the created estimation model will be described. At the time of actual estimation, feature data of columns 7002 to 7008 at the target time of estimation are created, and the estimator processes the feature data for each time to obtain an estimation result for each time. At this time, the likelihood of each estimation candidate is also obtained at the same time.

  FIG. 12A is a flowchart illustrating a process when the data acquired by the worker terminal 5 is transferred to the process estimation device 1 and the position estimation device 7. In step 1211, it is determined whether or not a unit time as a data acquisition interval has elapsed. If the unit time has elapsed, the process moves to step 1212. If the unit time has not elapsed, the process moves to step 1217. In step 1212, measured sensor information is obtained from various sensors of the portable information terminal 5. The sensor information includes acceleration obtained from an acceleration sensor, angular acceleration obtained from a gyro sensor, geomagnetism obtained from a geomagnetic sensor, and the like. In step 1213, the acquired sensor information is stored in the memory of the portable information terminal 5 together with the time information. In step 1214, it is determined whether or not communication with the process estimation device 1 and the position estimation device 7 is possible. If communication is possible, the process moves to step 1215. If communication is not possible, the process moves to step 1217. In step 1215, the acquired sensor information is transferred to the process estimating device 1 and the position estimating device 10 together with the time information. In step 1216, it is determined whether the process estimation has been completed. If the transfer has been completed, the transfer process is completed, and if not, the process proceeds to step 1217. In the step 1217, the processing is stopped for a predetermined time, and then the processing is shifted to the step 1211.

  FIG. 12B is a flowchart illustrating a process when transferring data included in the devices 2 to 4 to the process estimating device 1. In step 1221, it is detected whether or not each device has operated. If an operation has been detected, the process moves to step 1222. If no operation has been detected, the process moves to step 1227. In step 1222, a device log, a device sound, or a voltage applied to the device is obtained. In step 1223, the acquired device log and device sound are stored in the memory together with the time information. In step 1224, it is determined whether communication with the process estimation device 1 is possible. If communication is possible, the process moves to step 1225. If communication is not possible, the process moves to step 1227. In step 1225, the acquired device log and device sound are transferred to the process estimation device 1 together with the time information. In step 1226, it is determined whether the process estimation has been completed. If the transfer has been completed, the transfer process is completed, and if not, the process proceeds to step 1227. In step 1227, the process is stopped for a certain period of time, and then the process proceeds to step 1221.

  FIG. 13 is a flowchart illustrating a flow of a process of calculating a feature amount stored in the feature amount table 7000 by the device feature amount creating unit 104 and the worker feature amount creating unit 103 in the present embodiment. In step 1301, the device feature value creation unit 104 and the worker feature value creation unit 103 receive the information acquired by the acquisition unit 102. Specifically, information obtained from the worker terminal 5, the process estimating device 1, the position estimating device 7, the plan creating device 10, and the devices 2 to 4 is received. In step 1302, the device feature value creation unit 104 specifies a device for which a feature value is to be created. In the present embodiment, the device corresponding to the device ID stored in the column 1102 of the distribution pattern definition table 1100 is specified. In step 1303, the device feature value creation unit 104 creates feature values for all devices specified in step 1302. The feature quantity relating to the device is created by the method described above with reference to FIGS. In step 1304, the worker feature quantity creation unit 103 creates feature quantities related to the worker to be estimated. The feature quantity relating to the worker is created by the method described above with reference to FIGS. When the feature amount table 7000 is created according to the flow illustrated in FIG. 13, the process estimating unit 108 estimates the type of the work process of the worker at each time.

  FIG. 14 is a flowchart illustrating a detailed flow of the correction processing by the correction unit 109 in the present embodiment. The flow in FIG. 14 is performed at a timing when the estimation result by the process estimation unit 108 is sufficiently accumulated. In step 1401, the correction unit 109 removes noise from the results for each unit time arranged in the time series estimated by the process estimation unit 108. Specifically, a general median filter is applied to the likelihood calculated for each type of the process of the estimation candidate at each time. That is, the likelihood for each type of process at each time is calculated again. Then, the estimation result at each time is determined based on the likelihood of the recalculated result. When calculating the average value of the likelihood of each estimation candidate by the median filter, the size of the filter is treated as a hyper parameter, but an appropriate size is used for each estimation candidate. The purpose of applying the median filter is to remove noise from the estimation result. If the same purpose can be achieved, noise may be removed by a method other than the median filter.

  In step 1402, the correction unit 109 creates a group including one or more estimation units from the filtered estimation result. More specifically, continuous results that have the same estimation result with respect to the filtered estimation results for each unit time arranged in time series are assigned to the same group. Next, the processes of steps 1403 to 1412 are performed on all the created groups. In step 1403, the correction unit 109 determines whether the length of the target group is equal to or longer than the minimum work time. The minimum operation time is a value determined in advance for each process, and is an operation time that is always required when each process is performed. In step 1404, if the result in step 1403 is Yes, the process proceeds to step 1405; otherwise, the process proceeds to step 1406. In step 1405, a first True flag is set for the target group (the True flag is also set for the estimation result belonging to the target group). In step 1406, a first False flag is set for the target group (the False flag is also set for the estimation results belonging to the target group). In step 1407, the candidate of the estimation result of the target group is specified as the first process candidate. Specifically, the time zone of each group is specified, and the process / sub-process that the estimation target person is to perform in the same time zone is specified by referring to the worker planning table 4200. The time slot of each work scheduled to be performed by the worker to be estimated can be specified from columns 4208 and 4209. If the time slot of each work and the time slot of the target group partially overlap, the work schedule of the work is estimated. The type of the process is specified as a first candidate.

  In step 1408, from the first process candidates identified in step 1407, candidates that can be determined to be inappropriate are identified and excluded. In addition, the process candidate after the exclusion is defined as a second process candidate. Specifically, a work plan by a worker or a work plan by a device that is not reliably performed is specified from reliable information such as a device log. Further, other work plans that cannot be implemented unless the specified work plan is implemented are also specified in a potable manner. If there is a process that has been determined as a process candidate in step 1407 simply because it overlaps with the scheduled execution time of those identified work plans, the process is excluded from the process candidates. It should be noted that the constraints related to the execution of each process are specified by referring to the constraint process ID stored in the column 3509 in the process table 3500.

  In step 1409, it is determined whether the estimation result of the target group is included in the second process candidate. In step 1410, if the result in step 1409 is Yes, the process proceeds to step 1411; otherwise, the process proceeds to step 1412. In step 1411, a second True flag is set for the target group (the True flag is also set for estimation results belonging to the target group). In step 1412, a second False flag is set for the target group (the False flag is also set for the estimation results belonging to the target group). In step 1413, the estimation result is extended to a group (expansion target group) in which both the first False flag and the second True flag are set so as to have the minimum work time. That is, a part of the estimation result before and after adjacent to the expansion target group is converted into the estimation result of the expansion target group. At this time, a description will be given as to whether to convert the past or future result of the expansion target group or to convert the past or future result at what rate. An estimation result to be converted so as not to interfere with a group to which the first True flag is attached or another extension target group is determined. If interference occurs, conversion processing is performed so that the estimation result of the other group for which the first True flag is set is preferentially left as the final estimation result.

  In step 1414, among the estimation results belonging to the group in which the False flag is set in both the first flag and the second flag, the estimation results that were not converted to other estimation results in step 1413 (the final conversion target ) To other estimation results. Specifically, the following processing is performed for each final conversion target. First, it is determined whether the following condition 1 is satisfied in the order from the result before the final conversion target in the order of distance from the final conversion target (when there is no candidate when the condition 2 is satisfied). The type of the process of the estimation result that first satisfies the condition 1 is set as a past conversion destination candidate (candidate 1). Condition 1: Converted to another estimation result in step 1413, or both the first flag and the second flag are True. Condition 2: The first flag is True and the second flag is False. Similarly, it is determined whether the above condition 1 is satisfied in the order from the result closer to the future than the result (it is determined that there is no candidate when the above condition 2 is satisfied). The type of the process of the estimation result that first satisfies the condition 1 is set as a candidate for a future conversion destination (candidate 2). Next, a work area or an operable area in which the work of the estimation results of the candidates 1 and 2 can be performed is specified, respectively, and compared with the place where the worker to be estimated was at the time of the conversion target. Then, a candidate (candidate 1 or candidate 2) in which a work area or an operable area close to the place where the worker to be estimated is located is set as the estimation result of the final conversion target after conversion. In step 1415, for a group for which both the first True flag and the second False flag are set, the estimation result of the group is presented to the user of the process estimation device 1. Then, the user is prompted to determine the final estimation result for the group. Specifically, the estimation result at the time of step 1402 and the estimation result at the time of step 1414 are displayed on the output device 13 in a format that can be distinguished. Further, as to the estimation result at the time of step 1414, a time zone in which the first True flag is set and the second False is set is displayed in an identifiable format. Based on the above display, the user of the process estimating apparatus 1 uses the input device 12 to notify the correcting unit 109 of whether or not the estimation result at step 1415 is correct. The process estimating apparatus 1 may have a function that allows the user of the process estimating apparatus 1 to manually correct the estimation result when the estimation result is incorrect.

  In the present embodiment, the plan information on the worker and the device is used when the correction unit 109 corrects the result estimated by the process estimation unit 108. However, the feature amount may be created from the information on the plan and added to the feature amount table 7000, so that the process estimating unit 108 uses the feature amount when performing the estimation process.

  As described above, according to the technology described in the present embodiment, even in a situation where it is difficult to estimate only with the work area and the movement feature amount of the worker, the feature amount created from the time zone in which each device operates is used. Thus, the execution process of the worker can be accurately estimated.

<Second embodiment>
4 shows an example of a second embodiment of the present disclosure. In the present embodiment, a process estimating apparatus 1 for accurately estimating an execution process of a worker to be estimated when a plurality of workers work in the same printing factory will be described.

  FIG. 15 shows a functional block diagram for explaining main functions of the process estimating apparatus 1 according to the present embodiment. The functional block diagram according to the present embodiment is obtained by adding a relationship specifying unit 111 to the functional block diagram according to the first embodiment. The relation specifying unit 111 specifies a relation between two or more of the operation of the device, the work performed by the worker on the device, and the work performed by the worker other than the work on the device. The device feature value creation unit 104 determines a device for which a feature value is to be calculated based on the relationship specified by the relationship specification unit 111. The worker feature quantity creation unit 103 determines a worker whose feature quantity is to be calculated based on the relationship specified by the relationship specification unit 111.

  In particular, the relation specifying unit 111 specifies one or both of the operation of the apparatus and the work performed by the worker on the apparatus, which may be related to the worker to be estimated. The device feature value creation unit 104 calculates feature values relating to one or both of the operation of the specified device and the work performed by the worker on the device.

  FIG. 16 shows the concept of the feature amount table 7000 according to the present embodiment. Similar to the first embodiment, the feature amount table 7000 according to the present embodiment includes a feature amount (column 7002, column 7003, and column 7004) relating to the worker, a feature amount relating to the apparatus (column 7005 to column 7008), and the correct answer data (column 7009). ). In the present embodiment, the feature amount of each device has a column subdivided into columns 7014 to 7017 based on the relationship between the worker to be estimated identified by the relationship identifying unit 111 and each device. Hereinafter, an example will be described in which a distribution is created from a time zone during which the apparatus operates, and the apparatus characteristic amount is calculated based on the distribution. However, as described in the first embodiment, the worker performs work on the apparatus. The distribution and the feature amount may be created from the time zone.

  In column 7014, when the device targeted in column 7005 operates, the worker to be estimated performs a task related to the device operation, and the degree of the relationship with the device operation is indicated. The device characteristic amount when the size is large is stored. Specifically, when the order ID / work definition ID which is the target of the device operation is included in the work item planned as the estimation target, the distribution for creating the device feature amount is described in FIG. Is created based on the distribution pattern definition table. It should be noted that the work items planned for the estimation target (the process of the order ID / work definition ID which is the target of the device operation / sub-process) are only the estimation targets. That is, only when the worker ID to be estimated is described in the column 4202 of the worker planning table 4200, a distribution for creating a feature amount in the column 7014 due to the operation of the device is created. .

  In column 7015, when the device targeted in column 7005 operates, the worker to be estimated performs a task related to the device operation, and the degree of the relationship with the device operation is indicated. The device characteristic amount at the time of medium is stored. Specifically, when the work item planned as the estimation target includes the order ID / work definition ID that is the target of the device operation, the distribution for creating the device feature amount is described in FIG. Is created based on the distribution pattern definition table. However, there is a possibility that a plurality of workers including the estimation target may perform the work item (the order ID / operation definition ID process / sub-process targeted for the operation of the apparatus) planned for the estimation target. There is a condition that there is. That is, when the ID of the worker to be estimated and the ID of another worker are described in the column 4200 of the worker planning table 4200, the feature amount is created in the column 7015 due to the operation of the apparatus. A distribution is created. The difference between column 7014 and column 7015 is the degree to which the operation of the device is related to the estimation target. If the worker who performs the work related to the operation of the device is only the estimation target from the planning stage, it is very likely that the work is the estimation target. On the other hand, when one of the plurality of workers is scheduled to perform the operation related to the device operation, the possibility that the estimator performs the operation is reduced.

  In column 7016, when the device targeted in column 7005 operates, the estimation target worker does not plan to perform a work related to the device operation, but the degree of the relationship with the device operation is not described. The device characteristic amount when there is a medium level is stored. Specifically, when the following three conditions are satisfied, a distribution for creating the device feature amount is created in the column 7016 based on the distribution pattern definition table shown in FIG. The first condition is that the same work as the order ID / work definition ID that is the target of the device operation is not scheduled for the worker to be estimated. The second condition is that the worker who has scheduled the same work as the order ID / work definition ID which is the target of the device operation is not in a place where the work is to be performed in the scheduled time zone. . The third condition is that there is a worker to be estimated in a place where the work is scheduled to be performed during a time period when the worker described in the second condition is scheduled to work. In other words, if there is a possibility that the work related to the operation of the device has been performed by the worker to be estimated on behalf of the person in charge of planning, the feature amount is created in the column 7015 due to the operation of the device. Distribution is created.

  In a column 7017, when the device targeted in the column 7005 is operated, the estimation target worker is not planned to perform a work related to the device operation, and the degree of the relationship with the device operation is shown. The device feature value when the size is small is stored. Specifically, when the first condition shown below is satisfied and one or more of the second and third conditions are further satisfied, the distribution for creating the device feature amount is described in FIG. Is created in the column 7016 based on the distribution pattern definition table. The first condition is that the same work as the order ID / work definition ID that is the target of the device operation is not scheduled to be the estimation target. The second condition is that a worker who has scheduled the same work as the order ID / work definition ID that is the target of the device operation is in a place where the work is to be performed during the scheduled time zone. . The third condition is that there is no worker to be estimated at the place where the work was scheduled to be performed during the time period when the worker described in the second condition was scheduled to work. That is, when there is no possibility that the worker related to the operation of the apparatus has performed the work related to the operation of the apparatus on behalf of the person in charge of planning, the feature amount is created in the column 7015 due to the operation of the apparatus. Distribution is created. If the conditions in columns 7014 to 7016 are not satisfied, a distribution is created in column 7017.

  Note that when the device targeted by the column 7005 operates, a distribution due to the operation may be simultaneously created in a plurality of columns 7014 to 7017. For example, it is assumed that the target device ID in column 7005 is 7, and the device with the device ID = 7 is operating at time t1. Also, it is assumed that in the time period t2 to t3 including the time t1, the device with the device ID = 7 is scheduled to work on the work definition IDs 8 and 9 in the device plan table 4300. Further, it is assumed that a work process of work definition ID = 8 relating to the estimation target is planned in the worker plan table 4200 (the worker is only the estimation target), and a work process of work definition ID = 9 is not planned. In such a case, depending on whether the device with the device ID = 7 is working at the work definition ID = 8 or the work with the work definition ID = 9 at time t1, it is distributed in either column 7014 or column 7017. Is created. If either cannot be determined by using information such as the device log, a distribution is created in both the column 7014 and the column 7017 for the feature amount of the device with the device ID = 7 at the time t1, and the feature amount based on the distribution is created. Is calculated. In this case, the column 7017 also has a role of expressing the reliability of the distribution created in the column 7014. In other words, if a distribution is created only in column 7014 based on the operation of the device targeted in column 7005 based on the operation, it is determined that there is a high probability relationship between the device operation at that time and the estimation target. Show. On the other hand, when the distribution is created in the columns 7014 and 7017, it indicates that there is a possibility that there is a relationship between the device operation at that time and the estimation target, and there is a possibility that there is no relationship. By reflecting such a difference in the feature amount table 7000 and causing the computer to process it, an improvement in estimation accuracy can be expected. The method of calculating the values of the feature amounts stored in columns 7014 to 7017 will be described with reference to FIGS.

  FIG. 17 shows a distribution pattern definition table 1100 in the present embodiment. The distribution pattern definition table 1100 of the present embodiment has columns 1107 and 1108 in addition to the columns 1101 to 1106 described in the distribution pattern definition table 1100 shown in the first embodiment. A column 1107 stores information (completed, uncompleted) for distinguishing whether or not planned, as described with reference to FIG. Column 1108 stores information (large, medium, and small) for distinguishing the degree of the relationship described with reference to FIG. By referring to the information stored in columns 1107 and 1108 and the information of the device ID stored in column 1102, the distribution created in each of columns 7014 to 7017 described with reference to FIG. 16 is specified. be able to.

  FIGS. 18, 19, and 20 show a flow of a process related to the creation of the feature amount of the device by the device feature amount creation unit 104 and the relationship identification unit 111 in the present embodiment. Hereinafter, a procedure for creating a distribution for the operation of the device X will be described, assuming, for the sake of convenience, a case where the device X with the device ID = 9 operates during the time period T between times t4 and t5. FIG. 18 is a diagram showing the entire processing flow. In step 1801, the relationship specifying unit 111 defines an array SameID and an array DifferententID. In step 1802, the relation specifying unit 111 defines the variable Column1, the variable Column2, the variable Column3, and the variable Column4 in a Boolean type, and initializes them with False. Variables Column1 to Column4 are variables that hold information on whether to create a distribution in columns 7014 to 7017 in FIG. In step 1803, the relation specifying unit 111 specifies the device X performing the work process related to which work definition ID of which order in the time zone T. Specifically, it is specified from highly reliable information such as a device log. In step 1804, it is determined whether or not the relation specifying unit 111 has been specified. If the device X does not output a log, it cannot be specified, and depending on the output format of the log, it cannot be specified from the device log because a character string such as an order ID is not output. In step 1804, if the relationship specifying unit 111 has specified the set of the order ID and the work definition ID, the process proceeds to step 1806. If it cannot be identified, the process moves to step 1805.

  In step 1805, the relationship specifying unit 111 specifies, from the plan information of the device, whether the device X is performing a work process related to which work definition ID of which order in the time zone T. Specifically, by referring to the device plan table 4300, the plan information with the device ID of 9 in the column 4302 is extracted. The planned execution time zone of each extracted plan information is calculated from 4308 and column 4309, and the plan information including all or a part of the time zone T in the planned execution time zone is specified. From the columns 4303 and 4304 of the specified plan information, the order ID and the work definition ID can be specified. In step 1806, the relationship specifying unit 111 classifies the set of the order ID and the work definition ID specified in step 1803 or 1805 into a set included in the work plan to be estimated and a set not included in the work plan to be estimated. In step 1807, the relationship specifying unit 111 stores the set of the order ID and the work definition ID included in the work plan to be estimated in the array SameID. In step 1808, the relationship specifying unit 111 stores a set of the order ID and the work definition ID that are not included in the work plan to be estimated in the array DifferententID.

  In step 1809, the relationship identifying unit 111 determines the values of the variables Column1 and Column2 based on the information stored in the array SameID. That is, it is determined whether or not to create distributions in columns 7014 and 7015 in FIG. Detailed processing of step 1809 will be described later with reference to FIG. In step 1810, the relationship identifying unit 111 determines the values of the variables Column3 and Column4 based on the information stored in the array DifferentID. That is, it is determined whether or not to create distributions in columns 7016 and 7017 in FIG. The detailed processing of step 1810 will be described later with reference to FIG. In step 1811, the apparatus feature quantity creation unit 104 performs steps 1812 to 1813 on all columns for which True has been set in steps 1809 and 1810 among the variables Columns 1 to 4. In step 1812, the apparatus feature creating unit 104 creates a distribution based on the distribution pattern definition table 1100 described in FIG. In step 1813, the device feature value creation unit 104 calculates feature values related to the operation of the device X at each time based on the created distribution. A specific method of calculating the feature amount is the same as the procedure described in the first embodiment.

  FIG. 19 shows a specific processing flow of step 1809. In step 1814, the relation specifying unit 111 performs the processing of steps 1815 to 1819 for all sets of order IDs and work definition IDs stored in the array SameID. In step 1815, the relation specifying unit 111 performs the processing of steps 1816 to 1819 for all the steps / sub-steps associated with the set of the order ID and the work definition information ID. In step 1816, the relation specifying unit 111 specifies all the candidates of the performers of the work for the process (or sub-process) of the work plan to be estimated associated with the set of IDs. In step 1817, the relation specifying unit 111 determines whether or not the person who performs the work always includes the estimation target. Specifically, in step 1816, a plurality of persons are specified, and it is determined whether any one of the persons is a plan to execute the task or whether the estimation target is set as a worker (main person in charge) who always executes the task. . In step 1817, the relation specifying unit 111 proceeds to step 1818 if the estimation target is the main charge. If there is a possibility that the estimation target does not perform the operation, the process proceeds to step 1819. In step 1818, the relationship specifying unit 111 sets True to the variable Column1. In step 1818, the relationship specifying unit 111 sets True to the variable Column2.

  FIG. 20 shows a specific processing flow of step 1810. In step 1820, the relation specifying unit 111 performs the processing of steps 1821 to 1829 for all sets of order IDs and work definition IDs stored in the array DifferentID. In step 1821, the relation specifying unit 111 performs the processing of steps 1822 to 1827 for all the steps / sub-steps associated with the set of the order ID and the work definition information ID. In step 1822, the relationship identifying unit 111 determines whether the worker scheduled to perform the process (or sub-process) is alone in the work area or the operational area of the process (sub-process) during the scheduled work time zone. Identify. Specifically, the process of step 1822 is realized by specifying the position of each worker at each time with reference to the columns 7002 and 7003 of the feature amount table 7000. In step 1823, the relation specifying unit 111 proceeds to step 1824 when there is at least one worker in the work area or the operator-friendly area. If not, the process proceeds to step 1827. In step 1824, the relation specifying unit 111 specifies whether the worker to be estimated was located in the work area or the operable area of the process (sub-process) during the scheduled work time of the process (sub-process). In step 1825, when the worker to be estimated is in the scheduled work time slot in step 1824, the relation specifying unit 111 proceeds to step 1826. If not, the process proceeds to step 1827. In step 1826, the relation specifying unit 111 sets True to the variable Column3. In step 1827, the relationship specifying unit 111 ends the loop processing 1821 for the current step / sub-step, and shifts the processing to the loop processing 1821 for the next step / sub-step. In step 1830, the relation specifying unit 111 determines whether the variable Column3 is False. If false, the process proceeds to step 1828. In step 1828, the relationship specifying unit 111 sets True to the variable Column4. If the result of step 1830 is True, the relation identifying unit 111 advances the processing to 1829. In step 1829, the relation specifying unit 111 ends the loop processing 1820 for the current order ID / work definition ID, and shifts the processing to the loop processing 1820 for the next order ID / work definition ID.

  The processing of the correction unit 109 in the present embodiment will be described. The diagram of the processing flow of the correction by the correction unit 109 in the present embodiment is the same as FIG. 14 described in the first embodiment. However, in the present embodiment, the range of the plan information on the worker used in step 1407 in FIG. 14 is different from the first embodiment. In the present embodiment, first, the first estimation candidate is specified from the plan information of the worker to be estimated according to the procedure described in the first embodiment. Further, the processes and sub-processes which are the processing targets (planned by other workers) of the loop process 1821 when the step 1826 of FIG. 20 is performed are also set as the first estimation candidates.

  As described above, according to the technology described in the present embodiment, when estimating the execution process of the worker to be estimated, it is possible to appropriately specify the information of the device to be used for the estimation and the information of the other workers. That is, estimation can be performed with higher accuracy.

  The configuration and operation of the process estimation device according to the embodiment have been described above. It should be understood by those skilled in the art that these embodiments are exemplifications, and that various modifications can be made to the components and the combinations of the processes, and that such modifications are also within the scope of the present disclosure.

<Other embodiments>
The present disclosure can take embodiments as, for example, a system, an apparatus, a method, a program, or a recording medium (storage medium). Specifically, the present invention may be applied to a system including a plurality of devices (for example, a host computer, an interface device, an imaging device, a web application, and the like), or may be applied to a device including a single device. Good.

  Needless to say, the object of the present disclosure is achieved by the following. That is, a recording medium (or a storage medium) that records software program codes (computer programs) for realizing the functions of the above-described embodiments is supplied to a system or an apparatus. Needless to say, such a storage medium is a computer-readable storage medium. Then, the computer or the CPU or the MPU (Micro Processing Unit) of the system or the apparatus reads out and executes the program code stored in the recording medium. In this case, the program code itself read from the recording medium realizes the function of the above-described embodiment, and the recording medium on which the program code is recorded constitutes the present disclosure.

  The present disclosure supplies a program or a program that realizes one or more functions of the above-described embodiments to a system or an apparatus via a network or a storage medium, and one or more processors in a computer of the system or the apparatus read and execute the program. This processing can be realized. Further, it can also be realized by a circuit (for example, an ASIC) that realizes one or more functions.

  1 process estimation device, 2 printing machine, 3 bookbinding machine, 4 cutting machine, 5 portable information terminal, 6 wireless access point, 7 position estimation device, 8 work definition information DB, 9 order DB, 10 plan creation device, 11 local area network.

Claims (11)

Acquire two or more pieces of information among first information related to operation of the device, second information related to work performed by the worker on the device, and third information related to work performed by the worker other than the second information. Acquisition means for
An apparatus feature amount calculation unit that calculates a feature amount related to an apparatus from one or both of the first information and the second information,
From either or both of the second information and the third information, an operator feature amount calculation unit that calculates a feature amount for the worker,
Estimating means for estimating the work process of the worker based on the feature amount calculated by the device feature amount calculating means and the feature amount calculated by the worker feature amount calculating means,
The device feature value calculation means calculates a value of an objective variable calculated by a function using the time of estimation as an explanatory variable as a feature value at the time,
The process estimation device, wherein the function is a linear or non-linear function determined based on one or both of the first information and the second information. One or more of said functions are assigned to each device;
The process estimation device according to claim 1, wherein the device feature value calculation unit calculates one or more feature values for each device at the time of the estimation target. The first information includes at least one selected from a group including a device log of each device, a device sound, a scheduled operation of the device, and an estimation result of past operation,
The second information, the operation log of the device by the worker, including one or more selected from the group consisting of pre-processing related to the operation of the device, post-processing related to the operation of the device and their planning information,
The third information includes at least one selected from a group consisting of position information, orientation information, route information, posture information, motion information, and plan information for one or more workers including an estimation target. Item 3. The process estimation device according to item 1 or 2. The worker feature amount calculating means includes:
Using information having a combination of the region and the orientation obtained in advance, at each time, which combination the worker corresponds to, from one or both of the second information and the third information Identify as the first worker summary information,
If none of the combinations is applicable, at each time point, which route the worker is located and which direction of the route is directed to either or both of the second information and the third information. From the second worker summary information,
At each time, any of the first worker summary information or the second worker summary information specified for the worker is specified as the worker summary information at the time,
The worker summary information at each time, specifying the length of the time zone that continues continuously including each time, as the summary information continuation time at the time,
The latest time at which the value different from the worker summary information at each time is specified is specified from the past from the time, and the worker summary information at the latest time is regarded as the immediately preceding worker summary information at each time,
The latest time at which a value different from the worker summary information at each time is specified is specified from the future after the time, and the worker summary information at the latest time is set as the immediately following worker summary information at each time,
4. The worker feature amount calculating unit according to claim 3, wherein at least one of the worker summary information, the summary information duration time, the immediately preceding worker summary information, and the immediately following worker summary information is a feature amount related to the worker. A process estimating apparatus as described in the above. Operation of the device, the work on the device by the worker, the work by the worker other than the work on the device, further comprising relationship specifying means for specifying a relationship between two or more,
The device feature amount calculation unit determines a device for which a feature amount is to be calculated based on the relationship identified by the relationship identification unit,
The process estimation device according to any one of claims 1 to 4, wherein the worker characteristic amount calculating unit determines an operator whose characteristic amount is to be calculated based on the relationship specified by the relationship specifying unit. The relationship specifying means may be related to the worker of the estimation target, specifying one or both of the operation of the device and the work on the device by the worker,
6. The process estimating apparatus according to claim 5, wherein the apparatus characteristic amount calculating means calculates a characteristic amount relating to one or both of the operation of the specified apparatus and the work performed by the operator on the apparatus. The relation specifying means specifies the strength of the relation,
The process estimating apparatus according to claim 5, wherein the apparatus characteristic amount calculating unit and the worker characteristic amount calculating unit calculate a characteristic amount based on the relationship and the strength thereof.   The relation specifying unit is selected from a group consisting of a plan associated with the device, a plan associated with the worker, a past estimation result by the process estimation device, area information related to the device, and position information of the worker. The process estimation device according to any one of claims 5 to 7, wherein the relationship and the strength of the relationship are specified based on one or more. From the work process at each time of the worker estimated by the estimating means, further comprising a correcting means for correcting the erroneously estimated portion,
The correcting means is, in addition to the plan associated with the estimation target, plan information associated with a device associated with the estimation target identified by the relationship identification means, associated with another worker associated with the estimation target. 9. The process estimating apparatus according to claim 5, wherein the correction is performed using at least one selected from a group consisting of the planning information and the position information and the strength of the relationship. Acquire two or more pieces of information among first information related to operation of the device, second information related to work performed by the worker on the device, and third information related to work performed by the worker other than the second information. An acquisition process to
From either or both of the first information and the second information, a device feature value calculation step of calculating a feature value related to the device,
From one or both of the second information and the third information, an operator feature amount calculation step of calculating a feature amount for the worker,
An estimation step of estimating a work process of the worker based on the feature amount calculated by the device feature amount calculation step and the feature amount calculated by the worker feature amount calculation step,
The device feature value calculation step calculates a value of an objective variable calculated by a function using the time of estimation as an explanatory variable as a feature value at the time,
The process estimation method, wherein the function is a linear or non-linear function determined based on one or both of the first information and the second information.   A program for causing a computer to function as each unit of the device according to any one of claims 1 to 9.

Images