Information processing method, information processing device, and program

Abstract

The information processing device selects at least one representative action from all actions executable by the device, causes the device to execute the selected representative action, acquires measurement data of a prescribed evaluation index measured during execution of the representative action, and updates the control parameter on the basis of the acquired measurement data. Furthermore, during at least one of a first period in which a first representative operation is executed, a second period in which a second representative operation is executed, and a third period between the first period and the second period, the control unit causes the device to execute a predetermined evaluation operation for comprehensively evaluating the operation of the device. Measurement data of the evaluation index measured during execution of the evaluation operation is acquired, the comprehensive evaluation is performed on the basis of the acquired measurement data, and result data indicating an evaluation result of the comprehensive evaluation is output.

Description

Technical Field

[0001] This disclosure relates to information processing methods, information processing devices, and programs. Background Technology

[0002] The method for generating control parameters involved in the background art has been disclosed, for example, in Patent Document 1.

[0003] Previously, the goal was to generate appropriate control parameters.

[0004] Patent Document 1: International Patent Publication No. 2018 / 151215 Summary of the Invention

[0005] The purpose of this disclosure is to provide an information processing method, information processing device, and program that can efficiently generate appropriate control parameters.

[0006] One aspect of this disclosure relates to an information processing method for optimizing the control parameters in a device that performs multiple actions based on multiple control parameters. In the control parameter optimization process, the information processing device selects at least one representative action from all actions executable by the device, causes the device to execute the selected representative action, obtains measurement data of a predetermined evaluation index measured during the execution of the representative action, and updates the control parameters based on the obtained measurement data. The representative action includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output process, during at least one period—a first period during which the first representative action is executed, a second period during which the second representative action is executed, and a third period between the first and second periods—the device executes a predetermined evaluation action for comprehensively evaluating the actions of the device, obtains measurement data of the evaluation index measured during the execution of the evaluation action, performs the comprehensive evaluation based on the obtained measurement data, and outputs result data representing the evaluation result of the comprehensive evaluation. Attached Figure Description

[0007] Figure 1 This is a diagram that concisely illustrates the structure of the control parameter generation apparatus according to the embodiments of this disclosure.

[0008] Figure 2 This is a diagram that briefly represents an example of control parameters.

[0009] Figure 3 This is a diagram that briefly represents an example of a selection rule.

[0010] Figure 4 This is a schematic diagram illustrating an example of the change in the positional deviation of a driven object relative to a target position.

[0011] Figure 5 This is a simplified diagram illustrating an example of the measurement data output by the sensor.

[0012] Figure 6 This is a flowchart representing the processes performed by the information processing department.

[0013] Figure 7 It is a flowchart showing the details of the control parameter optimization process.

[0014] Figure 8 This is a flowchart showing the details of the evaluation result output processing.

[0015] Figure 9 It is a simplified representation of an example of an image containing an evaluation result related to the representative action and the evaluation of the action.

[0016] Figure 10 It is a simplified representation of a variation of an image containing an evaluation result related to the representative action and the evaluation of the action.

[0017] Figure 11 It is an enlarged representation Figure 9 A portion of the image shown.

[0018] Figure 12 This is a magnified representation of part of another example of an image containing the evaluation results. Detailed Implementation

[0019] (The insights that form the basis of this disclosure)

[0020] In recent years, a method has been proposed for generating appropriate control parameters related to the drive unit in a production apparatus. This method involves using a machine learning model or similar tool to search for appropriate control parameters, thereby generating appropriate control parameters. The production apparatus includes a drive unit such as a servo motor that drives an object (see, for example, Patent Document 1).

[0021] Generally speaking, the number of control parameters used in the drive unit of a production device can sometimes reach more than 50. In addition, the adjustment levels of these control parameters can sometimes reach more than 100 levels.

[0022] For example, if a production unit performs 80 actions, the number of control parameters for the drive source is 50, and the adjustment level of the control parameters is 100, the number of combinations of these parameters reaches 100. 50 ×80.

[0023] The inventors discovered the following problem: when generating control parameters for such a large number of combinations, when using methods such as machine learning models to search for appropriate control parameters, due to the excessively large search range, sometimes it takes a lot of time to find appropriate control parameters, or no appropriate control parameters can be found no matter how much time is spent.

[0024] Therefore, in order to realize the following parameter generation method, the inventors have conducted repeated experiments and research. This parameter generation method can efficiently generate appropriate control parameters even when dealing with a large number of combinations of control parameters.

[0025] Through the above experiments and research, the inventors have gained the following insights: Instead of searching for appropriate control parameters from the outset using a vast array of combinations, in the case of 80 actions performed by the production device, as an initial step, appropriate control parameters are searched only for a subset of these 80 actions, such as a single action. In the next step, the number of actions performed by the production device is increased, for example, to combinations related to two actions, and appropriate control parameters are searched using the results from the initial step as a guide. In the next step, the number of actions performed by the production device is further increased, for example, to combinations related to four actions, and appropriate control parameters are searched using the results from the previous step as a guide. These steps are repeated until finally, appropriate control parameters are searched using combinations related to all 80 actions. Thus, by focusing on combinations related to all 80 actions, appropriate control parameters can be generated more efficiently and reliably.

[0026] Subsequently, based on this insight, the inventors conducted further experiments and research, and came up with the following method for generating control parameters as disclosed herein.

[0027] Next, the various methods of this disclosure will be described.

[0028] The information processing method disclosed in the first aspect is used to optimize the control parameters in a device that performs multiple actions based on multiple control parameters. In the control parameter optimization process, the information processing device selects at least one representative action from all actions that the device can perform, causes the device to perform the selected representative action, obtains measurement data of a predetermined evaluation index measured during the execution of the representative action, and updates the control parameters based on the obtained measurement data. The representative action includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output process, during at least one period during the first period when the first representative action is executed, the second period when the second representative action is executed, and a third period between the first and second periods, the device performs a predetermined evaluation action for comprehensively evaluating the actions of the device, obtains measurement data of the evaluation index measured during the execution of the evaluation action, performs the comprehensive evaluation based on the obtained measurement data, and outputs result data representing the evaluation result of the comprehensive evaluation.

[0029] According to the first method, by having the device execute a representative action from all actions to update the control parameters, and by gradually increasing, for example, the number of actions included in the representative action, and repeatedly performing optimization processing, appropriate control parameters can be generated efficiently. Furthermore, during at least one period—during the first period when the first representative action is executed, the second period when the second representative action is executed, and the third period between the first and second periods—the device executes a prescribed evaluation action to perform a comprehensive evaluation and output result data representing the evaluation result of the comprehensive evaluation. Therefore, it is possible to inform the user of the optimization progress of the control parameters in the overall operation of the device during the optimization process.

[0030] The information processing method involved in the second aspect of this disclosure may be: in the first aspect, the evaluation action is all the actions.

[0031] According to the second method, the device is instructed to perform all actions as evaluation actions, thereby enabling an accurate comprehensive evaluation.

[0032] The information processing method involved in the third aspect of this disclosure may be: in the second aspect, in the control parameter optimization processing, based on the execution results of all the actions that have been performed as the evaluation action, the next representative action is selected.

[0033] According to the third method, the next representative action can be appropriately selected based on the execution results of all recent actions.

[0034] The information processing method involved in the fourth aspect of this disclosure may be: in the third aspect, in the control parameter optimization processing, all the actions that have been performed as the evaluation actions are clustered into multiple clusters, and the next representative action is selected from each cluster.

[0035] According to method 4, the representative action for the next time can be appropriately selected.

[0036] The information processing method involved in the fifth aspect of this disclosure may be: in the first aspect, the evaluation action is the differential action between all actions and the most recent representative action of the evaluation action.

[0037] According to the fifth method, the device executes the difference action between all actions and the nearest representative action of the evaluation action as the evaluation action, thereby maintaining the accuracy of the comprehensive evaluation and achieving high efficiency in the evaluation action.

[0038] The information processing method involved in the sixth aspect of this disclosure may be: in the first aspect, the evaluation action is a predetermined action that is pre-defined from all the actions.

[0039] According to the sixth method, the device is instructed to perform a predetermined action from all actions as an evaluation action, thereby enabling the evaluation action to be highly efficient.

[0040] The information processing method involved in the seventh aspect of this disclosure may be: in the first aspect, in the output of the result data, an image representing the evaluation result is generated as the result data, and the display unit displays the generated image.

[0041] According to the seventh method, the display unit displays an image representing the evaluation results, thereby providing the user with an easy-to-understand indication of the progress of the control parameter optimization.

[0042] The information processing method involved in the eighth aspect of this disclosure may be: in the seventh aspect, based on the measurement data obtained as the representative action is performed, calculating the evaluation value of the evaluation index related to the representative action, and based on the measurement data obtained as the evaluation action is performed, calculating the evaluation value of the evaluation index related to the evaluation action, wherein the image includes: time-series data of the evaluation value related to the representative action, and time-series data of the evaluation value related to the evaluation action.

[0043] According to method 8, the image representing the evaluation result includes time-series data of evaluation values ​​related to the representative action and time-series data of evaluation values ​​related to the evaluated action, thereby providing the user with an easy-to-understand indication of the progress of the optimization of the control parameters.

[0044] The information processing method involved in the ninth aspect of this disclosure may be: in the eighth aspect, by extracting the evaluation value corresponding to the representative action from the evaluation value of the evaluation index related to the evaluation action, and adding the time-series data of the evaluation value related to the representative action.

[0045] According to method 9, the evaluation value corresponding to the representative action is extracted from the evaluation value of the evaluation index related to the evaluation action. As a result, the time-series data of the evaluation value related to the representative action can be added, and its characteristics can be prompted to the user in a prolonged or retrospective manner.

[0046] The information processing method involved in the 10th aspect of this disclosure may be: in the 8th aspect, based on the input operation of the selection request made by the user to select an evaluation value from the time-series data of the evaluation values ​​related to the representative action, the control parameters for the next execution of the evaluation action are set according to the control parameters when the representative action corresponding to the evaluation value is executed.

[0047] According to the 10th method, the user can easily set the control parameters for the next evaluation action by performing input operations while observing the screen displayed on the display unit, thus improving the convenience for the user.

[0048] The information processing method involved in the 11th aspect of this disclosure may be: in the first aspect, determining whether the evaluation result of the comprehensive evaluation meets the prescribed termination condition, and if the evaluation result meets the termination condition, ending the optimization processing of the control parameters.

[0049] According to method 11, if the evaluation results of the comprehensive evaluation meet the specified termination conditions, the optimization process will end even before the representative action reaches the final stage, thus shortening the time required for optimization.

[0050] The information processing method involved in the 12th aspect of this disclosure may be: in the 1st aspect, based on the input operation of the execution request made by the user to perform the evaluation action, the device is instructed to perform the evaluation action.

[0051] According to the 12th method, the user can instruct the device to perform the evaluation action by inputting an evaluation action execution request, thereby improving the convenience for the user.

[0052] The information processing method involved in the 13th aspect of this disclosure may be: in the first aspect, based on the measurement data obtained as the representative action is executed, calculating the evaluation value of the evaluation index related to the representative action, calculating the variance of the multiple evaluation values ​​related to the multiple representative actions, and setting the number of times the evaluation action will be executed next based on the variance.

[0053] According to method 13, the number of times the evaluation action to be executed next can be appropriately set based on the variance of multiple evaluation values ​​related to multiple representative actions, thereby improving the evaluation accuracy of the comprehensive evaluation.

[0054] The information processing method involved in the 14th aspect of this disclosure may be: in the first aspect, based on the measurement data obtained as the representative action is performed, calculating the evaluation value of the evaluation index related to the representative action, and based on the multiple evaluation values ​​related to multiple representative actions, setting the control parameters for the next execution of the evaluation action.

[0055] According to method 14, it is possible to appropriately set the control parameters for the next execution of the evaluation action based on multiple evaluation values ​​related to multiple representative actions.

[0056] The information processing method involved in the 15th aspect of this disclosure may be: in the 14th aspect, in setting the control parameters, the control parameters when performing the representative action corresponding to the best evaluation value among the plurality of evaluation values ​​are set as the control parameters when performing the evaluation action next time.

[0057] According to method 15, the control parameters for the next execution of the evaluation action can be set based on the control parameters when executing the representative action corresponding to the best evaluation value among multiple evaluation values, thus improving the evaluation accuracy of the comprehensive evaluation.

[0058] The information processing method disclosed in the 16th aspect is used to optimize the control parameters in a device that performs multiple actions based on multiple control parameters. The information processing device selects at least one representative action from all actions that the device can perform, causes the device to perform the selected representative action, obtains measurement data of a predetermined evaluation index measured during the execution of the representative action, updates the control parameters based on the obtained measurement data, causes the device to perform a predetermined evaluation action for comprehensively evaluating the actions of the device during the execution of the representative action, obtains measurement data of the evaluation index measured during the execution of the evaluation action, performs the comprehensive evaluation based on the obtained measurement data, and selects the next representative action based on the execution result of the evaluation action.

[0059] According to method 16, it is possible to appropriately select the next representative action based on the execution results of the most recent evaluation action.

[0060] The information processing apparatus according to the 17th aspect of this disclosure is used to optimize the control parameters in an apparatus that performs multiple actions based on multiple control parameters. It includes a selection unit, a control unit, an acquisition unit, an update unit, and an output unit. In the control parameter optimization process, the selection unit selects at least one representative action from all actions that the apparatus can perform. The control unit instructs the apparatus to perform the representative action selected by the selection unit. The acquisition unit acquires measurement data of a predetermined evaluation index measured during the execution of the representative action. The update unit updates the control parameters based on the measurement data acquired by the acquisition unit. The representative action includes a first representative action, and... In the evaluation result output processing, and in the second representative action executed after the first representative action, the control unit causes the device to perform a prescribed evaluation action for comprehensively evaluating the operation of the device during at least one period during the first period when the first representative action is executed, during the second period when the second representative action is executed, and during the third period between the first period and the second period. The acquisition unit acquires measurement data of the evaluation index measured during the execution of the evaluation action. The output unit performs the comprehensive evaluation based on the measurement data acquired by the acquisition unit and outputs result data representing the evaluation result of the comprehensive evaluation.

[0061] According to the 17th method, by having the device execute a representative action from all actions to update the control parameters, and by gradually increasing, for example, the number of actions included in the representative action, and repeatedly performing optimization processing, appropriate control parameters can be generated efficiently. Furthermore, during at least one period—during the first period when the first representative action is executed, the second period when the second representative action is executed, and the third period between the first and second periods—the device executes a prescribed evaluation action to perform a comprehensive evaluation and outputs result data representing the evaluation result of the comprehensive evaluation. Therefore, it is possible to prompt the user, midway through the optimization process, the progress of the optimization of the control parameters in the overall operation of the device.

[0062] The procedure involved in the 18th aspect of this disclosure is used to instruct an information processing device to perform selection processing, control processing, acquisition processing, update processing, and output processing. The information processing device is used to optimize the control parameters in a device that performs multiple actions based on multiple control parameters. In the control parameter optimization processing, the selection processing selects at least one representative action from all actions that the device can perform; the control processing instructs the device to perform the representative action selected by the selection processing; the acquisition processing acquires measurement data of a predetermined evaluation index measured during the execution of the representative action; and the update processing updates the control parameters based on the measurement data acquired by the acquisition processing. The evaluation result output process includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output process, the control process causes the device to perform a prescribed evaluation action for comprehensively evaluating the actions of the device during at least one period, during a first period when the first representative action is executed, during a second period when the second representative action is executed, and during a third period between the first and second periods. The acquisition process acquires measurement data of the evaluation index measured during the execution of the evaluation action. The output process performs the comprehensive evaluation based on the measurement data acquired by the acquisition process and outputs result data representing the evaluation result of the comprehensive evaluation.

[0063] According to method 18, by having the device execute representative actions from all actions to update the control parameters, and by gradually increasing, for example, the number of actions included in the representative actions and repeatedly performing optimization processing, appropriate control parameters can be generated efficiently. Furthermore, during at least one period—during the first period when the first representative action is executed, the second period when the second representative action is executed, and the third period between the first and second periods—the device executes a prescribed evaluation action to perform a comprehensive evaluation and outputs result data representing the evaluation result of the comprehensive evaluation. Therefore, it is possible to prompt the user about the optimization progress of the control parameters in the overall operation of the device during the optimization process.

[0064] This disclosure also enables the implementation of programs that cause a computer to execute the characteristic structures contained in the above-described methods or apparatus, or systems that utilize such programs to perform actions. Furthermore, such computer programs can of course be distributed via computer-readable non-transitory storage media such as CD-ROMs (Compact Disc Read-Only Memory) or communication networks such as the Internet.

[0065] (Implementation of this disclosure)

[0066] Hereinafter, embodiments of the present disclosure will be described in detail using the accompanying drawings. It is assumed that elements with the same symbol in different drawings represent the same or corresponding elements. Furthermore, the constituent elements, their arrangement, connection methods, and order of operation shown in the following embodiments are examples and are not intended to limit the present disclosure. The present disclosure is limited only by the claims. Therefore, while it may not be necessary to achieve the objectives of this disclosure through the constituent elements of the following embodiments that are not described in the independent claims representing the highest-level concept of the present disclosure, such constituent elements will be described as constituent elements constituting a more preferred embodiment.

[0067] Figure 1 This is a simplified diagram illustrating the structure of the control parameter generation system according to an embodiment of the present disclosure. The control parameter generation system optimizes control parameters 31 by generating and updating them for the production apparatus 2. The control parameter generation system includes a control parameter generation device 1 and a sensor 3.

[0068] Production device 2 is a device used for production equipment, such as an installation device, processing device, manufacturing device, or handling device for installing, processing, manufacturing, or transporting equipment. Production device 2 is, for example, a production line set up in a factory.

[0069] Production device 2 can perform N actions (N is an integer greater than or equal to 3). N is, for example, 80. Furthermore, this disclosure is not limited to production device 2, but can be applied to any device capable of performing N actions.

[0070] The production apparatus 2 includes a storage unit 41, a control unit 42, a drive unit 43, and a drive object 44.

[0071] The storage unit 41 is composed of HDD (Hard Disk Drive), SSD (Solid State Disk), or semiconductor memory.

[0072] The control unit 42 is composed of processors such as CPU (Central Processing Unit).

[0073] The operation of the drive unit 43 is controlled by the control unit 42, which drives the object 44.

[0074] The drive unit 43 is, for example, a servo motor, a directional flow control valve for controlling the fluid in a pneumatic artificial muscle arm, or a directional flow control valve for controlling the fluid in a hydraulic arm. The servo motor can be, for example, a rotary motor or a linear motor.

[0075] The driven object 44 is an object driven by the drive unit 43. When the drive unit 43 is a servo motor, the driven object 44 is a head that transports the workpiece, or a nozzle mounted on the head for adsorbing the workpiece. Alternatively, when the drive unit 43 is a directional flow control valve, the driven object 44 is a pneumatic or hydraulic artificial muscle arm.

[0076] The control unit 42 controls the drive unit 43 by outputting a command to the drive unit 43 to move the driven object 44 to a predetermined target position. The command output from the control unit 42 to the drive unit 43 may be a position command indicating the position of the drive unit 43 or the driven object 44, or a torque command indicating the torque of the drive unit 43.

[0077] The storage unit 41 stores the control parameters 31 generated and updated by the control parameter generation device 1. The control unit 42 controls the drive unit 43 based on the control parameters 31 read from the storage unit 41. In other words, the control unit 42 uses the control parameters 31 when controlling the drive unit 43. The number of control parameters 31 is, for example, 50.

[0078] The control parameter generation device 1 includes an information processing unit 11, a storage unit 12, an input unit 13, a display unit 14, and a communication unit 15.

[0079] The information processing unit 11 is constructed using a processor such as a CPU. As a function implemented by the processor executing a program read from a non-volatile storage medium such as a computer-readable ROM (Read Only Memory), the information processing unit 11 includes a selection unit 21, a control unit 22, an acquisition unit 23, an update unit 24, and an output unit 25. In other words, the program described above is a program used to enable the information processing unit 11, which is an information processing device mounted on the control parameter generation device 1, to function as the selection unit 21 (selection unit), control unit 22 (control unit), acquisition unit 23 (acquisition unit), update unit 24 (update unit), and output unit 25 (output unit). Details of the processing performed by each processing unit will be described below.

[0080] The storage unit 12 is constructed using HDDs, SSDs, or semiconductor memory. The storage unit 12 stores control parameters 31, selection rules 32, measurement data 33, and inference models 34.

[0081] Figure 2This is a simplified diagram illustrating an example of control parameter 31. Control parameter 31 includes parameters a1 and a2 that adjust the vibration frequency of the driven object 44, parameters b1 and b2 that adjust the speed of the driven object 44, parameters c1 and c2 that adjust the depth of singularities in the vibration characteristics of the driven object 44, and parameters d1 and d2 that adjust the vibration amplitude of the driven object 44. Control parameter 31 may also include, for example, parameters b1, b2, c1, c2, and d1, d2 that have mutually trade-off relationships.

[0082] Figure 3 This diagram is a simplified representation of an example of selection rule 32. Selection rule 32 illustrates the rule by which the selection unit 21 selects at least one representative action from all N actions executable by the production unit 2. The representative action is the action that the production unit 2 executes during control parameter optimization processing by the control parameter generation device 1. Figure 3 In the example shown, the representative action is executed in eight stages, from stage 1 to stage 8. The number of actions X contained in the representative action increases as the stage progresses. For example, the number of actions X in stage 1 is 1, representing action 2. In stage 2, the number of actions X is 2, representing actions 3 and 6. In stage 3, the number of actions X is 4, representing actions 2, 7, 9, and so on. In stage 4, the number of actions X is 8, representing actions 1, 3, 7, 9, and so on. In stage 5, the number of actions X is 16, representing actions 2, 4, 6, 8, 10, and so on. In stage 6, the number of actions X is 32, representing actions 1, 3, 5, 7, 9, 10, and so on. In stage 7, the number of actions X is 64, representing actions 1, 2, 3, 5, 6, 7, 9, 10, and so on. Stage 8 represents all actions from 1 to 50.

[0083] Selection rule 32 is specified in advance by the user through operation input of input unit 13. In addition, it is not limited to user specification. In selection rule 32, the number and actions of representative actions of each stage can be preset according to the prescribed rules, or the number and actions of representative actions of each stage can be dynamically changed according to a specific method specified by the user.

[0084] Additionally, selection rule 32 may also include setting information for the actions executed in the evaluation actions among all N actions executable by production device 2. Evaluation actions are actions executed by control parameter generation device 1 in the evaluation result output processing to comprehensively evaluate the actions of production device 2. In this embodiment, the evaluation actions include all actions. Evaluation actions are executed during at least one period, including the period during which representative actions of each stage are executed and the transition period from one stage to the next. Furthermore, in Figure 3 In the example shown, the representative action of stage 8, which includes all actions, can also be performed as the evaluation action, which includes all actions.

[0085] Furthermore, instead of using selection rule 32, the representative action for the next step can also be selected based on the execution results of all the most recent actions that have been performed as evaluation actions (step SP062). For example, multiple actions can be identified from all actions in order of evaluation value from worst to best, and these multiple actions can be selected as the representative actions for the next step. Alternatively, all actions that have been performed as evaluation actions can be clustered into multiple clusters, and the action with the worst evaluation value in each cluster or an action randomly selected from each cluster can be selected as the representative action for the next step. Thus, the representative action for the next step can be appropriately selected based on the execution results of all the most recent actions that reflect the latest situation.

[0086] The evaluation action is not limited to all actions; it can also be the difference between all actions and the nearest representative action. For example, regarding the period during which the representative actions of stage 4 are executed, the representative actions include actions 1, 3, 7, 9, ... Therefore, the evaluation actions executed during this period include the differences between all actions and the representative actions, namely actions 2, 4, 5, 6, 8, 10, ... Similarly, regarding the period during which the representative actions of stage 5 are executed, the representative actions include actions 2, 4, 6, 8, 10, ... Therefore, the evaluation actions executed during the transition period from stage 5 to stage 6 include the differences between all actions and the representative actions, namely actions 1, 3, 5, 7, 9, ...

[0087] Furthermore, the evaluation action is not limited to all actions, but can also be predetermined actions from all actions. For example, multiple actions from all actions, ordered from largest to smallest impact on the performance of production unit 2, are defined as predetermined actions. In the evaluation action, the control parameter generation device 1 instructs production unit 2 to perform only the predetermined actions.

[0088] Reference Figure 1Measurement data 33 refers to the measured values ​​representing predetermined evaluation indicators obtained by sensor 3 during the execution of representative and evaluation actions. Multiple measurement data 33 are stored in storage unit 12, thereby forming a database. The evaluation indicator can be any indicator capable of quantifying the operation of production device 2; in this embodiment, a setpoint time is used.

[0089] Sensor 3 measures the position of the driven object 44 in the representative action and the evaluation action in chronological order. Then, the measurement data 33, which represents the measurement position corresponding to each action contained in the representative action and the evaluation action, is sent to the control parameter generation device 1.

[0090] Figure 4 This is a schematic diagram illustrating an example of the change in the positional deviation of the driven object 44 relative to the target position during the action of the production device 2 moving the driven object 44 to the target position.

[0091] exist Figure 4 In the diagram, the horizontal axis represents time, and the vertical axis represents the positional deviation of the driven object 44 relative to the target position.

[0092] like Figure 4 As shown in this specification, the allowable range refers to the range in which the positional deviation from the target position is within the required accuracy.

[0093] In addition, such as Figure 4 As shown in this specification, the moment when the target position is reached (hereinafter also referred to as the "setting moment") is the moment when the driven object 44 is last within the allowable range after it has reached the allowable range and has not exceeded the allowable range again.

[0094] In addition, such as Figure 4 As shown, in this specification, the setting time refers to the time from the start of the stop of the drive object 44 based on the instruction to move the drive object 44 to the end of the time until the drive object 44 reaches an allowable position that can be evaluated as having reached the target position; that is, the time from the start of the stop to the setting time. Alternatively, the setting time refers to the time from the start of the movement of the drive object 44 based on the instruction to move the drive object 44 to the end of the time until the drive object 44 reaches an allowable position that can be evaluated as having reached the target position; that is, the time from the start of the movement to the setting time.

[0095] Figure 5 This is a simplified diagram illustrating an example of the measurement data 33 output by sensor 3.

[0096] like Figure 5As shown, measurement data 33 is data that corresponds one-to-one with the elapsed time [ms] after the reference time and the deviation [mm] relative to the target position. Here, the reference time is the stop start time of the drive object 44 based on the command to move the drive object 44 to the target position, or the movement start time of the drive object 44 based on the command to move the drive object 44 to the target position.

[0097] Reference Figure 1 The inference model 34 is a learned model obtained through machine learning for optimizing the control parameters 31. The inference model 34 may also include a machine learning program. Alternatively, a machine learning program may be used instead of the inference model 34 to optimize the control parameters 31. The update unit 24 uses the inference model 34 to perform optimization that shortens the longest settling time among the settling times corresponding to each action included in the representative action. Thus, the control parameters 31 are updated. Furthermore, the optimization object is not limited to the longest settling time; it can be the average of multiple settling times, or multiple settling times or their average, ordered from longest to shortest.

[0098] It can use known algorithms such as Bayesian optimization, evolutionary strategy algorithm (CMA-ES), genetic algorithm (GA), or deep reinforcement learning as the optimization algorithm.

[0099] The information processing unit 11 and / or storage unit 12 may also be installed in an external terminal or server device capable of communicating with the control parameter generation device 1. The external terminal may include a personal computer, smartphone, or tablet computer, etc. The server device may include an edge server or cloud server, etc.

[0100] The input section 13 is constructed using any input device such as a mouse, keyboard, or touchscreen.

[0101] The display unit 14 is constructed using any display device such as a liquid crystal display or an organic EL (electroluminescence) display.

[0102] The communication unit 15 is constructed using a communication module that corresponds to any communication method such as Bluetooth (registered trademark) or Wi-Fi.

[0103] The communication unit 15 sends the control parameters 31 stored in the storage unit 12 to the production unit 2. If the production unit 2 receives the control parameters 31 from the control parameter generation device 1, the storage unit 41 uses the received new control parameters 31 to overwrite the stored control parameters 31. That is, the control parameters 31 are updated.

[0104] In addition, the communication unit 15 sends a control signal to the production unit 2 to execute the representative action or evaluation action. If the production unit 2 receives the control signal from the control parameter generation device 1, the control unit 42 controls the drive unit 43 based on the control signal and the control parameters 31 read from the storage unit 41, thereby executing the representative action or evaluation action.

[0105] Figure 6 This is a flowchart showing the processes performed by the information processing unit 11.

[0106] First, in step SP01, the update unit 24 sets the initial value of the control parameter 31. The initial value can be a predetermined value, a value input by the user using the input unit 13, or a value calculated by the update unit 24 using a calculation method input by the user using the input unit 13. The initial value of the control parameter 31 set by the update unit 24 is stored in the storage unit 12 and sent to the production unit 2 by the communication unit 15. The production unit 2 stores the received initial value of the control parameter 31 in the storage unit 41.

[0107] Next, in step SP02, the selection unit 21 sets the initial value of an integer variable representing the number of actions X. The selection unit 21 sets the initial value of the integer variable to 1 as the number of actions X for the first stage by referring to selection rule 32. Furthermore, although this embodiment describes an example where the initial value of the integer variable is 1, the initial value can be any integer between 1 and N, and is not necessarily limited to 1. Additionally, the initial value of the integer variable can also be specified by the user through operation input using the input unit 13.

[0108] Next, in step SP03, the information processing unit 11 performs control parameter optimization processing.

[0109] Figure 7 It is a flowchart showing the details of the control parameter optimization process.

[0110] First, in step SP031, the selection unit 21 selects X representative actions from all N actions by referring to selection rule 32. Specifically, the selection unit 21 selects action 2 as the representative action for the first stage. Furthermore, as described above, the selection is not limited to using selection rule 32; the next representative action can also be selected based on the execution results of all the most recent actions that have been performed as evaluation actions (step SP062). For example, multiple actions can be determined from all actions in order of evaluation value from worst to best, and these multiple actions can be selected as the next representative actions. Alternatively, all actions that have been performed as evaluation actions can be clustered into multiple clusters, and the action with the worst evaluation value in each cluster or an action randomly selected from each cluster can be selected as the next representative action.

[0111] Next, in step SP032, the control unit 22 generates a control signal that instructs the production device 2 to execute the representative action selected by the selection unit 21. The control signal generated by the control unit 22 is sent to the production device 2 by the communication unit 15. If the production device 2 receives the control signal, the control unit 42 controls the drive unit 43 based on the control signal and the control parameters 31 read from the storage unit 41, thereby executing the representative action comprising X actions. If the production device 2 executes X actions, the sensor 3 sends X measurement data 33 corresponding to the X actions included in the representative action to the control parameter generation device 1. The communication unit 15 receives the X measurement data 33 sent from the sensor 3, stores the X measurement data 33 in the storage unit 12, and inputs it to the information processing unit 11.

[0112] Next, in step SP033, the acquisition unit 23 acquires X measurement data 33 from the communication unit 15, each corresponding to one of the X actions included in the representative action.

[0113] Next, in step SP034, the update unit 24 calculates the evaluation value of the evaluation index for each of the X measurement data 33 acquired by the acquisition unit 23. In this embodiment, the update unit 24 calculates the settling time for each of the X measurement data 33.

[0114] Next, in step SP035, the update unit 24 uses the inference model 34 to perform optimization that shortens the longest settling time among the X settling times related to the X measurement data 33, thereby updating the control parameter 31. The updated control parameter 31 is stored in the storage unit 12 and sent to the production unit 2 by the communication unit 15. The production unit 2 stores the received updated control parameter 31 in the storage unit 41.

[0115] Next, in step SP036, the output unit 25 generates image data containing the evaluation results related to the representative action based on the X settling times related to the X measurement data 33 calculated in step SP034, and outputs the generated image data. For example, the output unit 25 uses the longest settling time among the X settling times as the evaluation value related to the representative action and generates image data containing the evaluation result representing that evaluation value. The image data output from the output unit 25 is input to the display unit 14, and the display unit 14 displays an image based on the input image data.

[0116] In addition, Figure 7The example shown illustrates the selection of a representative action (step SP031) before each execution of the representative action (step SP032), but it is not limited to this. The selection of a representative action can also be performed whenever the representative action is executed a predetermined number of times (e.g., tens to hundreds of times) or at predetermined time intervals. This allows for faster operation compared to the case where the selection of a representative action is performed each time.

[0117] Reference Figure 6 In step SP04, following step SP03, the control unit 22 determines whether the specified representative action termination condition is met. The target value of the evaluation value in each stage can be set as the representative action termination condition, as can the upper limit of the number of iterations of the optimization process in each stage, or the upper limit of the execution time of the optimization process in each stage. For example, if the target value of the evaluation value is set as the representative action termination condition, the control unit 22 determines that the representative action termination condition is not met if the evaluation value of the most recent representative action (the longest settling time among X settling times) exceeds the target value, and determines that the representative action termination condition is met if the evaluation value of the most recent representative action is below the target value.

[0118] If the representative action termination condition is not met (step SP04: No), then in step SP05, the control unit 22 determines whether the prescribed evaluation execution conditions are met. The evaluation execution conditions include the condition that a certain amount of time has elapsed since the reference time. The reference time includes the start time of the execution of the first representative action in each stage and the start time of the execution of the last evaluation action. The control unit 22 has a timer that resets its value at the reference time. When the timer's value reaches or exceeds a threshold, it determines that a certain amount of time has elapsed since the reference time. Additionally, the evaluation execution conditions include the condition that the number of times the representative actions in each stage have been executed has reached a certain number. The control unit 22 has a counter that resets its value before the execution of the first representative action in each stage begins, and increments the counter each time a representative action is executed. When the counter's value reaches or exceeds a threshold, the control unit 22 determines that the number of times the representative actions in each stage have been executed has reached a certain number.

[0119] If the evaluation execution conditions are not met (step SP05: No), the processing after step SP03 is executed.

[0120] If the evaluation execution conditions are met (step SP05: Yes), then in step SP06, the information processing unit 11 performs evaluation result output processing. The processing in step SP06 is equivalent to the evaluation result output processing performed during the period when the representative actions of each stage are executed.

[0121] Figure 8 This is a flowchart showing the details of the evaluation result output processing.

[0122] First, in step SP061, the control unit 22 sets the evaluation action conditions for performing the evaluation action.

[0123] The evaluation action conditions include action specification information, which specifies the action to be performed among all N evaluation actions. The evaluation action can also be all actions. By having production unit 2 perform all actions as the evaluation action, accurate comprehensive evaluation can be performed. Alternatively, the evaluation action can be the difference between all actions and the nearest representative action of that evaluation action. By having production unit 2 perform the difference between all actions and the nearest representative action of that evaluation action as the evaluation action, the accuracy of the comprehensive evaluation can be maintained, and the efficiency of the evaluation action can be improved. Furthermore, the evaluation action can also be a pre-defined action from all actions. By having production unit 2 perform a pre-defined action from all actions as the evaluation action, the efficiency of the evaluation action can be improved.

[0124] Furthermore, the evaluation action conditions include setting information for the value of control parameter 31 when performing the evaluation action. Alternatively, the latest control parameter 31 stored in the storage unit 41 at the current time point can be used directly as the control parameter 31 for performing the evaluation action. Or, the control parameter 31 for the next evaluation action can be set based on multiple evaluation values ​​related to multiple representative actions performed during the period from the execution time of the previous evaluation action to the current time point. For example, the control unit 22 determines the optimal evaluation value from multiple evaluation values ​​related to multiple representative actions performed during the period from the execution time of the previous evaluation action to the current time point, and sets the control parameter 31 set when performing the representative action corresponding to the optimal evaluation value as the control parameter 31 for the next evaluation action. This improves the evaluation accuracy of the comprehensive evaluation performed by performing the evaluation action.

[0125] Next, in step SP062, the control unit 22 generates a control signal that instructs the production unit 2 to perform the evaluation actions set in step SP061. The control signal generated by the control unit 22 is sent to the production unit 2 by the communication unit 15. If the production unit 2 receives the control signal, the control unit 42 controls the drive unit 43 based on the control signal and the control parameters 31 read from the storage unit 41, thereby performing the evaluation actions. If the production unit 2 performs the evaluation actions, the sensor 3 sends measurement data 33 corresponding to each action included in the evaluation actions to the control parameter generation device 1. The communication unit 15 receives the measurement data 33 sent from the sensor 3, stores the measurement data 33 in the storage unit 12, and inputs it to the information processing unit 11.

[0126] Next, in step SP063, the acquisition unit 23 acquires measurement data 33 corresponding to each action included in the evaluation action from the communication unit 15.

[0127] Next, in step SP064, the output unit 25 calculates the evaluation value of the evaluation index for each measurement data 33 acquired by the acquisition unit 23. In this embodiment, the output unit 25 calculates the settling time for each measurement data 33 and takes the longest settling time among the multiple settling times associated with the multiple measurement data 33 as the evaluation value related to this evaluation operation.

[0128] Next, in step SP065, the output unit 25 generates image data containing the evaluation results related to the current evaluation action and outputs the generated image data. The image data output from the output unit 25 is input to the display unit 14, and the display unit 14 displays an image based on the input image data.

[0129] Reference Figure 6 In step SP07, following step SP06, the control unit 22 determines whether the prescribed optimization termination conditions are met. The optimization termination condition can be set as a target value of the evaluation value related to the evaluation action, an upper limit on the total number of iterations of optimization processes related to all stages up to the current time point, or an upper limit on the execution time of optimization processes related to all stages up to the current time point. For example, if the target value of the evaluation value is set as the optimization termination condition, the control unit 22 determines that the optimization termination condition is not met if the evaluation value (maximum settling time) of the most recent evaluation action exceeds the target value, and determines that the optimization termination condition is met if the evaluation value of the most recent evaluation action is below the target value.

[0130] If the optimization termination condition is not met (step SP07: No), the processing after step SP03 is executed.

[0131] If the optimization termination condition is met (step SP07: Yes), the optimization process for control parameter 31 ends. Even if the representative action reaches its final stage, the optimization process ends when the evaluation result of the comprehensive evaluation obtained by performing the evaluation action meets the optimization termination condition, thus shortening the time required for optimization.

[0132] If the representative action termination condition is met in step SP04 (step SP04: Yes), then in step SP08, the selection unit 21 determines whether the number of actions X included in the representative action is equal to the total number of actions N.

[0133] When X equals N (step SP08: Yes), it means the action has reached the final stage (in Figure 3 In the example, this is stage 8, therefore, the optimization process for control parameter 31 ends.

[0134] If X is less than N (step SP08: No), it means the action has not yet reached the final stage. Therefore, in step SP09, the selection unit 21 updates X by referring to selection rule 32. That is, the selection unit 21 transitions the selection of X representative actions from all N actions from the current stage to the next stage.

[0135] Next, in step SP10, the control unit 22 determines whether the prescribed evaluation execution conditions are met. The evaluation execution conditions include the condition that a certain amount of time has elapsed since the reference time. The reference time includes the start time of the previous evaluation action. The control unit 22 has a timer; it resets the timer's value at the reference time, and determines that a certain amount of time has elapsed since the reference time when the timer's value reaches or exceeds a threshold.

[0136] If the evaluation execution conditions are not met (step SP10: No), the processing after step SP03 is executed.

[0137] If the evaluation execution conditions are met (step SP10: Yes), then in step SP11, the information processing unit 11 performs evaluation result output processing. The processing in step SP11 is equivalent to the evaluation result output processing performed during the transition period from one stage to the next. Details of the evaluation result output processing in step SP11 are provided by... Figure 8 The flowchart shown is an example. Alternatively, the following structure can be used: the decision-making process in step SP10 is omitted, and step SP11 is always executed if X is updated.

[0138] Next, in step SP12, the control unit 22 determines whether the prescribed optimization termination condition is met. The details of the optimization termination condition in step SP12 are the same as those in step SP07.

[0139] If the optimization termination condition is not met (step SP12: No), the processing after step SP03 is executed.

[0140] If the optimization termination condition is met (step SP12: Yes), the optimization process for control parameter 31 ends. Even if the representative action reaches its final stage, the optimization process ends when the evaluation result of the comprehensive evaluation obtained by performing the evaluation action meets the optimization termination condition, thus shortening the time required for optimization.

[0141] Figure 9This is a simplified representation of an example image containing evaluation results related to the representative action and the evaluated action. The image includes time-series data of the evaluation value Q related to the representative action and the evaluation value P related to the evaluated action. Figure 9 In the example shown, the evaluation value is the settling time on the vertical axis and the elapsed time on the horizontal axis. Feature K1 is a line graph connecting the optimal values ​​of multiple evaluation values ​​Q associated with multiple representative actions repeatedly executed in the first stage; feature K2 is a line graph connecting the optimal values ​​of multiple evaluation values ​​Q associated with multiple representative actions repeatedly executed in the second stage; and feature K3 is a line graph connecting the optimal values ​​of multiple evaluation values ​​Q associated with multiple representative actions repeatedly executed in the third stage.

[0142] Evaluation value P01 is the evaluation value related to the evaluation action executed at time T02, a certain time elapsed after the execution start time T01 of the initial representative action in Phase 1. Evaluation value P02 is the evaluation value related to the evaluation action executed at time T03 during the transition period from Phase 1 to Phase 2. Evaluation value P03 is the evaluation value related to the evaluation action executed at time T05, a certain time elapsed after the execution start time T04 of the initial representative action in Phase 2. Evaluation values ​​P04 to P07 are the evaluation values ​​related to the evaluation actions executed at times T06 to T09, respectively, a certain time elapsed after times T05 to T08. Evaluation value P08 is the evaluation value related to the evaluation action executed at time T10 during the transition period from Phase 2 to Phase 3. Evaluation value P09 is the evaluation value related to the evaluation action executed at time T12, a certain time elapsed after the execution start time T11 of the initial representative action in Phase 3. The evaluation value P10 is the evaluation value related to the evaluation action performed at time T13, which is a certain time after time T12.

[0143] Additionally, the image includes an icon 50 labeled "Evaluation Execution". If a user clicks on the icon 50 using the input unit 13, a request to execute the evaluation action is sent to the information processing unit 11. After receiving the execution request, the information processing unit 11 performs the same evaluation result output processing as steps SP06 and SP11.

[0144] Figure 10 This is a simplified representation of a variation of an image containing evaluation results related to the representative action and the evaluated action. The characteristic KA is a line graph connecting the optimal values ​​among multiple evaluation values ​​P related to the evaluated action. Figure 10In the example, the evaluation value P04 is worse than the evaluation value P03; therefore, the evaluation value P04 is excluded from characteristic KA. Furthermore, if the evaluation actions performed after time T02 include a representative action corresponding to characteristic K1, by extracting the evaluation value Q corresponding to characteristic K1 from the evaluation values ​​P after time T02, characteristic K1 can be displayed retrospectively after time T02. The same applies to other characteristics K. Moreover, if the evaluation actions performed before time T04 include a representative action corresponding to characteristic K2, by extracting the evaluation value Q corresponding to characteristic K2 from the evaluation values ​​P before time T04, characteristic K2 can be displayed retrospectively before time T04. The same applies to other characteristics K.

[0145] Figure 11 It is an enlarged representation Figure 9 The image shown is a partial diagram. The line graph of characteristic K can also be expanded to depict multiple evaluation values ​​Q associated with multiple representative actions repeatedly performed in each stage. Alternatively, only the depiction of the multiple evaluation values ​​Q can be included, omitting the line graph of characteristic K.

[0146] Users can select an evaluation value from the timing data of the evaluation value Q related to the representative action by using the operation input of the input unit 13. For example, if the user clicks on the evaluation value Q100 on feature K2 (or the time T100 on the horizontal axis), the selection information of the evaluation value Q100 is input to the information processing unit 11. After receiving the selection information, the information processing unit 11 reads the control parameter 31 for executing the representative action corresponding to the evaluation value Q100 from the storage unit 12, and sets the control parameter 31 for the next execution of the evaluation action based on the control parameter 31. For example, the control parameter 31 for the next execution of the evaluation action is set to be equal to the control parameter 31 for executing the representative action corresponding to the evaluation value Q100. Users can easily set the control parameter 31 for the next execution of the evaluation action by performing input operations while observing the screen displayed on the display unit 14, thus improving user convenience.

[0147] Figure 12 This is a magnified view of a portion of another example containing the evaluation results. In the example above, the evaluation action is performed once at each execution time, but the number of times the evaluation action is performed is not limited to once, but can be multiple times. The control unit 22 calculates multiple evaluation values ​​Q( ) related to the multiple representative actions performed during the period from the execution time T05 of the last evaluation action to the current time T06. Figure 11The area indicated by the two dashed lines contains the variance of multiple evaluation values ​​(Q). Based on the calculated variance, the control unit 22 sets the number of times the evaluation action will be executed in the next iteration. For example, the larger the variance, the more times the control unit 22 sets the number of executions; the smaller the variance, the fewer times the control unit 22 sets the number of executions. Figure 11 In the example shown, three evaluation actions are performed at time T06, and the evaluation values ​​P04a, P04b, and P04c associated with each evaluation action are recorded. Therefore, the number of times the next evaluation action will be performed can be appropriately set, thus improving the accuracy of the comprehensive evaluation.

[0148] As explained above, according to this embodiment, the control parameters 31 are updated by having the production device 2 execute a representative action from all N actions, and the optimization process is repeated by gradually increasing the number of actions included in the representative action. Therefore, appropriate control parameters 31 can be generated efficiently.

[0149] Furthermore, according to this embodiment, for example, during at least one period in which the representative action of the first stage (the first representative action) is executed (times T01 to T02), during the second period in which the representative action of the second stage (the second representative action) is executed (times T04 to T09), and during the third period between the first and second periods (times T02 to T04), the production device 2 is instructed to perform a prescribed evaluation action, thereby performing a comprehensive evaluation. The output unit 25 outputs result data representing the evaluation result of the comprehensive evaluation. Therefore, it is possible to prompt the user about the optimization progress of the control parameter 31 in the overall operation of the production device 2 during the middle stage of the optimization process.

[0150] Industrial availability

[0151] This disclosure can be widely applied to systems for generating control parameters, etc.

Claims

1. An information processing method for optimizing the control parameters in a device that performs multiple actions based on multiple control parameters, characterized in that, Information processing device In the control parameter optimization process, at least one representative action is selected from all actions that the device can perform. The device is then instructed to execute the selected representative action, and measurement data of a predetermined evaluation index measured during the execution of the representative action is obtained. Based on the obtained measurement data, the control parameters are updated. The representative action includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output processing, during at least one period—during the first period when the first representative action is executed, during the second period when the second representative action is executed, and during the third period between the first and second periods—the device is instructed to perform a prescribed evaluation action for comprehensively evaluating the actions of the device, measurement data of the evaluation index measured during the execution of the evaluation action is obtained, the comprehensive evaluation is performed based on the obtained measurement data, and result data representing the evaluation result of the comprehensive evaluation is output.

2. The information processing method according to claim 1, characterized in that, The evaluation action refers to all of the actions.

3. The information processing method according to claim 2, characterized in that, In the control parameter optimization process, the next representative action is selected based on the execution results of all actions that have been performed as the evaluation action.

4. The information processing method according to claim 3, characterized in that, In the control parameter optimization process, all actions that have been performed as the evaluation actions are clustered into multiple clusters, and the next representative action is selected from each cluster.

5. The information processing method according to claim 1, characterized in that, The evaluated action is the difference action between all actions and the most recent representative action of the evaluated action.

6. The information processing method according to claim 1, characterized in that, The evaluation action is a pre-defined action from all the actions.

7. The information processing method according to claim 1, characterized in that, In the output of the result data, an image representing the evaluation result is generated as the result data, and the generated image is displayed on the display unit.

8. The information processing method according to claim 7, characterized in that, Based on the measurement data obtained during the execution of the representative action, the evaluation value of the evaluation index related to the representative action is calculated. Based on the measurement data obtained during the execution of the evaluation action, the evaluation value of the evaluation index related to the evaluation action is calculated. The image contains: Time-series data of the evaluation values ​​associated with the representative action, and The time-series data of the evaluation values ​​associated with the evaluation action.

9. The information processing method according to claim 8, characterized in that, By extracting the evaluation value corresponding to the representative action from the evaluation value of the evaluation index related to the evaluation action, the time-series data of the evaluation value related to the representative action is added.

10. The information processing method according to claim 8, characterized in that, The input operation of a selection request made by a user to select an evaluation value from the time-series data of the evaluation values ​​associated with the representative action is a selection request. Based on the control parameters when performing the representative action corresponding to the evaluation value, the control parameters are set for the next execution of the evaluation action.

11. The information processing method according to claim 1, characterized in that, Determine whether the evaluation result of the comprehensive evaluation meets the specified termination condition. If the evaluation result meets the termination condition, end the optimization process of the control parameters.

12. The information processing method according to claim 1, characterized in that, The input operation of the request to perform the evaluation action is based on the reason given by the user. Based on the execution request, the device is instructed to perform the evaluation action.

13. The information processing method according to claim 1, characterized in that, Based on the measurement data obtained during the execution of the representative action, the evaluation value of the evaluation index related to the representative action is calculated. Calculate the variance of the multiple evaluation values ​​associated with the multiple representative actions. Based on the variance, the number of times the evaluation action will be executed next is set.

14. The information processing method according to claim 1, characterized in that, Based on the measurement data obtained during the execution of the representative action, the evaluation value of the evaluation index related to the representative action is calculated. Based on the multiple evaluation values ​​associated with the multiple representative actions, the control parameters are set for the next execution of the evaluation action.

15. The information processing method according to claim 14, characterized in that, In setting the control parameters, the control parameters for performing the representative action corresponding to the best evaluation value among the multiple evaluation values ​​are set as the control parameters for performing the evaluation action the next time.

16. An information processing method for optimizing the control parameters in a device that performs multiple actions based on multiple control parameters, characterized in that, Information processing device Select at least one representative action from all actions that the device can perform, and have the device perform the selected representative action. Obtain measurement data of a predetermined evaluation index measured during the execution of the representative action. Update the control parameters based on the obtained measurement data. During the period in which the representative action is performed, the device is instructed to perform a prescribed evaluation action for comprehensively evaluating the action of the device, and measurement data of the evaluation index measured during the execution of the evaluation action are obtained. The comprehensive evaluation is then performed based on the obtained measurement data. Based on the results of the evaluation action, the next representative action is selected.

17. An information processing apparatus for optimizing the control parameters in an apparatus for performing multiple actions based on multiple control parameters, characterized in that, It includes a selection unit, a control unit, an acquisition unit, an update unit, and an output unit. In the control parameter optimization process, the selection unit selects at least one representative action from all actions that the device can perform, the control unit instructs the device to execute the representative action selected by the selection unit, the acquisition unit acquires measurement data of a predetermined evaluation index measured during the execution of the representative action, and the updating unit updates the control parameters based on the measurement data acquired by the acquisition unit. The representative action includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output processing, the control unit causes the device to perform a prescribed evaluation action for comprehensively evaluating the operation of the device during at least one period, during the first period when the first representative action is executed, during the second period when the second representative action is executed, and during the third period between the first period and the second period. The acquisition unit acquires measurement data of the evaluation index measured during the execution of the evaluation action. The output unit performs the comprehensive evaluation based on the measurement data acquired by the acquisition unit and outputs result data representing the evaluation result of the comprehensive evaluation.

18. A program for instructing an information processing device to perform selection processing, control processing, acquisition processing, update processing, and output processing, said information processing device being used to optimize said control parameters in a device that performs multiple actions based on multiple control parameters, the program being characterized in that... In the control parameter optimization process, the selection process selects at least one representative action from all actions that the device can perform; the control process instructs the device to execute the representative action selected by the selection process; the acquisition process acquires measurement data of a predetermined evaluation index measured during the execution of the representative action; and the update process updates the control parameters based on the measurement data acquired by the acquisition process. The representative action includes a first representative action and a second representative action executed after the first representative action. In the evaluation result output processing, the control processing causes the device to perform a prescribed evaluation action for comprehensively evaluating the actions of the device during at least one period, during the first period when the first representative action is executed, during the second period when the second representative action is executed, and during the third period between the first period and the second period. The acquisition processing acquires measurement data of the evaluation index measured during the execution of the evaluation action. The output processing performs the comprehensive evaluation based on the measurement data acquired by the acquisition unit and outputs result data representing the evaluation result of the comprehensive evaluation.

Images