Support device for introducing robot control ai

Abstract

To support rental of a machine learning model or the like for controlling a robot by a fixed amount contract.SOLUTION: The robot control AI introduction assisting device 1 according to the present invention includes a flat-rate contract managing part 111 that manages a rental period and a payment status of a flat-rate contract related to rental of a plurality of robots 2, an operation data-acquiring part 112 that acquires operation data-of an acquisition-source robot when payment of the flat-rate contract related to the acquisition-source robot included in the plurality of robots 2 has been made, a machine-learning part 113 that performs machine-learning of a control machine-learning model used for control of the robot 2 using learning data-based on the operation data-and a machine-learning model-providing part 114 that provides the renting robot with the control machine-learning model machine-learned by the machine-learning part 113 during the rental period of the flat-rate contract related to the renting robot included in the plurality of robots 2.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a support device for introducing robot control AI. [Background technology]

[0002] Robots provide a variety of functions through the cooperation of sensors and mechanical drives. Advances in robotics have led to the development of a variety of robots for manufacturing, retail, welfare, medical care, logistics, hospitality, agriculture, construction, security, and other fields. However, the economic burden of deploying advanced robots, which require sophisticated control, has hindered their adoption. The economic burden of deploying robots can be reduced by renting them under a fixed-price contract instead of purchasing them.

[0003] Regarding technologies for supporting the rental of advanced equipment such as robots in conventional services and conventional technologies, Patent Document 1 discloses a sharing system that enables multiple users to share equipment that is under the management of others and cannot be used without usage authorization, in which the system receives and records rental conditions specified in advance from the owner of the equipment, receives a loan application from the user that includes the specified borrowing conditions, extracts the equipment that matches the borrowing conditions and is available for rental, and transfers or grants usage authorization for the equipment from the owner to the user.

[0004] The technology described in Patent Document 1 makes it possible to mutually lend and borrow equipment such as drones owned and managed by multiple owners as needed, without the need for a central management entity. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2018-132794 Summary of the Invention [Problem to be solved by the invention]

[0006] Meanwhile, robots achieve desired functions by controlling their hardware with software that coordinates sensors and mechanical drive units. The technology described in Patent Document 1 merely reduces the economic burden associated with introducing hardware by mutually lending and borrowing robot hardware between suitable parties as needed. Therefore, the technology described in Patent Document 1 has room for further improvement in that it also allows artificial intelligence (AI) that controls robots that provide advanced functions to be rented under a fixed-price contract, thereby reducing the burden associated with its introduction and renewal.

[0007] The objective of this invention is to support the rental of machine learning models that control robots under fixed-price contracts in order to reduce the financial burden associated with the introduction of AI that controls robots that provide advanced functions. [Means for solving the problem]

[0008] As a result of intensive research into solving the above-mentioned problems, the inventors have found that the above-mentioned object can be achieved by a device that manages the provision of machine learning models and performs processing such as improving the machine learning models using operational data provided by the robots to which the models are loaned, so that artificial intelligence (AI) that performs advanced control over robots can be provided on a subscription basis. As a result, the inventors have completed the present invention.

[0009] One aspect of the present invention provides a robot control AI introduction support device comprising: a fixed-price contract management unit that manages data regarding the rental period and payment status of fixed-price contracts for the rental of multiple robots; an operation data acquisition unit that acquires operation data of an acquiring robot included in the multiple robots when payment for the fixed-price contract for the robot has been made; a machine learning unit that performs machine learning of a control machine learning model used to control the robot using learning data based on the operation data; and a machine learning model provision unit that provides a rental robot with a control machine learning model that has been machine-learned by the machine learning unit when the rental period for the fixed-price contract for a rental robot included in the multiple robots is in progress; wherein the operation data acquisition unit acquires operation data including sensor data and control data of the acquiring robot, the machine learning unit performs machine learning using learning data that includes the sensor data as an explanatory variable and the control data as a target variable, and the machine learning model provision unit provides a control machine learning model that has been machine-learned using learning data based on multiple operation data including operation data acquired from other robots different from the single rental robot.

[0010] By using machine learning, software that controls robots can provide more advanced functions. However, providing machine learning models requires significant development costs for designing the machine learning model, collecting training data, and performing machine learning using the training data. Such high development costs increase the financial burden of purchasing machine learning models and may hinder the widespread adoption of machine learning models.

[0011] On the other hand, in a fixed-price contract, also known as a subscription contract, the payment amount per period is generally smaller than that of a one-time purchase. In this aspect of the present invention, a fixed-price contract management unit manages the rental period and payment status. As part of the above management, in this aspect, a machine learning model providing unit provides a control machine learning model used to control the robot to the robot for which payment has been made, based on the above-mentioned payment status. Therefore, this aspect can appropriately provide a control machine learning model despite being a fixed-price contract that reduces the financial burden associated with providing the robot and the control machine learning model.

[0012] However, when a robot providing advanced functions is provided in various forms, such as through a fixed-price contract or purchase, the operator of the robot may be concerned that operational data of the robot outside the rental period, such as that of a purchased robot, may also be passed on to the provider.

[0013] As part of the management described above, in this aspect, the operation data acquisition unit acquires operation data, including robot sensor data and control data, from the robot rented under the fixed-price contract based on the rental period described above. When the payment for the fixed-price contract has been made and the rental period is within, the machine learning model provision unit has provided the control machine learning model to the robot from which the data was acquired, allowing the operation data acquisition unit to acquire appropriate operation data.

[0014] In this aspect, the machine learning unit performs machine learning on the control machine learning model based on the acquired operational data, thereby improving its performance, etc. By performing this series of processes, this aspect can both eliminate the above-mentioned concerns and improve the machine learning model using learning data acquired from the loaned robot.

[0015] Furthermore, this aspect reduces the economic burden associated with using robots controlled by machine learning models through the above-described process. This helps promote the widespread adoption of robots controlled by machine learning models. Furthermore, this aspect improves the machine learning model to provide better control by using operational data collected from widespread robots in this way instead of training data, which is typically obtained at great expense. Thus, this aspect of the present invention can reduce the economic burden associated with introducing software to control robots that provide advanced functions by enabling the rental of machine learning models that control robots under a fixed-fee contract.

[0016] In addition, the present invention can take various forms, exemplified by an aspect in which operational data associated with a region is used to improve a machine learning model according to the region, an aspect in which frequently used command items are extracted and provided based on the operational data, an aspect in which the linguistic expressions used by a robot to a user are improved based on the operational data, an aspect in which an instruction manual is generated based on the operational data, etc. These various forms, each with their own unique configuration, contribute to reducing the burden associated with installing and updating software that controls a robot that provides advanced functions. [Effects of the Invention]

[0017] As described above, the present invention can help reduce the financial burden associated with introducing AI to control robots that provide advanced functions by enabling machine learning models that control robots to be rented under a fixed-price contract. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is a block diagram showing an example of the hardware and software configuration of a system S according to this embodiment. [Figure 2] FIG. 2 is an example of the fixed-price contract database 131. [Figure 3] FIG. 3 is an example of the robot database 132. [Figure 4] FIG. 4 is an example of the robot 2. [Figure 5] FIG. 5 is a main flowchart showing an example of a preferable flow of the support process of this embodiment. [Figure 6] FIG. 6 is a continuation of the previous figure. [Figure 7] FIG. 7 is a continuation of the previous figure. [Figure 8] FIG. 8 is a continuation of the previous figure. [Figure 9] FIG. 9 is a continuation of the previous figure. DETAILED DESCRIPTION OF THE INVENTION

[0019] First, although the following disclosure, diagrams, and / or claims may be described as being presented alone or in combination with one or more other aspects, the subject matter of the immediate disclosure is not intended to be so limited. That is, the immediate disclosure, diagrams, and claims are intended to encompass the various aspects described herein, each alone or in one or more combinations with each other. For example, even if the immediate disclosure describes and illustrates a first, second, and third embodiment in such a way that the first embodiment is described and illustrated specifically in conjunction with the second embodiment, or the second embodiment is described and illustrated only in conjunction with the third embodiment, the immediate disclosure and illustrations are not so limited and may include only the first embodiment, only the second embodiment, only the third embodiment, or one or more combinations of the first, second, and / or third embodiments, such as the first and second embodiments, the first and third embodiments, the second and third embodiments, or the first, second, and third embodiments.

[0020] The use of the phrase "or" in this document shall mean a "non-exclusive" arrangement unless expressly specified otherwise. For example, when we say "item x is A or B," we mean either: (1) item x is either A or B, but not both; or (2) item x is both A and B. In other words, the word "or" is not used to define an "exclusive" arrangement.

[0021] Additionally, the phrases "comprising at least one of" and "comprising at least one of the following," when used in conjunction with a system or element, mean that the system or element includes one or more of the elements listed after the phrase. For example, if there are three types of elements, element 1 through element 3, the phrases "comprising at least one of" and "comprising at least one of the following" are to be interpreted as any of the following structural arrangements: a device including the first element, a device including the second element, a device including the third element, a device including the first and second elements, a device including the first and third elements, a device including the second and third elements, or a device including the first, second, and third elements.

[0022] A similar interpretation is intended when the phrase "used in at least one of the following" is used in this context. Furthermore, as used in this context, "and / or" is used as a verbal conjunction to indicate that one or more of the listed elements or conditions are included or occur. For example, a device including a first element, a second element, and / or a third element is to be interpreted as any of the following structural arrangements: a device including the first element, a device including the second element, a device including the third element, a device including the first element and the second element, a device including the first element and the third element, a device including the second element and the third element, or a device including the first element, the second element, and the third element.

[0023] In addition, the use of the phrase "and / or" in the text means a "non-exclusive" agreement, as stipulated in the Japanese Industrial Standards (JIS) "Format and preparation method of standard sheets JIS Z 8301."

[0024] Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings.

[0025] <System S> FIG. 1 is a block diagram showing an example of the hardware configuration and software configuration of a system S of this embodiment. The system S of this embodiment is a support system for introducing a robot-controlling AI. The system S is configured to include at least a support device 1 for introducing a robot-controlling AI. The support device 1 is preferably configured to communicate with a robot 2 via a network N. The support device 1 is also preferably configured to communicate with a terminal T used by an operator of the robot 2, etc., via the network N.

[0026] [Support device 1] The support device 1 includes a control unit 11, a storage unit 13, and a communication unit 14. The type of the support device 1 is not particularly limited, and may be, for example, a server device, a cloud server, or the like.

[0027] [Control unit 11] The control unit 11 includes a central processing unit (CPU), a random access memory (RAM), a read only memory (ROM), and the like.

[0028] The control unit 11 cooperates with at least one of the storage unit 13 and the communication unit 14 as necessary. The control unit 11 then realizes the software components of the program of this embodiment executed by the support device 1. Specifically, the control unit 11 includes a fixed-price contract management unit 111, an operational data acquisition unit 112, a machine learning unit 113, a machine learning model provision unit 114, a frequently used item extraction unit 115, a frequently used item provision unit 116, an instruction manual generation unit 117, and the like.

[0029] [Storage section 13] The memory unit 13 is a device for storing data and / or files, and has a storage unit for non-temporarily storing data using a hard disk, semiconductor memory, recording medium, memory card, etc. The memory unit 13 stores programs executed by the microcomputer, a fixed-price contract database 131, a robot database 132, and other data.

[0030] (Fixed-price contract database 131) The fixed-price contract database 131 is a database that stores data (fixed-price contract information) related to the rental of robots based on fixed-price contracts. The data stored in the fixed-price contract database 131 is managed by the fixed-price contract management unit 111. This data includes at least data related to the rental period and payment status of the fixed-price contract. In order to link with the robot database 132 described below, this data preferably includes information that identifies the robot related to the fixed-price contract (for example, a robot ID).

[0031] For efficient data management, the data is preferably stored in association with information identifying the fixed-price contract (e.g., a contract ID). For managing fixed-price contracts for each customer, the data preferably includes customer name, customer address, customer contact information, or customer representative.

[0032] 2 is an example of the fixed-price contract database 131. In this example, fixed-price contract information relating to contract IDs "C0001," "C0002," "C0101," etc. is shown as data stored in the fixed-price contract database 131.

[0033] The fixed-price contract information for contract ID "C0001" includes data such as customer name "△△ Kogyo Tokyo Office," contract period "20△△ year △△ month - 20△△ year △△ month," payment status "20△△ year △△ month unpaid," and robot ID "R0001." This information indicates a fixed-price contract with some unpaid amounts.

[0034] The fixed-price contract information for contract ID "C0002" includes data such as customer name "△△ Kogyo Osaka Office," contract period "20△△ year △△ month - 20△△ year △△ month," payment status "paid up to 20△△ year △△ month," and robot ID "R0002." This information indicates a fixed-price contract for a corporation in the same industry that belongs to the contractor associated with the fixed-price contract information for contract ID "C0001," but at a business location in a different region.

[0035] The fixed-price contract information for contract ID "C0101" includes data such as customer name "Osaka △△ Medical Corporation," contract period "20△△ year △△ month - 20△△ year △△ month," payment status "paid up to 20△△ year △△ month," and robot IDs "R0101" and "R0102." This information indicates a fixed-price contract for a corporation in a different industry that is located in the same region as the fixed-price contract information for contract ID "C0002."

[0036] By storing this data in the example, the assistance device 1 can extract fixed-price contracts related to contractors, industries, regions, and robots that have the same or similar relationship to a certain fixed-price contract. This allows the assistance device 1 to identify the rental robot for which a control machine learning model is to be provided and operation data related to the contractor, industry, region, or robot based on the extracted fixed-price contract, and use this data for machine learning in the control machine learning model (described below) to be provided to the rental robot. Therefore, by using operation data related to the fixed-price contract, the assistance device 1 can reduce the financial burden associated with introducing software to control the robot, while providing a control machine learning model that can achieve advanced functions through high-quality machine learning that uses operation data related to the rental robot as learning data.

[0037] Furthermore, by storing these data in this example, when a fixed-price contract payment is made, the assistance device 1 acquires operation data from the source robot related to the fixed-price contract. Then, when the fixed-price contract rental period is in progress, the assistance device 1 can provide the control machine learning model to the rental robot related to the fixed-price contract.

[0038] (Robot Database 132) The robot database 132 is a database that stores data (robot information) on robots rented under fixed-price contracts. To ensure efficient data management in conjunction with the above-mentioned fixed-price contract database 131, the data is preferably stored in association with information identifying the robot (e.g., a robot ID). The data includes at least the robot's operational data and a control machine learning model used to control the robot. To collect operational data for each type of robot and provide a control machine learning model, the data preferably includes information identifying the type of robot.

[0039] (Operational Data) In order to contribute to machine learning for a control machine learning model that controls a robot in which sensors and mechanical drivers work together to achieve advanced functions, the operation data preferably includes at least sensor data and control data of the source robot from which the operation data is obtained. In order to optimize the control machine learning model according to region, the operation data preferably includes data indicating the region in which the source robot is operated. In order to perform machine learning for each command given by a user, the operation data preferably includes data indicating the command item. In order to reflect changes in the operation data over time in the control machine learning model, the operation data is preferably stored in association with information that specifies time (e.g., date and time).

[0040] (sensor data) The sensor data is not particularly limited. The sensor data includes, for example, data on the rotational position of a servo motor, data on the load applied to the driving unit 25, data on a weight sensor, data on an optical sensor (such as a camera), data on a vibration sensor, data on an acoustic sensor (such as a microphone), data on a temperature sensor, data on an obstacle detection device (such as a radar, sonar, or proximity sensor), data on the current position measured by a positioning system, or data obtained from a device that acquires biometric information. If the robot 2 includes an input unit 26 (described below), the sensor data may include data obtained from the input device.

[0041] (control data) The control data includes, for example, data related to control of a servo motor (e.g., control signal), data related to control of an engine, data related to control of a steering mechanism, data related to control of a motor (e.g., voltage or frequency), data related to control of a vibration drive device, data related to hydraulic control, data related to water pressure control, data related to air pressure control, etc. If the robot 2 is equipped with a display unit 27 (described below), the control data may include data related to display control.

[0042] For example, if the robot is a barista robot, the sensor data preferably includes data on the rotational position of a servo motor provided on a robot arm that operates various devices related to serving beverages, data on an optical sensor (such as a camera) that reads containers and operation panels of various devices located around the tip of the robot arm, and data on a device that detects obstacles around the robot arm. This allows the assistance device 1 to perform machine learning so that, for example, the complex relationship between the rotational position of the servo motor and the optical sensor data on the containers or various devices is reflected in the mechanical drive of the robot arm using the servo motor.

[0043] Furthermore, if the robot is a barista robot that uses multiple servo motors to move its robot arm, the control data preferably includes data related to the control of the servo motors. By using this data as learning data together with the sensor data related to the barista robot described above, the assistance device 1 can train a control machine learning model to learn appropriate control for moving the robot arm in a given situation, based on learning data including, for example, optical sensor data indicating a situation in which "if the robot arm were to be extended a little further it would be in a position to grasp the container," sensor data indicating the "rotational position of the servo motor that would allow the robot arm to be extended a little further," and data related to the control of the servo motor that successfully moved the robot arm to a position in which it could grasp the container at that time.

[0044] The control machine learning model is a machine learning model used to control a robot. The control machine learning model is subjected to machine learning in the machine learning unit 113 described below using learning data that includes at least sensor data as explanatory variables and control data as objective variables. In order to contribute to the control of a robot that provides multiple functions in response to commands, it is preferable that the control machine learning model is subjected to machine learning using learning data that further includes the above-mentioned command items as explanatory variables. In order to perform optimization according to a region, it is preferable that the control machine learning model is subjected to machine learning using learning data that further includes the above-mentioned data indicating the region as explanatory variables.

[0045] (Language model for linguistic communication) In order to provide a machine learning model related to linguistic communication performed by a robot, it is preferable that the robot information further includes a language model for generating linguistic expressions to be output from the aforementioned source robot to a user of the source robot. In this case, in order to optimize the linguistic expressions, it is preferable that the robot operation data further include the linguistic expressions output from the aforementioned source robot to a user of the source robot, situations related to the linguistic expressions, and the user's evaluation of the linguistic expressions.

[0046] The situation relating to the linguistic expression includes a situation in which a specific command is input to the robot and a specific operating situation of the robot. The language model is configured to output a linguistic expression that is expected to be highly evaluated by the user when data indicating the situation is input.

[0047] The language model is preferably configured to output a linguistic expression corresponding to a specific command when the robot receives the command. An example of the command and linguistic expression is when a command to serve an espresso is input and the linguistic expression corresponding to the command is "Espresso, right? Understood. Please press the button to select single or double."

[0048] Preferably, the language model is configured to output a linguistic expression corresponding to a specific operating situation when the robot reaches the specific operating situation. An example of the command and linguistic expression is when, when the robot reaches the operating situation where extraction of espresso is completed, the linguistic expression corresponding to the operating situation is output, such as "Espresso is ready. Please be careful when drinking it as it is hot. Thank you for using our service."

[0049] In order to generate language expressions that can be evaluated by users depending on the situation, it is preferable that the language model has been pre-trained using training data that includes the situation and the user's evaluation of the generated language expression as explanatory variables, and that includes the language expression corresponding to the situation as a target variable.

[0050] In order to generate language expressions that will be evaluated by users in a specific region, it is preferable that the language model has been pre-trained using training data that includes the region, situation, and the user's evaluation of the generated language expression as explanatory variables, and that includes the language expression corresponding to the region and situation as a target variable.

[0051] The data related to the language model may be stored in an external server in addition to being stored in the storage unit 13. A language model in which data is stored in an external server provides generated text or the like to the assistance device 1 via, for example, an API.

[0052] (Language model for generating instructions) In order to provide a machine learning model related to a robot instruction manual, the robot information preferably further includes a language model that generates an instruction manual based on operation data. The language model is configured to generate an instruction manual indicating how to use the robot when a command item for the robot and operation data associated with the command item are input. In this case, the operation data acquired by the operation data acquisition unit 112 includes sensor data and control data associated with the command item for the robot.

[0053] In order to generate instructions that will be evaluated by users, it is preferable that the language model has been pre-trained using training data that includes command items, operational data, and the user's evaluation of the generated instructions as explanatory variables, and that includes instructions corresponding to the command items and operational data as objective variables.

[0054] 3 is an example of the robot database 132. In this example, robot information related to robot IDs "R0001," "R0002," "R0101," "R0102," etc. is shown as data stored in the robot database 132.

[0055] The robot information associated with robot ID "R0001" includes the robot type "barista robot" and region "Tokyo," as well as operational data including "date and time, command items, sensors, mechanical drive," etc., a control machine learning model, and a language model (not shown). In other words, the robot information includes operational data for when a barista robot operated in Tokyo serves coffee, and a control machine learning model for controlling the barista robot.

[0056] The robot information associated with robot ID "R0002" and the robot information associated with robot ID "R0101" both include the robot type "barista robot" and region "Osaka," operational data including "date and time, command items, sensors, mechanical drive," etc., and a control machine learning model and language model (not shown). In other words, this robot information is data about multiple robots that are the same type as the robot information associated with robot ID "R0001" but are operated in different regions.

[0057] The robot information associated with robot ID "R0102" includes the robot type "medical robot" and region "Osaka," operational data including "date and time, command items, sensors, mechanical drive," etc., and a control machine learning model and language model (not shown). In other words, this robot information is data about a robot that is in the same region as the robot information associated with robot ID "R0002" and the robot information associated with robot ID "R0101," but is of a different type.

[0058] By storing this data in this example, the support device 1 can provide, for example, a barista robot with robot ID "R0001" with a control machine learning model in which machine learning has been performed based on the operational data of other barista robots with robot IDs "R0002" and "R0101." Furthermore, the support device 1 can provide, for example, a barista robot with robot ID "R0002" with a control machine learning model in which machine learning aimed at regional optimization has been performed based on the operational data of robot ID "R0101" of a barista robot of the same type operated in the same region.

[0059] [Communications Section 14] The communication unit 14 is not particularly limited as long as it includes a component that connects the support device 1 to the network N and enables communication. Examples of such a component include a network card that complies with the Ethernet standard and a communication device that complies with wireless LAN.

[0060] [Robot 2] The robot 2 includes a control unit 21, a memory unit 22, a communication unit 23, a sensor 24, and a drive unit 25. The robot 2 is a mechanical device realized by the control unit 21, the memory unit 22, etc. working together, and controls the drive unit 25 based on data from the sensor 24 using software components that utilize a control machine learning model provided from the assistance device 1 via the communication unit 23.

[0061] The robot 2 is not particularly limited as long as it is an industrial robot equipped with a sensor 24 and a drive unit 25. Examples of such robots 2 include medical robots, agricultural robots, mobile robots, self-driving robots, inspection robots, household robots, engineering robots, security robots, service industry robots, logistics sorting support robots, aging society support robots, daily life support robots, fishing support robots, disaster relief support robots, and construction support robots.

[0062] Service industry robots include, for example, barista robots equipped with robotic arms that operate various devices to serve drinks. Drone robots include, for example, transport robots that control rotors to transport cargo based on wind direction and speed data and position information obtained from sensors 24. Agricultural robots include, for example, harvesting robots that control a movement mechanism and a robotic arm for harvesting fruit based on the status of a field or orchard acquired by sensors 24.

[0063] The robot 2 may also be an automatic processing device whose main function is to process information on data from the sensor 24, etc. Examples of automatic processing devices whose main function is to process information include emotion recognition robots, marketing robots, health management robots, sales robots, personnel evaluation robots, education robots, livestock support robots, aquaculture support robots, dairy farming support robots, entertainment support robots, management consulting robots, and management strategy planning support robots.

[0064] In the following, we will mainly explain examples where the robot 2 is a barista robot (an example of a service industry robot), a transport robot (an example of a drone robot), and a harvesting robot (an example of an agricultural robot), but a person skilled in the art will be able to implement the system S of this embodiment for other types of robot 2 based on these examples.

[0065] The robot 2 includes a control unit 21, a memory unit 22, and a communication unit 23 as components that realize software components related to the control of the drive unit 25, the display unit 27, etc. The robot 2 preferably includes a sensor 24 to acquire information related to the operation of the robot 2. The robot 2 preferably includes a drive unit 25 to provide various functions through mechanical operation. The robot 2 preferably includes an input unit 26 to acquire input from the user. The robot 2 preferably includes a display unit 27 to convey information to the user.

[0066] [Control unit 21] The hardware configuration of the control unit 21 may be the same as that of the control unit 11. The control unit 21 cooperates with other members as necessary to realize various software components that control the robot 2. The control unit 21 then executes processing to control the drive unit 25 and the like by processing using a control machine learning model and the like provided by the assistance device 1 based on data acquired from the sensor 24 and the like.

[0067] [Storage section 22] The hardware configuration of the storage unit 22 may be the same as that of the storage unit 13. The storage unit 22 stores a control machine learning model provided by the assistance device 1, a program for controlling the robot 2, and the like.

[0068] [Communications Department 23] The hardware configuration of the communication unit 23 may be the same as that of the communication unit 14. The communication unit 23 connects the robot 2 to the network N to enable communication with the support device 1 and the like.

[0069] [Sensor 24] The sensors 24 include various devices that acquire information related to the operation of the robot 2 and output it as sensor data. The sensors 24 include, for example, a sensor that acquires the rotational position of a servo motor, a sensor that acquires the load applied to the drive unit 25, etc., a weight sensor, an optical sensor (such as a camera), a vibration sensor, an acoustic sensor (such as a microphone), a temperature sensor, a device that detects obstacles (such as a radar, sonar, or proximity sensor), a positioning system (such as a global positioning system (GPS) or a global navigation satellite system (GNSS)), or a device that acquires biometric information. The sensor data may include either a digital signal or an analog signal. When the sensor data includes an analog signal, the control unit 21 preferably processes the analog signal to convert it into digital information.

[0070] [Driver 25] The driving unit 25 is a component that mechanically moves the robot 2. The driving unit 25 is controlled by control data generated by processing using the above-mentioned control machine learning model or the like. The control data includes digital signals or analog signals. The type of the driving unit 25 is not particularly limited. The type of the driving unit 25 includes, for example, a servo motor, an engine, a steering mechanism, a motor, a vibration driving device, a hydraulic control device, a water pressure control device, or a pneumatic control device.

[0071] A servo motor is a motor that moves, for example, a robot arm, an opening / closing section, a conveyor, a cropper, or a sensor dome in a commanded direction. An engine is an internal combustion engine or an external combustion engine that generates rotational motion that is transmitted to drive wheels, rotors, or other rotating members. A steering mechanism is a device that changes the direction of a car's wheels, a ship's rudder, or an aircraft's rotors. A motor is an electric motor that rotates drive wheels, rotors, fans, pulleys, or gears. A vibration drive device is a device that rotates or moves using vibration. A hydraulic control device is a device that uses hydraulic pressure to move, for example, a robot arm, an opening / closing section, a conveyor, a cropper, or a sensor dome. A water pressure control device is similar to a hydraulic control device, except that it uses water pressure instead of hydraulic pressure. A pneumatic control device is similar to a hydraulic control device, except that it uses air pressure instead of hydraulic pressure.

[0072] [Input section 26] The input unit 26 is not particularly limited as long as it is a component that converts input from a user into an electrical signal and transmits it to the control unit 21, etc. The type of the input unit 26 may be, for example, a touch panel, a keyboard, various pointing devices, a switch, a microphone, or a camera.

[0073] [Display section 27] The display unit 27 is not particularly limited as long as it is a component that conveys information to the user. The type of the display unit 27 is, for example, a liquid crystal display (LCD), an organic electroluminescence display (OLED), electronic paper, or a micro light-emitting diode display (micro LED). Note that in this embodiment, a component that conveys information by non-optical means, such as a speaker, is also treated as providing the same function as the display unit 27.

[0074] [Robot 2, the barista robot] 4 is an example of the robot 2. This example shows the robot 2, which is a barista robot, and peripheral devices operated by the robot 2.

[0075] In this example, robot 2, which is a barista robot, is equipped with a control unit 21, a memory unit 22, a communication unit 23, a sensor 24 which is a camera that captures images of the area around the robot arm, multiple drive units 25 which are servo motors that move each rotating part of the robot arm in a specified direction, and an LCD display with a touch panel that also serves as an input unit 26 and a display unit 27.

[0076] The robot 2, which is a barista robot, moves its robot arm by using the various components described above in cooperation to operate the mill M that grinds the coffee beans, the water server W that provides water and ice, and the espresso machine E that brews the coffee. In addition, the robot arm of the robot 2 moves a serving container (not shown) to the designated position of the various peripheral devices or to the customer as needed.

[0077] Peripheral devices such as mill M, water server W, and espresso machine E, which are designed for human use, require the operator to visually read the positions of meters and levers, place containers in the appropriate positions while looking around the peripheral devices, and then operate dials, buttons, levers, etc., which require delicate application of force. In order to operate such peripheral devices with a robotic arm and furthermore to grasp and carry containers made of flexible materials such as plastic or paper without crushing them, it is necessary to appropriately process sensor data from sensor 24 and accurately move the robotic arm through control of drive unit 25.

[0078] The assistance device 1 of this embodiment provides a control machine learning model that has achieved such processing through machine learning based on operational data including both sensor data and control data, and further performs further machine learning using operational data collected from multiple robots 2, including other robots 2, during a fixed-price contract, to assist the robot 2 in serving beverages such as coffee using these peripheral devices. In addition, the assistance device 1 of this embodiment can provide a control machine learning model that provides services tailored to each customer through machine learning based on the operational data. For example, the control machine learning model can learn customer preferences and control the robot 2 to suggest recommended menus or services when the customer visits.

[0079] [Robot 2, a transport robot] Robot 2, which is a transport robot, includes, for example, a multicopter with multiple rotors and a gripping mechanism that grips and transports luggage, a loading platform, etc. Robot 2, which is a transport robot, controls drive unit 25 that rotates the rotors and another drive unit 25 that operates the gripping mechanism by processing using a control machine learning model based on sensor data from sensors 24, including a camera that photographs the surrounding area, an anemometer that measures wind speed, etc., and a positioning system, etc., to transport luggage.

[0080] The functions of determining whether to fly or not based on the status of each part of the aircraft obtained from sensor data, flying an appropriate route by using operational data including past flight data based on meteorological conditions such as wind direction and surrounding conditions, and delivering luggage to a convenient destination require complex decisions based on sensor data. The assistance device 1 of this embodiment provides a control machine learning model that achieves such processing through machine learning based on operational data, and further performs further machine learning using the operational data to assist the robot 2 in making such decisions and appropriately transporting luggage.

[0081] [Robot 2, a harvesting robot] Robot 2, a harvesting robot, is equipped with a movement mechanism constituted by drive wheels, etc., and a harvesting robot arm moved by a servo motor or air pressure control device, etc. Robot 2, a harvesting robot, controls drive unit 25, which moves the movement mechanism, and another drive unit 25, which moves the harvesting robot arm, by processing using a control machine learning model based on sensor data from sensors 24, which include a camera that photographs fruit trees, etc. to be harvested, and a positioning system, etc., to harvest fruits, etc.

[0082] The functions of appropriately applying fertilizer or pesticides according to the condition of the crop or soil, determining the appropriate time for harvesting according to the growth conditions, efficiently moving along an appropriate route according to the growth conditions of each area, determining which fruits are suitable for harvesting, and harvesting such fruits without damaging them require complex judgments based on sensor data. The assistance device 1 of this embodiment provides a control machine learning model that achieves such processing through machine learning based on operational data, and further performs further machine learning using the operational data to assist the robot 2 in making such judgments and appropriately harvesting the fruits.

[0083] [Other robots with mechanical movements 2] In addition to the above, the system S of this embodiment may include various robots 2 that provide advanced functions through mechanical operations in which the sensor 24 and the drive unit 25 work together. The following is one example.

[0084] In the robot 2, which is an industrial robot, the support device 1 of this embodiment can improve production efficiency, reduce costs, improve quality, or reduce downtime by applying data-driven optimization using operational data to a control machine learning model.

[0085] In the robot 2, which is a medical robot, the assistance device 1 of this embodiment can provide feedback to a control machine learning model based on analysis of operational data during surgery, or optimize routine tasks based on operational data, etc. In addition, the assistance device 1 of this embodiment can also contribute to proposing rehabilitation plans based on operational data related to rehabilitation, early diagnosis of illness based on operational data related to diagnostic devices, and providing personalized care based on operational data related to patient care.

[0086] In the agricultural robot 2, the support device 1 of this embodiment can optimize the movement path of the robot 2 based on operation data, early detection of pests and diseases through machine learning based on sensor data from cameras and the like and spraying pesticides and the like based on the early detection, understanding of growth conditions through machine learning based on sensor data from cameras and the like and spraying fertilizer and the like based on that understanding, or creating an irrigation plan through machine learning based on sensor data from soil moisture sensors and controlling the irrigation robot based on that plan, etc. In addition, the support device 1 of this embodiment can also contribute to optimizing worker schedules through machine learning based on operation data related to agricultural work, and optimizing planting plans or harvest plans through machine learning based on operation data related to market demand, etc.

[0087] In the robot 2, which is a mobile robot, the assistance device 1 of this embodiment can realize optimization of movement plans using real-time traffic information by a control machine learning model that has undergone machine learning based on operational data including past movement data, suitable obstacle avoidance based on sensor data by a control machine learning model that has undergone machine learning based on operational data related to obstacle avoidance, improvement of energy efficiency by a control machine learning model that has undergone machine learning based on operational data including remaining battery or fuel, or navigation according to the user by a control machine learning model that has undergone machine learning based on operational data related to navigation.

[0088] In the robot 2, which is an autonomous driving robot, the assistance device 1 of this embodiment can realize real-time route calculation or traffic condition learning through machine learning using operational data related to route optimization. Additionally, the assistance device 1 of this embodiment can also contribute to obstacle detection and avoidance or weather response through machine learning based on operational data related to environmental recognition and safety improvement, battery management or energy efficiency optimization based on operational data related to energy management, and preventive maintenance or automatic diagnostic systems based on operational data related to maintenance and preventive maintenance. Furthermore, the assistance device 1 of this embodiment can also contribute to high-precision navigation or dynamic route change based on autonomous navigation, collaboration between multiple robots or demand forecasting and vehicle allocation planning based on fleet management, personalized services or real-time support based on improved user experience, etc.

[0089] In the robot 2, which is an inspection robot, the support device 1 of this embodiment can enhance image analysis or detect minute defects through machine learning using operational data related to inspection accuracy. Additionally, the support device 1 of this embodiment can contribute to providing efficient inspection routes through machine learning based on operational data related to optimizing motion paths, providing real-time analysis or feedback loops based on analysis and feedback of inspection data, and detecting equipment anomalies or automatically diagnosing equipment based on preventive maintenance. The support device 1 of this embodiment can also contribute to resource optimization or work schedule proposals based on efficient resource management, continuous learning or improvement proposals based on continuous learning and improvement of data, and detecting abnormal patterns based on early detection of anomalies and countermeasures.

[0090] In the robot 2, which is a household robot, the assistance device 1 of this embodiment can optimize the path of the cleaning robot using a control machine learning model that has undergone machine learning based on operational data related to the cleaning status, optimize the operation of each robot using a control machine learning model that has undergone machine learning based on operational data related to household power consumption, and improve household security using a control machine learning model that has undergone machine learning based on operational data related to sensor data from security devices, etc. In addition, the assistance device 1 of this embodiment can also contribute to the creation of a health promotion plan using machine learning based on operational data related to family health data.

[0091] In the robot 2, which is an engineer robot, the assistance device 1 of this embodiment can realize automatic calibration or optimization of motion paths through machine learning using operational data related to work accuracy and efficiency. Additionally, the assistance device 1 of this embodiment can contribute to providing an anomaly detection or automatic diagnosis system through machine learning based on operational data related to preventive maintenance and maintenance, process optimization or dynamic scheduling based on operational data related to work processes, and risk assessment or automatic application of safety protocols based on operational data related to the safety of the work environment. The assistance device 1 of this embodiment can also contribute to data-driven improvement or feedback loops based on continuous learning and improvement of technology, optimization of collaborative work or interaction through machine learning based on operational data related to collaboration with humans, and adaptation of the latest technology or prediction of technological trends based on the introduction and adaptation of new technology.

[0092] In the robot 2, which is a security robot, the assistance device 1 of this embodiment can detect abnormal behavior or provide real-time responses through machine learning using operational data related to anomaly detection and response. Additionally, the assistance device 1 of this embodiment can contribute to efficient route calculation or dynamic route adjustment through machine learning based on operational data related to patrol routes, providing a data integration platform or long-term data analysis based on operational data related to integrated management of surveillance data, and real-time environmental recognition or adaptive movement based on operational data related to environmental recognition and obstacle avoidance. The assistance device 1 of this embodiment can also contribute to strengthening security protocols or protecting privacy through machine learning based on operational data related to security enhancement and data privacy, and building data-driven improvements or feedback loops through machine learning based on operational data related to continuous learning and improvement.

[0093] In the robot 2, which is a service industry robot, the assistance device 1 of this embodiment can realize customer profiling or the provision of recommended services through machine learning using operational data related to personalizing customer service. Additionally, the assistance device 1 of this embodiment can contribute to automating tasks or dynamic scheduling through machine learning based on operational data related to improving business efficiency, enhancing natural language processing or providing real-time feedback based on operational data related to interaction quality, and providing demand forecasts or trend analyses based on operational data related to trend forecasting and demand management. The assistance device 1 of this embodiment can also contribute to data security or privacy management through machine learning based on operational data related to security and privacy, and building data-driven improvements or feedback loops through machine learning based on operational data based on continuous learning and improvement.

[0094] In the robot 2, which is a logistics sorting support robot, the support device 1 of this embodiment can achieve route optimization or real-time dynamic scheduling through machine learning using operational data related to the sorting process. Additionally, the support device 1 of this embodiment can contribute to improving barcode and QR code (registered trademark) reading accuracy or recognition accuracy through image analysis through machine learning based on operational data to reduce errors and improve accuracy, automating tasks or providing robot-human collaboration based on operational data related to labor efficiency, and providing anomaly detection or automatic diagnosis through machine learning based on operational data related to preventive maintenance and maintenance. Furthermore, the support device 1 of this embodiment can contribute to analyzing or predicting business data based on data analysis and improvement suggestions, applying safety protocols or responding to regulations based on safety and regulatory compliance, and building data-driven improvements or feedback loops based on continuous learning and improvement.

[0095] In the robot 2, which is an aging society support robot, the assistance device 1 of this embodiment can monitor vital signs or manage medication through machine learning using operational data related to health management and preventive care. Additionally, the assistance device 1 of this embodiment can also contribute to providing fall detection or automatic locking systems using machine learning based on operational data related to enhanced safety and security, providing housework assistance or mobility assistance based on operational data related to support for daily life, and providing communication assistance or recreational activities based on operational data related to social interaction and mental care. The assistance device 1 of this embodiment can also contribute to telemedicine or notifications to family members based on remote monitoring and support, and building data-driven improvements or feedback loops based on continuous learning and improvement.

[0096] In the robot 2, which is a daily life support robot, the assistance device 1 of this embodiment can optimize cleaning or automate laundry through machine learning using operational data related to improving the efficiency of housework. Additionally, the assistance device 1 of this embodiment can also contribute to vital sign monitoring or fitness support through machine learning based on operational data related to health management and fitness, recipe suggestions and cooking support or ingredient management based on operational data related to meal management and cooking support, and automatic monitoring systems or elderly care based on operational data related to security and monitoring. The assistance device 1 of this embodiment can also contribute to schedule optimization or personalized reminders based on schedule management and reminders, and data-driven optimization or the use of feedback based on continuous learning and improvement.

[0097] In the robot 2, which is a fishing support robot, the support device 1 of this embodiment can analyze fish detection data through machine learning using operational data related to fish detection and identifying optimal fishing grounds, or integrate environmental data. Additionally, the support device 1 of this embodiment can also contribute to automating fishing operations or providing real-time monitoring and adjustment through machine learning based on operational data related to improving harvesting efficiency, managing catch volume or optimizing species selection based on operational data related to sustainable fisheries management, collecting marine environmental data or providing climate change impact forecasts based on operational data related to environmental monitoring and adaptation, etc. The support device 1 of this embodiment can also contribute to automating work or providing a safety monitoring system based on labor efficiency and safety improvements, analyzing collected data based on data-driven improvements, or building a feedback loop.

[0098] In the robot 2, which is a disaster relief support robot, the support device 1 of this embodiment can realize aerial surveillance by drones or on-site surveys by ground robots through machine learning using operational data related to real-time situation assessment and information collection. Additionally, the support device 1 of this embodiment can contribute to locating disaster victims or providing medical support through machine learning based on operational data related to rescue and support of disaster victims, providing delivery or inventory management based on operational data related to logistics and supply management, and removing obstacles or monitoring infrastructure status based on operational data related to infrastructure restoration support. Furthermore, the support device 1 of this embodiment can contribute to restoring communication infrastructure or providing information based on communication support, and building data-driven improvements or feedback loops based on continuous learning and improvement.

[0099] In the robot 2, which is a construction support robot, the support device 1 of this embodiment can optimize work plans or improve resource management efficiency through machine learning using operational data related to the construction process. Additionally, the support device 1 of this embodiment can contribute to real-time quality monitoring or highly accurate construction through machine learning based on operational data related to construction quality, automatic operation of heavy machinery or safety monitoring systems based on operational data related to labor efficiency and safety improvement, and collection of environmental data or sustainable construction based on operational data related to environmental monitoring and adaptation. The support device 1 of this embodiment can also contribute to data-driven improvement or feedback loop construction based on continuous learning and improvement, cost analysis and forecasting based on cost reduction and budget management, or optimal material procurement.

[0100] [Information processing robot 2] The assistance device 1 of this embodiment can also be effective in providing a machine learning model in which machine learning has been performed based on operational data collected in accordance with a fixed-price contract, even for a robot 2 whose main function is to display the results of information processing instead of performing mechanical operations based on sensor data.

[0101] In the robot 2, which is an emotion-recognition robot, the assistance device 1 of this embodiment can realize multimodal data analysis or individual customization through machine learning using operational data related to the accuracy of emotion recognition. Additionally, the assistance device 1 of this embodiment can also contribute to context understanding or real-time feedback through machine learning based on operational data related to interaction optimization, building user profiles based on personalized responses or providing learning and adaptation, monitoring stress levels based on stress detection and mental care, or providing mental health support. Furthermore, the assistance device 1 of this embodiment can also contribute to continuous data collection or building a feedback loop based on continuous learning and improvement.

[0102] In the robot 2, which is a marketing robot, the support device 1 of this embodiment can realize analysis by a generative AI using behavioral data related to improving targeting accuracy and the provision of personalized campaigns. Additionally, the support device 1 of this embodiment can also contribute to providing content generation by a generative AI and automated A / B testing related to content optimization, optimizing chatbot responses and providing real-time support related to improving customer engagement, analyzing campaign effectiveness and optimizing ROI related to analyzing and optimizing marketing effectiveness, optimizing customer journeys and providing cross-selling and up-selling related to improving customer experience, etc. Furthermore, the support device 1 of this embodiment can also contribute to building a feedback loop based on continuous learning and improvement and providing trend forecasts, etc.

[0103] In the robot 2, which is a health management robot, the assistance device 1 of this embodiment can analyze health data using a generative AI to create an individual health plan, track habits, and provide improvement suggestions. Additionally, the assistance device 1 of this embodiment can also contribute to anomaly detection or risk assessment using a generative AI for preventive medicine and early detection, real-time monitoring or integrated data management for improving health monitoring efficiency, feedback loops or health advice for providing feedback and advice, and stress level monitoring or mental health plans for mental health support. The assistance device 1 of this embodiment can also contribute to data-driven improvement or strengthening of feedback loops based on continuous learning and improvement.

[0104] In the robot 2, which is a sales robot, the support device 1 of this embodiment can realize the analysis of behavioral data using generative AI to improve the accuracy of customer targeting or the provision of a personalized approach. Additionally, the support device 1 of this embodiment can also contribute to the provision of lead generation automation or follow-up automation using generative AI to automate and streamline sales processes, performance analysis or predictive analysis to analyze and optimize sales performance, chatbot optimization or real-time support to improve the quality of interactions, and market trend analysis or demand forecasting to predict trends and manage demand. The support device 1 of this embodiment can also contribute to the utilization of feedback based on continuous learning and improvement or the provision of data-driven decision-making.

[0105] In the robot 2, which is a personnel evaluation robot, the assistance device 1 of this embodiment can realize the analysis of performance data by a generation AI to improve evaluation accuracy or the provision of a 360-degree evaluation. Additionally, the assistance device 1 of this embodiment can also contribute to the detection and correction of bias by a generation AI to eliminate bias or the provision of standardized evaluation criteria, skill gap analysis or training suggestions to create individual growth plans, the provision of feedback or engagement measurement to improve engagement and motivation, and the provision of performance prediction or optimal personnel placement related to performance prediction and personnel placement. Furthermore, the assistance device 1 of this embodiment can also contribute to the creation of a feedback loop based on continuous learning and improvement or the provision of data-driven decision-making.

[0106] In the robot 2, which is an educational robot, the assistance device 1 of this embodiment can realize, for example, analysis of learning progress by a generative AI for optimizing individual learning or provision of personalized content. Additionally, the assistance device 1 of this embodiment can also contribute to providing instant evaluation or automatic grading by a generative AI for real-time feedback and evaluation, providing interactive learning support or improving engagement for promoting effective interactions, monitoring concentration or adjusting the physical environment for optimizing the learning environment, and providing outcome prediction or improvement suggestions for learning methods for predicting and improving learning outcomes. Furthermore, the assistance device 1 of this embodiment can also contribute to building a feedback loop based on continuous learning and improvement or providing data-driven educational strategies.

[0107] In the robot 2, which is a livestock support robot, the support device 1 of this embodiment can monitor health conditions or provide individual care plans using a generation AI related to health management. Additionally, the support device 1 of this embodiment can also contribute to tracking reproductive cycles or managing genetic information using a generation AI related to improving the efficiency of breeding management, monitoring feed consumption or adjusting nutritional balance using a generation AI related to feed management, and providing environmental monitoring or an automatic cleaning system using a generation AI related to environmental management. Furthermore, the support device 1 of this embodiment can also contribute to automating tasks or optimizing work schedules based on labor efficiency, and building data-driven improvements or feedback loops based on continuous learning and improvement.

[0108] In the robot 2, which is an aquaculture support robot, the support device 1 of this embodiment can provide water quality monitoring or an automatic feeding system using generative AI related to the aquaculture environment. Additionally, the support device 1 of this embodiment can also contribute to real-time health monitoring or preventive treatment using generative AI related to health management and disease prevention, optimization of breeding timing or genetic diversity management using generative AI related to efficient breeding management, and automation of tasks or optimization of work schedules using generative AI related to labor efficiency. Furthermore, the support device 1 of this embodiment can also contribute to optimal resource utilization or reduction of environmental impact based on the realization of sustainable aquaculture, and the creation of data-driven improvements or feedback loops based on continuous learning and improvement.

[0109] In the robot 2, which is a dairy farming support robot, the support device 1 of this embodiment can provide real-time health monitoring or individual care plans using a generative AI for health management. Additionally, the support device 1 of this embodiment can also contribute to reproductive cycle management or genetic information analysis using a generative AI for breeding management, feed consumption monitoring or nutritional balance adjustment using a generative AI for feed management, milk production monitoring or milking robot optimization using a generative AI for milk production, etc. Furthermore, the support device 1 of this embodiment can also contribute to environmental monitoring or an automatic cleaning system using a generative AI for environmental management, task automation or work schedule optimization using a generative AI for labor efficiency, and the creation of data-driven improvements or feedback loops based on continuous learning and improvement.

[0110] In the robot 2, which is an entertainment support robot, the support device 1 of this embodiment can recommend personalized content based on a generation AI's analysis of the user's viewing history and preferences, and suggest events based on the user's interests and past participation history. Additionally, the support device 1 of this embodiment can contribute to optimizing interactions with a conversational AI assistant using the generation AI, analyzing audience reactions and providing feedback during live performances, and other such activities. The support device 1 of this embodiment can also contribute to improving the efficiency of content production through the generation AI's creation of first drafts of scripts and optimization of visual effects, optimizing marketing through the optimization of targeted advertising and the analysis of marketing campaign effectiveness, and improving user engagement through social media analysis and user feedback analysis. Furthermore, the support device 1 of this embodiment can contribute to continuous learning and improvement of functions through data-driven improvements, customization based on feedback from the field, and other such activities.

[0111] In the robot 2, which is a management consulting robot, the support device 1 of this embodiment can perform market analysis, trend forecasting, competitive analysis, etc. using a generative AI. Additionally, the support device 1 of this embodiment can contribute to business process optimization by having the generative AI analyze business data to propose improvements for efficiency and promote business automation. The support device 1 of this embodiment can also contribute to financial management and cost reduction by having the generative AI analyze real-time financial data to propose cost reductions and optimize budgeting and management. Furthermore, the support device 1 of this embodiment can contribute to human resource management and development by having the generative AI analyze employee performance data to provide fair and objective evaluations and analyze skill sets to assign employees to optimal projects and roles. The support device 1 of this embodiment can also contribute to risk management and ensuring compliance with laws and regulations through risk prediction and compliance monitoring by the generative AI. Furthermore, the support device 1 of this embodiment can contribute to continuous learning and improvement by having the generative AI build a feedback loop to continuously improve service quality and learn the latest knowledge and technologies to provide optimal solutions. The support device 1 of this embodiment utilizes generation AI and data accumulated after use to optimize the functions of the management consulting robot, enabling data-driven management strategy planning, business process optimization, financial management and cost reduction, human resource management and development, risk management and compliance, and continuous learning and improvement, enabling companies to manage their businesses more effectively and efficiently and increase their competitiveness.

[0112] In a business strategy planning support robot, the support device 1 of this embodiment can realize strategies based on market analysis and trend forecasts, securing competitive advantages through competitor analysis and benchmarking, optimizing financial planning and risk management, improving the efficiency of human resource management and organizational development, improving the accuracy of strategy simulation and evaluation, continuous learning and improvement, etc. In addition, the support device 1 of this embodiment can also contribute to feedback to management and improving employee engagement.

[0113] [Network N] The type of the network N is not particularly limited as long as it enables communication between the support device 1, etc. The type of the network N is, for example, the Internet, a mobile phone network, a wireless LAN, etc.

[0114] [Terminal T] The terminal T is used by an administrator of the assistance device 1 or a user of the robot 2. The type of the terminal T is, for example, a smartphone, a tablet terminal, a mobile terminal such as a laptop computer, or a fixed terminal such as a desktop computer.

[0115] [Main flowchart of support processing] FIG. 5 is a main flowchart showing an example of a preferred flow of the support processing of this embodiment. FIG. 6 is a diagram continuing from the previous diagram. FIG. 7 is a diagram continuing from the previous diagram. FIG. 8 is a diagram continuing from the previous diagram. FIG. 9 is a diagram continuing from the previous diagram. Below, an example of a preferred flow of the support processing of this embodiment will be explained using FIGS. 5 to 9.

[0116] The support device 1 executes a series of processes for updating the contract period based on the payment made in response to the invoice related to the fixed-price contract through the fixed-price contract management unit 111. Steps S1 to S6 are an example of this process.

[0117] [Step S1: Determine whether to send a request] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the fixed-price contract management unit 111. Then, the control unit 11 executes a process to determine whether to send a bill related to the fixed-price contract for the rental of the robot using the fixed-price contract management unit 111 (bill sending determination step). If it is determined that a bill should be sent, the control unit 11 shifts the process to step S2. If it is not determined that a bill should be sent, the control unit 11 shifts the process to step S3.

[0118] In the billing transmission determination step, the fixed-price contract management unit 111 implements a procedure for determining whether to send a bill, for example, by determining that a bill should be sent if there is an unpaid fixed-price contract that has passed the scheduled billing date.

[0119] [Step S2: Send request] The control unit 11 executes a process of transmitting a bill related to the above-mentioned bill transmission determination step to the user's terminal T through the fixed-rate contract management unit 111 (bill transmission execution step). The control unit 11 moves the process to step S3.

[0120] [Step S3: Determine whether to update payment status] The control unit 11 executes a process to determine whether to update the payment status related to the billing in the billing transmission execution step described above, using the fixed-rate contract management unit 111 (payment status update determination step). If it is determined that the status should be updated, the control unit 11 proceeds to step S4. If it is not determined that the status should be updated, the control unit 11 proceeds to step S5.

[0121] In the payment status update determination step, the fixed-price contract management unit 111 performs a process to determine whether to update the payment status, for example, when it receives data indicating that a transfer has been made corresponding to the invoice, by determining whether to update the payment status.

[0122] [Step S4: Update payment status] The control unit 11 executes processing to update the payment status related to the above-mentioned payment status update determination step in the fixed-rate contract database 131 by the fixed-rate contract management unit 111 (payment status update execution step). The control unit 11 moves the processing to step S5.

[0123] [Step S5: Determine whether to renew the contract period] The control unit 11 executes a process to determine whether to renew the contract period related to the payment status update execution step described above, using the fixed-rate contract management unit 111 (contract period renewal determination step). If it is determined that renewal should be performed, the control unit 11 proceeds to step S6. If it is not determined that renewal should be performed, the control unit 11 proceeds to step S7.

[0124] In the contract period renewal determination step, the fixed-price contract management unit 111 realizes a process of determining whether to renew the contract period, for example, by a procedure of determining that the contract period related to the payment status update execution step should be renewed if there are no unpaid invoices related to the contract period before and after renewal.

[0125] [Step S6: Renew contract period] The control unit 11 executes processing to update the contract period related to the above-mentioned contract period renewal determination step in the fixed-rate contract database 131 by the fixed-rate contract management unit 111 (contract period renewal execution step). The control unit 11 moves the processing to step S7.

[0126] When a fixed-price contract for one of the robots is paid for, the operation data acquisition unit 112 of the support device 1 executes a series of processes for acquiring operation data of the robot. Steps S7 to S8 are an example of the processes.

[0127] [Step S7: Determine whether to acquire operational data] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the operation data acquisition unit 112. Then, the control unit 11 executes a process to determine whether to acquire operation data from the robot 2 using the operation data acquisition unit 112 (operation data acquisition determination step). If it is determined that the data should be acquired, the control unit 11 shifts the process to step S8. If it is not determined that the data should be acquired, the control unit 11 shifts the process to step S9.

[0128] In the operation data acquisition determination step, the operation data acquisition unit 112 realizes the process of determining whether to acquire operation data by, for example, a procedure of determining that operation data should be acquired from one robot 2 (acquiring robot) included in the multiple robots 2 related to the system S when a fixed-price contract payment has been made for that robot 2. This process may further include a procedure of determining that operation data should be acquired from the acquiring robot when the latest operation data for the acquiring robot has not been acquired at or after the timing when the operation data for the acquiring robot is periodically acquired and when a fixed-price contract payment has been made.

[0129] [Step S8: Obtain operational data] The control unit 11 executes a process of acquiring operation data from the robot 2 (acquisition source robot) by the operation data acquisition unit 112 (operation data acquisition execution step). The control unit 11 moves the process to step S9.

[0130] The operation data preferably includes at least sensor data and control data of the acquiring robot. In order to optimize the control machine learning model according to the region, the operation data preferably includes data indicating the region in which the acquiring robot is operated. In order to perform machine learning for each command given by the user, the operation data preferably includes data indicating the command item. In order to reflect changes in the operation data over time in the control machine learning model, the operation data is preferably associated with information specifying time (e.g., date and time).

[0131] The assistance device 1 executes a series of processes for performing machine learning of a control machine learning model used to control the robot using learning data based on operation data, by the machine learning unit 113. Steps S9 to S10 are an example of this process.

[0132] [Step S9: Determine whether to perform machine learning on the control machine learning model] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the machine learning unit 113. Then, the control unit 11 executes a process of determining whether to perform machine learning of the control machine learning model using the machine learning unit 113 (control machine learning determination step). If it is determined that machine learning should be performed, the control unit 11 shifts the process to step S10. If it is not determined that machine learning should be performed, the control unit 11 shifts the process to step S11.

[0133] In the control machine learning determination step, the machine learning unit 113 realizes a process of determining whether to perform machine learning of the control machine learning model, for example, by a procedure of determining that machine learning of the control machine learning model should be performed when the accumulation of new operational data exceeds a given threshold.

[0134] [Step S10: Perform machine learning of the control machine learning model] The control unit 11 executes a process of performing machine learning of a control machine learning model used for controlling the robot by the machine learning unit 113 (control machine learning execution step). The control unit 11 moves the process to step S11.

[0135] In the control machine learning execution step, the machine learning unit 113 performs machine learning of a control machine learning model using learning data based on the above-mentioned operation data. Specifically, the machine learning unit 113 performs machine learning using learning data that includes the sensor data included in the operation data as explanatory variables and the control data included in the operation data as objective variables.

[0136] In order to perform machine learning according to a region, the machine learning unit 113 preferably includes a procedure for performing machine learning of a control machine learning model associated with a specified region using learning data based on operational data related to the region. At this time, the memory unit 13 stores the control machine learning model associated with each region.

[0137] In order to perform machine learning according to the type of robot 2, the machine learning unit 113 preferably includes a procedure for performing machine learning of a control machine learning model associated with a specified type of robot 2 using learning data based on operational data related to the type. At this time, the memory unit 13 stores a control machine learning model associated with each type of robot 2.

[0138] When the above-described machine learning is performed, the support device 1 executes a series of processes in which the machine learning model providing unit 114 provides, to a rental robot included in a plurality of robots, a control machine learning model that has been machine-learned by the machine learning unit 113, when the rental robot is within the rental period of a fixed-price contract for the rental robot. Steps S11 to S12 are an example of this process.

[0139] [Step S11: Determine whether to provide a control machine learning model] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the machine learning model providing unit 114. Then, the control unit 11 executes a process of determining whether to provide the control machine learning model on which the above-mentioned machine learning has been performed to the robot 2 (rented robot) rented under a fixed-price contract by the machine learning model providing unit 114 (control machine learning model provision determination step). If it is determined that the control unit 11 should provide the control machine learning model, the control unit 11 proceeds to step S12. If it is not determined that the control unit 11 should provide the control machine learning model, the control unit 11 proceeds to step S13.

[0140] In the control machine learning model provision determination step, the machine learning model provision unit 114 realizes a process of determining whether to provide a control machine learning model to a rented robot by providing the rented robot with a control machine learning model that has been machine learned by the machine learning unit 113 when the rental period for one of the robots 2 (rented robot) included in the multiple robots 2 related to the system S is in the fixed-price contract. This process may further include a step of determining to provide a control machine learning model to the rented robot when the rental period is in progress and the latest control machine learning model has not been provided to the rented robot at or after the timing of periodically updating the control machine learning model of the rented robot.

[0141] [Step S12: Provide the control machine learning model] The control unit 11 executes a process of providing the control machine learning model, which has undergone the above-mentioned machine learning by the machine learning unit 113, to the rental robot by the machine learning model providing unit 114 (control machine learning model providing execution step). The control unit 11 proceeds to step S13.

[0142] When the control machine learning model is managed according to the region or type of robot 2, etc., it is preferable that in the control machine learning model provision execution step, the machine learning model provision unit 114 provides the loaned robot with a control machine learning model related to the region or type of robot 2, etc. that is suitable for the loaned robot.

[0143] The support device 1 preferably executes a series of processes in which the frequently used item extraction unit 115 extracts one or more frequently used items from among a plurality of command items for the robot, and the frequently used item provision unit 116 provides the one or more frequently used items to the loaned robot. Steps S13 to S15 are an example of the process. At this time, the operation data acquisition unit 112 preferably acquires operation data further including the frequency of use of each command item for the source robot.

[0144] [Step S13: Determine whether to provide frequently used items] The control unit 11 cooperates with the memory unit 13 and the communication unit 14 to execute the frequently used item extraction unit 115. Then, the control unit 11 executes a process in which the frequently used item extraction unit 115 extracts one or more frequently used items from among a plurality of command items for the robot 2 and determines whether to provide the extracted items to the rental robot (frequently used item provision determination step). If it is determined that the items should be provided, the control unit 11 shifts the process to step S14. If it is not determined that the items should be provided, the control unit 11 shifts the process to step S16.

[0145] In the frequently used item provision determination step, the frequently used item extraction unit 115 may further include a procedure for determining that the frequently used items should be provided to the loaned robot, for example, when the loaned robot is in the loan period and the latest frequently used items have not been provided to the loaned robot at or after the time when the loaned robot's frequently used items are regularly updated.

[0146] [Step S14: Extract frequently used items] The control unit 11 executes a process of extracting one or more frequently used items from among a plurality of command items for the robot 2 using the frequently used item extraction unit 115 (frequently used item extraction step). The control unit 11 moves the process to step S15.

[0147] In the frequently used item extraction step, the frequently used item extraction unit 115 extracts the one or more frequently used items described above based on the operation data that further includes the frequency of use of the command items. In order to reflect the operation data of robots 2 of similar types or operating areas in other robots 2, it is preferable that the frequently used item extraction unit 115 compiles the operation data by type of robot 2 or by area related to the robot 2, and extracts the one or more frequently used items described above based on the compiled operation data.

[0148] [Step S15: Provide frequently used items] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the frequently used item providing unit 116. Then, the control unit 11 executes a process of providing one or more frequently used items extracted in the frequently used item extraction step to the lending robot by the frequently used item providing unit 116 (frequently used item providing execution step). The control unit 11 shifts the process to step S16.

[0149] The assistance device 1 executes a series of processes in which the machine learning unit 113 performs machine learning on a language model using training data that includes, as explanatory variables, a situation related to a linguistic expression output from the source robot to the user of the source robot and the user's evaluation of the situation, and that includes the linguistic expression as a target variable, and provides the robot 2 with a language model that outputs a linguistic expression that is expected to be highly evaluated when the situation is input. Steps S16 to S18 are an example of the processes. At this time, it is preferable that the operation data acquisition unit 112 acquires operation data that further includes the linguistic expression output from the source robot to the user of the source robot, the situation related to the linguistic expression, and the user's evaluation of the linguistic expression.

[0150] [Step S16: Determine whether to perform machine learning on the language model] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the machine learning unit 113. Then, the control unit 11 executes a process of determining whether to perform machine learning of a language model using the machine learning unit 113 (language machine learning determination step). If it is determined that machine learning should be performed, the control unit 11 proceeds to step S17. If it is not determined that machine learning should be performed, the control unit 11 proceeds to step S19.

[0151] In the control machine learning determination step, the machine learning unit 113 realizes a process of determining whether to perform machine learning of a language model, for example, by a procedure of determining that machine learning of a language machine learning model should be performed when the accumulation of new operational data exceeds a given threshold.

[0152] [Step S17: Perform machine learning of the language model] The control unit 11 executes a process of performing machine learning of a language model by the machine learning unit 113 (language machine learning execution step). The control unit 11 moves the process to step S18. This language model is a model that outputs a language expression intended for the user of the robot 2 when a situation related to the robot 2 is input.

[0153] In the language learning execution step, the machine learning unit 113 performs machine learning of a language model using learning data based on the above-mentioned operational data. Specifically, the machine learning unit 113 performs machine learning using learning data that includes, as explanatory variables, situations related to linguistic expressions included in the operational data and the user's evaluation of the linguistic expressions, and that includes, as a target variable, the linguistic expressions related to the situations.

[0154] In order to perform machine learning according to a region, the machine learning unit 113 preferably includes a procedure for performing machine learning of a language model associated with a specified region using learning data based on operation data related to the region. At this time, the storage unit 13 stores the language model associated with each region.

[0155] In order to perform machine learning according to the type of robot 2, the machine learning unit 113 preferably includes a procedure for performing machine learning of a language model associated with a specified type of robot 2 using learning data based on operation data related to the specified type of robot 2. At this time, the storage unit 13 stores a language model associated with each type of robot 2.

[0156] [Step S18: Provide a language model] The control unit 11 cooperates with the storage unit 13 and the communication unit 14 to execute the machine learning model providing unit 114. Then, the control unit 11 executes a process of providing the language model on which the above-mentioned machine learning has been performed by the machine learning unit 113 to the rental robot by the machine learning model providing unit 114 (language model provision execution step). The control unit 11 moves the process to step S19.

[0157] [Optimizing visual communication] In order to optimize visual communication, it is preferable that the support device 1 acquires operational data including visual expressions and evaluations of the visual expressions instead of linguistic expressions and evaluations of the linguistic expressions, and performs machine learning of a visual expression selection model using learning data that includes, as explanatory variables, a situation related to the visual expressions included in the operational data and the user's evaluation of the visual expressions, and that includes, as a target variable, the visual expressions related to the situation (visual expression machine learning step).

[0158] As a result, the assistance device 1 can execute a visual expression specifying step of specifying a visual expression suitable for a situation by inputting data indicating a situation related to the operation of the robot 2 into the visual expression selection model. As a result, the visual expressions performed by the robot 2 via the drive unit 25 or the display unit 27 are optimized in various situations related to the operation of the robot 2. For example, the assistance device 1 can optimize the operation of the robot 2 so as to perform a visual expression using a specific male avatar that is expected to be more highly rated, based on operation data collected in an industry and region with a high proportion of female customers.

[0159] The assistance device 1 preferably executes a series of processes using a language model with command items for the robot and the operation data as inputs, by the instruction generation unit 117, to generate instructions showing how to use the command items. Steps S19 to S21 are an example of this process. At this time, the operation data acquisition unit 112 preferably acquires operation data including sensor data of the acquisition source robot associated with the command items and control data of the acquisition source robot.

[0160] [Step S19: Determine whether to generate instructions] The control unit 11 cooperates with the memory unit 13 and the communication unit 14 to execute the instruction generation unit 117. Then, the control unit 11 executes a process of determining whether to generate an instruction showing how to use the command items for the robot 2 using the instruction generation unit 117 (instruction generation determination step). If it is determined that an instruction should be generated, the control unit 11 moves the process to step S20. If it is not determined that an instruction should be generated, the control unit 11 returns the process to step S1 and repeats the processes from step S1 to step S21.

[0161] In the instruction manual generation determination step, the instruction manual generation unit 117 realizes a process of determining whether to generate an instruction manual, for example, by a procedure of determining that an instruction manual should be generated when an instruction manual generation is instructed from terminal T.

[0162] [Step S20: Generate instructions] The control unit 11 executes a process in which the instruction generating unit 117 receives command items and operation data for the robot 2 and uses a language model to generate instructions indicating how to use the command items (instruction generating execution step). The control unit 11 then proceeds to step S21.

[0163] In order to generate instructions in a given style based on operational data, in the instruction generation execution step, it is preferable that the instruction generation unit 117 generates instructions by processing using a language model that has been pre-trained using training data that includes command items and operational data as explanatory variables and instruction data as a target variable.

[0164] [Step S21: Provide instructions] The control unit 11 executes a process of providing the instruction manual relating to the instruction manual generation execution step to a terminal T or the like used by the user of the robot 2 by the instruction manual generation unit 117 (instruction manual provision step). The control unit 11 returns the process to step S1 and repeats the processes from step S1 to step S21.

[0165] [Cost-effectiveness calculation step] It is preferable that the assistance device 1 further executes a cost-effectiveness calculation step of estimating the economic effect of introducing the robot 2 based on the business performance both before and after the introduction of the robot 2 and the above-mentioned operation data, and calculating the cost-effectiveness of the robot 2 based on the usage fee related to the fixed-price contract for the robot 2 and the estimated economic effect. The cost-effectiveness calculation step is executed, for example, by the control unit 11 in cooperation with the storage unit 13 and the communication unit 14 executing a cost-effectiveness calculation unit (not shown), which estimates the economic effect of introducing the robot 2 based on the business performance both before and after the introduction of the robot 2 and the above-mentioned operation data, and calculates the cost-effectiveness of the robot 2 based on the usage fee related to the fixed-price contract for the robot 2 and the estimated economic effect.

[0166] In order to improve accuracy by estimating the cost-effectiveness for each industry, if the operation data includes the industry of the user who uses the robot 2, the cost-effectiveness calculation step preferably estimates the economic effect and cost-effectiveness for each industry. In order to improve accuracy by estimating the cost-effectiveness for each type of robot 2, the cost-effectiveness calculation step preferably estimates the economic effect and cost-effectiveness for each type of robot 2. In order to improve accuracy by estimating the cost-effectiveness for each region in which the robot 2 is operated, the cost-effectiveness calculation step preferably estimates the economic effect and cost-effectiveness for each region. In order to further improve accuracy by estimating based on multiple conditions, estimates for each industry, type, or region may be made by combining two or more of these.

[0167] The support device 1 visualizes the positive impact that the introduction of the robot 2 under a fixed-price contract will have on users through the cost-effectiveness calculation step, encouraging the introduction of the robot 2. This allows the support device 1 to further reduce the economic burden associated with the use of the robot 2 controlled by the control machine learning model, further facilitating its widespread adoption. Therefore, the support process of this embodiment can reduce the economic burden associated with the introduction of AI that controls the robot 2 that provides advanced functions through the above-described process.

[0168] [Effect of support processing] The robot 2 of this embodiment provides various functions through cooperation between the sensor 24 and the drive unit 25. The robot 2 of this embodiment includes various robots 2 developed in line with advances in robotics, aimed at fields such as manufacturing, retail, welfare, medical care, logistics, hospitality, agriculture, construction, security, and others.

[0169] The robot 2 achieves desired functions by controlling its hardware with software that coordinates the sensors 24 and the drive unit 25. By using machine learning, the software that controls the robot 2 can provide more advanced functions. However, providing a machine learning model related to machine learning requires significant development costs for designing the machine learning model, collecting training data, and performing machine learning using the training data. Such large development costs increase the financial burden of purchasing a machine learning model and may hinder the widespread use of machine learning models.

[0170] In the support process of this embodiment, the fixed-price contract management unit 111 manages the rental period and payment status of the robot 2 (steps S1 to S6). As a result, the machine learning model providing unit 114 of this embodiment provides the robot 2 (rented robot) for which payment has been made with a control machine learning model used to control the robot 2, based on the above-mentioned payment status (steps S11 to S12). Therefore, the support device 1 of this embodiment can appropriately provide a control machine learning model despite being a fixed-price contract that reduces the economic burden associated with providing the robot and the control machine learning model.

[0171] Robots that provide advanced functions require improving their machine learning models using daily operational data. In the assistance process of this embodiment, the operational data acquisition unit 112 acquires operational data, including sensor data and control data, from the robot 2 rented under a fixed-price contract based on the rental period (steps S7 to S8). The machine learning unit 113 uses this operational data to perform machine learning on a control machine learning model (steps S9 to S10). Then, if payment for the fixed-price contract has been made and the rental period is still in progress, the machine learning model provision unit 114 provides the robot 2 with a control machine learning model that has been updated through the machine learning and is now capable of providing more advanced functions (steps S11 to S12). While control software in an initialized state is often difficult to use, the assistance device 1 of this embodiment provides a control machine learning model that has been updated through machine learning and is now capable of achieving more advanced functions, allowing the user of the robot 2 to use the robot 2 through easier-to-use software.

[0172] When the robot 2 is provided in various forms, such as by fixed-price contract or purchase, the operator of the robot 2 may be concerned that even operational data of the robot outside the rental period, such as by a purchased robot, may be passed on to the provider. The support process of this embodiment dispels this concern by acquiring operational data from the robot 2 rented under a fixed-price contract (steps S7 to S8).

[0173] The support process of this embodiment reduces the economic burden associated with using a robot 2 controlled by a control machine learning model through processing related to a fixed-price contract for the robot 2, thereby promoting its widespread use. Furthermore, the support process of this embodiment improves the control machine learning model to provide better control by using operational data collected from widespread robots in this way instead of training data, which is expensive to arrange. Thus, the support process of this embodiment can reduce the economic burden associated with introducing AI to control a robot 2 that provides advanced functions through processing that allows the control machine learning model that controls the robot 2 to be rented under a fixed-price contract.

[0174] The assistance process of this embodiment can be configured to extract frequently used items from operation data that further includes data related to command items for the robot 2 and provide the extracted items to the robot 2 (steps S13 to S15). This allows the assistance device 1 of this embodiment to improve the operation of the robot 2 so that frequently used items extracted based on a large amount of operation data collected under a fixed-fee contract are displayed preferentially. By further improving operation in this way in addition to the reduced burden provided by the fixed-fee contract and the provision of an excellent machine learning model, it is expected that operation data related to frequently used items will be more easily collected, and the assistance device 1 can further improve its control of frequently used items.

[0175] The assistance process of this embodiment can be configured to improve the language model by machine learning based on operation data including situation-appropriate language expressions and data related to the user's evaluation of the same, and provide the improved language model to the robot 2 (steps S16 to S18). This allows the assistance device 1 of this embodiment to improve the behavior of the robot 2 so that communication is performed using an improved language model based on a large amount of operation data collected through a flat-fee contract. The synergistic effect of the reduced burden due to the flat-fee contract, the provision of an excellent machine learning model, and such operation improvements is expected to further expand the use of the robot 2, making it easier to collect more operation data, and therefore the assistance device 1 can further improve control of command items related to improved communication.

[0176] The assistance process of this embodiment can be configured to generate instructions by machine learning based on operation data that further includes data related to command items for the robot 2, and provide the instructions to the user of the robot 2, etc. (steps S19 to S21). This allows the assistance device 1 of this embodiment to assist the user of the robot 2 with instructions generated based on a large amount of operation data collected through a fixed-price contract. In this way, it is expected that the robot 2 will be used more appropriately and operation data under appropriate use will be more easily collected, allowing the assistance device 1 to further improve control related to command items in the instructions.

[0177] <Example of use of the support device 1 of this embodiment> The following is an example of how the support device 1 of this embodiment is used.

[0178] [Registering Robot 2 and Fixed Price Contract] The administrator of the support device 1 registers data related to the robot 2 to be lent in the robot database 132. The administrator of the support device 1 also registers data related to the fixed-price contract for the robot 2 in the fixed-price contract database 131.

[0179] [Payment of fees] When it is time for billing, the support device 1 sends a bill for the usage fee to the user who is renting the robot 2 under a fixed-price contract. The user receives the bill via the terminal T and pays the usage fee. The support device 1 updates the payment status and renews the contract period according to the payment status.

[0180] [Providing control machine learning models and collecting operational data] The support device 1 provides a control machine learning model for controlling a rental robot to a rental robot that is currently under a fixed-price contract. The support device 1 also acquires operational data from a source robot for which payment is being made under a fixed-price contract.

[0181] [Improvement of control machine learning models] The assistance device 1 improves the control machine learning model through machine learning based on the acquired operation data. The assistance device 1 also improves the language model that outputs linguistic expressions from the robot 2 to the user through similar machine learning. The robot 2 improves its behavior using the improved control machine learning model or language model. In addition, the assistance device 1 extracts frequently used items from the operation data and provides them to the robot 2. The robot 2 improves the user interface provided to the user using the frequently used items.

[0182] [Generate instructions] The support device 1 generates an instruction manual for the robot 2 by processing using a language model based on the acquired operation data, and provides it to the user.

[0183] [Use in business situations] Support Device 1 can also be used in a variety of business scenarios, including data analytics platforms that collect and analyze data to help companies make decisions; SaaS (Software as a Service) businesses that provide industry-specific software to help customers analyze and optimize their data; data management consulting services where expert teams provide consulting services for data collection, analysis, and optimization; marketplace platforms that allow companies to share and trade their data with other companies; AI-based predictive analytics services that use generative AI to forecast future trends and demand; intelligent automation solutions that use AI to automate and streamline business processes; data security and privacy management solutions that enable companies to safely manage data and ensure compliance with regulations; multi-channel marketing platforms that use AI to optimize marketing campaigns across multiple channels (online advertising, social media, email marketing, etc.); data education and training services that offer online courses and workshops to help corporate employees acquire data analysis and AI skills; and customer data platforms (CDPs) that centrally manage customer data, analyze customer behavior, and develop personalized marketing strategies. These business models, which can be customized to meet different industries and needs, contribute to building sustainable businesses by providing diverse revenue streams.

[0184] It should be noted that within the scope of the concept of the present invention, those skilled in the art may conceive of various modifications and alterations. Therefore, it is understood that such modifications and alterations fall within the scope of the present invention. For example, even if a person skilled in the art appropriately adds, deletes, or modifies components of the above-described embodiment, or adds, omits, or modifies the conditions of a process, such modifications are also included within the scope of the present invention as long as they maintain the gist of the present invention. [Explanation of symbols]

[0185] S System 1 Support equipment 11 Control section 111 Fixed Price Contract Management Department 112 Operational Data Acquisition Unit 113 Machine Learning Department 114 Machine Learning Model Provision Department 115 Frequently used item extraction part 116 Frequently used item provision section 117 Instruction Generation Unit 13 Storage section 131 Fixed-price contract database 132 Robot Database 14 Communications Department 2. Robot 21 Control Unit 22 Memory section 23 Communications Department 24 sensors 25 Drive unit 26 Input section 27 Display section N Network T-Terminal

Claims

1. a fixed-price contract management unit that manages data relating to the rental period and payment status of fixed-price contracts for the rental of a plurality of robots; an operation data acquisition unit that acquires operation data of one acquisition source robot included in the plurality of robots when payment for the fixed-price contract for the acquisition source robot is being made; a machine learning unit that performs machine learning on a control machine learning model used to control a robot using learning data based on the operational data; a machine learning model providing unit that provides a control machine learning model that has been machine-learned by the machine learning unit to a rental robot included in the plurality of robots during the rental period of the fixed-price contract for the rental robot; Equipped with the operation data acquisition unit acquires operation data including sensor data and control data of the acquisition source robot, the machine learning unit performs machine learning using learning data that includes the sensor data as an explanatory variable and the control data as a target variable; the machine learning model providing unit provides a control machine learning model in which machine learning is performed using learning data based on a plurality of operation data including operation data acquired from other robots different from the one rental robot; A support device for introducing robot control AI.

2. the operation data acquisition unit acquires operation data including data indicating an area in which the acquisition source robot is operated; the machine learning unit performs machine learning of a control machine learning model associated with a specified region using learning data based on operation data related to the specified region; The machine learning model providing unit provides the rental robot with a control machine learning model associated with the area in which the rental robot is operated. The support device according to claim 1 .

3. a frequently used item extraction unit that extracts one or more frequently used items from among a plurality of command items for the robot; a frequently used item providing unit that provides the one or more frequently used items to the rental robot; Furthermore, the operation data acquisition unit acquires operation data further including a frequency of use of each command item for the acquisition source robot; the frequently used item extraction unit extracts the one or more frequently used items based on operation data further including a frequency of use of the command item; The support device according to claim 1 .

4. the operation data acquisition unit acquires operation data further including a linguistic expression output from the acquisition source robot to a user of the acquisition source robot, a situation related to the linguistic expression, and an evaluation by the user of the linguistic expression; the machine learning unit performs machine learning using learning data that includes the situation and the evaluation as explanatory variables and the linguistic expression as a target variable, and performs machine learning of a language model that outputs a linguistic expression that is expected to be highly evaluated when the situation is input; the machine learning model providing unit provides the language model to the rental robot; The support device according to claim 1 .

5. an instruction generation unit that receives an instruction item for the robot and the operation data as an input and generates an instruction manual indicating a method of use related to the instruction item by processing using a language model; the operation data acquisition unit acquires operation data including sensor data and control data of the acquisition source robot associated with the command item; The support device according to claim 1 .

Images