Generation method for control parameter, program, recording medium, and generation device for control parameter
The method uses a machine learning model to update control parameters for drive machines and production equipment based on vibration data, addressing the inefficiency of conventional methods by automating the process and reducing the time required for parameter adjustment.
Patent Information
- Application Number
- JP2024139256
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-10-29
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-09-14
AI Technical Summary
Conventional methods for generating control parameters for drive machines and production equipment are time-consuming and require skilled engineers, often taking several days, and there is a need for a more efficient and automated approach.
A method using a machine learning model to update control parameters based on evaluation index data representing vibration of a driven object after reaching a permissible position, which includes generating measurement data, evaluation index data, and updating control parameters using a Bayesian optimization algorithm.
Enables the generation of appropriate control parameters in a relatively short time without the need for skilled engineers, effectively reducing the adjustment time and improving efficiency.
Smart Images

Figure 2024159859000001_ABST
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a method, a program, a recording medium, and a device for generating control parameters, and more particularly to a method, a program, a recording medium, and a device for generating control parameters used in a production device. [Background technology]
[0002] Conventionally, for a driving machine using a motor, a method has been proposed in which a machine learning model is used to generate control parameters for the motor and the driving machine (see, for example, Patent Document 1). [Prior art documents] [Patent documents]
[0003] [Patent Document 1] International Publication No. 2018 / 151215 Summary of the Invention
[0004] 2. Description of the Related Art Conventionally, there has been a demand for generating appropriate control parameters for controlling driving machines and production equipment.
[0005] In view of the above, an object of the present disclosure is to provide a control parameter generating method, a program, a recording medium, and a control parameter generating device that are capable of generating appropriate control parameters.
[0006] A method for generating control parameters according to an aspect of the present disclosure is a method for generating control parameters for use in a production device including a servo motor, a control circuit for controlling the servo motor, a memory for storing control parameters used by the control circuit when controlling the servo motor, and a driven object driven by the servo motor, and includes the following actions: After a time when the driven object reaches an allowable position that can be evaluated as having reached a predetermined target position based on a command to move the position of the driven object to the predetermined target position, measurement data representing the position is obtained from a sensor that measures the position of the driven object; Based on the measurement data, evaluation index data representing vibration of the driven object after the time of reaching the predetermined target position is generated; Based on the evaluation index data, the control parameters are updated using a machine learning model that learns the relationship between the evaluation index data and the control parameters; and, Based on the evaluation index data, the updated control parameters are output to the production device to be stored in the memory.
[0007] A program according to one aspect of the present disclosure is a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0008] A recording medium according to another aspect of the present disclosure is a recording medium having recorded thereon a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0009] A control parameter generating device according to another aspect of the present disclosure is a control parameter generating device for use in a production device including a servomotor, a first control circuit for controlling the servomotor, a memory for storing control parameters used when the first control circuit controls the servomotor, and an object to be driven by the servomotor. This control parameter generating device includes an input unit and a second control circuit. The input unit is configured to input a command to move the object to be driven to a predetermined target position based on a command to move the object to be driven to the predetermined target position. After the object reaches an allowable position that can be evaluated as having reached a position, measurement data representing the position is acquired from a sensor that measures the position of the object to be driven. The second control circuit generates evaluation index data representing vibration of the object to be driven after the arrival time based on the measurement data, and updates the control parameters based on the evaluation index data using a machine learning model that learns the relationship between the evaluation index data and the control parameters. The output unit includes an output unit that outputs the updated control parameters to the production device.
[0010] A method for generating control parameters according to another aspect of the present disclosure is a method for generating control parameters for use in a production device including a servomotor, a control circuit for controlling the servomotor, a memory for storing control parameters used by the control circuit when controlling the servomotor, and an object to be driven by the servomotor, and includes the following actions: After a time when the object to be driven reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command for positioning the object to be driven to a predetermined target position, evaluation index data representing vibration of the object to be driven after the time of reaching the predetermined target position is obtained from a sensor with a processing device that measures the position of the object to be driven. Based on the evaluation index data, the control parameters are updated using a machine learning model that learns the relationship between the evaluation index data and the control parameters.
[0011] A program according to another aspect of the present disclosure is a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0012] A recording medium according to another aspect of the present disclosure is a recording medium having recorded thereon a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0013] A control parameter generating device according to another embodiment of the present disclosure is a control parameter generating device for use in a production device including a servo motor, a first control circuit for controlling the servo motor, a memory for storing control parameters used by the first control circuit when controlling the servo motor, and a driven object driven by the servo motor. The control parameter generating device includes an input unit and a second control circuit. The input unit acquires evaluation index data representing vibration of the driven object after the time of arrival from a sensor with a processing device that measures the position of the driven object after the time of arrival, the sensor measuring the position of the driven object based on a command to move the position of the driven object to a predetermined target position. The second control circuit includes a second control circuit that updates the control parameter based on the evaluation index data using a machine learning model that learns the relationship between the evaluation index data and the control parameter.
[0014] According to the control parameter generating method, program, recording medium, and control parameter generating device according to the above aspects of the present disclosure, it is possible to generate appropriate control parameters. [Brief description of the drawings]
[0015] [Figure 1] FIG. 1 is a schematic diagram showing an overview of a control parameter generating system according to the first embodiment. [Diagram 2] FIG. 2 is a block diagram showing a configuration of the control parameter generating system according to the first embodiment. [Diagram 3] FIG. 3 is a schematic diagram illustrating an example of control parameters according to the first embodiment. [Figure 4] FIG. 4 is a schematic diagram showing an example of the transition of the position deviation of the object to be driven with respect to the target position. [Diagram 5] FIG. 5 is a schematic diagram showing an example of the measurement data according to the first embodiment. [Figure 6]FIG. 6 is a schematic diagram showing how the machine learning model 14 according to the first embodiment outputs a control parameter based on the result of learning the relationship between the evaluation index data and the control parameter. [Figure 7] FIG. 7 is a diagram showing an example of an image generated by the image generating unit according to the first embodiment. [Figure 8] FIG. 8 is a diagram showing an example of an image generated by the image generating unit according to the first embodiment. [Figure 9] FIG. 9 is a sequence diagram of the first control parameter adjustment process according to the first embodiment. [Figure 10] FIG. 10 is a flowchart of the first control parameter adjustment process according to the first embodiment. [Figure 11] FIG. 11 is a schematic diagram showing an overview of a control parameter generating system according to the second embodiment. [Figure 12] FIG. 12 is a block diagram showing a configuration of a control parameter generating system according to the second embodiment. [Figure 13] FIG. 13 is a sequence diagram of the second control parameter adjustment process according to the second embodiment. [Figure 14] FIG. 14 is a flowchart of the second control parameter adjustment process according to the second embodiment. [Figure 15] FIG. 15 is a block diagram showing a configuration of a generating device according to a modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0016] (How one aspect of the present disclosure was achieved) Conventionally, the adjustment of control parameters used in a production device having a servo motor for driving a driven object is done manually. However, the number of control parameters used in such a production device can reach 80, and the number of parameters that actually need to be adjusted can reach 30 to 60. In addition, the parameters to be adjusted differ depending on the driven object and the application to be implemented in the production device.
[0017] For these reasons, conventionally, adjustment of control parameters has been handled individually by skilled engineers, but even so, it may take several days to adjust the control parameters.
[0018] Such adjustment of the control parameters is necessary to suppress the generation of abnormal noise due to vibration, or to suppress the adverse effect of the vibration of the previous operation on the next operation when one operation is stopped and then started.
[0019] For example, Patent Document 1 proposes adjusting such control parameters using a machine learning model. However, a skilled engineer adjusts the control parameters using the five human senses, such as checking abnormal sounds with his ears and checking vibrations with his eyes and touch. Therefore, a separate study is required as to what kind of data should be input to the machine learning model to effectively adjust the control parameters.
[0020] In this context, the inventors have conducted extensive research and experiments to determine what data should be input to a machine learning model to effectively adjust the control parameters. As a result, the inventors have found that there is a correlation between the control parameters and evaluation index data that represents the vibration of the driven object after the object reaches an allowable position that can be evaluated as having reached a predetermined target position. The inventors have then found that the control parameters can be effectively adjusted by inputting the evaluation index data into a machine learning model.
[0021] Based on this knowledge, the inventors further conducted research and experiments, and developed the control parameters according to the present disclosure below. The present inventors have come up with a method for generating control parameters, a program, a recording medium, and an apparatus for generating control parameters.
[0022] A method for generating control parameters according to one aspect of the present disclosure is a method for generating control parameters for use in a production device including a servomotor, a control circuit for controlling the servomotor, a memory for storing control parameters used by the control circuit when controlling the servomotor, and an object to be driven by the servomotor, and includes the following actions: After a time when the object to be driven reaches an allowable position that can be evaluated as having reached a predetermined target position based on a command to move the position of the object to be driven to a predetermined target position, measurement data representing the position is obtained from a sensor that measures the position of the object to be driven; Based on the measurement data, evaluation index data representing vibration of the object to be driven after the time of reaching the predetermined target position is generated; Based on the evaluation index data, the control parameters are updated using a machine learning model that learns the relationship between the evaluation index data and the control parameters.
[0023] According to the above-described method for generating control parameters, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the allowable position is reached, appropriate control parameters can be generated without the involvement of a skilled technician and in a relatively short time.
[0024] In the above-mentioned method for generating a control parameter, the updated control parameter may be output to the production device for storage in the memory.
[0025] The evaluation index data may indicate a sum of areas surrounded by a deviation waveform indicating vibration of the driven object after the time reached and a reference axis.
[0026] This makes it possible to generate appropriate control parameters based on the sum of the areas surrounded by the reference axis and the deviation waveform indicating the vibration of the driven object after it has reached the permissible position.
[0027] The evaluation index data may represent a degree of variation in a deviation waveform indicating vibration of the driven object after the time reached.
[0028] This makes it possible to generate appropriate control parameters based on the degree of variation in the deviation waveform that indicates the vibration of the driven object after the time when the tolerable position is reached.
[0029] The evaluation index data may indicate an effective value of a deviation waveform indicating vibration of the driven object after the time reached.
[0030] This makes it possible to generate appropriate control parameters based on the effective value of the deviation waveform that indicates the vibration of the driven object after the time when the tolerable position is reached.
[0031] The production device may include one of a mounting device, a processing device, and a machining device.
[0032] This makes it possible to generate appropriate control parameters for a production device, which is any one of a mounting device, a processing device, and a machining device.
[0033] A program according to another aspect of the present disclosure is a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0034] According to the above program, by updating the control parameters based on evaluation index data indicating the vibration after the time when the allowable position is reached, appropriate control parameters can be generated without the involvement of an experienced technician and in a relatively short time.
[0035] A recording medium according to another aspect of the present disclosure is a recording medium having recorded thereon a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0036] According to the above-mentioned recording medium, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the allowable position is reached, appropriate control parameters can be generated without the need for an experienced technician and in a relatively short time.
[0037] A control parameter generating device according to another embodiment of the present disclosure is a control parameter generating device for use in a production device including a servo motor, a first control circuit for controlling the servo motor, a memory for storing control parameters used when the first control circuit controls the servo motor, and a driven object driven by the servo motor. The control parameter generating device includes an input unit and a second control circuit. The input unit acquires measurement data representing the position from a sensor that measures the position of the driven object after a time when the driven object reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command to move the position of the driven object to the predetermined target position. The second control circuit generates evaluation index data representing vibration of the driven object after the time when the driven object reaches the predetermined target position based on the measurement data, and updates the control parameter based on the evaluation index data using a machine learning model that learns the relationship between the evaluation index data and the control parameter.
[0038] According to the control parameter generating device, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the tolerable position is reached, it is possible to generate appropriate control parameters without the need for a skilled technician and in a relatively short time. The control parameter generating device may further include an output unit that outputs the updated control parameters to the production device for storage in the memory.
[0039] A method for generating control parameters according to another aspect of the present disclosure is a method for generating control parameters for use in a production device including a servomotor, a control circuit for controlling the servomotor, a memory for storing control parameters used by the control circuit when controlling the servomotor, and an object to be driven by the servomotor, and includes the following actions: After a time when the object to be driven reaches an allowable position that can be evaluated as having reached a predetermined target position based on a command to move the position of the object to be driven to a predetermined target position, evaluation index data representing vibration of the object to be driven after the time of reaching the predetermined target position is obtained from a sensor with a processing device that measures the position of the object to be driven. Based on the evaluation index data, the control parameters are updated using a machine learning model that learns the relationship between the evaluation index data and the control parameters.
[0040] According to the above-mentioned method for generating control parameters, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the tolerable position is reached, it is possible to generate appropriate control parameters without the need for a skilled technician and in a relatively short time. In the above-mentioned method for generating control parameters, the updated control parameters may be output to the production device for storage in the memory.
[0041] The evaluation index data may indicate a sum of areas surrounded by a deviation waveform indicating vibration of the driven object after the time reached and a reference axis.
[0042] This makes it possible to generate appropriate control parameters based on the sum of the areas surrounded by the reference axis and the deviation waveform indicating the vibration of the driven object after it has reached the permissible position.
[0043] The evaluation index data may represent a degree of variation in a deviation waveform indicating vibration of the driven object after the time reached.
[0044] This makes it possible to generate appropriate control parameters based on the degree of variation in the deviation waveform that indicates the vibration of the driven object after the time when the tolerable position is reached.
[0045] The evaluation index data may indicate an effective value of a deviation waveform indicating vibration of the driven object after the time reached.
[0046] This makes it possible to generate appropriate control parameters based on the effective value of the deviation waveform that indicates the vibration of the driven object after the time when the tolerable position is reached.
[0047] The production device may include one of a mounting device, a processing device, and a machining device.
[0048] This makes it possible to generate appropriate control parameters for a production device, which is any one of a mounting device, a processing device, and a machining device.
[0049] A program according to another aspect of the present disclosure is a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0050] According to the above program, by updating the control parameters based on evaluation index data indicating the vibration after the time when the allowable position is reached, appropriate control parameters can be generated without the involvement of an experienced technician and in a relatively short time.
[0051] A recording medium according to another aspect of the present disclosure is a recording medium having recorded thereon a program for causing a computer of an information processing device connected to the production device to execute the control parameter generation method.
[0052] According to the above-mentioned recording medium, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the allowable position is reached, appropriate control parameters can be generated without the need for an experienced technician and in a relatively short time.
[0053] A control parameter generating device according to another embodiment of the present disclosure is a control parameter generating device for use in a production device including a servo motor, a first control circuit for controlling the servo motor, a memory for storing control parameters used by the first control circuit when controlling the servo motor, and a driven object driven by the servo motor. The control parameter generating device includes an input unit and a second control circuit. The input unit acquires evaluation index data representing vibration of the driven object after the time of arrival from a sensor with a processing device that measures the position of the driven object after the time of arrival, the sensor having the processing device measuring the position of the driven object. The second control circuit updates the control parameter based on the evaluation index data using a machine learning model that learns the relationship between the evaluation index data and the control parameter.
[0054] According to the control parameter generating device, by updating the control parameters based on the evaluation index data indicating the vibration after the time when the tolerable position is reached, it is possible to generate appropriate control parameters without the need for a skilled technician and in a relatively short time. The control parameter generating device may further include an output unit that outputs the updated control parameters to the production device for storage in the memory.
[0055] A specific example of a control parameter generation system according to an embodiment of the present disclosure will be described below with reference to the drawings. Each embodiment shown here shows a specific example of the present disclosure. Therefore, the numerical values, shapes, components, the arrangement and connection of the components, as well as the steps (processes) and the order of the steps shown in the following embodiments are merely examples and are not intended to limit the present disclosure. In addition, each figure is a schematic diagram and is not necessarily an exact drawing. In each figure, the same reference numerals are used for configurations that are substantially the same as those in other figures, and duplicated explanations are omitted or simplified.
[0056] (Embodiment 1) The following describes a control parameter generation system 1 according to embodiment 1. This control parameter generation system is a system that generates control parameters used in a production device that includes a servo motor that drives an object to be driven.
[0057] <Configuration> FIG. 1 is a schematic diagram showing an overview of a control parameter generating system 1 according to the first embodiment.
[0058] FIG. 2 is a block diagram showing the configuration of the control parameter generating system 1. As shown in FIG.
[0059] As shown in FIG. 1, the control parameter generating system 1 includes a generating device 10, a production device 20, and a sensor 30.
[0060] The production apparatus 20 is used to produce devices, and processes, mounts, transports, etc. the devices. The production apparatus 20 is installed, for example, on a production line in a factory. Specifically, the production apparatus 20 is, for example, an LED (Light Emitting Diode) bonder, a mounter, a processing machine, a take-out robot, etc.
[0061] As shown in FIG. 2, the production device 20 includes a memory 21, a control circuit 22, a servo motor 23, and an object to be driven 24.
[0062] The operation of the servo motor 23 is controlled by the control circuit 22, and drives the driven object 24. The servo motor 23 may be, for example, a rotary motor or a linear motor.
[0063] The driven object 24 is an object that is driven by the servo motor 23. The driven object 24 is, for example, an object to be processed in the production device 20, an object to be mounted on, an object to be transported, etc.
[0064] The control circuit 22 controls the servo motor 23 by outputting a command to the servo motor 23 to set the position of the driven object 24 to a predetermined target position. The command output by the control circuit 22 to the servo motor 23 may be, for example, a position command that commands the position of the servo motor 23 or the driven object 24, or may be, for example, a torque command that commands the torque of the servo motor 23.
[0065] The control circuit 22 controls the servo motor 23 based on the control parameters stored in the memory 21. In other words, the control circuit 22 uses the control parameters stored in the memory 21 when controlling the servo motor 23.
[0066] The memory 21 stores control parameters used by the control circuit 22 to control the servo motor 23. The control parameters stored in the memory 21 are the control parameters output from the generation device 10.
[0067] FIG. 3 is a schematic diagram showing an example of the control parameters stored in the memory 21. As shown in FIG.
[0068] As shown in FIG. 3, the control parameters stored in the memory 21 include, for example, parameters a1, a2 for adjusting the vibration frequency of the object to be driven 24, parameters b1, b2 for adjusting the speed of the object to be driven 24, parameters c1, c2 for adjusting the depth of the singular point in the vibration characteristics of the object to be driven 24, parameters d1, d2 for adjusting the vibration amplitude of the object to be driven 24, etc.
[0069] In general, the control parameters include parameters that have a trade-off relationship with each other, such as parameters b1 and b2 that adjust the speed, parameters c1 and c2 that adjust the depth of the singularity in the vibration characteristics, and parameters d1 and d2 that adjust the vibration amplitude.
[0070] Returning again to FIG. 1 and FIG. 2, the description of the control parameter generating system 1 will continue.
[0071] The sensor 30 measures the position of the driven object 24. Then, the sensor 30 outputs to the generating device 10 measurement data representing the position of the driven object 24 after the driven object 24 reaches an allowable position at which it can be evaluated that the driven object 24 has reached a predetermined target position.
[0072] FIG. 4 is a schematic diagram showing an example of the transition of the position deviation of the driven object 24 from the target position when the production device 20 drives the driven object 24 to the target position.
[0073] In FIG. 4, the horizontal axis indicates time, and the vertical axis indicates the position deviation of the driven object 24 from the target position.
[0074] As shown in Fig. 4, in this specification, the permissible position refers to a position where the position deviation from the target position is within the required accuracy. Also, as shown in Fig. 4, in this specification, the time at which the permissible position is reached, which can be evaluated as having reached the target position (hereinafter, also referred to as "settling time"), refers to the time at which the driven object 24 finally reaches the permissible position (time sett_t in Fig. 4) in the case where the driven object 24 reaches the permissible position and then does not deviate from the permissible position again.
[0075] FIG. 5 is a data configuration diagram showing an example of the measurement data output by the sensor 30. As shown in FIG.
[0076] As shown in FIG. 5, the measurement data is, for example, data in which the elapsed time [ms] after the time when the tolerable position is reached is associated one-to-one with the deviation amount [mm] from the target position.
[0077] Returning again to FIG. 1 and FIG. 2, the description of the control parameter generating system 1 will continue.
[0078] The generating device 10 generates control parameters (hereinafter also referred to as "update control parameters") for updating the control parameters (hereinafter also referred to as "stored control parameters") stored in the memory 21. More specifically, the generating device 10 sequentially acquires measurement data output from the sensor 30 as a result of the production device 20 driving the driven object 24 to a target position using the stored control parameters, sequentially generates update control parameters based on the acquired measurement data, and sequentially outputs the generated update control parameters to the production device 20.
[0079] As shown in FIG. 2, the generating device 10 includes an input unit 11, an output unit 12, a control circuit 13, and It comprises a machine learning model 14, an operation receiving unit 15, an image generating unit 16, and a display unit 17.
[0080] The generation device 10 is realized, for example, by a computer device including a processor, a memory, and an input / output interface, in which the processor executes a program stored in the memory. Such a computer device is, for example, a personal computer.
[0081] The input unit 11 acquires the measurement data output from the sensor 30 .
[0082] The control circuit 13 generates evaluation index data indicating the vibration of the driven object 24 after the settling time based on the measurement data acquired by the input unit 11, and updates the control parameters based on the generated evaluation index data using a machine learning model 14 that learns the relationship between the evaluation index data and the control parameters.
[0083] The evaluation index data generated by the control circuit 13 has one of the following (1) to (3) or (4): (1) the sum of the areas surrounded by the deviation waveform indicating the vibration of the driven object 24 after the settling time and the reference axis (hereinafter also referred to as "total area evaluation index data"); (2) the degree of variation of the deviation waveform indicating the vibration of the driven object 24 after the settling time (hereinafter also referred to as "variation degree evaluation index data"); (3) the effective value of the deviation waveform indicating the vibration of the driven object 24 after the settling time (hereinafter also referred to as "effective value evaluation index data"); and (4) the time from the time when the driven object 24 starts to be driven to the settling time (hereinafter also referred to as "settling time evaluation index data").
[0084] The evaluation index data generated by the control circuit 13 is specified by a user who uses the control parameter generating system 1.
[0085] The total area evaluation index data, the degree of variation evaluation index data, and the effective value evaluation index data generated by the control circuit 13 will be described in more detail below with reference to FIG.
[0086] As shown in FIG. 4, more specifically, the total area data generated by the control circuit 13 is calculated by the formula shown in (Formula 1), (Formula 2), or (Formula 3) when the position deviation f of the driven object with respect to the target position is expressed as f=E(t) with time t as a variable.
[0087]
number
[0088]
number
[0089]
number
[0090] For example, cal_e may be set to the time when the position deviation of the driven object 24 with respect to the target position finally reaches within 1 / 10 of the required accuracy in a case where the position deviation does not deviate from within 1 / 10 of the required accuracy again after reaching within 1 / 10 of the required accuracy. Also, cal_e may be set to the time when the number of data included in the measurement data acquired by the input unit 11 reaches a predetermined number, or may be set to the time when a specified time has elapsed from the settling time.
[0091] More specifically, the variation degree data generated by the control circuit 13 is calculated by the formula shown in (Formula 4) where def=cal_e-sett_t.
[0092]
number
[0093]
number
[0094] The machine learning model 14 is a machine learning model that learns the relationship between the evaluation index data and the control parameters, and learns the relationship between the evaluation index data and the control parameters each time evaluation index data is sequentially input from the control circuit 13, and sequentially outputs the control parameters to the control circuit 13 based on the learning results.
[0095] More specifically, the machine learning model 14 utilizes a Bayesian optimization algorithm to learn the relationship between the evaluation index data and the control parameters corresponding to the evaluation index data (hereinafter also referred to as “corresponding control parameters”) each time evaluation index data is sequentially input from the control circuit 13, and based on the learning results, outputs to the control circuit 13 the control parameters predicted to minimize the evaluation index data.
[0096] FIG. 6 is a schematic diagram showing how the machine learning model 14 learns the relationship between the evaluation index data and the corresponding control parameters, and outputs the control parameters that are predicted to minimize the evaluation index data based on the learning results.
[0097] 6, the horizontal axis is the value of the control parameter, and the vertical axis is the value of the evaluation index data. Also, in FIG. 6, the points plotted as circles indicate the values of the control parameters corresponding to the evaluation index data for each value of the evaluation index data previously input to the machine learning model 14.
[0098] As shown in FIG. 6, the machine learning model 14 uses a Bayesian optimization algorithm to predict the relationship between the evaluation index data and the corresponding control parameters to a certain extent, and uses the predicted relationship to output the control parameters that are predicted to minimize the evaluation value data.
[0099] Returning again to FIG. 1 and FIG. 2, the description of the control parameter generating system 1 will continue.
[0100] The output unit 12 outputs the control parameters updated by the control circuit 13 to the production device 20 so that the control parameters are stored in the memory 21 .
[0101] The operation receiving unit 15 receives an input operation to the generation device 10 by a user who uses the control parameter generation system 1.
[0102] The display unit 17 displays an image to be provided to a user who uses the control parameter generating system 1.
[0103] The image generating unit 16 generates an image to be displayed on the display unit 17 .
[0104] Fig. 7 is an example of an image generated by the image generating unit 16. Fig. 7 is an example of an image (hereinafter also referred to as an "evaluation index data designation image") that prompts a user who uses the control parameter generating system 1 to designate which evaluation index data should be generated by the control circuit 13 when the control parameter generating system 1 starts a first control parameter adjustment process described later.
[0105] As shown in FIG. 7, the generation device 10 displays an evaluation index data designation image on the display unit 17, thereby prompting the user to designate at least one of the total area evaluation index data, the degree of variation evaluation index data, and the effective value evaluation index data (hereinafter, these three evaluation index data are also referred to as “vibration evaluation index data”), and the settling time evaluation value data, as the evaluation index data to be generated by the control circuit 13.
[0106] For example, the user specifies settling time evaluation index data as the evaluation index data by checking "settling time" in the evaluation index data specification image. For example, the user specifies area sum evaluation index data as the evaluation index data by checking "vibration" and putting a circle next to "area" in the evaluation index data specification image. For example, the user specifies variation degree evaluation index data as the evaluation index data by checking "vibration" and putting a circle next to "variance" in the evaluation index data specification image. For example, the user specifies effective value evaluation index data as the evaluation index data by checking "vibration" and putting a circle next to "RMS (root mean square)" in the evaluation index data specification image.
[0107] Also, the user may specify either the settling time evaluation index data or the vibration evaluation index data as the evaluation index data by checking both "settling time" and "vibration" in the evaluation index data specification image. In this case, the control circuit 13 may specify the evaluation index data whose value is the sum of the settling time evaluation index data and the vibration evaluation index data, or may specify the evaluation index data whose value is the sum of values obtained by multiplying the settling time evaluation index data and the vibration evaluation index data by predetermined weighting coefficients.
[0108] FIG. 8 is another example of an image generated by the image generating unit 16. FIG. 8 shows an example of an image generated by the image generating unit 16 while the control parameter generating system 1 is performing a first control parameter adjustment process described later. This is an example of an image (hereinafter also referred to as a "situation presenting image") that presents the status of the evaluation index data values in the parameter adjustment process to a user who uses the parameter generation system 1.
[0109] 8, the horizontal axis represents the number of control parameters sequentially output by the generating device 10, and the vertical axis represents the values of the evaluation index data corresponding to each of the control parameters sequentially output by the generating device 10. Also, in the graph in FIG. 8, the solid broken line represents the transition of the minimum value of the evaluation index data corresponding to the control parameters output by the generating device 10.
[0110] As shown in FIG. 8, an icon labeled "Pause" (hereinafter also referred to as the "Pause icon") and an icon labeled "Resume" (hereinafter also referred to as the "Resume icon") are located in the upper right corner of the regular status image.
[0111] The interrupt icon is an icon that allows the user to interrupt the first control parameter adjustment process by clicking the interrupt icon when, for example, the user visually checks the situation presentation image and determines that there is no need to continue the first control parameter adjustment process any further because the value of the evaluation index value data corresponding to the control parameter output by the generation device 10 has become sufficiently small.
[0112] The resume icon is an icon that allows the user to continue the first control parameter adjustment process by clicking the resume icon when, for example, the user visually checks the situation presentation image and determines that it is necessary to continue the first control parameter adjustment process because the value of the evaluation index value data corresponding to the control parameter output by the generation device 10 is not sufficiently small.
[0113] <Operation> The operation of the control parameter generating system 1 having the above configuration will be described below.
[0114] The control parameter generation system 1 executes a first control parameter adjustment process for adjusting to appropriate values the control parameters stored in the memory 21. The first control parameter adjustment process is started, for example, when a user who uses the control parameter generation system 1 performs an operation on the generation device 10 to start the first control parameter adjustment process.
[0115] FIG. 9 is a sequence diagram of the first control parameter adjustment process, and FIG. 10 is a flowchart of the first control parameter adjustment process.
[0116] As shown in FIGS. 9 and 10, when the first control parameter adjustment process is started, the generating device 10 starts a predetermined program for executing the first control parameter adjustment process (step S10).
[0117] When the predetermined program is started, the output unit 12 outputs the initial value of the control parameter to the memory 21 (step S15). At this time, the output unit 12 may output, for example, the initial value of the control parameter consisting of a predetermined value, or may output, for example, the initial value of the control parameter consisting of a value designated by the user, or may output, for example, the initial value of the control parameter consisting of a value calculated by a calculation method designated by the user.
[0118] When the control parameters are output from the output unit 12, the memory 21 stores the control parameters output from the output unit 12 (step S20). Then, the control circuit 22 controls the position of the driven object 24 to a predetermined target position based on the control parameters stored in the memory 21. The control circuit 22 generates a command to drive the servo motor 23 and outputs it to the servo motor 23. In response to this, the control circuit 22 controls the servo motor 23 (step S25). Then, the servo motor 23 drives the driven object 24 based on the command output from the control circuit 22 (step S30).
[0119] When the driven object 24 is driven by the servo motor 23, the sensor 30 measures the position of the driven object 24 (step S35) and outputs measurement data representing the position of the driven object 24 after the driven object 24 reaches an allowable position at which it can be evaluated that the driven object 24 has reached a predetermined target position, to the input unit 11. Then, the input unit 11 acquires the measurement data output from the sensor 30 (step S40).
[0120] When the input unit 11 acquires the measurement data, the control circuit 13 generates evaluation index data indicating the vibration of the driven object 24 after the settling time based on the measurement data (step S45). Then, the control circuit 13 checks whether the evaluation index data satisfies a predetermined condition (step S50). Here, the predetermined condition is, for example, a condition that the evaluation index data has a value indicating that an appropriate control parameter has been generated.
[0121] In the process of step S50, if the generated evaluation index data does not satisfy a predetermined condition (step S50: No), the control circuit 13 outputs the generated evaluation index data to the machine learning model 14. Then, the machine learning model 14 learns the relationship between the evaluation index data and the corresponding control parameter (step S55), and outputs the control parameter predicted to have the smallest value of the evaluation index data. Then, the control circuit 13 updates the control parameter previously output by the output unit 12 with the control parameter newly output from the machine learning model 14 (step S60). Then, the output unit 12 transmits the control parameter newly updated by the control circuit 13 to the memory 21 (step S65).
[0122] When the process of step S65 ends, the control parameter generating system 1 proceeds to the process of step S20, and repeats the processes from step S20 onwards.
[0123] In the process of step S50, if the generated evaluation index data satisfies a predetermined condition (step S50: Yes), the control parameter generating system 1 ends the first control parameter adjustment process.
[0124] <Consideration> As described above, the control parameter generating system 1 configured as above executes the first control parameter adjustment process to repeatedly adjust the evaluation index value data until the evaluation index value data satisfies a predetermined condition.
[0125] Therefore, the control parameter generation system 1 can generate appropriate control parameters without the involvement of a skilled engineer and in a relatively short time.
[0126] (Embodiment 2) Hereinafter, a control parameter generating system according to the second embodiment, which is configured by partially modifying the configuration of the control parameter generating system 1 according to the first embodiment, will be described.
[0127] The control parameter generating system 1 is an example of a system in which the sensor 30 measures the position of an object to be driven and outputs the measurement data, and the measurement data generating device 10 acquires the measurement data output from the sensor 30.
[0128] In contrast, in the control parameter generating system according to the second embodiment, the sensor with the processing function according to the second embodiment measures the position of the object to be driven, outputs evaluation index data, and The measurement data generating device according to the second aspect is an example of a system that acquires evaluation index data output from the sensor with processing function according to the second embodiment.
[0129] Here, regarding the control parameter generation system of embodiment 2, components similar to those of control parameter generation system 1 have already been explained, so they are given the same symbols and their detailed explanations are omitted, and the explanation focuses on the differences from control parameter generation system 1.
[0130] FIG. 11 is a schematic diagram showing an overview of a control parameter generating system 1A according to the second embodiment.
[0131] FIG. 12 is a block diagram showing a configuration of a control parameter generating system 1A.
[0132] 11, the control parameter generation system 1A is configured by changing the generation device 10 from the control parameter generation system 1 according to the first embodiment to a generation device 10A, and changing the sensor 30 to a sensor with a processing function 30A. Also, as shown in Fig. 12, the generation device 10A is configured by changing the input unit 11 from the generation device 10 to an input unit 11A, and changing the control circuit 13 from the generation device 10 to a control circuit 13A.
[0133] The sensor with processing function 30A measures the position of the driven object 24. Then, the sensor with processing function 30A generates measurement data representing the position of the driven object 24 after the time when the driven object 24 reaches an allowable position at which it can be evaluated that the driven object 24 has reached a predetermined target position. Furthermore, the sensor with processing function 30A generates evaluation index data representing vibration of the driven object 24 after the settling time based on the generated measurement data, and outputs the generated evaluation index data to the generating device 10A.
[0134] The sensor with processing function 30A is realized, for example, by a computer device including a sensing device for measuring the position of the driven object 24, a processor, a memory, and an input / output interface, in which the processor executes a program stored in the memory. Such a computer device is, for example, a personal computer.
[0135] The input unit 11A acquires the evaluation index data output from the sensor with processing function 30A.
[0136] The control circuit 13A updates the control parameters based on the evaluation index data acquired by the input unit 11A, using a machine learning model 14 that learns the relationship between the evaluation index data and the control parameters.
[0137] <Operation> The operation of the control parameter generating system 1A having the above configuration will be described below.
[0138] The control parameter generation system 1A performs a second control parameter adjustment process in which some of the processes in the first control parameter adjustment process according to the first embodiment are changed.
[0139] FIG. 13 is a sequence diagram of the second control parameter adjustment process, and FIG. 14 is a flowchart of the second control parameter adjustment process.
[0140] As shown in FIGS. 13 and 14, in the second control parameter adjustment process, the process of step S35 is changed to the process of step S135, the process of step S40 is changed to the process of step S140, and the process of step S45 is changed to the process of step S The processing has been changed to processing 145.
[0141] When the process of step S30 is completed, the sensor with processing function 30A measures the position of the driven object 24 (step S135) and generates measurement data representing the position of the driven object 24 after the time when the driven object 24 reaches an allowable position at which it can be evaluated that the driven object 24 has reached a predetermined target position. Then, the sensor with processing function 30A generates evaluation index data representing the vibration of the driven object 24 after the settling time based on the generated measurement data (step S140) and outputs the generated evaluation index data to the input unit 11A. Then, the input unit 11A acquires the evaluation index data output from the sensor with processing function 30A (step S145).
[0142] When the process of step S145 ends, the control parameter generating system 1A proceeds to the process of step S50.
[0143] In the process of step S50, if the generated evaluation index data satisfies a predetermined condition (step S50: Yes), the control parameter generating system 1A ends the second control parameter adjustment process.
[0144] <Consideration> As described above, the control parameter generation system 1A having the above configuration, like the control parameter generation system 1 according to embodiment 1, executes a second control parameter adjustment process to repeatedly adjust the evaluation index value data until the evaluation index value data satisfies a predetermined condition.
[0145] Therefore, according to the control parameter generation system 1A, similarly to the control parameter generation system 1, appropriate control parameters can be generated without the assistance of a skilled technician and in a relatively short time.
[0146] (supplement) As described above, the technology disclosed in this application has been described based on the first and second embodiments. However, the present disclosure is not limited to the first and second embodiments. As long as it does not deviate from the spirit of the present disclosure, various modifications conceived by a person skilled in the art to the present embodiment or the modifications, or forms constructed by combining components in different embodiments or modifications, may also be included within the scope of one or more aspects of the present disclosure.
[0147] (1) In the first embodiment, the generating device 10 has been described as being realized by one computer device as an example. However, the generating device 10 is not necessarily limited to being realized by one computer device as long as the generating device 10 can realize the same functions. The generating device 10 may be realized by, for example, a plurality of computer devices capable of communicating with each other.
[0148] FIG. 15 is a block diagram showing an example of the configuration of generation device 10 in the case where generation device 10 is realized by a plurality of computer devices capable of communicating with each other.
[0149] As shown in FIG. 15, the generation device 10 may be realized by (1) a first computer device 100 having an input unit 11, an output unit 12, an operation reception unit 15, an image generation unit 16, a display unit 17, and a communication unit 18, (2) a second computer device 110 having a control circuit 13, a machine learning model 14, and a communication unit 19, and (3) a network 120 that communicatively connects the communication unit 18 and the communication unit 19.
[0150] In this case, the first computing device 100 is, for example, a personal computer. , the second computing device 110 may be a server.
[0151] (2) A comprehensive or specific aspect of the present disclosure may be realized in a system, an apparatus, a method, an integrated circuit, a program, or a non-transitory recording medium such as a computer-readable CD-ROM. Also, the present disclosure may be realized in any combination of a system, an apparatus, a method, an integrated circuit, a program, and a non-transitory recording medium. For example, the present disclosure may be realized as a program for causing a computer device to execute the processing performed by the generating device. [Industrial Applicability]
[0152] The present disclosure is widely applicable to systems for generating control parameters, etc. [Explanation of symbols]
[0153] 1. 1A control parameter generation system 10, 10A generator 11, 11A input section 12 Output section 13, 13A Control circuit 14 Machine Learning Models 15 Operation reception section 16 Image Generation Unit 17 Display section 18, 19 Communications Department 20 Production Equipment 21 Memory 22 Control circuit 23 Servo motor 24 Drive object 30 Sensors 30A Sensor with Processing Function 100 First computer device 110 Second computer device 120 Network
Claims
1. A method for generating control parameters for use in a production device that drives a driven object by the servo motor, the production device comprising: a servo motor; a control circuit that controls the servo motor; and a memory that stores control parameters used when the control circuit controls the servo motor, the production device comprising: acquiring measurement data representing the position of the object to be driven from a sensor that measures the position of the object to be driven after a time when the object to be driven reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command to move the position of the object to be driven to the predetermined target position; generating evaluation index data representing vibration of the driven object after the arrival time based on the measurement data, the arrival time representing a time after the arrival time at which the driven object no longer deviates from the allowable position; updating the control parameters based on the evaluation index data, using a machine learning model that learns a relationship between the evaluation index data and the control parameters, so as to output the control parameters that are predicted to reduce the evaluation index data; How to generate control parameters.
2. Outputting the updated control parameters to the production device for storage in the memory. The method for generating a control parameter according to claim 1.
3. 2. The method for generating a control parameter according to claim 1, wherein the evaluation index data indicates a sum of an area surrounded by a deviation waveform indicating vibration of the driven object after the time reached and a reference axis.
4. 2. The method for generating a control parameter according to claim 1, wherein the evaluation index data indicates a degree of variation in a deviation waveform indicating vibration of the driven object after the time reached.
5. 2. The method for generating a control parameter according to claim 1, wherein the evaluation index data indicates an effective value of a deviation waveform indicating vibration of the driven object after the time reaches the target time.
6. The method for generating a control parameter according to claim 1 , wherein the production device includes any one of a mounting device, a processing device, and a machining device.
7. A program for causing a computer of an information processing device connected to the production device to execute the method for generating a control parameter according to any one of claims 1 to 5.
8. 6. A recording medium having recorded thereon a program for causing a computer of an information processing device connected to the production device to execute the control parameter generating method according to any one of claims 1 to 5.
9. A control parameter generating device for use in a production device that drives a driven object by the servo motor, the control parameter generating device comprising: a servo motor; a first control circuit that controls the servo motor; and a memory that stores control parameters used when the first control circuit controls the servo motor, the control parameter generating device comprising: an input unit that acquires measurement data representing a position of the object to be driven from a sensor that measures the position of the object to be driven after a time when the object to be driven reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command to move the position of the object to be driven to the predetermined target position; generating evaluation index data representing vibration of the driven object after the arrival time based on the measurement data, the arrival time representing a time after the arrival time at which the driven object no longer deviates from the allowable position; Based on the evaluation index data, a relationship between the evaluation index data and the control parameter is calculated. A second control circuit updates the control parameters by using a machine learning model to output the control parameters that are predicted to reduce the evaluation index data; A control parameter generating device comprising:
10. and an output unit that outputs the updated control parameters to the production device so that the updated control parameters are stored in the memory. The control parameter generating device according to claim 9.
11. A method for generating control parameters for use in a production device that drives a driven object by the servo motor, the production device comprising: a servo motor; a control circuit that controls the servo motor; and a memory that stores control parameters used when the control circuit controls the servo motor, the production device comprising: After a time when the driven object reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command to move the position of the driven object to the predetermined target position, evaluation index data representing vibration of the driven object after the time when the driven object reaches the allowable position is obtained from a sensor with a processing device that measures the position of the driven object, and the time when the driven object reaches the allowable position represents a time when the driven object does not deviate from the allowable position again after the time when the driven object reaches the predetermined target position, updating the control parameters based on the evaluation index data, using a machine learning model that learns a relationship between the evaluation index data and the control parameters, so as to output the control parameters that are predicted to reduce the evaluation index data; How to generate control parameters.
12. Outputting the updated control parameters to the production device for storage in the memory. The method for generating a control parameter according to claim 11.
13. 12. The method for generating a control parameter according to claim 11, wherein the evaluation index data indicates a sum of an area surrounded by a deviation waveform indicating vibration of the driven object after the time reached and a reference axis.
14. 12. The method for generating a control parameter according to claim 11, wherein the evaluation index data indicates a degree of variation in a deviation waveform indicating vibration of the driven object after the time reached.
15. The method for generating a control parameter according to claim 11, wherein the evaluation index data indicates an effective value of a deviation waveform indicating vibration of the driven object after the time reaches the target time.
16. The method for generating a control parameter according to claim 11 , wherein the production device includes any one of a mounting device, a processing device, and a machining device.
17. A program for causing a computer of an information processing device connected to the production device to execute the method for generating a control parameter according to any one of claims 11 to 15.
18. A recording medium storing a program for causing a computer of an information processing device connected to the production device to execute the method for generating a control parameter according to any one of claims 11 to 15.
19. A control parameter generating device for use in a production device that drives a driven object by the servo motor, the control parameter generating device comprising: a servo motor; a first control circuit that controls the servo motor; and a memory that stores control parameters used when the first control circuit controls the servo motor, the control parameter generating device comprising: The object to be driven after the time when the object to be driven reaches an allowable position that can be evaluated as having reached the predetermined target position based on a command to move the position of the object to be driven to the predetermined target position. an input unit for acquiring evaluation index data representing vibration of the driven object after the arrival time from a sensor equipped with a processing device that measures a position; a second control circuit that updates the control parameter based on the evaluation index data by using a machine learning model that learns a relationship between the evaluation index data and the control parameter, so as to output the control parameter that is predicted to reduce the evaluation index data; The reached time indicates a time after the time when the driven object no longer deviates from the permissible position. A generator of control parameters.
20. and an output unit that outputs the updated control parameters to the production device so that the updated control parameters are stored in the memory.
20. The control parameter generating device according to claim 19.
Citation Information
Patent Citations
Positioning control device and positioning method
JP6664561B1
Control device and control method
WO2018151215A1
