Control method of article conveying device and article conveying device

Abstract

The present invention relates to a control method for an article conveying device and an article conveying device. The control method for an article conveying device (1) according to one embodiment comprises a step A of using information related to the input amount indicating the weight value of articles conveyed by a conveying unit (20) to a component arranged on the downstream side, information indicating the state of articles on the conveying unit (20), and control parameters of the conveying unit (20) as learning data, and generating a learning model that infers control parameters set for conveying articles of a target weight; a step B of controlling the conveying of articles based on the learning model; a step C of selectively switching between a production mode that participates in actual production and a non-production mode that does not participate in production to operate the article conveying device (1); and a step D of collecting and storing, as learning data, information related to the input amount and control parameters actually obtained when the article conveying device (1) is operated in the non-production mode.

Description

Technical Field

[0001] The present invention relates to a control method of an article conveying device and the article conveying device. Background Art

[0002] In recent years, the development of technologies for controlling various devices using deep learning has been progressing. To achieve this control, it is necessary to acquire a large amount of learning data while changing various control parameters of the device.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 2020-11326 Summary of the Invention

[0006] Technical problem to be solved by the invention

[0007] However, when actual production is carried out using the above-described apparatus, there is a problem that various control parameters cannot be freely changed to avoid a decrease in yield, and sufficient learning data cannot be obtained to perform the above-described control.

[0008] Therefore, in view of the above-mentioned problems, an object of the present invention is to provide a control method and an article conveyance device that can obtain sufficient learning data for control while avoiding a decrease in yield.

[0009] Solutions for solving technical problems

[0010] A control method for an article conveying device according to one embodiment has the following key points: step A, which uses information related to the input amount representing the weight value of the article conveyed to a component arranged on the downstream side through a conveying unit, information representing the state of the article on the conveying unit, and control parameters of the conveying unit as learning data, and generates a learning model, which infers the control parameters set for conveying the article of target weight; step B, which controls the conveyance of the article based on the learning model; step C, which selectively switches between a production mode that participates in actual production and a non-production mode that does not participate in the production to operate the article conveying device; and step D, which collects and stores the actually obtained information related to the input amount and the control parameters as the learning data when the article conveying device is operated in the non-production mode.

[0011] The article conveying device involved in one embodiment has the following key points: it includes a conveying part for conveying the articles to be put in; a learning model generating part for using information related to the input amount representing the weight value of the articles conveyed by the conveying part to the component arranged on the downstream side, information representing the state of the articles on the conveying part, and control parameters of the conveying part as learning data, and generating a learning model, which infers the control parameters set for conveying the articles of target weight; and a control part for controlling the conveying of the articles based on the learning model, the control part selectively switching between a production mode participating in actual production and a non-production mode not participating in the production to operate the article conveying device, and when the article conveying device is operated in the non-production mode, the control part collects and stores the actually obtained information related to the input amount and the control parameters as the learning data.

[0012] Effects of the Invention

[0013] According to the present invention, a method for controlling an article conveyance device and an article conveyance device can be provided, which can obtain sufficient learning data for control while avoiding a decrease in yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a perspective view showing an example of the entire combination metering device 1 according to one embodiment.

[0015] Figure 2 This is a diagram showing an example of functional blocks of the combination metering device 1 according to one embodiment.

[0016] Figure 3 This is a flowchart showing an example of a control method of the combination metering device 1 according to one embodiment.

[0017] Description of Reference Numerals

[0018] 1…article conveying device, combined metering device; 10…distribution table; 11…feeding trough; 12…storage hopper; 12a…gate; 13…metering hopper; 14…boost hopper; 15…collecting discharge chute; 20…conveyor unit; 30…learning model generation unit; 40…control unit; 50…image acquisition unit. DETAILED DESCRIPTION

[0019] Below, the preferred embodiment of the present invention is described in detail with reference to the accompanying drawings. It should be noted that in the description of the following drawings, the same or similar parts are marked with the same or similar figure marks. However, it should be noted that the drawings are schematic, and the proportions of each dimension are different from the actual ones. Therefore, the specific dimensions should be judged with reference to the following description. In addition, the drawings may also contain parts with different dimensional relationships or proportions. In this specification and the drawings, the elements with substantially the same function and structure are marked with the same figure marks to omit repeated descriptions. In addition, the illustrations of elements that are not directly related to the present invention are omitted.

[0020] (First embodiment)

[0021] Below, refer to Figures 1 to 3 In the first embodiment of the present invention, a combination weighing device 1 will be described as an example of an article transport device 1 . Figure 1 1 is a perspective view showing an example of the entire combination metering device 1 according to the present embodiment. Figure 2 1 is a diagram showing an example of functional blocks of a combination metering device 1 according to an embodiment. Figure 3 This is a flowchart showing an example of a control method of the combination metering device 1 according to one embodiment.

[0022] like Figure 1 As shown, the combined weighing device 1 according to this embodiment includes a distributing table 10 , a feed trough 11 , a storage hopper 12 , a weighing hopper 13 , a booster hopper 14 , and a collecting and discharging chute 15 .

[0023] The dispersing table 10 is configured to disperse objects dropped from a supply device (not shown) toward a feed trough 11. For example, the dispersing table 10 in this embodiment is configured to be driven to rotate about a rotation axis C extending in the vertical direction, thereby dispersing the objects in a circumferential direction while conveying them radially outward. However, the dispersing table 10 is not limited to this configuration and may be of any configuration as long as it has the function of conveying the objects to the feed trough 11.

[0024] For example, the object to be measured is soft and sticky food such as raw chicken. However, the object to be measured is not limited to this.

[0025] The feed trough 11 is configured to supply the objects to be weighed supplied from the distributing table 10 to the corresponding storage hoppers 14 provided downstream. For example, the feed trough 11 may be configured to convey the objects to be weighed downstream by vibration.

[0026] The storage hopper 12 is configured to temporarily retain the objects supplied from the feed chute 11. A gate 12a is provided below each storage hopper 12. The objects retained in the storage hopper 12 are discharged into the weighing hopper 13 provided below the gate 12a by opening and closing the gate 12a.

[0027] The weighing hopper 13 is configured to temporarily retain and store the objects discharged from the hopper 12. The weighing hopper 13 is connected to a mass detector (not shown).

[0028] Furthermore, gates (not shown) are provided below the weighing hoppers 13 , and objects retained in the weighing hoppers 13 are discharged into the booster hopper 14 disposed below the gates by opening and closing the gates.

[0029] The booster hopper 14 is configured to receive and temporarily hold the objects supplied from the weighing hopper 13. A gate (not shown) is provided below each booster hopper 14. By opening and closing the gate, the objects retained in the booster hopper 14 are discharged into a collecting discharge chute 15 located below the gate.

[0030] The collecting and discharging chute 15 is configured to collect the objects discharged from the plurality of booster hoppers 14 and cause the objects to fall downward.

[0031] like Figure 2 As shown, the combined metering device 1 includes a transport unit 20 , a learning model generation unit 30 , and a control unit 40 as functional blocks. Furthermore, the combined metering device 1 may include an image acquisition unit 50 .

[0032] The conveying unit 20 is configured to convey the thrown articles. For example, the conveying unit 20 is composed of the distributing table 10, the feeding trough 11, etc. Hereinafter, in this specification, the case where the feeding trough 11 is used as the conveying unit 20 is described as an example.

[0033] The image acquisition unit 50 is configured to acquire an image of an article placed on the feed chute 11 (the conveyor unit 20). For example, the image acquisition unit 50 may be configured by any type of camera.

[0034] The learning model generation unit 30 is configured to generate a learning model that estimates control parameters set for conveying articles of a target weight through the feed chute 11 .

[0035] In this learning model, information related to the input amount, which represents the weight value of the items transported through the feed trough 11 to the component arranged on the downstream side (such as the storage hopper 12), information representing the state of the items on the feed trough 11, and control parameters of the feed trough 11 (conveying unit 20) are used as learning data.

[0036] For example, the learning model generation unit 30 is configured to generate a learning model through machine learning, deep learning, or the like.

[0037] For example, the weight value of the item transported to the component arranged on the downstream side through the feed trough 11 is the weight value of the item supplied from the feed trough 11 to the measuring hopper 13 through the storage hopper 12, and is the weight value obtained by the weight detector (weighing sensor) connected to the measuring hopper 13.

[0038] It should be noted that the information related to the input amount may be the weight value of the above-mentioned item itself, or may be any value corresponding to the weight value of the above-mentioned item.

[0039] The control parameters of the feed trough 11 may include the time for vibrating the feed trough 11 or the voltage applied to vibrate the feed trough 11. As described below, the control parameters include a first control parameter and a second control parameter.

[0040] Furthermore, the information indicating the state of the articles on the feed chute 11 (the conveying unit 20 ) may be, for example, an image of the articles acquired by the image acquisition unit 50 .

[0041] Specifically, the learning model generation unit 30 is configured to generate the learning model using the learning data collected and stored by the control unit 40 .

[0042] Furthermore, the learning model generation unit 30 may be configured to use the image of the article acquired by the image acquisition unit 50 as the information related to the input amount.

[0043] The control unit 40 is configured to control the conveyance of articles by the conveyance unit 20 based on the learning model generated by the learning model generation unit 30 .

[0044] Furthermore, the control unit 40 is configured to operate the combination weighing device 1 by selectively switching between a production mode in which the device participates in actual production and a non-production mode in which the device does not participate in actual production.

[0045] Among them, the production mode is a mode in which the combination weighing device 1 performs a weighing operation of non-weighed objects in a production line or the like.

[0046] On the other hand, the non-production mode is a mode of the combination weighing device 1 in which all objects retained in the combination weighing device 1 are discharged (i.e., a mode of the combination weighing device 1 in which all objects are discharged), or a test mode in which the combination weighing device 1 is driven.

[0047] The control unit 40 is configured to collect and store information on actually obtained input amounts and control parameters as learning data when the combination weighing device 1 is operated in the non-production mode.

[0048] According to this configuration, when actual production is not being performed in the combination metering device 1, that is, when the combination metering device 1 is operating in a non-production mode, control parameters can be freely changed to collect learning data without being aware of a deterioration in yield.

[0049] Furthermore, the control unit 40 may be configured to collect and store information on actually obtained input amounts and control parameters as learning data even when the combination weighing device 1 is operated in the production mode.

[0050] According to this configuration, when actual production is performed in the combination metering device 1, that is, when the combination metering device 1 operates in the production mode, learning data related to control parameters used in actual production are collected, thereby enabling efficient generation of a learning model.

[0051] Specifically, the control unit 40 may be configured to collect and store the first control parameter and information on the input amount when the first control parameter is used as learning data when the combination metering device 1 is operated in the production mode.

[0052] It should be noted that the first control parameter may be a control parameter collected from the combination metering device 1 operating in production mode during a predetermined period, or may be a parameter within a predetermined range determined in advance. The predetermined range is the range of control parameters that can be used in the combination metering device 1 operating in production mode.

[0053] According to this configuration, the combination weighing device 1 can collect learning data while performing actual production.

[0054] When operating the combination weighing device 1 in the non-production mode, the control unit 40 collects and stores, as learning data, control parameters in a second range different from the first range and information on the input amount when the control parameters in the second range are used.

[0055] It should be noted that the second control parameter is a parameter other than the first control parameter among all the control parameters that the combination metering device 1 can obtain. The second control parameter may be all the parameters other than the first control parameter among all the control parameters that the combination metering device 1 can obtain, or may be a portion of the parameters other than the first control parameter among all the control parameters that the combination metering device 1 can obtain.

[0056] According to this configuration, by collecting only the learning data related to the control parameters that cannot be collected when the combination weighing device 1 operates in the production mode, sufficient learning data can be efficiently acquired for conveyance control.

[0057] Below, refer to Figure 3An example of the operation of the combination metering device 1 according to one embodiment will be described.

[0058] like Figure 3 As shown, in step S101 , the control unit 40 of the combination metering device 1 operates the combination metering device 1 in the production mode according to an instruction from the operator, and collects and stores learning data related to the first control parameter.

[0059] In step S102 , the control unit 40 of the combination metering device 1 switches the mode of the combination metering device 1 from the production mode to the non-production mode in accordance with an instruction from the operator.

[0060] In step S103 , the control unit 40 of the combination metering device 1 collects and stores learning data related to the second control parameter from the combination metering device 1 operating in the non-production mode.

[0061] In step S104 , the learning model generation unit 30 of the combination metering device 1 generates a learning model using the learning data collected and stored by the control unit 40 .

[0062] In step S105 , the control unit 40 of the combination weighing device 1 controls the conveyance of the article by the conveyance unit 20 based on the learning model generated by the learning model generation unit 30 .

[0063] According to the present embodiment, it is possible to obtain sufficient learning data for performing conveyance control in the combination weighing device 1 while avoiding deterioration in yield.

[0064] While the present invention has been described in detail using the above-described embodiments, it will be apparent to those skilled in the art that the present invention is not limited to the embodiments described in this specification. The present invention can be implemented in various modifications and variations without departing from the spirit and scope of the present invention as defined by the claims. Therefore, the description in this specification is for illustrative purposes only and is not intended to limit the present invention in any way.

[0065] For example, in the above embodiment, the combination weighing device 1 is described as the article conveying device 1. However, the article conveying device 1 is not limited to the combination weighing device 1, and may be, for example, a food conveying device that supplies food to the combination weighing device 1.

Claims

1. A method for controlling an article conveying device, comprising: Step A, using, as learning data, information related to the amount of input indicating the weight of the articles conveyed by a conveyor to a component disposed downstream, information indicating the state of the articles on the conveyor, and control parameters of the conveyor, to generate a learning model that estimates the control parameters set for conveying the articles of a target weight; Step B: controlling the conveyance of the article based on the learning model; Step C, selectively switching between a production mode in which the article conveying device participates in actual production and a non-production mode in which the article conveying device does not participate in the production, to operate the article conveying device, wherein the non-production mode is a mode in which the article conveying device discharges all articles retained in the article conveying device or a test mode in which the article conveying device is driven; and In step D, when the article transport device is operated in the non-production mode, actually obtained information on the input amount and the control parameters are collected and stored as the learning data.

2. The control method of the article conveying device according to claim 1, wherein: In the step D, when the article transport device is operated in the production mode, the actually obtained information on the input amount and the control parameters are also collected and stored as the learning data.

3. The control method of the article conveying device according to claim 2, wherein: The control parameters include a first control parameter and a second control parameter. In the step D, When the article conveying device is operated in the production mode, the first control parameter and information related to the input amount when the first control parameter is used are collected and stored as the learning data; When the article transport device is operated in the non-production mode, the second control parameter and information related to the input amount when the second control parameter is used are collected and stored as the learning data.

4. The control method of the article conveying device according to claim 1, wherein: In the step A, an image obtained by photographing the article is used as information indicating the state of the article.

5. An article conveying device comprising: Conveying part, transporting the input items; a learning model generating unit that uses, as learning data, information related to the input amount indicating the weight value of the article conveyed by the conveying unit to a component arranged on the downstream side, information indicating the state of the article on the conveying unit, and control parameters of the conveying unit, and generates a learning model that estimates the control parameters set for conveying the article of a target weight; and A control unit controls the conveyance of the article based on the learning model. The control unit selectively switches between a production mode that participates in actual production and a non-production mode that does not participate in the production to operate the article conveying device. The non-production mode is a mode of the article conveying device in which all articles retained in the article conveying device are discharged or a test mode of the article conveying device is driven. When operating the article transport device in the non-production mode, the control unit collects and stores actually obtained information on the input amount and the control parameters as the learning data.

6. The article conveying device according to claim 5, wherein: When operating the article transport device in the production mode, the control unit also collects and stores actually obtained information on the input amount and the control parameters as the learning data.

7. The article conveying device according to claim 6, wherein: The control parameters include a first control parameter and a second control parameter. When the control unit operates the article conveying device in the production mode, the control unit collects and stores the first control parameter and information related to the input amount when the first control parameter is used as the learning data. When operating the article transport device in the non-production mode, the control unit collects and stores the second control parameter and information related to the input amount when the second control parameter is used as the learning data.

8. The article conveying device according to claim 5, wherein: The article transport device includes an image acquisition unit that acquires an image of the article placed on the transport unit. The learning model generation unit uses the image of the article as information indicating the state of the article.

Images