Information processing device, polishing robot, and polishing system

Abstract

To make a robot properly execute polishing operation according to a polishing material and the type of desired finishing.SOLUTION: A processor (10) of an information processing device (1A) includes an acquisition step (S11) that acquires material information and finishing information as input information, an identifying step (S12) that identifies polishing conditions on the basis of the material information and finishing information, and teaching steps (S13-14) that control a polishing robot (2A) to polish the material under the identified polishing conditions. The polishing conditions include parameters showing the force of pressing the material against a polishing machine (3).SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an information processing device, a polishing robot, and a polishing system. [Background technology]

[0002] Various technologies have been developed for allowing robots to polish resin materials, metal materials, etc. For example, Patent Document 1 discloses a technology related to a polishing device that polishes semiconductor wafers. Also, Patent Document 2 discloses a technology related to force control of a processing tool during finish processing. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-053550 [Patent Document 2] Japanese Patent Application Publication No. 2019-089144 Summary of the Invention [Problem to be solved by the invention]

[0004] The technology described in Patent Document 1 is a technology for precisely polishing a specific object, namely a semiconductor wafer. In other words, with the technology described in Patent Document 1, the material of the object to be polished and the type of polishing finish are almost fixed. On the other hand, there is also a demand for a robot that can polish various materials to a desired finish.

[0005] An object of one aspect of the present invention is to allow a robot to perform appropriate polishing depending on the material to be polished and the type of finish desired. [Means for solving the problem]

[0006] An information processing device according to one aspect of the present disclosure is an information processing device including one or more processors. The processors execute the following steps: an acquisition step of acquiring, as input information, material information, which is information about a material to be polished, and finishing information, which indicates a type of polishing finish; a specification step of specifying polishing conditions, which are information indicating how the polishing will be performed, based on the input information; and a control step of controlling a polishing robot to polish the material under the polishing conditions specified in the specification step. The polishing conditions include a parameter indicating a pressing force of the material against the polishing machine. [Effects of the Invention]

[0007] According to one aspect of the present invention, it is possible to cause a robot to perform appropriate polishing depending on the material to be polished and the type of finish desired. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram showing an overview of a grinding system according to a first embodiment. [Figure 2] 1 is a diagram showing an example of the structure of a polishing robot according to a first embodiment. [Figure 3] 1A and 1B are diagrams illustrating a structure of a force sensor and parameters detected by the force sensor; [Figure 4] 1 is a block diagram showing the internal configuration of an information processing device and a polishing robot according to a first embodiment. [Figure 5] FIG. 2 is a diagram showing data stored in a memory of the information processing device according to the first embodiment. [Figure 6] 5 is a flowchart showing an example of a processing flow in a processor of the information processing device according to the first embodiment. [Figure 7] FIG. 10 is a diagram showing an overview of a learning system according to a second embodiment. [Figure 8] FIG. 10 is a diagram showing data stored in a memory of an information processing device according to a second embodiment. [Figure 9]10 is a flowchart showing an example of a processing flow in a processor of an information processing device according to a second embodiment. [Figure 10] FIG. 10 is a diagram showing an outline of a grinding system according to a third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] [Embodiment 1] The polishing system according to this embodiment is a system for specifying appropriate polishing conditions according to the object to be polished and the finish state. The polishing system according to this embodiment is also a system for controlling a polishing robot so that polishing is performed under the specified polishing conditions. Hereinafter, a first embodiment of the present invention will be described.

[0010] <Polishing System Overview> FIG. 1 is a diagram showing an overview of a polishing system 100 according to this embodiment. The polishing system 100 includes an information processing device 1A, a polishing robot 2A, and a polishing machine 3. FIG. 1 also shows a user operating the information processing device 1A and a workpiece W to be polished. The information processing device 1A and the polishing robot 2A are connected by wire or wirelessly. The information processing device 1A and the polishing machine 3 may also be connected by wire or wirelessly.

[0011] (Information processing device 1A) The information processing device 1A, from the information input thereto, identifies various conditions for the polishing method (i.e., polishing conditions) that allow the polishing robot 2A to properly polish the workpiece W using the polishing machine 3. In this embodiment, material information and finishing information are input to the information processing device 1A. "Material information" is information about the material to be polished (i.e., the workpiece W). Information about the material is, for example, the type and dimensions of the raw material of the material. "Finishing information" is information indicating the type of polishing finish. The type of finish indicates the type of surface condition of the workpiece W after polishing is completed (i.e., after finishing), such as rough finish, matte finish, mirror finish, etc. In addition to the above, various other information that affects the identification of the polishing conditions may be input to the information processing device 1A. Such information input for identifying the polishing conditions will hereinafter be referred to as "input information."

[0012] Here, "completion of polishing of workpiece W" means that a series of procedures specified by the teaching data described below has been completed for the workpiece W. For example, if the teaching data specifies the polishing operations for one process of polishing the workpiece W, "completion of polishing of the workpiece W" means that the polishing for that one process has been completed. Also, for example, if the teaching data specifies the polishing operations for all processes up to the shipment of the workpiece W, "completion of polishing of the workpiece W" means that all processes of polishing the workpiece W have been completed.

[0013] The information processing device 1A is a device that directly or indirectly controls the polishing robot 2A. When the information processing device 1A is connected to a polishing machine 3, the information processing device 1A may control the operation of the polishing machine 3. In this embodiment, as an example, the information processing device 1A transmits teaching data to the polishing robot 2A. This allows the information processing device 1A to indirectly control the polishing robot 2A.

[0014] "Teaching data" refers to a program and / or parameters for causing the polishing robot 2A to perform a specific operation. In this embodiment, the teaching data refers to a program and / or parameters for causing the polishing robot 2A to perform a series of operations for polishing the workpiece W. The information processing device 1A creates teaching data according to the identified polishing conditions and transmits it to the polishing robot 2A. The polishing robot 2A then executes an operation based on this teaching data. This allows the information processing device 1A to cause the polishing robot 2A to polish the workpiece W as specified in the teaching data.

[0015] (Polishing robot 2A and polishing machine 3) The polishing robot 2A is a robot that polishes a workpiece W under the control of the information processing device 1A. The specific structure of the polishing robot 2A will be described later. The polishing machine 3 is a machine to which an abrasive material for polishing the workpiece W is attached. In this embodiment, as shown in FIG. 1, the polishing machine 3 has a structure in which an abrasive material such as a metal file, sandpaper, or polisher is attached to the surface of a cylindrical core material. The polishing machine 3 rotates the core material using power from a motor or the like, thereby rotating the abrasive material together with the core material. The polishing robot 2A polishes the workpiece W by pressing the workpiece W attached to the tip of its robot arm 24 against the rotating abrasive material.

[0016] (Overview of Operation of the Polishing System 100) As shown in Figure 1, when a user inputs material information and finishing information to information processing device 1A, information processing device 1A accepts the input and acquires the material information and finishing information. Information processing device 1A identifies appropriate polishing conditions from the acquired material information and finishing information. "Appropriate polishing conditions" refer to the conditions for polishing a workpiece having the properties indicated by the input material information until it passes a quality check for the type of finish indicated by the input finishing information.

[0017] The information processing device 1A creates teaching data according to the identified polishing conditions and transmits it to the polishing robot 2A. Upon receiving the teaching data, the polishing robot 2A executes the teaching data. As a result, the polishing robot 2A executes a polishing operation suited to the material information and finishing information input by the user, and polishes the workpiece W.

[0018] <Structure of Polishing Robot 2A> Fig. 2 is a diagram showing an example of the structure of a polishing robot 2A. As shown in Fig. 2, the polishing robot 2A includes a base 22, a base 23, a robot arm 24, and a joint 25. In addition, a force sensor 40 is incorporated between the robot arm 24 and the joint 25.

[0019] The base 22 is the base of the polishing robot 2A. The base 22 may be rotatable. When the base 22 rotates, the base 23 fixed to the base 22, and the communication IF 26, force sensor 40, and joint 25 connected to the base 23 rotate together. This allows the polishing robot 2A to change direction. The base 22 may be fixed to a self-propelled mechanism such as an AGV (Automatic Guided Vehicle). In other words, the polishing robot 2A may be a self-propelled robot.

[0020] The base 23 is a structure that connects the base 22 and the robot arm 24. The control unit 20, memory unit 21, and communication IF 26, which will be described later, may be built into the base 23 or the base 22. The base 23 may also have a built-in power source, such as a motor, for moving the robot arm 24.

[0021] The robot arm 24 is an arm that serves as the "hand" of the polishing robot 2A when it performs its work. The robot arm 24 may be configured with multiple joints, as shown in FIG. 2. The position and orientation of the tip of the joint 25 of the robot arm 24 can be adjusted by adjusting the bending angle of the joint. The tip of the robot arm 24 is structured to allow a force sensor 40 to be attached. The joint 25 is a structure for setting the workpiece W on the robot arm 24. In this embodiment, the joint 25 has a structure for fixing the workpiece W and a structure for attaching the force sensor 40. The force sensor 40 is then attached between the robot arm 24 and the joint 25.

[0022] The force sensor 40 is a sensor for detecting the force and moment acting on the workpiece W. In the example of Fig. 2, the force sensor 40 is attached between the robot arm 24 and the joint 25, but the attachment position of the force sensor 40 is not limited to the example of Fig. 2. The force sensor 40 may be attached at any position on the polishing robot 2A as long as it is a position where the force and moment acting on the workpiece W can be detected.

[0023] <Structure of force sensor 40> Fig. 3 is a diagram showing the structure of the force sensor 40 shown in Fig. 2 and parameters detected by the force sensor 40. As in Fig. 2, in the example of Fig. 3, the force sensor 40 is attached between the robot arm 24 and the joint portion 25.

[0024] 3, the force sensor 40 has a first member having a first surface 401, a second member having a second surface 402, and a strain element (not shown) disposed between the first and second members. The first surface 401 of the force sensor 40 is attached to the robot arm 24.

[0025] The second surface 402 is attached to the joint portion 25. The force sensor 40 detects the deformation of the internal strain-generating body, thereby detecting the magnitude of the force acting in each of the three axes (x-axis, y-axis, z-axis) in the force sensor 40 (fx, fy, fz in FIG. 3) and the magnitude of the moment around each axis (mx, my, mz in FIG. 3). In this embodiment, the force sensor 40 is connected to the workpiece W via the joint portion 25. Therefore, the various forces and moments detected by the force sensor 40 can be considered to be substantially the same as the various forces and moments acting on the workpiece W.

[0026] Of the above-mentioned forces and moments, as long as the force sensor 40 can detect at least the magnitude of the force (fz) relative to the Z axis, it is not essential to detect other forces and moments. When the robot arm 24 operates to press the workpiece W against an object, a force is applied to the workpiece W in the fz direction. In other words, in this embodiment, at least fz is involved in the parameters indicating the pressing force of the workpiece W against the grinder 3.

[0027] <Internal configuration of information processing device 1A and polishing robot 2A> 4 is a block diagram showing the internal configuration of the information processing device 1A and the polishing robot 2A. The information processing device 1A includes a processor 10, a memory 11, an input interface (IF) 12, an output IF 13, and a communication IF 14. The processor 10, the memory 11, the input IF 12, the output IF 13, and the communication IF 14 are interconnected via a bus.

[0028] The polishing robot 2A includes a control unit 20, a memory unit 21, a communication IF 26, and a force sensor 40. The control unit 20, the memory unit 21, and the communication IF 26 are interconnected via a bus. If the base 22 has an electrical component such as a motor, the base 22 may also be interconnected with the above-mentioned components via the bus.

[0029] The sanding robot 2A is connected to a force sensor 40 attached to the sanding robot 2A. The connection IF between the sanding robot 2A and the force sensor 40 is not particularly limited. For example, the force sensor 40 may be configured to be connected to the bus of the sanding robot 2A via the end of the robot arm 24.

[0030] (Internal configuration of information processing device 1A) The memory 11 stores data and programs necessary for the operation of the information processing device 1 A. An example of a device that can be used as the memory 11 is a semiconductor RAM (Random Access Memory).

[0031] 5 is a diagram showing data stored in memory 11. In this embodiment, memory 11 stores, for example, trained model M1 and teaching data T1. Note that memory 11 may also store various data and programs for operating information processing device 1A, such as an operating system (OS) program for information processing device 1A. Memory 11 may also store various parameters detected by force sensor 40 in chronological order.

[0032] The trained model M1 is a trained model that has been machine-learned to understand the relationship between various pieces of information included in the input information (e.g., material information and finishing information) and various pieces of information included in the polishing conditions (e.g., pressing force fz). In this embodiment, the means and method by which the information processing device 1A acquires the trained model M1 are not particularly limited.

[0033] The data structure of the trained model M1 is not particularly limited as long as the trained model M1 can estimate the polishing conditions. For example, the trained model M1 may be a neural network (NN) in which the weighting coefficients for the output of each node are optimized by machine learning. If the trained model M1 is a NN, it is preferable that the trained model M1 be a multi-layer NN (e.g., four or more layers) that can be expected to have high identification accuracy.

[0034] The teaching data T1 is teaching data for operating the polishing robot 2 A. The teaching data T1 is created by the processor 10.

[0035] The processor 10 is an arithmetic unit that performs various calculations by executing instructions written in a program stored in the memory 11. Devices that can be used as the processor 10 include, for example, a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a microcontroller, or a combination of these.

[0036] The input IF 12 is an interface through which the information processing device 1A receives input from an input device. Specific standards for the input IF 12 include, for example, Universal Serial Bus (USB), Advanced Technology Attachment (ATA), Small Computer System Interface (SCSI), and Peripheral Component Interconnect (PCI). Examples of input devices include a keyboard, a mouse, a touchpad, a microphone, or a combination of these. For example, in the example shown in FIG. 1, the information processing device 1A receives input operations via the keyboard and / or mouse, which are input devices. A user inputs material information and finishing information using the keyboard and / or mouse.

[0037] The output IF13 is an interface through which the information processing device 1A outputs data to an output device. Specific examples of the output IF13 are the same as those of the input IF12. Examples of the output device include a display and a speaker. For example, in the example shown in FIG. 1, the information processing device 1A may display the material information, finishing information, and polishing conditions input by the user on a display, which is an output device.

[0038] The communication IF 14 is an interface for connecting the information processing device 1A and the polishing robot 2A. Specific standards for the communication IF 14 include, for example, interfaces such as Ethernet (registered trademark), LAN, and Wi-Fi (registered trademark).

[0039] (Internal structure of polishing robot 2A) The storage unit 21 stores data and programs necessary for the operation of the sanding robot 2 A. In this embodiment, the storage unit 21 stores the teaching data T1 that the sanding robot 2 A receives from the information processing device 1A.

[0040] The control unit 20 comprehensively controls the sanding robot 2A by executing commands written in a program stored in the memory unit 21. More specifically, the control unit 20 reads and executes the teaching data T1 stored in the memory unit 21, thereby operating drive mechanisms such as various motors and gears provided in the base 22, the base 23, the robot arm 24, and the joint unit 25 as specified by the teaching data T1. This allows the control unit 20 to move the entire sanding robot 2A as instructed by the teaching data T1.

[0041] The communication IF 26 is a communication IF that connects the polishing robot 2 A and the information processing device 1 A. Specific examples of the communication IF 26 are the same as those of the communication IF 14.

[0042] The force sensor 40 transmits the force and moment parameters it detects to the control unit 20. The force sensor 40 may detect and transmit the parameters at any time. The control unit 20 may also monitor the detection values ​​from the force sensor 40 and perform feedback control of the entire polishing robot 2A in accordance with the detection values.

[0043] <Processing flow> (Overall processing of information processing device 1A) 6 is a flowchart showing an example of the flow of processing in the processor 10 of the information processing device 1 A. When the processor 10 receives input of material information and finishing information from the user, it executes a series of processing shown in FIG.

[0044] First, the processor 10 acquires input information via the input IF 12 (S11, acquisition step). In this embodiment, the processor 10 acquires at least material information and finishing information. The processor 10 may store the acquired input information in the memory 11.

[0045] Next, the processor 10 identifies the polishing conditions based on the input information (S12, identification step). The method for identifying the polishing conditions will be described in detail later. As described above, the identified polishing conditions include at least parameters indicating the pressing force of the workpiece W against the polishing machine 3 (parameters indicating the force and moment acting on the workpiece W, including at least fz). The processor 10 may display the identified polishing conditions on a display or the like via the output IF13.

[0046] Next, the processor 10 controls the polishing robot 2A to polish the workpiece W under the specified polishing conditions (S13 to S14, control step). In this embodiment, the processor 10 first creates teaching data T1 for the polishing robot 2A to perform polishing under the specified polishing conditions (S13, teaching data creation step). For example, the processor 10 creates teaching data T1 in which the pressing force threshold value when the polishing robot 2A performs polishing is set to a predetermined range from the pressing force specified as the polishing condition. Next, the processor 10 transmits the created teaching data T1 to the polishing robot 2A via the communication IF 14 (S14, teaching data transmission step). The processor 10 may store the created teaching data T1 in the memory 11.

[0047] Upon receiving the teaching data T1, the polishing robot 2A stores the teaching data T1 in the memory unit 21 and executes the memory unit 21 at any timing. As a result, the polishing robot 2A can, for example, perform polishing while monitoring the detection value from the force sensor 40 and adjusting the pressing force on the workpiece W so that it is a pressing force appropriate for the polishing conditions. Note that the processor 10 of the information processing device 1A may instruct the polishing robot 2A on the timing of executing the teaching data T1 when transmitting the teaching data T1. Alternatively, the teaching data T1 may include information specifying the timing of executing the teaching data T1.

[0048] According to the above process, appropriate polishing conditions can be identified based on information about the material to be polished and the type of polishing finish. The identified polishing conditions can then be taught to the polishing robot. This allows the polishing robot 2A to perform appropriate polishing based on the material to be polished and the type of desired finish. Furthermore, according to the above process, by transmitting teaching data T1 to the polishing robot 2A, the polishing robot can perform appropriate polishing based on the material to be polished and the type of desired finish.

[0049] (Method for determining polishing conditions) The method for identifying the polishing conditions in S12 of Fig. 6 will be described in further detail. In the polishing system 100, the processor 10 estimates the polishing conditions using the trained model M1 (estimation step). As described above, the trained model M1 is a trained model that has been machine-learned to determine the relationship between various pieces of information contained in the input information and various pieces of information contained in the polishing conditions.

[0050] The processor 10 inputs at least a portion of the input information into the trained model M1. For example, the processor 10 inputs material information and finishing information into the trained model M1. If the input information includes other information and the trained model M1 has also learned the relationship between the other information and the polishing conditions, the processor 10 may also input the other information into the trained model M1.

[0051] The trained model M1 outputs polishing conditions according to the input information. That is, the trained model M1 estimates polishing conditions. For example, when material information and finishing information are input, the trained model M1 estimates polishing conditions for the material and dimensions of the workpiece W indicated by the material information to satisfy the finishing state indicated by the finishing information.

[0052] In addition, if there are multiple patterns of polishing conditions that satisfy the conditions indicated by the input information, the trained model M1 may output all of the multiple patterns of polishing conditions, or may output only one of them. If multiple patterns of polishing conditions are output, the processor 10 identifies one of the polishing conditions as the polishing condition to be used in creating the teaching data T1. The method of identification is not particularly limited. If only one pattern of polishing conditions is output from the trained model M1, the processor 10 identifies that one polishing condition as the polishing condition to be used in creating the teaching data T1.

[0053] In this way, according to the dressing system 100 of this embodiment, appropriate dressing conditions can be estimated from input information using the trained model M1.

[0054] First Modification of First Embodiment In the example of FIG. 1, one polishing robot 2A is connected to one information processing device 1A. However, in the polishing system 100, the information processing device 1A and the polishing robot 2A do not need to have a one-to-one relationship. For example, the information processing device 1A may communicate with multiple polishing robots 2A. The information processing device 1A may then create teaching data T1 corresponding to input information for each polishing robot 2A and transmit the teaching data T1 to each polishing robot 2A. The polishing robot 2A may also communicate with multiple information processing devices 1A. For example, the polishing robot 2A may be configured to operate using the teaching data T1 upon receiving teaching data T1 from an information processing device 1A owned by a certain user. Furthermore, the polishing system 100 may be configured such that multiple information processing devices 1A each communicate with multiple polishing robots 2A.

[0055] <<Second Modification of First Embodiment>> In the above description, an example has been described in which polishing conditions are estimated and identified from input information using the trained model M1. However, the information processing device 1A according to this embodiment may be configured to identify polishing conditions from input information without using the trained model M1.

[0056] For example, the memory 11 of the information processing device 1A may store a data table in which optimal polishing conditions are associated with combinations of various pieces of information included in the input information. Then, in S12 of Fig. 6, the processor 10 may identify the polishing conditions corresponding to the input information obtained in S11 by referring to the data table. Note that this data table may be a combination of multiple tables or may be a multidimensional table.

[0057] [Embodiment 2] The trained model M1 described in the first embodiment may be a model constructed by supervised learning using the training data, with the following (i) or (ii) as the training data: (i) Combination of input information and polishing conditions when workpiece W is actually polished (ii) A combination of input information and polishing conditions created by simulating the polishing of the workpiece W. A second embodiment of the present disclosure will be described below with reference to Figures 7 to 9. For ease of explanation, members having the same functions as those described in the above embodiment will be denoted by the same reference numerals, and their description will not be repeated. This also applies to the following embodiments.

[0058] FIG. 7 is a diagram showing an overview of a learning system 200 according to this embodiment. The learning system 200 is a system for constructing the trained model M1 described above. A second embodiment of the present invention will be described below. The learning system 200 includes an information processing device 1B, a polishing robot 2B, and a polishing machine 3.

[0059] The hardware configurations of the information processing device 1B and the polishing robot 2B are the same as those of the information processing device 1A and the polishing robot 2A, respectively. However, the polishing robot 2B is a robot that allows direct teaching. That is, a user can manually operate the polishing robot 2B to polish a workpiece W. The polishing robot 2B is a robot that can record, in the memory unit 21, parameters related to the drive of the robot arm 24 and detection values ​​of the force sensor 40 when manually operated.

[0060] When the polishing robot 2B is operated manually, the aforementioned parameters and detected values ​​obtained change moment by moment as the polishing operation progresses (i.e., as time passes). The control unit 20 of the polishing robot 2B records the parameters and detected values ​​that change during a series of polishing operations in the memory unit 21 in chronological order. In other words, the polishing robot 2B is a robot that can record log data of various parameters related to its own operation and the detected values ​​of the force sensor 40 when operated manually. Hereinafter, this log data will also be referred to as a "teaching log."

[0061] Furthermore, the input IF 12 and the output IF 13 are not essential components of the information processing device 1B. The information stored in the memory 11 according to this embodiment is different from that stored in the memory 11 according to the first embodiment. The information stored in the memory 11 will be described later.

[0062] (Overview of the operation of the learning system 200) In the learning system 200 shown in FIG. 7, the polishing robot 2B receives direct teaching. At this time, the polishing robot 2B records a teaching log. The polishing robot 2B transmits the teaching log to the information processing device 1B. The information processing device 1B receives the teaching log from the polishing robot 2B. The information processing device 1B also acquires various setting information and quality information that affect the polishing conditions when the polishing operation indicated in the received teaching log is performed.

[0063] Here, the "various setting information that affects the polishing conditions" is the same type of information as the input information in embodiment 1. That is, the "various setting information that affects the polishing conditions" is, for example, material information and finish information. Hereinafter, in this embodiment, the "various setting information that affects the polishing conditions" will be referred to as "input information" for convenience.

[0064] Furthermore, "quality information" is information indicating the quality evaluation of polishing by the polishing robot 2B. The quality evaluation of polishing is, for example, an evaluation of the extent to which the surface of the workpiece W meets the target surface characteristics, based on human visual inspection or measurement values ​​from a sensor device. In this embodiment, the quality information indicates one of two types: "polishing OK" and "polishing NG." Polishing OK indicates that the workpiece W after polishing is of acceptable quality. Polishing NG indicates that there is a problem with the quality of the workpiece W after polishing, such as when polishing is insufficient or excessive.

[0065] The method for acquiring the input information and quality information in the information processing device 1B is not particularly limited. However, the input information and quality information acquired by the information processing device 1B are information obtained when polishing indicated by at least one of the teaching logs is performed. In other words, it is desirable for the information processing device 1B to acquire the input information and quality information in a manner that allows the teaching log corresponding to this information to be identified.

[0066] For example, the information processing device 1B may acquire the input information and / or quality information via the polishing robot 2B. In this case, the polishing robot 2B is provided with an input IF that accepts input operations from a user. Then, either before or after direct teaching, the polishing robot 2B acquires the input information manually input via the input IF as input information corresponding to the direct teaching. Furthermore, after direct teaching, the polishing robot 2B acquires the quality information manually input via the input IF as quality information corresponding to the direct teaching. The polishing robot 2B associates the acquired input information and / or quality information with a teaching log and transmits the acquired input information and / or quality information to the information processing device 1B.

[0067] Furthermore, for example, the information processing device 1B may allow a user to manually input the input information and / or quality information via the input IF 12. At that time, the user may specify which teaching log the input information and / or quality information the user inputs corresponds to.

[0068] In this way, the information processing device 1B acquires the teaching log, input information during the polishing operation indicated by the teaching log, and quality information indicating the results of the polishing indicated by the teaching log. The information processing device 1B creates learning data D1 based on the acquired teaching log, input information, and quality information. Thereafter, at any timing, the information processing device 1B causes the unlearned model to perform machine learning using the learning data D1. This results in the construction of a trained model. The trained model constructed here may be the same as the trained model M1 described in embodiment 1. In the following description, it is assumed that the information processing device 1B constructs the trained model M1.

[0069] FIG. 8 is a diagram showing data stored in the memory of the information processing device 1B. As shown in FIG. 8, at least learning data D1 and a trained model M1 are stored in the memory 11 of the information processing device 1B. Note that FIG. 8 shows the memory 11 after the trained model has been constructed. Before the trained model M1 is constructed, at least an untrained model is stored in the memory 11. The untrained model is data before machine learning of the trained model M1.

[0070] The training data D1 is one or more pieces of training data used in machine learning to build the trained model M1. The training data D1 is data in which at least polishing conditions and input information are associated with quality information (e.g., whether polishing is OK or NG). In addition, the memory 11 may also store a teaching log, input information, and quality information, each associated with the other.

[0071] The polishing robot 2B records the operating parameters and detected values ​​when direct teaching is performed as a teaching log. The direct teaching here is the teaching of the movement of "the polishing robot 2B polishing the workpiece W using the polishing machine 3." The memory unit 21 of the polishing robot 2B records various polishing conditions, including the detected values ​​of the force sensor 40, and input information when polishing is performed.

[0072] The method for collecting the learning data D1 in this embodiment is not limited to the above-described method. For example, the control unit 20 of the polishing robot 2B may create the learning data D1 by associating (i) the polishing conditions included in the teaching log, (ii) the input information acquired by the polishing robot 2B, and (iii) the quality information acquired by the polishing robot 2B. In this way, when the polishing robot 2B creates the learning data D1, the information processing device 1B acquires the learning data D1 instead of acquiring teaching data or the like from the polishing robot 2B. Then, machine learning is performed based on the acquired learning data D1.

[0073] Furthermore, for example, the information processing device 1B may create the learning data D1 by associating (i) input information and (ii) polishing conditions included in the teaching log when the polishing robot 2B is operated using teaching data created by online teaching and / or offline teaching with (iii) quality information. Furthermore, the learning system 200 may sense the polishing conditions (e.g., pressing force) when a human polishes various materials using the polisher 3 without using the polishing robot 2B. Then, a combination of the sensed polishing conditions and input information with the quality information may be stored in the information processing device 1B as the learning data D1.

[0074] <Processing flow> 9 is a flowchart showing an example of the flow of processing in the processor 10 of the information processing device 1 B. Note that the start timing of the processing in FIG. 9 is not particularly limited.

[0075] The processor 10 of the information processing device 1B creates learning data D1 from the teaching log, input information, and quality information acquired from the polishing robot 2B (S21, learning data preparation step). Alternatively, the information processing device 1B acquires the learning data D1 from the polishing robot 2B (S21, learning data preparation step). The processor 10 stores the created or acquired learning data D1 in the memory 11.

[0076] Next, the processor 10 constructs a trained model M1 by performing machine learning on the untrained model using the training data D1 (S22, trained model construction step). The structure of the trained model M1 and the machine learning method are not particularly limited. For example, assume that the trained model M1 is constructed as a neural network that, when at least a portion of input information is input, outputs polishing conditions for which the quality information indicates "polishing OK." In this case, the weighting coefficients for the output of each node of the untrained neural network are optimized by machine learning using the training data D1 as training data. This makes it possible to create the trained model M1 from the untrained model. Note that, as described above, the group of training data D1 may include data whose quality information indicates "polishing NG." Such data, which represents so-called failed polishing, may or may not be used in machine learning.

[0077] According to the above processing, a trained model M1 can be constructed that can estimate, from the input information, the polishing conditions under which the quality information indicates that polishing is OK (i.e., the polishing conditions under which the polishing result will be of good quality).

[0078] [Embodiment 3] The dressing system 100 according to the first embodiment and the learning system 200 according to the second embodiment may be implemented in combination. Hereinafter, a third embodiment of the present disclosure will be described with reference to FIG. 10. FIG. 10 is a diagram showing an overview of a dressing system 300 according to this embodiment. As shown in FIG. 10, the dressing system 300 is a system that combines the dressing system 100 and the learning system 200.

[0079] Information processing device 1C is an information processing device that has the functions of both information processing device 1A and information processing device 1B. That is, information processing device 1C has the following functions (1) to (3).

[0080] (1) Function to build a trained model M1 (2) A function to estimate polishing conditions using the constructed trained model M1 (3) A function to identify one polishing condition from the estimated polishing conditions and create teaching data corresponding to that polishing condition. The polishing robot 2A and the polishing robot 2B may be the same device or different devices. Also, part or all of the learning data D1 in Fig. 10 may be obtained through online teaching and offline teaching, similar to the learning system 200.

[0081] According to the configuration shown in FIG. 10, polishing conditions can be identified in the same manner as in the polishing system 100, using a trained model M1 constructed in the same manner as in the learning system 200.

[0082] [Modification] (Variations of polishing techniques) Furthermore, the workpiece W may be polished by a method other than the methods described in the above-described embodiments. For example, a camera capable of photographing the workpiece W may be installed in advance (or mounted on the polishing robot 2A or 2B), and the workpiece W may be photographed by the camera. The control unit 20 may then analyze the image from the camera to identify areas on the workpiece W that are particularly uneven, and operate the robot arm 24 to polish (i.e., press against) those areas intensively. This allows for more efficient polishing. Furthermore, it becomes possible to identify polishing conditions for such efficient polishing.

[0083] (Modification of the grinder 3) In the above-described embodiments, polishing is performed by pressing the workpiece W attached to the tip of the robot arm 24 against the fixed polisher 3. However, the positional relationship between the polisher 3 and the workpiece W may be reversed. In other words, the polisher may be attached to the tip of the robot arm 24, and the polisher may be vibrated, rotated, or the like while being pressed against the fixed workpiece W, thereby polishing the workpiece W.

[0084] Although the grinder 3 is pressed against the workpiece W, rather than the grinder 3 being pressed against the workpiece W, the parameter of the pressing force may be measured by the same measuring method as in the above-described embodiment.

[0085] (Variations of various information included in input information) In each of the above-described embodiments, the input information may include polishing tool information and / or polishing count information. By constructing the trained model M1 based on a wider variety of input information, it is possible to construct a trained model M1 that can more accurately estimate appropriate polishing conditions. Furthermore, by inputting a wider variety of input information into the trained model M1, it is possible to have the trained model M1 more accurately estimate appropriate polishing conditions.

[0086] More specifically, "polishing tool information" is information indicating the type of polishing machine 3 and / or the type of abrasive attached to the polishing machine 3. Depending on the type of polishing machine 3, the appropriate polishing conditions (e.g., the pressing force on the workpiece W) may change. Furthermore, the appropriate polishing conditions may also change depending on the type of abrasive attached to the polishing machine 3. Therefore, by including polishing tool information in the input information, it is possible to identify more appropriate polishing conditions according to the type of polishing machine 3 and / or the type of abrasive.

[0087] Furthermore, "polishing count information" is information that indicates the type and number of polishing operations to be performed until polishing of the material to be polished is complete (i.e., finishing). For example, when polishing metal to a mirror finish, rough polishing may be performed several times, followed by polishing with fine sandpaper to reduce surface irregularities, and then polishing to a mirror finish. In this way, it may be necessary to perform multiple types of polishing and / or multiple times of polishing to achieve a certain finish on a certain material. The polishing count information is information that specifies the type and number of polishing operations. In other words, the polishing count information can be said to indicate the polishing procedure.

[0088] (Variations of various information included in polishing conditions) In each of the above-described embodiments, the polishing conditions may include at least one of polishing tool information, polishing count information, and polishing speed information. Note that if the polishing tool information and polishing count information are included in the input information, they are not included in the polishing conditions.

[0089] The "polishing speed information" is information indicating at least one of the polishing speed of the polishing robot and the polishing speed of the polishing machine 3. In other words, the polishing speed information is information indicating the time that the workpiece W is in contact with the polishing machine 3.

[0090] If the polishing conditions include polishing tool information, an appropriate type of polishing machine can be identified according to the input information. This allows the polishing robot to polish the material more appropriately. If the polishing conditions include polishing count information, an appropriate type and number of polishings to be performed until polishing is complete can be identified according to the input information. This allows the polishing robot to polish the material more appropriately. If the polishing conditions include polishing speed information, an appropriate polishing speed for the polishing robot and / or polishing machine can be identified according to the input information. This allows the polishing robot to polish the material more appropriately.

[0091] (Modification of parameters indicating pressing force) The parameter indicating the pressing force of the workpiece W is not limited to fz. For example, if the polishing surface of the workpiece W is not flat, that is, if part of the workpiece to be polished is inclined and / or curved, when the workpiece W is pressed against the polishing machine 3, rattles will occur in the workpiece W and the robot arm 24. The mx and my detected by the force sensor 40 indicate the moments acting on the workpiece W and the robot arm 24 in this way. When the force sensor 40 detects mx and my, the pressing force of the workpiece W may be expressed as a combination of the parameters fz, mx, and my.

[0092] Similarly, Mz detected by the force sensor 40 may be included in the parameters indicating the pressing force of the workpiece W. For example, when a polishing machine is attached to the tip of the robot arm 24 and is pressed against a fixed workpiece W to polish it, the polishing machine at the tip of the robot arm 24 may rotate (vibrate) irregularly. This vibration can be detected as the value of Mz.

[0093] [Software implementation example] The functions of information processing devices 1A, 1B, and 1C (hereinafter collectively referred to as "devices") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device.

[0094] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program, thereby realizing the functions described in each of the above embodiments.

[0095] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0096] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0097] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0098] 100, 300 Polishing System 200 Learning System 1A, 1B, 1C Information processing equipment 2A, 2B Polishing Robot 3 Polishing machine 10 processors 11. Memory 12 Input IF 13 Output IF 14 Communication Interface 20 Control Unit 21 Memory section 22 Foundation 23 Base 24 Robot Arm 25 Joint 26 Communication Interface 40 Force sensor D1 Learning data M1 pre-trained model S11 Acquisition step S12 Specific step S13 Teaching step T1 Teaching Data

Claims

1. An information processing device for controlling the operation of a polishing robot, the information processing device comprising: a robot arm for setting a material to be polished; a control unit for controlling the operation of the robot arm; and a force sensor for measuring the pressing force of the material set on the robot arm against a polishing machine, one or more processors; an acquisition step in which the processor acquires, as input information, material information that is information about the material to be polished and finishing information that indicates the type of polishing finish; a specifying step of specifying polishing conditions, which are information indicating how to polish the object, based on the input information; a control step of controlling the polishing robot so as to polish the material under the polishing conditions specified in the specifying step; the polishing conditions include a parameter indicating a pressing force of the material against the polishing machine; An information processing device characterized in that either the input information or the polishing conditions includes polishing count information indicating the type and number of polishings to be performed until polishing of the material is completed.

2. In the control step, the processor a teaching data creating step of creating teaching data for the polishing robot to perform polishing under the specified polishing conditions; 2. The information processing apparatus according to claim 1, further comprising: a teaching data transmitting step of transmitting the teaching data created in the teaching data creating step to the polishing robot.

3. 3. The information processing device according to claim 1, wherein either the input information acquired in the acquisition step or the polishing conditions identified in the identification step includes polishing tool information indicating the type of the polishing machine.

4. The information processing device according to any one of claims 1 to 3, characterized in that the polishing conditions identified in the identification step include polishing speed information indicating at least one of the polishing speed of the polishing robot and the polishing speed of the polishing machine.

5. In the specifying step, the processor The information processing device according to any one of claims 1 to 4, characterized in that an estimation step of estimating the polishing conditions is executed by inputting at least a portion of the input information into a trained model that has undergone machine learning to determine the relationship between various information contained in the input information and various information contained in the polishing conditions.

6. The processor: a learning data preparation step of acquiring or creating learning data that combines the input information and the polishing conditions; The information processing device according to claim 5, further comprising: a trained model construction step for constructing a trained model by machine learning the relationship between various information contained in the input information and various information contained in the polishing conditions through supervised learning using the learning data.

7. one or more processors; The processor: a learning data preparation step of acquiring or creating learning data that combines input information and polishing conditions; a trained model construction step of constructing a trained model by machine learning the relationship between various information included in the input information and various information included in the polishing conditions through supervised learning using the training data; the input information includes at least material information, which is information about a material to be polished, and finishing information, which indicates a type of finishing of the polishing, and the polishing conditions include at least a parameter, which indicates a pressing force of the material against the polishing machine by a robot arm for setting the material; 10. An information processing device, wherein either the input information or the polishing conditions includes polishing count information indicating the type and number of polishing operations to be performed until polishing of the material is completed.

8. A polishing robot that communicates with the information processing device according to any one of claims 1 to 6, A robot arm for setting the material to be polished; a control unit that controls the operation of the robot arm; a force sensor for measuring a pressing force of the material set on the robot arm against the polishing machine, The polishing robot is characterized in that the control unit operates the robot arm under the control of the information processing device.

9. An information processing device according to any one of claims 1 to 6; The polishing robot according to claim 8 ; and the polishing machine.

Images