Product Design System

Abstract

We provide a product design system that can autonomously design products. [Solution] The product design system 30 includes a component design AI unit 31 that infers the components of a product by utilizing the AI's internal knowledge of components based on the environmental conditions under which the product will be used; an auxiliary design AI unit 32 that assists in the design of a product by utilizing the AI's internal knowledge of design; an autonomous design AI unit 33 that autonomously determines the knowledge lacking in the design of the product and cooperates with an external device possessing that knowledge; and a design agent AI unit 34 that designs the product by coordinating the component design AI unit 31, the auxiliary design AI unit 32, and the autonomous design AI unit 33.

Description

Technical Field

[0001] The present invention relates to a product design system.

Background Art

[0002] Conventionally, product design has been performed by a computer using a CAD (Computer-Aided Design) system. In recent years, various CADs utilizing artificial intelligence (AI: Artificial Intelligence) have also been developed (for example, Patent Document 1).

[0003] Recent AI is mainly based on a neural network that imitates the information processing mode of the brain, that is, the neurons and synapses of the brain. The neural network is composed of two-stage processing of learning and discrimination. In the learning stage, features are learned from a large number of inputs, and a learning model as a neural network is constructed. In the discrimination stage, a new input is discriminated using the constructed learning model. The neural network can improve the discrimination accuracy for an input by improving the accuracy of the learning model, but a large number of input information is required to improve the accuracy of the learning model, and the learning takes time. In addition, an AI system using a conventional neural network cannot operate by autonomously judging according to its own intention.

[0004] In contrast, the inventors of the present invention have invented an AI system capable of autonomously making decisions and operating according to its own will (Patent Document 2). This system comprises an information vessel and one or more knowledge units. Each of the one or more knowledge units comprises a knowledge storage unit for storing knowledge information, a knowledge management unit for managing the knowledge information stored in the knowledge storage unit so as to associate concepts with data values, an activation information determination means for determining a concept to be activated in the knowledge information when the information effectively held in the information vessel matches the data value of the concept in the knowledge information, based on an intention attribute that determines the concept to be activated in correspondence with the concept to be awakened in the knowledge information, and an information transmission means for providing the data value of the determined concept to the information vessel. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Patent No. 7355605 [Patent Document 2] Patent No. 7193870 [Overview of the project] [Problems that the invention aims to solve]

[0006] The CAD system disclosed in Patent Document 1 uses a conventional neural network-based AI system to predict manufacturing costs from design data. However, the design of the structure and mechanisms required for a product based on the environmental conditions in which the product will be used (structural design), and the design of the parts necessary to realize that structure and mechanism (parts design), have traditionally been performed manually. In contrast, there is a desire for a CAD system that autonomously performs structural and parts design based on the environmental conditions in which the product will be used, but neural networks, which require learning based on a vast amount of input information, are not suitable for such design.

[0007] This invention has been made in view of the above circumstances, and aims to provide a product design system that can autonomously design products. [Means for solving the problem]

[0008] To achieve this objective, a first aspect of the present invention is a product design system comprising: a component design AI unit that infers the components of a product by utilizing the knowledge of components that the artificial intelligence possesses internally, based on the environmental conditions under which the product will be used; a support design AI unit that assists in the design of the product by utilizing the knowledge of design that the artificial intelligence possesses internally; an autonomous design AI unit that determines, based on autonomous will, the knowledge that the artificial intelligence lacks in the design of the product and cooperates with an external device that possesses that knowledge; and a design agent AI unit that designs the product by coordinating the component design AI unit, the support design AI unit, and the autonomous design AI unit.

[0009] A second aspect of the present invention is a product design system according to the first aspect, wherein the component design AI unit comprises structural design reasoning means for inferring the structural specifications of the product based on the characteristics and / or features of the product to be designed.

[0010] A third aspect of the present invention is a product design system according to the second aspect, wherein, if the structural design reasoning means does not have knowledge of the structure of the product to be designed within the product design system, the autonomous design AI unit acquires knowledge of the structure of the product to be designed from the external device and infers the specifications of the structure of the product.

[0011] A fourth aspect of the present invention is a product design system according to any of the first to third aspects, wherein the component design AI unit infers the components of the product based on structural conditions, which are conditions that the structure of the product should have, including the environmental conditions.

[0012] A fifth aspect of the present invention is a product design system according to the fourth aspect, wherein the component design AI unit includes a condition change receiving means for receiving a change of the structural conditions from a designer if the inferred component of the product does not satisfy the structural conditions.

[0013] A sixth aspect of the present invention is a product design system according to any of the first to fifth aspects, wherein the component design AI unit links the design data of the product based on the inferred components of the product to a CAD system to generate drawings.

[0014] A seventh aspect of the present invention is a product design system according to any of the first to sixth aspects, wherein the support design AI unit internally holds design checklists and / or design rules as design knowledge.

[0015] An eighth aspect of the present invention is a product design system according to the seventh aspect, wherein the support design AI unit includes design checking means for identifying parts that deviate from the checklist and / or design rules in the design of the product using the parts of the product inferred by the parts design AI unit.

[0016] A ninth aspect of the present invention is a product design system according to the eighth aspect, wherein the support design AI unit includes a component modification means for modifying the components of the product to satisfy the checklist and / or design rules when the design checking means identifies a deviation from the checklist and / or design rules in the design of the product using the components of the product inferred by the component design AI unit.

[0017] A tenth aspect of the present invention is a product design system according to any one of the first to ninth aspects, wherein the autonomous design AI unit can be linked with at least one of the following external devices: a design support tool providing system that provides a design support tool for supporting the design of the product; an analysis tool providing system that provides an analysis tool for analyzing the designed product; an estimation system for estimating the manufacturing and / or processing costs of the designed product; and a knowledge database for storing technical knowledge. [Effects of the Invention]

[0018] According to the product design system of the first aspect of the present invention, a component design AI unit infers the components of a product based on the environmental conditions under which the product will be used, utilizing the knowledge of components that the artificial intelligence possesses. Furthermore, an auxiliary design AI unit provides support for product design by utilizing the design knowledge that the artificial intelligence possesses. In addition, an autonomous design AI unit autonomously determines the knowledge that the artificial intelligence lacks in product design and collaborates with an external device that possesses that knowledge. Finally, a design agent AI unit coordinates the component design AI unit, the auxiliary design AI unit, and the autonomous design AI unit to perform structural design and component design of the product while making autonomous decisions. This has the effect of enabling autonomous product design.

[0019] The product design system according to the second aspect of the present invention provides the following effects in addition to those of the product design system according to the first aspect. Specifically, the structural design reasoning means of the component design AI unit infers the structural specifications of the product based on the characteristics and / or features of the product to be designed. This has the effect of enabling the autonomous inference of structural specifications according to the characteristics and / or features of the product.

[0020] According to the product design system according to the third aspect of the present invention, in addition to the effects achieved by the product design system according to the second aspect, the following effects are achieved. That is, when the product design system does not have knowledge about the structure of the product to be designed, the autonomous design AI unit acquires knowledge about the structure of the product to be designed from an external device, and the structure design inference means infers the specifications of the product structure. As a result, if the product design system does not have knowledge about the structure of the product to be designed, it is possible to autonomously acquire that knowledge from an external device and perform the structure design of the product.

[0021] According to the product design system according to the fourth aspect of the present invention, in addition to the effects achieved by the product design system according to any one of the first to third aspects, the following effects are achieved. That is, based on the structural conditions that are the conditions that the structure of the product derived including environmental conditions should have, the parts of the product are inferred by the part design AI unit. As a result, it is possible to autonomously perform part design so as to satisfy the structural conditions that the product structure should have.

[0022] According to the product design system according to the fifth aspect of the present invention, in addition to the effects achieved by the product design system according to the fourth aspect, the following effects are achieved. That is, when the inferred parts of the product do not satisfy the structural conditions, the change of the structural conditions by the designer is received by the condition change reception means of the part design AI unit. As a result, it is possible to perform part design while interacting with the designer so as to satisfy the structural conditions.

[0023] According to the product design system according to the sixth aspect of the present invention, in addition to the effects achieved by the product design system according to any one of the first to fifth aspects, the following effects are achieved. That is, the design data of the product based on the inferred parts of the product is drawn in cooperation with the CAD system by the part design AI unit. As a result, the designer can easily confirm on the spot from the CAD system the product that has been autonomously structurally designed and part designed by the product design system.

[0024] According to the product design system according to the seventh aspect of the present invention, in addition to the effects achieved by the product design systems according to any of the first to sixth aspects, the following effects are achieved. That is, as knowledge related to design, it is possible to autonomously determine whether the product designed in terms of structure and components satisfies the checklist and / or design rules by utilizing the checklist and / or design rules related to design. As a result, there is an effect that product design can be performed in accordance with the checklist and / or design rules.

[0025] According to the product design system according to the eighth aspect of the present invention, in addition to the effects achieved by the product design system according to the seventh aspect, the following effects are achieved. That is, by the design check means of the support design AI unit, the parts of the product that deviate from the checklist and / or the design rules are specified for the design of the product using the parts of the product inferred by the parts design AI unit. As a result, there is an effect that the parts design AI unit and the support design AI unit can cooperate to enable parts design in accordance with the checklist and / or design rules.

[0026] According to the product design system according to the ninth aspect of the present invention, in addition to the effects achieved by the product design system according to the eighth aspect, the following effects are achieved. That is, when the design check means of the design agent AI unit specifies the parts of the product that deviate from the checklist and / or design rules for the design of the product using the parts of the product inferred by the parts design AI unit, the parts of the product are changed by the parts change means of the support design AI unit so as to satisfy the checklist and / or design rules. As a result, there is an effect that it is possible to autonomously enable parts design in accordance with the checklist and / or design rules.

[0027] The product design system according to the tenth aspect of the present invention provides the following effects in addition to the effects of the product design system according to any one of the first to ninth aspects. Specifically, the autonomous design AI unit can be linked with at least one of the following external devices: a design support tool providing system that provides design support tools for supporting product design; an analysis tool providing system that provides analysis tools for analyzing the designed product; an estimation system that estimates the manufacturing and / or processing costs of the designed product; and a knowledge database that stores technical knowledge. As a result, if there is a lack of knowledge in at least one of the following areas: product design support, analysis of the designed product, estimation of the manufacturing costs of the designed product, and technical knowledge, the system can autonomously link with the external device to acquire the missing knowledge and reliably provide support for product design. [Brief explanation of the drawing]

[0028] [Figure 1] This block diagram shows the configuration of an application system, including an AI unit. [Figure 2] This is a block diagram showing the configuration of the AI ​​unit. [Figure 3] Figures (a) to (d) are examples of knowledge management tables stored in the knowledge units of an AI unit. [Figure 4] This flowchart shows the processing flow executed in the processing unit of each knowledge unit within the AI ​​unit. [Figure 5] (a) is a flowchart showing the flow of the activation process, and (b) is a flowchart showing the flow of the task-dependent learning process. [Figure 6] (a) is a flowchart illustrating the flow of a self-regulating learning process, and (b) and (c) are diagrams showing how a new concept is added to the knowledge management table. [Figure 7] This figure schematically shows the configuration of a product design system according to one embodiment of the present invention. [Figure 8](a) is a flowchart showing the flow of the design agent AI processing performed by the design agent AI unit of the product design system, and (b) is a schematic diagram of the selection screen displayed by the design agent AI processing. [Figure 9] This visualizes the flow of component design AI processing performed by the component design AI unit of the product design system. [Figure 10] (a) and (b) are schematic diagrams of the environmental condition input screen displayed by the AI ​​processing for the part design. [Figure 11] (a) to (c) are schematic diagrams of the AI-generated part design dialogue screen displayed by the AI-generated part design process. [Figure 12] (a) is a schematic diagram of the contents of a CSV file used when registering knowledge information about parts, and (b) is a schematic diagram of a part of the display of the environmental conditions input screen when new knowledge about parts is registered. [Figure 13] This flowchart shows the flow of the support design AI processing performed by the support design AI unit of the product design system. [Figure 14] (a) and (b) are schematic diagrams of the AI-assisted design dialogue screen displayed by the AI-assisted design processing. [Figure 15] This is a schematic diagram of the AI-assisted design dialogue screen displayed by the AI-assisted design processing. [Figure 16] This is a schematic diagram of the AI-assisted design dialogue screen displayed by the AI-assisted design processing. [Figure 17] This flowchart shows the flow of autonomous design AI processing performed by the autonomous design AI unit of the product design system. [Figure 18] (a) and (b) are schematic diagrams of the autonomous design AI dialogue screen displayed by the autonomous design AI processing. [Figure 19] (a) and (b) are schematic diagrams of the autonomous design AI dialogue screen displayed by the autonomous design AI processing. [Figure 20](a) and (b) are schematic diagrams of the environmental condition input screen when designing a product with a new structure using the knowledge registration button in the component design AI unit, and (c) is a schematic diagram of the contents of the CSV file used to make those design instructions. [Figure 21] (a) and (b) are schematic diagrams of the autonomous design AI dialogue screen when the autonomous design AI unit acquires structural analysis of the product from external knowledge as missing knowledge. [Figure 22] This is a schematic diagram of the autonomous design AI dialogue screen when the autonomous design AI unit acquires structural analysis of the product from external knowledge sources due to a lack of knowledge. [Modes for carrying out the invention]

[0029] Hereinafter, embodiments for carrying out the present invention will be described with reference to the accompanying drawings. The embodiments described below are all preferred specific examples of the present invention. Therefore, the numerical values, shapes, materials, components, arrangement positions of components, and connection configurations shown in the following embodiments are examples and are not intended to limit the present invention. Accordingly, among the components in the following embodiments, those not described in the independent claims representing the highest-level concept of the present invention will be described as optional components. Furthermore, in each figure, substantially identical components are denoted by the same reference numerals, and redundant explanations are omitted or simplified.

[0030] <1. About AI Unit 110> <1.1. Basic Configuration of AI Unit 110> Before describing a product design system 30 (see Figure 7) according to one embodiment of the present invention, the basic configurations of the component design AI unit 31, the support design AI unit 32, the autonomous design AI unit 33, and the design agent AI unit 34 that constitute this product design system 30 will be explained with reference to Figures 1 to 6, using the AI ​​unit 110 shown in Figures 1 and 2 as an example. First, Figure 1 is a block diagram showing the configuration of the application system 100 including the AI ​​unit 110. Figure 2 is a block diagram showing the configuration of the AI ​​unit 110.

[0031] This AI unit 110 is an artificial intelligence system disclosed in Patent Document 2, which gives intelligence itself active functions and volitional attributes, and realizes an inference model in which intelligence itself performs volitional activation and autonomous learning based on the information it perceives. The basic concept is as described in Patent Document 2.

[0032] This section primarily describes the general configuration of the AI ​​unit 110 necessary for constructing the product design system 30. Furthermore, as an example of an application system 100 using the AI ​​unit 110, we will explain the "Japanese Flower Encyclopedia System," which operates based on knowledge of flowers. The product design system 30 utilizes the AI ​​unit 110 to construct the component design AI unit 31, the support design AI unit 32, the autonomous design AI unit 33, and the design agent AI unit 34, as part of the application system 100.

[0033] In Figure 1, the application system 100 includes an AI unit 110, as well as an interface device 120 (PC, smartphone, etc.) which serves as the interface (UI) between the AI ​​unit 110 and the user. The interface device 120 and the AI ​​unit 110 are connected via wired and / or wireless connections, enabling information to be sent and received directly or via a network such as the Internet.

[0034] The interface device 120 has an input analysis function that analyzes user input and extracts information from that input, and an output generation function that generates output for the user from the information output from the AI ​​unit 110. The interface device 120 displays the input information and the output information from the AI ​​unit 110 in relation to that input information, or outputs it as audio, etc.

[0035] The AI ​​unit 110 includes an information processing device 111 and an information storage device 112. The information processing device 111 has the functions of knowledge reasoning, which is the process by which knowledge evaluation is performed through the awakening and activation of knowledge, and active learning, which is the process by which learning requests are made based on the learning motivation of knowledge intelligence. The information processing device 111 can be configured using various known hardware resources, such as a von Neumann or non-von Neumann computer (including a CPU), and the software contained therein.

[0036] The information storage device 112 stores various types of knowledge information (knowledge database), a knowledge management table used for managing that knowledge information, and other information necessary for processing by the information processing device 111, as described later.

[0037] The AI ​​unit 110 is configured as shown in Figure 2, specifically through the functions of the information processing device 111 and the information storage device 112. That is, the AI ​​unit 110 has an information vessel 10 and a plurality of knowledge units 20(1), 20(2), 20(3), ... 20(n). Each knowledge unit 20(i) (where i is a natural number from 1 to n) is coupled to the information vessel 10 in a manner that allows for the transmission and reception of information.

[0038] The information vessel 10 holds the provided information as valid information in a state where it can be output. The information vessel 10 is connected to the interface device 120 via an information input unit 11, an information output unit 12, an educational input unit 13, and a learning output unit 14. Information from the interface device 120 is provided to the information vessel 10 through the information input unit 11, and the information effectively held in the information vessel 10 is output to the interface device 120 through the information output unit 12.

[0039] Furthermore, when learning request information representing a learning request for a certain concept constituting knowledge is provided to the information vessel 10 from any of the knowledge units 20(i), that learning request information is output to the interface device 120 through the learning output unit 14. The interface device 120 provides educational information to the information vessel 10 through the educational input unit 13. The knowledge unit 20(i) that provided the learning request information to the information vessel 10 can then take in the educational information provided to the information vessel 10.

[0040] The information vessel 10 can be configured, for example, using a database, and the information registered in this database (information vessel 10) will be retained as valid information.

[0041] Each knowledge unit 20(i) corresponds to one piece of knowledge (for example, knowledge about Japanese flowers) and includes a processing unit 21, a knowledge storage unit 22, and a communication unit 23. The processing unit 21 is configured as one function of the information processing device 111 (see Figure 1) and performs processing related to the activation / learning of concepts based on volitional attributes. The knowledge storage unit 22 stores a knowledge management table that manages knowledge information, including data values ​​representing each of the multiple concepts that constitute a piece of knowledge. The knowledge storage unit 22 is also used to store information necessary for processing by the information processing device 111.

[0042] Here, we will explain the outline of the knowledge management table with reference to Figure 3. Figures 3(a) to 3(d) are diagrams showing examples of knowledge management tables. For example, one piece of knowledge information, "The flower that is pink in the season of spring is a cherry blossom," is managed by a knowledge management table (knowledge management unit) that associates the concept "flower" with the data value "cherry blossom," the concept "season" with the data value "spring," and the concept "color" with the data value "pink," as shown in Figure 3(a). In addition, one piece of knowledge information is managed as knowledge about a representative concept. For example, the above-mentioned knowledge, "The flower that is pink in the season of spring is a cherry blossom," is managed as knowledge about the representative concept "flower" by the knowledge management table shown in Figure 3(a).

[0043] Furthermore, knowledge information representing the representative concept "flower," such as "The flower that is purple in the spring season is wisteria," can be managed by a knowledge management table that associates the concept "flower" with the data value "wisteria," the concept "season" with the data value "spring," and the concept "color" with the data value "purple," as shown in Figure 3(b). Similarly, knowledge information representing the representative concept "flower," such as "The flower that is pink in the spring season is lotus," can be managed by a knowledge management table that associates the concept "flower" with the data value "lotus," the concept "season" with "spring," and the concept "color" with the data value "pink," as shown in Figure 3(c).

[0044] Furthermore, knowledge information representing the representative concept "event," such as "the event that takes place in spring and features cherry blossoms is the entrance ceremony," can be managed by a knowledge management table that associates the concept "event" with the data value "entrance ceremony," the concept "season" with the data value "spring," and the concept "flower" with the data value "cherry blossom," as shown in Figure 3(d).

[0045] Returning to Figure 2, the communication unit 23, under the control of the processing unit 21, inputs information effectively held in the information vessel 10 (receiving unit) and also outputs information to the information vessel 10 (providing unit).

[0046] <1.2. Operation of AI Unit 110> <1.2.1. The willpower activation effect of AI unit 110> Next, the operation of the application system 100, including the AI ​​unit 110 described above, will be explained. When a user inputs information related to a question (for example, "What is a pink spring flower?") into the interface device 120, the interface device 120 displays the question, outputs it as audio, and analyzes the question to extract the information contained in it. For example, the information "pink," "spring," and "flower" is extracted from the question. The interface device 120 then outputs the extracted information (for example, information related to "flower," "pink," and "spring") to the AI ​​unit 110. The information output from the interface device 120 (for example, "pink" and "spring") is provided to the information vessel 10 through the information input unit 11 in the AI ​​unit 110 and is held (registered) in the information vessel 10 as valid information.

[0047] Now, referring to Figures 4 to 6, we will explain the processing performed in the processing unit 21 of each knowledge unit 20(i) of the AI ​​unit 110. First, Figure 4 is a flowchart showing the flow of that processing.

[0048] When information is effectively stored (registered) in the information vessel 10, the processing unit 21 takes in the information stored in the information vessel 10 via the communication unit 23 (receiver) and determines whether the information matches a data value of any of the concepts managed in the knowledge management table (see Figures 3(a) to (d)) (S11).

[0049] If the information held in the information vessel 10 matches the data value of any concept managed in the knowledge management table (S11: Yes), the processing unit 21 confirms that the concept whose data value matches the information held in the information vessel 10 has been awakened (S12), and determines whether or not other concepts should be activated as a result of the awakening of that concept (S13). Each knowledge unit 20(i) has a will attribute that determines which concept to activate when a certain concept is awakened (for example, "When the concepts "season" and "color" are awakened, activate the concept "flower""). This will attribute may be defined in the program executed by the processing unit 21, in the form of a decision table in the information storage device 112, or in any other form.

[0050] If, based on the determination in S13, it is determined that a certain concept included in the knowledge information should be activated according to the activation conditions (for example, when the concepts "season" and "color" are awakened) (S13: Yes), the processing unit 21 executes the activation process (S14). Figure 5(a) is a flowchart showing the flow of this activation process (S14). In the activation process (S14), based on the will attribute, the concept to be activated in correspondence with the awakened concept in one piece of knowledge is determined (S141). Then, the processing unit 21 reads the data value corresponding to the concept to be activated from the knowledge storage unit 22 by referring to the knowledge management table and provides that data value to the information vessel 10 via the communication unit 23 (provider) (S142).

[0051] Returning to Figure 4, let's continue the explanation. After the process in S14, the processing unit 21 repeatedly executes the process described above, while confirming that the predetermined stop conditions (e.g., loss of knowledge, termination of knowledge function, etc.) are not met (S17: No). If, during this process, the aforementioned stop conditions such as loss of knowledge or termination of function are met (S17: Yes), the process described above is terminated.

[0052] As described above, the activated conceptual data values ​​provided by the knowledge unit 20(i) and effectively held in the information vessel 10 are output from the information vessel 10 to the interface device 120 through the information output unit 12. When the interface device 120 receives the data values ​​as input, it generates information representing the answer to the question using those data values ​​and outputs the answer as text, voice, etc. This allows the user to obtain the answer to the question from the AI ​​unit 110.

[0053] This operation will be explained using a concrete example. For instance, when a user inputs information related to the question "What is a pink spring flower?" into the interface device 120, the interface device 120 extracts the information "pink," "spring," and "flower" from the question. The interface device 120 then inputs the extracted information "pink" and "spring" into the AI ​​unit 110 as information related to the "flower." This information is provided to the information vessel 10 through the information input unit 11 in the AI ​​unit 110 and is held (registered) in the information vessel 10 as valid information.

[0054] Here, in knowledge unit 20(i), knowledge information is managed in the knowledge management table shown in Figure 3(a), and the intention attribute "When the concepts "season" and "color" are awakened, activate the concept "flower"" is defined. As described above, when the information vessel 10 holds the information for "spring" and "pink," the knowledge information (see Figure 3(a)) confirms that the concept "season" for the data value "spring" and the concept "color" for the data value "pink" are awakened concepts (S11, S12 in Figure 4). Based on the will attribute "When the concepts 'season' and 'color' are awakened, activate the concept 'flower'", the knowledge unit 20(i) determines that the concept "flower" is the concept to be activated (S13, S14 in Figure 4 (S141 in Figure 5(a))), and the data value "cherry blossom" of the activated concept "flower" in the knowledge information is provided from the knowledge unit 20(i) to the information vessel 10 (S142 in Figure 5(a)), and is effectively held in the information vessel 10.

[0055] At this time, knowledge information is managed in the knowledge management table shown in Figure 3(c), and if there is another knowledge unit 20(j) with the will attribute "When the concepts 'season' and 'color' are awakened, activate the concept 'flower'", the same procedure (S11-S14 in Figure 14) is performed in that knowledge unit 20(j) as well. As a result, when the information vessel 10 effectively holds the information for "spring" and "pink", the concept for "flower" is activated in accordance with the awakened concepts for "season" and "color", and the data value "lotus" for that concept for "flower" is provided to the information vessel 10 and effectively held in that information vessel 10.

[0056] As described above, the data values ​​"Sakura" and "Rengesou," which are effectively held in the information vessel 10, are output to the interface device 120 through the information output unit 12. When the interface device 120 receives the data values ​​"Sakura" and "Rengesou" as input, it generates information representing the answer to the aforementioned question using those data values, for example, "It's a Sakura" and "It's a Rengesou," and outputs the sentences representing the answer, "It's a Sakura" and "It's a Rengesou," by displaying them, speaking them, etc. In this way, the user can obtain the information "Sakura" and "Rengesou" from the AI ​​unit 110 as the answer to the question, "What is a pink spring flower?"

[0057] Furthermore, in addition to the knowledge units 20(i) and 20(j) mentioned above, if there is another knowledge unit 20(k) whose knowledge information is managed in the knowledge management table shown in Figure 3(d) and whose will attribute "When the concepts 'season' and 'flower' are awakened, activate the concept 'event'" is defined, the same procedure (see Figure 4) is followed for that knowledge unit 20(k) as well. However, the awakening of concepts is based on data values ​​that are effectively held in the information vessel 10 by the activation of concepts in other knowledge units 20(i), as will be explained below.

[0058] In other words, as mentioned above, even if the information "spring" and "pink" is effectively stored in the information vessel 10 based on the question "What is a pink spring flower?" from the interface device 120, the knowledge unit 20(k) corresponding to knowledge that does not have the concept of "color" will not respond to the information "pink" (S11: No in Figure 4). Also, at this stage, the information stored in the information vessel 10 does not contain (is not respond to) the data value of the concept "flower," so the activation condition (when the concepts "season" and "flower" are awakened) is not satisfied (S13: No in Figure 4), and the activation process based on the will attribute (S14 in Figure 4) is not performed.

[0059] However, as mentioned above, the data value "Sakura" for the activated concept "Flower" is provided to the information vessel 10 from knowledge unit 20(i), and this data value "Sakura" is effectively held in the information vessel 10. Then, knowledge unit 20(k), which corresponds to the knowledge that has the concept of "Flower," responds to "Sakura" provided by knowledge unit 20(i) along with "Spring," which is effectively held in the information vessel 10 (S11: Yes in Figure 4), and the concepts "Season" and "Flower" are confirmed as awakened concepts (S12 in Figure 4). Here, the activation condition (when the concepts "season" and "flower" are awakened) is satisfied (S13: Yes in Figure 4), so the activation process (S14 in Figure 4, S141 and S142 in Figure 5(a)) is executed, and based on the will attribute "Activate the concept "event" when the concepts "season" and "flower" are awakened," the concept "event" is activated in the managed knowledge information (see Figure 3(d)), and the corresponding data value "entrance ceremony" is provided to the information vessel 10 and is effectively held in that information vessel 10.

[0060] Then, the data value "Sakura" from knowledge unit 20(i) and the data value "Entrance Ceremony" from knowledge unit 20(k), which are effectively held in the information vessel 10, are output to the interface device 120 through the information output unit 12. When the interface device 120 receives the data values ​​"Sakura" and "Entrance Ceremony" as input, it generates information representing the answer to the aforementioned question using those data values, for example, "It's Sakura" and "There is an entrance ceremony," and outputs the sentences representing the answer, "It's Sakura" and "There is an entrance ceremony," by displaying them, speaking them, etc. In this way, the user can obtain from the AI ​​unit 110 the information "Sakura" along with the information about the event "Entrance Ceremony," which is related to the spring flower, Sakura, as an answer to the question "What is a pink spring flower?".

[0061] <1.2.2. Task-dependent learning of AI unit 110> By the way, from the information extracted from the question provided by the user to the interface device 120 (for example, "What are spring flowers?") and effectively stored in the information vessel 10 (for example, "spring"), the knowledge unit 20(i) (for example, a knowledge unit having the will attribute "Activate the concept "flower" when the concepts "season" and "color" are awakened") cannot determine which concept to activate based on the will attribute due to a lack of awakened concepts (for example, the concept "color"). In processing according to the procedure shown in Figure 4, the processing unit 21 of such a knowledge unit 20(i) determines that the information effectively stored in the information vessel 10 does not satisfy the conditions for activating a concept (S13: No), and further determines whether the learning conditions based on the learning will attribute, which will be described later, are met (S15). If the learning conditions are met (S15: Yes), the processing unit 21 executes a task-dependent learning process (S16) related to learning (active learning).

[0062] Task-dependent learning is learning about a missing task, specifically, learning about a missing concept (the missing task) when there is a lack of other concepts to activate a particular concept within the knowledge information.

[0063] Each knowledge unit 20(i) can define a learning intention attribute that instructs it to learn a missing concept when it is not possible to determine which concept to activate based on intention attributes due to a lack of awakened concepts. For example, in a knowledge unit 20(i) defined with the aforementioned intention attribute "Activate the concept 'flower' when the concepts 'season' and 'color' are awakened," the learning intention attribute "Learn the concept 'color' when the concept 'season' is awakened but the concept 'color' is not" is defined. This learning intention attribute, like the aforementioned intention attributes, may be defined in the program executed by the processing unit 21, in the form of a decision table in the information storage device 112, or in any other form.

[0064] In such a knowledge unit 20(i), if the conditions for task-dependent learning (for example, the condition that the concept "season" is awakened but the concept "color" is not) are met based on the learning intention attribute (S15:Yes), the processing unit 21 executes task-dependent learning processing based on that learning intention attribute (S16).

[0065] Here, Figure 5(b) is a flowchart showing the flow of the task-dependent learning process (S16). In the task-dependent learning process (S16), the processing unit 21 confirms the missing concept (for example, the concept "color") as a concept to be learned based on the learning intention attribute (for example, if the concept "season" is awakened but the concept "color" is not awakened, learn the concept "color") (S161), and provides learning request information representing the learning request for that concept to the information vessel 10 via the communication unit 23 (provider) (S162). After that, the processing unit 21 enters a waiting state for educational information, which will check again whether the information provided as educational information from the interface device 120 is effectively held in the information vessel 10 (S163).

[0066] Meanwhile, learning request information provided from the knowledge unit 20(i) to the information vessel 10 (for example, a learning request about the concept "color") is output to the interface device 120 through the learning output unit 14. When the interface device 120 receives learning request information, it generates a sentence related to the information request (for example, "What color is the flower?") based on that learning request information, and outputs the sentence related to the information request by displaying it, speaking it, or the like.

[0067] When a user inputs information that answers this information request (for example, "The color is pink") into the interface device 120, the interface device 120 extracts the information contained in the input (for example, "color" and "pink") and outputs the extracted information (for example, the word "pink" related to "color") to the AI ​​unit 110.

[0068] The information output to the AI ​​unit 110 (for example, "pink") is provided to the information vessel 10 via the educational input unit 13 in the AI ​​unit 110, and that information (for example, "pink") is effectively stored (registered) in the information vessel 10 as educational information.

[0069] When the information as educational information is effectively stored in the information vessel 10 (S163:Yes), the processing unit 21, which was in a waiting state for educational information as described above, takes in the information as educational information (for example, "pink") from the information vessel 10 via the communication unit 23 (receiver). The processing unit 21 then refers to the knowledge management table and determines whether the educational information matches the data value of the missing concept (for example, the data value "pink" for the missing concept "color") (S164). If the information as educational information matches the data value of the missing concept (S164:Yes), the processing unit 21 starts the activation process (S14) and submits the missing concept as an awakened concept for the activation process (S14) (S165). As a result, the processing unit 21, with the missing awakened concept resolved, determines the concept to be activated based on the will attribute and transmits the data value of the activated concept (S14 in Figure 4, S141 and S142 in Figure 5(a)).

[0070] On the other hand, if the information as educational information does not match the data value of the missing concept (for example, the data value "pink" for the missing concept "color") (S164: No), the processing unit 21 terminates the processing related to learning (task-dependent learning) without initiating the activation process (S4). In this case, the knowledge unit 20(i) does not perform the activation process and does not return information related to the answer to the user's question to the user.

[0071] <1.2.3. Self-regulating learning of AI unit 110> Next, we will explain the process related to self-regulated learning, which involves learning about a concept when a concept for which no data value exists arises in the knowledge information (when the number of concepts constituting the knowledge increases). Self-regulated learning is the learning of a changed concept that occurs by self-requirement when a concept constituting a piece of knowledge changes (for example, when the number of concepts constituting the knowledge increases). Specifically, it is the learning of a concept when a concept for which data value is missing arises in the knowledge information (when the number of concepts constituting the knowledge increases).

[0072] For example, a user may want to add a concept to an existing piece of knowledge (for instance, knowledge about the representative concept "flower"). In this case, the user inputs information representing the concept to be added to that knowledge (for example, the concept "country" to be added to the knowledge about flowers) into the interface device 120. The interface device 120 then outputs definition information representing the concept to be added to that knowledge (for example, the knowledge about the representative concept "flower" includes the concept "country") to the AI ​​unit 110. The AI ​​unit 110 then effectively stores (registers) the definition information from the interface device 120 in the information vessel 10.

[0073] In the processing steps of each knowledge unit 20(i), if the information held in the information vessel 10 does not match any of the conceptual values ​​managed in the knowledge management table (S11: No), the processing unit 21 in each knowledge unit 20(i) determines whether the definition information effectively held in the information vessel 10 is definition information representing a new concept in the knowledge of a representative concept (for example, "flower") managed in its own knowledge management table (see Figures 3(a) to (d)) (S18).

[0074] Here, if the definition information held in the information vessel 10 is definition information representing a new concept in its own knowledge (S18: Yes), the processing unit 21 takes in the definition information via the communication unit 23 (receiver) and adds the concept contained in that definition information (for example, the concept "country") as a concept that constitutes knowledge represented by knowledge information managed in the knowledge management table (see Figures 3(a) to (d)) (for example, knowledge of the representative concept "flower") (S19). Specifically, the concept represented by the definition information is added to the knowledge management table. For example, Figure 6(b) shows the state in which the concept "country" has been added to the knowledge management table in Figure 3(a) according to the definition information. As a result of the addition of the concept by the process in S19, as shown in Figure 6(b), a state arises in which a concept (for example, the concept "country") with a corresponding data value is missing in the knowledge information managed in the knowledge management table.

[0075] When a concept with missing data values ​​occurs in the knowledge information, the processing unit 21 executes a self-adjusting learning process (S20). Figure 6(a) is a flowchart showing the flow of the self-adjusting learning process (S20). In this self-adjusting learning process (S20), the processing unit 21 refers to the knowledge management table and confirms that the concept with missing data values ​​in the knowledge information (for example, the concept "country") is a concept to be learned (S201).

[0076] Then, the processing unit 21 provides learning request information to the information vessel 10 via the communication unit 23 (provider) that represents the learning request for the concept that has been confirmed as the concept to be learned (for example, "country") (S202). The learning request information indicates that it requests the data value of a concept (for example, "country") that has been added to the knowledge (for example, knowledge about the representative concept "flower" for the data value "cherry blossom"). Subsequently, the processing unit 21 enters a waiting state for educational information, repeatedly checking whether the educational information is effectively held in the information vessel 10 in response to the provision of the learning request information (S203).

[0077] Meanwhile, the learning request information provided from the knowledge unit 20(i) to the information vessel 10 is input to the interface device 120 through the learning output unit 14. When the interface device 120 receives the learning request information, it generates information related to the information request (for example, "What is the country of cherry blossoms?") based on that learning request information and outputs the text related to that information request by displaying it, speaking it, etc. When the user inputs information that answers this learning request (for example, "The country is Japan") to the interface device 120, the interface device 120 extracts the information contained in that input (for example, "country" and "Japan") and outputs the extracted information (for example, the word "Japan" related to "country") to the AI ​​unit 110. This information (for example, "Japan") is provided to the information vessel 10 through the educational input unit 13 in the AI ​​unit 110, and the information of that word (for example, "Japan") is effectively held (registered) in the information vessel 10 as educational information.

[0078] When the information vessel 10 effectively holds information as educational information (S203: Yes), the processing unit 21 of the knowledge unit 20(i), which was in a waiting state for educational information, takes in the information as educational information (for example, "Japan") from the information vessel 10 via the communication unit 23 (receiver). The processing unit 21 then stores the taken-in information (for example, "Japan") as a data value for an additional concept (for example, the concept "Country") in the knowledge storage unit 22, and rewrites the knowledge management table so that the taken-in information (for example, "Japan") is associated with the additional concept (for example, "Country") as a data value (S204). This updates the knowledge information. For example, Figure 6(c) shows the state in which a data value (for example, "Japan") has been added to the concept "Country" in the knowledge management table of Figure 6(b) through self-adjusting learning.

[0079] Subsequently, the processing unit 21 outputs completion information for knowledge updating to the information vessel 10 via the communication unit 23 (provider) (S205). This completion information is provided to the interface device 120, which outputs completion information (for example, "Registered"). This allows the user to know that a new concept (for example, the concept "Country" for the data value "Japan") has been added to the specified knowledge (for example, knowledge about cherry blossoms).

[0080] According to the AI ​​unit 110 described above, each knowledge unit 20(i) responds to the data values ​​of one or more concepts provided to the information vessel 10 (awakening of concepts) based on its will attribute by transmitting data values ​​to the information vessel 10 (activation) as an activity based on one piece of knowledge (knowledge activity). This enables autonomous actions (knowledge reasoning function) that are judged by the unit's own will (will attribute).

[0081] Furthermore, if each knowledge unit 20(i) cannot transmit (activate) a data value using only the data value of a concept provided to the information vessel 10, it makes a learning request for the missing data value of the concept based on the learning intention attribute, and obtains the missing data value from the educational information provided in response to that learning request. Then, taking into account the missing data value of the concept, it transmits (activates) a data value to the information vessel 10. Thus, according to the AI ​​unit 110 described above, autonomous task-dependent learning (active learning) becomes possible based on one's own will (learning intention attribute).

[0082] Furthermore, in each knowledge unit 20(i), if a concept with missing data values ​​occurs in the knowledge it manages, it assesses the situation, makes a learning request for the concept with missing data values, and supplements the data values ​​of the concept related to that learning with educational information provided in response to that learning request. Thus, according to the AI ​​unit 110 described above, autonomous, self-regulated learning (active learning) becomes possible at the user's own discretion.

[0083] <2. Configuration of Product Design System 30> <2.1. Overview of Product Design System 30> Next, with reference to Figure 7, the configuration of a product design system 30 according to one embodiment of the present invention will be described. Figure 7 is a schematic diagram showing the configuration of the product design system 30.

[0084] The product design system 30 is a system that designs products by coordinating and operating three design AI units, namely a component design AI unit 31, an assistance design AI unit 32, and an autonomous design AI unit 33, through a single design agent AI unit 34.

[0085] The component design AI unit 31, the support design AI unit 32, the autonomous design AI unit 33, and the design agent AI unit 34 are each composed of the AI ​​unit 110 shown in Figure 2. The component design AI unit 31, the support design AI unit 32, the autonomous design AI unit 33, and the design agent AI unit 34 each realize intention activation (knowledge reasoning) and active learning (task-dependent learning and self-regulated learning).

[0086] <2.2. Component Design AI Unit 31> The component design AI unit 31 is a design AI unit that, based on the environmental conditions under which the product will be used, utilizes the AI ​​unit 110, which functions as artificial intelligence, to infer product components that satisfy those environmental conditions by leveraging knowledge about the components (component knowledge), etc.

[0087] The component design AI unit 31 possesses component knowledge as internal knowledge. In addition to component knowledge, the component design AI unit 31 may also possess product knowledge as internal knowledge. That is, each of the knowledge units 20(i) (see Figure 2) possessed by the component design AI unit 31 corresponds to knowledge about a single component (component knowledge) or knowledge about a single product (product knowledge).

[0088] The component design AI unit 31 displays an environmental condition input screen 50 (see Figure 10) on a display device (not shown) of the product design system 30, which accepts input of the environmental conditions under which the product will be used when designing the product.

[0089] Furthermore, before inferring the components of a product, the component design AI unit 31 performs structural design inference to infer the specifications of the structure and mechanism necessary to realize the characteristics and features of the product being designed and / or to satisfy the input environmental conditions. Alternatively, instead of inferring the specifications of the structure and mechanism, structural design may be performed to directly design the specifications of the structure and mechanism. This structural design inference or structural design does not have to be performed by artificial intelligence, including the AI ​​unit 110, but may also be performed by processing on a von Neumann type computer.

[0090] For example, the component design AI unit 31 may be equipped with a conceptual design API as an API (Application Programming Interface), and structural design reasoning or structural design may be performed in a structural design application via this conceptual design API. Alternatively, structural design reasoning or structural design may be performed by a structural design program pre-installed by the component design AI unit 31. Furthermore, the component design AI unit 31 may perform structural design reasoning using product knowledge it possesses as internal knowledge.

[0091] Furthermore, structural design reasoning or structural design may be performed by acquiring technical knowledge of the product's structure, which is provided as external knowledge by the autonomous design AI unit 33 (described later) in cooperation with a knowledge database, which is one of the external devices located outside the product design system 30, or it may be performed in cooperation with a design support system, which is one of the external devices. In this case, the component design AI unit 31 will perform structural design reasoning or structural design in cooperation with the autonomous design AI unit 33 via the design agent AI unit 34.

[0092] Furthermore, if the structural and mechanical specifications of the product to be designed are fixed (for example, if only long umbrellas are to be designed), the component design AI unit 31 may omit structural design reasoning or structural design by pre-memorizing the structural and mechanical specifications of that product as knowledge.

[0093] The component design AI unit 31 performs structural calculations of the product to be designed based on the environmental conditions entered by the designer on the environmental conditions input screen 50, the structural design reasoning, or the specifications of the product's structure and mechanism obtained through structural design, or the specifications of the product's structure and mechanism that it already possesses as knowledge. Then, based on the conditions (structural conditions) that each structure and mechanism should have derived from the structural calculations, the component design AI unit 31 performs component design reasoning to infer the combination of parts to be used in the product from the internal knowledge of parts, thereby inferring a combination of parts that can satisfy the environmental conditions entered by the designer. This component design reasoning is performed based on the intention attributes of each knowledge unit 20(i) that constitutes the component design AI unit 31.

[0094] Furthermore, this component design reasoning method infers not just one, but multiple combinations of components that satisfy the structural conditions. This allows the designer to select the most suitable combination of components for the product from among the inferred combinations. In addition, this component design reasoning method can also infer combinations of components that do not necessarily satisfy the structural conditions. This allows the designer to reconsider the structural conditions themselves based on the combinations of components that do not satisfy the conditions, and then perform the component design reasoning again to find a more suitable combination of components.

[0095] The component design AI unit 31 controls the component design AI dialogue screen 60 (see Figure 11) to be displayed on a display device (not shown) of the product design system 30, which presents the inferred component combinations to the designer. At this time, the component design AI unit 31 determines whether the inferred component combination satisfies the structural conditions for each component combination, and displays the result of this determination on the component design AI dialogue screen 60. The structural conditions that formed the basis of the component design inference are also displayed on the component design AI dialogue screen 60.

[0096] The component design AI unit 31 is configured to allow the designer to change structural conditions via the component design AI dialogue screen 60, based on information displayed on the component design AI dialogue screen 60, such as the content of the inferred component combinations and whether or not those component combinations satisfy the structural conditions. The component design AI unit 31 is then configured to re-infer (re-infer) the component combinations to be used in the product based on the changed structural conditions.

[0097] In this way, if the component design AI unit 31 infers a combination of components that does not satisfy the structural conditions calculated for the product, it autonomously decides to change those structural conditions and seeks instructions from the designer. Then, when the component design AI unit 31 receives instructions from the designer regarding the structural conditions, it continues the design by re-inferring the combination of components for the product based on the changed structural conditions.

[0098] Furthermore, the component design AI unit 31 is configured to allow designers to register new component knowledge via the environmental condition input screen 50, and can use this newly registered component knowledge to infer which components to use in the product.

[0099] The component design AI unit 31 also has a CAD-API as an API, and it causes the CAD system to create drawings of the component design inferred by the CAD-API. The designer can check whether there are any problems with the component design inference by reviewing the CAD drawings of the product created by the CAD system. If there are no problems with the component design inference, the component design AI unit 31 saves the design data of the component design inferred by the AI ​​unit 31 based on the instructions given to the designer.

[0100] <2.3. AI Unit 32 for Design Support> The support design AI unit 32 is a design AI unit that assists in product design by utilizing the design knowledge that the AI ​​unit 110, which functions as artificial intelligence, internally possesses as internal knowledge. Specifically, the support design AI unit 32 internally possesses design checklists and / or design rules as design knowledge.

[0101] The checklist is used to verify whether the design conditions and constraints required for the parts used in the product design are met. For example, a bolt requires a relief groove of a predetermined length, depending on the bolt's diameter and thread pitch, in order to achieve a complete mating with the mating part down to the root. The checklist checks whether the bolt registered as a part, or a bolt inferred as one of the parts to be used in the product, has a relief groove of the predetermined length. Design rules define the rules that must be followed in the design of a product. For example, they specify which parts can be mated and which cannot. Each of the knowledge units 20(i) (see Figure 2) of the support design AI unit 32 corresponds to one of the check items specified in the checklist and one of the rules specified in the design rules.

[0102] The AI ​​support design unit 32 utilizes design checklists and / or design rules as design knowledge (internal knowledge) to autonomously determine whether the structurally and componentally designed product satisfies the checklists and / or design rules. This enables product design in accordance with the checklists and / or design rules.

[0103] The supportive design AI unit 32 reads design data of products whose component design inference was performed by the component design AI unit 31, or products whose design assistance was provided by the autonomous design AI unit 33 using external knowledge from an external device, and has an evaluation tool that evaluates whether the product satisfies the checklist and / or design rules based on the read design data. This evaluation tool is constructed based on the intention attributes of each knowledge unit 20(i) that constitutes the supportive design AI unit 32.

[0104] Furthermore, the support design AI unit 32 has a dialogue tool for interacting with the designer. If the evaluation tool identifies any deviations from the checklist and / or design rules, the dialogue tool autonomously determines that the correction of those deviations is an item requiring instructions from the designer, and presents the correction to the designer via the support design AI dialogue screen 70.

[0105] When a correction to that section is accepted as an instruction from the designer, the conversational tool continues the product design by executing the correction. Here, the conversational tool offers two correction methods: AI-based correction and CAD-based correction. If the designer instructs for an AI-based correction, the conversational tool autonomously corrects the design of the section to be corrected based on a checklist and / or design rules. In this way, it can assist in the product design to satisfy the checklist and / or design rules while interacting with the designer.

[0106] Furthermore, if the designer instructs the AI ​​to make modifications in CAD, the dialogue tool accepts the modifications made by the designer to the CAD system and reflects them in the design data. In this case, the AI ​​support design unit 32 uses an evaluation tool to assess whether the modified design data satisfies the checklist and / or design rules. If there are no deviations from the checklist and / or design rules, the AI ​​support design unit 32 terminates the support design based on the designer's instructions. On the other hand, if there are still deviations from the checklist and / or design rules, the AI ​​support design unit 32 again uses the dialogue tool to ask the designer for instructions on how to make further modifications.

[0107] <2.4. Autonomous Design AI Unit 33> The autonomous design AI unit 33 is a design AI unit that, based on its autonomous will, determines the knowledge gaps in the product design of the AI ​​unit 110, which functions as artificial intelligence, and collaborates with external devices that possess that knowledge as external knowledge. The autonomous design AI unit 33 reads the product design data inferred by the component design AI unit 31 and the design data modified by the support design AI unit 32, and determines the knowledge gaps. The determination of whether or not there are knowledge gaps is made based on the knowledge information and will attributes of each knowledge unit 20(i) (see Figure 2) possessed by the autonomous design AI unit 33. In addition, the autonomous design AI unit 33 can also determine the knowledge gaps by receiving the knowledge gaps output from the component design AI unit 31 and the support design AI unit 32 via the design agent AI unit 34.

[0108] The autonomous design AI unit 33 is configured to be able to connect with external devices, including a design support tool provision system that provides design support tools to assist in product design as external knowledge, an analysis tool provision system that provides analysis tools including 3D-CAD analysis tools to analyze the designed product as external knowledge, an estimation system that estimates the manufacturing and / or processing costs of the designed product as external knowledge, and a knowledge database that stores technical knowledge such as book data including automotive engineering-related books and conference proceedings as external knowledge. Connection with external devices is made via a network such as the internet.

[0109] As a design support tool, it can be integrated with various well-known tools such as computer-aided engineering (CAE), computer-aided manufacturing (CAM), product lifecycle management (PLM), project management software, finite element analysis (FEA), computational fluid dynamics (CFD), electronic design automation (EDA), and 3D printing software.

[0110] If the autonomous design AI unit 33 does not possess the necessary analytical knowledge for product design within the product design system 30, it will autonomously determine that the analytical knowledge is lacking and, in cooperation with the analysis tool provision system that provides the analytical tools, perform the analysis of the designed product. Furthermore, if the autonomous design AI unit 33 does not possess the necessary knowledge for structural design or structural design inference for the product being designed within the product design system 30, it will autonomously determine that the necessary knowledge is lacking and, in cooperation with the design support tool provision system that provides design support tools for structural design or structural design inference for the product, enable the designer to perform structural design or structural design inference for the product. Alternatively, the autonomous design AI unit 33 may autonomously determine and execute a procedure to acquire sufficient structural knowledge of the product to perform structural calculations through cooperation with a knowledge database.

[0111] Thus, the autonomous design AI unit 33 can be linked with at least one of the following external devices: a design support tool provision system, an analysis tool provision system, an estimation system, and a knowledge database. This allows it to autonomously link with the external device and acquire the necessary knowledge to reliably support product design if it lacks knowledge in at least one of the following areas: product design support, analysis of designed products, estimation of manufacturing costs for designed products, or technical knowledge.

[0112] Furthermore, if the autonomous design AI unit 33 determines that it lacks knowledge in the product design, it will decide that it needs to obtain that missing knowledge from an external device and will ask the designer for instructions via the autonomous design AI dialogue screen 85 whether it is possible to obtain that missing knowledge. If the designer instructs that the missing knowledge be obtained from an external device, the autonomous design AI unit 33 will obtain that missing knowledge in cooperation with the external device, and the design will continue. In this way, it can autonomously obtain missing knowledge from an external device while interacting with the designer in the product design.

[0113] <2.5. Design Agent AI Unit 34> The design agent AI unit 34 is an AI unit 110 that designs a product by coordinating the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33. When the product design system 30 starts executing the program that operates the product design system 30, it first executes a design agent AI process (see Figure 8(a)) to operate the design agent AI unit 34, and displays a selection screen 40 (see Figure 8(b)) on a display device (not shown) provided in the product design system 30.

[0114] The selection screen 40 is a screen that allows the designer to select which of the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33 to operate. The design AI unit selected by the designer on the selection screen 40 is then executed by the design agent AI unit 34.

[0115] Through the work of this design agent AI unit 34, for example, the design data of a product whose part design has been inferred by the part design AI unit 31 can be evaluated by the support design AI unit 32 to see if it satisfies a checklist and / or design rules, thereby supporting the product design. Furthermore, through the work of the design agent AI unit 34, for example, in part design inference by the part design AI unit 31 or in design support by the support design AI unit 32, any missing knowledge can be autonomously determined by the autonomous design AI unit 33, and the missing knowledge can be acquired in cooperation with an external device.

[0116] Thus, the presence of the design agent AI unit 34 enables the autonomous design of a product by coordinating with the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33.

[0117] <3. Operation of the product design system 30> Next, the operation of the product design system 30 will be explained with reference to Figures 8 to 22. Here, the explanation will be given using the example of a "umbrella design system" that designs umbrellas.

[0118] <3.1. Design Agent AI Processing> Figure 8(a) is a flowchart showing the flow of design agent AI processing performed by the design agent AI unit 34. When the application of the product design system 30 is launched by the designer, the product design system 30 first starts executing design agent AI processing using the design agent AI unit 34.

[0119] When the execution of the design agent AI processing begins, the design agent AI unit 34 first displays the selection screen 40 on the display device (not shown) of the product design system 30 (S31).

[0120] Here, Figure 8(b) is a schematic diagram of the selection screen 40. The selection screen 40 displays a component design button 41, an assisted design button 42, an autonomous design button 43, and an exit button 44. These buttons are configured to be selectable by the designer operating an input device (mouse, stylus, etc.), and the design agent process executes processing according to the operated button.

[0121] In other words, during the design agent processing, the design agent AI unit 34 first determines whether the part design button 41 has been selected by the designer (S32). If the part design button 41 has been selected by the designer (S32: Yes), the design agent AI unit 34 instructs the part design AI unit 31 to execute the part design AI processing (S33).

[0122] Details of the component design AI processing (S33) will be described later with reference to Figures 9 to 12, but to briefly explain, the component design AI processing (S33) involves the component design AI unit 31 receiving input from the designer regarding the conditions of the environment in which the product will be used (environmental conditions), and based on these environmental conditions, performing component design inference by utilizing the knowledge of components held within the component design AI unit 31 while interacting with the designer to infer which components will be used in the product.

[0123] In the design agent processing, after the component design AI processing (S33) is completed, or if, as a result of the judgment in S32, it is determined that the component design button 41 has not been selected by the designer (S32: No), the design agent AI unit 34 then determines whether the support design button 42 has been selected by the designer (S34). If the support design button 42 has been selected by the designer (S34: Yes), the design agent AI unit 34 instructs the support design AI unit 32 to execute the support design AI processing (S35).

[0124] Details of the AI-assisted design processing (S35) will be described later with reference to Figures 13 to 16, but to briefly explain, the AI-assisted design processing (S35) involves the AI-assisted design unit 32 analyzing whether the part design reasoning in the part design AI unit 31 satisfies the checklist and / or design rules stored as internal knowledge, and making necessary modifications to the design while interacting with the designer.

[0125] In the design agent processing, after the support design AI processing (S35) is completed, or if, as a result of the judgment in S34, it is determined that the support design button 42 has not been selected by the designer (S34: No), the design agent AI unit 34 then determines whether the autonomous design button 43 has been selected by the designer (S36). If the result is that the autonomous design button 43 has been selected by the designer (S36: Yes), the design agent AI unit 34 instructs the autonomous design AI unit 33 to execute the autonomous design AI processing (S37).

[0126] Details of the autonomous design AI processing (S37) will be described later with reference to Figures 17 to 19, but to briefly explain, the autonomous design AI processing (S37) involves the autonomous design AI unit 33 autonomously determining the knowledge lacking in the product design, and collaborating with an external device that possesses that missing knowledge as external knowledge while interacting with the designer.

[0127] In the design agent processing, after the autonomous design AI processing (S37) is completed, or if, as a result of the judgment in S36, it is determined that the autonomous design button 43 has not been selected by the designer (S36: No), the design agent AI unit 34 then determines whether the exit button 44 has been selected by the designer (S38). If the exit button 44 has been selected by the designer (S38: Yes), the design agent AI processing is terminated and the application of the product design system 30 is closed. On the other hand, if, as a result of the judgment in S38, the exit button 44 has not been selected by the designer (S38: No), the process returns to S32. The design agent AI processing then repeatedly executes the processes from S32 to S38 until the exit button 44 is selected.

[0128] <3.2. AI Processing for Part Design> Next, the part design AI processing (S33) will be explained with reference to Figures 9 to 12. First, Figure 9 is a flowchart showing the flow of the part design AI processing (S33). As described above, the part design AI processing (S33) is a process executed by the part design AI unit 31 when the designer selects the part design button 41 on the selection screen 40.

[0129] When the execution of the component design AI processing (S33) begins, the component design AI unit 31 first displays the environmental condition input screen 50 on the display device (not shown) of the product design system 30 (S331). The environmental condition input screen 50 will now be described with reference to Figure 10. Figure 10 is a schematic diagram of the environmental condition input screen 50, where (a) shows the state in which the environmental conditions for product use have been entered by the designer into the environmental condition input area 51, and (b) shows the state in which structural calculations have been performed based on the entered environmental conditions.

[0130] The environmental conditions input screen 50 includes an environmental conditions input area 51, a structural calculation button 52, a structural conditions display area 53, a knowledge registration button 54, an inference button 55, a registered parts display area 56, a parts modification button 57, and an exit button 58.

[0131] The environmental condition input area 51 is an area for receiving input from the designer regarding the environmental conditions under which the product to be designed will be used. The environmental condition input area 51 lists environmental condition items, and for each environmental condition item, the designer can input the environmental conditions using a known input device (not shown), such as a keyboard, provided in the product design system 30. For example, in the example shown in Figure 10(a), the environmental condition items for the umbrella to be designed are listed as "rain angle" expected when using the umbrella, "height" of the person expected to use the umbrella, "weight limit" of the umbrella, "bending limit strength" of the umbrella's central shaft, "vertical strength" of the umbrella's ribs, and "central shaft coefficient," which is the ratio of the lengths of the central shaft and the ribs.

[0132] When the environmental condition input screen 50 is displayed during the S331 process in Figure 9, the initial value of the environmental condition is displayed for each environmental condition item in the environmental condition input area 51. This initial value may always be a constant value, or it may be a value entered by the designer when the structural design or component design reasoning was performed last time. The designer can use the initial value displayed in the environmental condition input area 51 as the environmental condition as is. On the other hand, if the designer wants to change the environmental condition from the initial value displayed in the environmental condition input area 51, they can change the initial value displayed in association with the environmental condition item they wish to change to the desired environmental condition value.

[0133] The structural calculation button 52 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to start the structural calculation of the product to be designed. When the structural calculation button 52 is selected, the component design AI unit 31 first performs structural design inference, which infers the specifications of the structure and mechanism necessary to realize the characteristics and features of the product to be designed, or structural design, which directly designs these specifications. However, as mentioned above, if the specifications of the structure and mechanism of the product to be designed already exist, structural design inference or structural design is omitted. Next, once the specifications of the structure and mechanism of the product to be designed are determined, the structural calculation of the product to be designed is performed based on the environmental conditions entered by the designer on the environmental conditions input screen 50 and the specifications of the structure and mechanism.

[0134] As shown in Figure 10(b), the structural condition display area 53 is an area that displays the structural conditions that each structure or mechanism should have, derived from structural calculations, for each structural condition item (an item indicating which structure or mechanism the structural condition applies to). Here, for at least some of the structural condition items displayed in the structural condition display area 53 (in the example in Figure 10(b), the "weight" of the "umbrella," the "number" of the "ribs," and the "material" and "length" of the "handle"), the designer can change the values ​​of the structural conditions using a known input device such as a keyboard (not shown). In other words, the structural condition display area 53 presents the designer with structural conditions that have been autonomously inferred based on environmental conditions, and for changeable environmental condition items, the values ​​of those structural conditions can be changed in dialogue with the designer.

[0135] The knowledge registration button 54 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, in order to register new knowledge to the part design AI unit 31. The part design AI unit 31 can register new knowledge information, such as knowledge about the material, size (length, width, height, diameter, etc.), and weight of a new part, as well as knowledge about cost, difficulty of processing, and ease of assembly. In addition to knowledge about parts, the part design AI unit 31 can also register instructions to the part design AI unit 31 (for example, instructions on which product to design, such as "long umbrella" or "folding umbrella") as new "autonomous will" knowledge information. The knowledge registration button 54 is used when registering such knowledge information. Details of knowledge registration will be described later with reference to Figure 12.

[0136] The inference button 55 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to infer the parts to be used in the product based on the structural conditions of the product inferred by structural design (structural calculation) based on the input environmental conditions. When this inference button 55 is selected, the parts that satisfy the structural conditions for each structural condition item displayed in the structural condition display area 53 are inferred using the knowledge of parts contained within the parts design AI unit 31. The inference button 55 may be made active (selectable by the designer) once the structural calculation has been performed.

[0137] Furthermore, even if the component design AI unit 31 cannot infer a component that satisfies the structural conditions for each structural condition item displayed in the structural condition display area 53 based solely on its internal knowledge of components, it will infer a component for the product that satisfies at least some of the structural conditions, or a component that closely matches the structural conditions. Then, as will be described later, the component design AI unit 31 presents the inferred component to the designer while explicitly indicating the structural conditions that were not met. This allows the designer to perform component design inference for the product while modifying the structural conditions. In other words, the component design AI unit 31 can perform component design inference for the product while interacting with the designer.

[0138] The registered parts display area 56 is an area that displays a list of parts (registered parts) that the parts design AI unit 31 has internally as knowledge. After performing structural design (structural calculation) of a product, the parts design AI unit 31 displays the results in the structural conditions display area 53, and at the same time, displays the list of registered parts in the registered parts display area 56. The parts design AI unit 31 is configured so that the designer can select one of the parts displayed in the registered parts display area 56 and modify the knowledge about the selected part. The designer can also decide to register another new part as knowledge information for the parts design AI unit 31 in relation to the parts displayed in the registered parts display area 56.

[0139] The part modification button 57 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, when modifying the knowledge of registered parts that the part design AI unit 31 internally holds as knowledge. When one of the parts displayed in the registered part display area 56 is selected and the part modification button 57 is selected, a screen (not shown) for modifying the knowledge of the selected part is displayed, and the knowledge of the part can be modified through that screen. The modification of the part's knowledge is performed by changing the data values ​​associated with each concept in the knowledge management table of the selected part, or by adding a new concept and its data value.

[0140] The exit button 58 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, to terminate the part design AI processing (S33). When the exit button 58 is selected, the part design AI processing (S33) terminates its own process and returns to the design agent AI processing.

[0141] Returning to Figure 9, let's continue the explanation of the part design AI processing (S33). After processing in S331, the part design AI unit 31 determines whether the structural calculation button 52 on the environmental conditions input screen 50 has been selected (S332). If the structural calculation button 52 has been selected (S332: Yes), the unit first performs structural design inference to infer the specifications of the structure and mechanism that will realize the characteristics and features of the product to be designed and / or satisfy the input environmental conditions (S334). This S334 process constitutes the structural design inference means of the present invention. Through this S334 process, the structural specifications according to the characteristics and / or features of the product can be autonomously inferred.

[0142] As mentioned above, the S334 process may involve directly designing the structural and mechanical specifications, or the S334 process may be skipped if the structural and mechanical specifications of the product being designed are fixed.

[0143] Next, the component design AI unit 31 performs a structural calculation of the structure of the product to be designed based on the environmental conditions entered (or pre-set) in each environmental condition item in the environmental condition input area 51 and the specifications of the product structure and mechanism obtained (or pre-set) through the processing in S334 (S335). Then, the component design AI unit 31 displays the structural conditions (calculation results) derived from the structural calculation performed in the processing in S334 in the structural condition display area 53 of the environmental condition input screen 50, as shown in Figure 10(b) (S336). At this time, the designer can change the values ​​of at least some of the structural condition items displayed in the structural condition display area 53. This allows the structural conditions to be determined in dialogue with the designer and the process to proceed to component design inference.

[0144] In addition, during the S336 process, the component design AI unit 31 also executes a process to display a list of components (registered components) that it internally possesses as knowledge, in the registered component display area 56 of the environmental condition input screen 50, as shown in Figure 10(b). Based on this list of registered components, the designer can register knowledge information about new components or modify the knowledge of registered components.

[0145] If the structural calculation button 52 is not selected after processing in S336 or in the judgment in S332 (S332: No), then it is determined whether the inference button 55 is selected (S337). If the inference button 55 is selected as a result (S337: Yes), the part design AI unit 31 performs part design inference, which infers the parts to be used in the product from the internal knowledge of parts, based on the structural conditions estimated or determined, including environmental conditions (S338). This enables autonomous part design to satisfy the structural conditions that the product structure should have. As mentioned above, part design inference infers multiple combinations of parts to be used in the product. In addition, part design inference may also infer combinations of parts that do not satisfy the structural conditions.

[0146] In the S338 process, the component design AI unit 31 also determines, for each inferred combination of components, whether that combination satisfies the structural conditions displayed in the structural condition display area 61. This S338 process constitutes the "instruction awaiting item determination means" of the present invention. That is, through the S338 process, if the inferred components of the product do not satisfy the structural conditions, the component design AI unit 31 determines that the change in structural conditions is an item requiring instructions from the designer.

[0147] Then, the component design AI unit 31 displays the component design AI dialogue screen 60, which shows the result of the component design inference performed in S338, on the display device (not shown) of the product design system 30 (S339).

[0148] Here, the part design AI dialogue screen 60 will be explained with reference to Figure 11. Figure 11(a) is a schematic diagram of the part design AI dialogue screen 60, Figure 11(b) shows the state in the part design AI dialogue screen 60 where the value of a structural condition item has been changed, and Figure 11(c) shows the state where part design inference has been performed again for the changed structural condition.

[0149] The component design AI dialogue screen 60 includes a structural condition display area 61, a re-inference button 62, an initial condition button 63, an inference result display area 64, a selected product display area 65, a save button 66, a drawing button 67, and an exit button 68.

[0150] The structural conditions display area 61 is an area that displays the structural conditions of the product (the conditions that each structure and mechanism constituting the product should have) for each structural condition item, which are derived from the results of the component design inference displayed in the inference result display area 64. In other words, the structural conditions display area 61 displays the content that was displayed in the structural conditions display area 53 of the environmental conditions input screen 50 immediately before the inference button 55 was selected for the environmental conditions input screen 50, or the content that was displayed in the structural conditions display area 61 immediately before the re-inference button 62 described later was selected for the component design AI dialogue screen 60.

[0151] The structural condition display area 61, like the structural condition display area 53, is configured to allow the designer to change the values ​​of at least some structural condition items (in the example in Figure 11, the "number" of "ribs" and the "material" and "length" of the "handle") using a known input device such as a keyboard (not shown). For example, in the example in Figure 11(b), the "material" of the "handle" is changed from "wood" to "plastic".

[0152] The structural condition display area 61 presents the structural conditions used in the part design reasoning to the designer, and for modifiable structural condition items, the designer can change the values ​​of those structural conditions in dialogue with the designer, thereby allowing the part design reasoning to be performed again as described later. For example, if the reasoning result display area 64, described later, indicates that the combination of parts in the product reasoned by the part design reasoning does not satisfy the structural conditions (in the example shown in Figure 11(a), the structural condition regarding the weight of the umbrella is 500g or less, while the weight of the reasoned combination of parts is 507g), the designer's change of structural conditions is accepted via the structural condition display area 61. This allows the part design to satisfy the structural conditions in dialogue with the designer.

[0153] For example, in the examples shown in Figures 11(b) and (c), by changing the structural condition regarding the material of the handle from "wood" to "plastic" and re-inferring the component design, the weight of the re-inferred combination of components becomes 394g, demonstrating that a combination of components that satisfies the structural condition has been inferred.

[0154] Furthermore, this structural condition display area 61 corresponds to the "condition change receiving means" and "input receiving means" of the present invention.

[0155] The re-inference button 62 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to re-perform component design inference, which infers the combination of parts to be used in the product based on the structural conditions set in the structural conditions display area 61 (structural conditions after the designer's modification). When this re-inference button 62 is selected, the combination of parts is re-inferred based on the structural conditions after the designer's modification, and the product design continues. In other words, the re-inference button 62 corresponds to the "design continuation means" of the present invention. The re-inference button 62 may be made active (selectable by the designer) when at least one of the structural conditions displayed in the structural conditions display area 61 is changed.

[0156] The initial condition button 63 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to return at least one of the structural conditions displayed in the structural condition display area 61 to its initial state after that condition has been changed. The initial condition may be a predetermined fixed value, or it may be the structural condition when part design inference was performed by selecting the inference button 55 on the environmental condition input screen 50 before the part design AI dialogue screen 60 is displayed. When the designer repeatedly re-infers the part design inference while changing the environmental conditions displayed in the structural condition display area 61, there are often times when they want to return the structural condition to its initial state. In such cases, the designer can easily return the structural condition to its initial state by selecting the initial condition button 63.

[0157] The inference result display area 64 is the area that displays the results of the part design inference performed by the processing in S338. In the example shown in Figure 11(a), the inference button 55 on the environmental condition input screen 50 shown in Figure 10(b) is selected, and the results of the part design inference performed based on the structural conditions displayed in the structural condition display area 53 of the environmental condition input screen 50 are displayed in the inference result display area 64. Also, in the example shown in Figure 11(b), the re-inference button 62 on the part design AI dialogue screen 60 is selected, and the results of the part design inference performed based on the structural conditions changed in the structural condition display area 61 are displayed in the inference result display area 64.

[0158] The inference result display area 64 displays all possible combinations of parts used in the product obtained through part design inference. In the example in Figure 11, the part combination indicated by No. 1 and the part combination indicated by No. 2 are displayed. Furthermore, for each inferred part combination, whether or not that part combination satisfies the structural conditions displayed in the structural conditions display area 61 is indicated by "OK" or "NG" in the "Judgment" column of the inference result display area 64.

[0159] For example, in the example shown in Figure 11(b), the combination of parts indicated as No. 1 is marked "OK" in the "Judgment" column, allowing the designer to understand that the combination of parts indicated as No. 1 satisfies the structural conditions displayed in the structural conditions display area 61. Similarly, the combination of parts indicated as No. 2 is marked "NG" in the "Judgment" column, allowing the designer to understand that the combination of parts indicated as No. 2 does not satisfy the structural conditions displayed in the structural conditions display area 61.

[0160] The selected product display area 65 is an area that displays the combination of parts selected by the designer from among the combinations of parts displayed in the inference result display area 64. When the designer selects a position in the inference result display area 64 by operating a known input device (not shown), such as a mouse or stylus, the combination of parts displayed at the selected position is selected, and that selected combination of parts is displayed in the selected product display area 65. For example, in the example shown in Figure 11(c), the selected product display area 65 shows that the combination of parts No. 1 has been selected.

[0161] The save button 66 is a button that can be selected by the designer using a known input device (not shown) such as a mouse or stylus to save the design data of the combination of parts displayed in the selected product display area 65, that is, the combination of parts selected by the designer. The drawing button 67 is a button that can be selected by the designer using a known input device (not shown) such as a mouse or stylus to have the CAD system draw the product based on the combination of parts displayed in the selected product display area 65, that is, the combination of parts selected by the designer. The exit button 68 is a button that can be selected by the designer using a known input device (not shown) such as a mouse or stylus to terminate the part design AI processing (S33).

[0162] Returning to Figure 9, we continue the explanation of the part design AI processing (S33). After processing in S339, the part design AI unit 31 determines whether the re-inference button 62 was selected (S340). If the re-inference button 62 was selected (S340: Yes), the process returns to S338 and the part design inference is executed again. At this time, if the structural conditions displayed in the structural conditions display area 61 of the part design AI dialogue screen 60 have been changed by the designer, or if the initial conditions button 63 has been selected by the designer and the structural conditions displayed in the structural conditions display area 61 have been returned to the initial conditions, the combination of parts used in the product is re-inferred based on the changed structural conditions or the initial conditions. In this way, part design inference is performed while interacting with the designer through the part design AI dialogue screen 60.

[0163] On the other hand, if the re-inference button 62 is not selected as a result of the judgment in S340 (S340: No), the component design AI unit 31 then determines whether the drawing button 67 has been selected (S341). If the drawing button 67 has been selected (S341: Yes), the AI ​​unit has the CAD system, in cooperation with the CAD-API, create a drawing of the product based on the combination of parts displayed in the selected product display area 65, that is, the combination of parts selected by the designer, and displays the CAD screen on the display device (not shown) of the product design system 30 (S342). As a result, the design data of the product based on the inferred product parts is converted into a drawing. Therefore, the designer can easily check the product whose structure and parts have been autonomously designed by the product design system 30 from the CAD system on the spot, and can confirm whether there are any problems with the component design inference by checking the CAD drawing of the drawn product.

[0164] After processing in S342, or as a result of the judgment in S341, if the drawing button 67 is not selected (S341: No), the component design AI unit 31 then determines whether the save button 66 has been selected (S343). If the save button 66 has been selected (S343: Yes), the design data of the component combination displayed in the selected product display area 65, that is, the component combination selected by the designer, is saved to a predetermined recording medium such as a hard disk drive (HDD) or solid state drive (SSD) provided in the product design system 30 (S344). Using this saved design data, the design agent AI unit 34 can coordinate the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33.

[0165] If, after processing in S344, or as a result of the judgment in S343, the save button 66 is not selected (S343: No), the component design AI unit 31 then determines whether the exit button 68 is selected (S345). If the exit button 68 is not selected (S345: No), the process returns to S340. The processes from S340 to S345 (including the processes in S338 and S339 if an affirmative judgment is made in the judgment in S340) are then repeatedly executed until the exit button 68 is selected. On the other hand, if, as a result of the judgment in S345, it is determined that the exit button 68 is selected (S345: Yes), the component design AI process ends and the process returns to the design agent AI process (see Figure 8(a)). At this time, the selection screen 40 is displayed on the display device (not shown) of the product design system 30.

[0166] Let's return to the explanation of the S337 decision. When the S337 decision is made, the environmental conditions input screen 50 (see Figure 10) is displayed on the product design system 30's display device (not shown). If the inference button 55 is not selected as a result of the S337 decision (S337: No), the component design AI unit 31 then determines whether the knowledge registration button 54 is selected (S346). If the knowledge registration button 54 is selected as a result (S346: Yes), new knowledge is registered (S347).

[0167] Here, we will explain the registration of new knowledge with reference to Figure 12. Figure 12(a) is a schematic diagram illustrating the contents of a CSV (comma-separated values) file used when registering knowledge information about parts as new knowledge, and Figure 12(b) is a schematic diagram illustrating a part of the display on the environmental conditions input screen 50 when new knowledge about parts has been registered.

[0168] New knowledge is registered using a CSV file. That is, the designer pre-records the knowledge information they want to register in a CSV file. For example, if you want to register knowledge about a part called "handle," as shown in Figure 12(a), you would first list "handle" as the representative value for the representative concept "part" in the CSV file, and then list the value of that part (handle) as the dependent value for each dependent concept. For example, in the example shown in Figure 12(a), the dependent value "M001" is listed for the dependent concept "code," the dependent value "wood" is listed for the dependent concept "material," the dependent value "200" and its unit "mm" are listed for the dependent concept "length," and the dependent value "260" and its unit "g" are listed for the dependent concept "weight." In addition, the CSV file also includes information on dependent concepts whose dependent values ​​can be variable (indicated by "*" in Figure 12(a)).

[0169] In the S347 process, by reading such a CSV file, the knowledge information contained in the CSV file is registered as new knowledge in the part design AI unit 31. When new knowledge about a part is registered, as shown in Figure 12(b), dependent concepts related to the structure of the registered part (material, length, etc.) are added as structural condition items in the structural condition display area 53. In addition, the newly registered part is also added and displayed in the registered part display area 56.

[0170] Returning to Figure 9, let's continue the explanation of the part design AI process. After the process in S347, or as a result of the judgment in S346, if the knowledge registration button 54 is not selected (S346: No), the part design AI unit 31 then determines whether the part modification button 57 has been selected (S348). If the part modification button 57 has been selected (S348: Yes), the AI ​​unit modifies the knowledge of the previously selected part in the registered part display area 56 (S349). In other words, as described above, in the process of S349, a screen (not shown) for modifying the knowledge of the selected part is displayed, allowing the designer to modify the knowledge of the part through that screen. Alternatively, in the process of S349, the AI ​​may load a CSV file containing the modified knowledge of the part to enable modification of the part knowledge.

[0171] If, after processing in S349 or as a result of the determination in S348, the part modification button 57 is not selected (S348: No), the part design AI unit 31 then determines whether the exit button 58 is selected (S350). If the exit button 58 is not selected (S350: No), the process returns to S332. The processes from S332 to S350 are then repeatedly executed until the exit button 58 is selected. On the other hand, if, as a result of the determination in S350, it is determined that the exit button 58 is selected (S350: Yes), the part design AI process ends and the process returns to the design agent AI process (see Figure 8(a)). At this time, the selection screen 40 is displayed on the display device (not shown) of the product design system 30.

[0172] As described above, the component design AI unit 31 can autonomously design product components while interacting with the designer via the component design AI dialogue screen 60 by executing the component design AI processing (S33). Specifically, if the component design AI unit 31 infers a product component based on structural conditions and that component does not satisfy those structural conditions, it determines that a change in those structural conditions is required, and seeks instructions from the designer. When a change in structural conditions is accepted as an instruction from the designer, the design process continues by re-inferring the product component based on the changed structural conditions. This allows for product component design based on the product's structural conditions while interacting with the designer.

[0173] <3.3. AI-assisted design processing> Next, the support design AI processing (S35) will be explained with reference to Figures 13 to 16. First, Figure 13 is a flowchart showing the flow of the support design AI processing (S35). As mentioned above, the support design AI processing (S35) is a process executed by the support design AI unit 32 when the support design button 42 is selected by the designer on the selection screen 40.

[0174] When the execution of the support design AI processing (S35) begins, the support design AI unit 32 first displays the support design AI dialogue screen 70 on a display device (not shown) of the product design system 30 (S351).

[0175] Here, the AI-assisted design dialogue screen 70 will be explained with reference to Figures 14-16. Figure 14(a) is a schematic diagram showing the initial screen of the AI-assisted design dialogue screen 70 displayed on the display device of the product design system 30 as a result of processing in S335, and Figure 14(b) is a schematic diagram showing the AI-assisted design dialogue screen 70 displayed on the display device of the product design system 30 after the design data has been read and analysis has been performed in accordance with the checklist and / or design rules. Furthermore, Figure 15 is a schematic diagram showing the AI-assisted design dialogue screen 70 displayed on the display device of the product design system 30 as a result of the analysis, before any design modifications are made, and Figure 16 is a schematic diagram showing the AI-assisted design dialogue screen 70 displayed on the display device of the product design system 30 after the modifications have been made.

[0176] The AI-assisted design dialogue screen 70 is a screen for reading design data of products whose component design has been inferred by the component design AI unit 31 or products whose design has been assisted by the autonomous design AI unit 33 using external knowledge from an external device. The AI ​​then uses an evaluation tool to assess whether the product satisfies the checklist and / or design rules, which are internal knowledge, and autonomously determines areas that seem to require correction. The AI ​​then makes design modifications in dialogue with the designer. The AI-assisted design dialogue screen 70 includes a design data loading button 71, an analysis result display area 72, a correction instruction display area 73, an AI correction button 74, a CAD correction button 75, a correction content display area 76, a design data overwrite save button 77, a drawing button 78, and an exit button 79.

[0177] The design data loading button 71 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, in order to load design data. The design data is data stored in the component design AI unit 31 and the autonomous design AI unit 33, and as described above, it is the design data of products whose component design was inferred by the component design AI unit 31, or products whose design was assisted by the autonomous design AI unit 33 using external knowledge from an external device. Design data stored in the support design AI unit 32 may be loaded again. Furthermore, the design data of products whose component design was inferred by the component design AI unit 31 includes not only the data stored in the component design AI unit 31 itself, but also data stored after design assistance was provided by the support design AI unit 32 to the data stored in the component design AI unit 31, and data stored after missing knowledge was supplemented by the autonomous design AI unit 33 in cooperation with an external device.

[0178] The analysis results display area 72 is the area that displays the results of the evaluation performed by the evaluation tool on the loaded design data, showing whether the product satisfies the checklist and / or design rules. In the example shown in Figure 14(b), the evaluation tool applies the checklist and / or design rules to each part of the product whose part design inference was performed, and indicates in the "Status" column whether the checklist and / or design rules are satisfied or not as "OK" (satisfied) or "NG" (not satisfied) for each part. This allows the designer to understand which parts satisfy the checklist and / or design rules and which do not.

[0179] In the analysis results display area 72, items that do not satisfy the checklist and / or design rules (in the example shown in Figure 15, the item "center bar" whose "status" is "NG") can be selected by the designer using a known input device (not shown), such as a mouse or stylus. If there are multiple items that do not satisfy the checklist and / or design rules, each item can be selected individually. When the designer selects one of the items that does not satisfy the checklist and / or design rules in the analysis results display area 72, as shown in Figure 15, the content of the correction instructions to satisfy the checklist and / or design rules for that item is displayed in the correction instruction display area 73, and the CAD system is launched and the CAD drawing 80 of the product before correction is also displayed. The designer can check the correction instructions and CAD drawing 80 displayed in the correction instruction display area 73 and decide whether or not to make the correction.

[0180] The AI ​​correction button 74 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to allow the support design AI unit 32 to autonomously perform the correction instructions displayed in the correction instruction display area 73. When the AI ​​correction button 74 is selected by the designer, the support design AI unit 32 performs the correction on the item to be corrected based on the correction instructions displayed in the correction instruction display area 73, that is, to make the item satisfy the checklist and / or design rules. As a result, even if the product based on the loaded design data does not satisfy the checklist and / or design rules, the support design AI unit 32 will autonomously correct it to satisfy the checklist and / or design rules while interacting with the designer, allowing the designer to easily continue product design.

[0181] The CAD modification button 75 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to allow the designer to make modifications to the modification instructions displayed in the modification instruction display area 73. When the CAD modification button 75 is selected by the designer, the support design AI unit 32 starts the CAD system and, as shown in Figure 16, displays the CAD drawing 81 that can be modified, along with the CAD drawing 80 before modification, enabling the designer to make modifications such as design changes. The support design AI unit 32 then accepts the modifications made by the designer to the CAD drawing 81 and reflects them in the design data.

[0182] In this case, the support design AI unit 32 uses an evaluation tool to assess whether the modified design data satisfies the checklist and / or design rules. The evaluation results (analysis results) are then displayed in the analysis result display area 72, allowing the designer to understand whether the manually made modifications were effective or not.

[0183] Furthermore, when a correction is made using the AI ​​correction button 74 or the CAD correction button 75, the details of the correction are displayed in the correction details display area 76, as shown in Figure 16. In addition, the CAD system displays the CAD drawing 80 before the correction and the CAD drawing 81 after the correction. The designer can determine whether the correction was appropriate based on the CAD drawing 80 before the correction, the CAD drawing 81 after the correction, and the details of the correction displayed in the correction details display area 76.

[0184] Furthermore, the AI ​​correction button 74 and the CAD correction button 75 correspond to the "input receiving means" of the present invention.

[0185] The design data overwrite save button 77 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to overwrite the design data of a product that has been modified by the AI ​​modification button 74 or the CAD modification button 75 with the design data loaded by the design data load button 71.

[0186] The drawing button 78 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to have the CAD system draw a product based on the design data. For example, after the modified CAD drawing 81 has been closed by the designer, the designer can select this drawing button 78 to display the modified CAD drawing based on the design data on the display device (not shown) of the product design system 30. This allows the designer to check the modified CAD drawing as many times as needed to ensure there are no problems.

[0187] The exit button 79 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, to terminate the AI-assisted design processing (S35).

[0188] Returning to Figure 13, and continuing the explanation of the support design AI processing (S35), after the processing in S351, the support design AI unit 32 determines whether the design data loading button 71 on the support design AI dialogue screen 70 has been selected (S352). If the design data loading button 71 has been selected (S352: Yes), the support design AI unit 32 prompts the designer to select the design data file to be loaded, and then loads the design data from the selected file (S353). Next, the support design AI unit 32 performs an analysis of the product based on internal knowledge by having the evaluation tool implemented in the support design AI unit 32 evaluate whether the product shown in the loaded design data satisfies the checklist and / or design rules (S354).

[0189] Then, as shown in Figure 14(b), the support design AI unit 32 displays the analysis results in the analysis result display area 72 of the support design AI dialogue screen 70 (S355). By displaying the results in this analysis result display area 72, the designer can identify any parts (components) in the loaded design data that deviate from the checklist and / or design rules.

[0190] In this way, the support design AI unit 32 identifies areas that deviate from the checklist and / or the design rules in the design of a product using parts of the product inferred by the parts design AI unit 31. This enables the parts design AI unit 31 and the support design AI unit 32 to work together to design parts in accordance with the checklist and / or the design rules. The processing in S354 constitutes the "design checking means" and the "instruction pending item determination means" of the present invention.

[0191] After processing in S355, or as a result of the judgment in S352, if the design data loading button 71 is not selected (S352: No), the support design AI unit 32 then determines whether the AI ​​correction button 74 is selected (S356) when one of the items that does not satisfy the checklist and / or design rules is selected by the designer in the analysis result display area 72, a correction instruction for that item is displayed in the correction instruction display area 73, and the CAD drawing 80 before correction is displayed. If the AI ​​correction button 74 is selected (S356: Yes), the support design AI unit 32 autonomously performs the correction based on the correction instruction displayed in the correction instruction display area 73 for the item to be corrected, i.e., the item satisfies the checklist and / or design rules (S357). In processing S357, the support design AI unit 32 displays the content of the correction in the correction content display area 76, and also displays the corrected CAD drawing 81.

[0192] Thus, when the AI ​​component design unit 31 identifies a deviation from the checklist and / or design rules in the design of a product using the product's components inferred by the AI, selecting the AI ​​correction button 74 modifies the product's components to satisfy the checklist and / or design rules. This enables autonomous component design in accordance with the checklist and / or design rules. The process in S357 constitutes the "component modification means" and "design continuation means" of the present invention.

[0193] Furthermore, after processing in S357, or as a result of the judgment in S356, if the AI ​​correction button 74 is not selected (S356: No), the support design AI unit 32 then determines in the analysis result display area 72 whether the CAD correction button 75 was selected, given that one of the items that does not satisfy the checklist and / or design rules has been selected by the designer and the CAD drawing 80 before correction is displayed (S358). If the CAD correction button 75 is selected (S358: Yes), the CAD drawing 81 that can be corrected is displayed, and the CAD system accepts manual corrections from the designer (S359). In processing S359, the support design AI unit 32 displays the content of the accepted corrections in the correction content display area 76, and uses an evaluation tool to evaluate (analyze) whether the product satisfies the checklist and / or design rules in the design data into which the corrections have been reflected, and then displays the evaluation result (analysis result) in the analysis result display area 72. This allows the designer to understand whether the manual corrections made were effective or not. Furthermore, the process in S359 constitutes the "design continuation means" of the present invention.

[0194] Then, after processing in S359, or as a result of the judgment in S358, if the CAD modification button 75 is not selected (S358: No), the support design AI unit 32 then determines whether the design data overwrite save button 77 is selected (S360). If the design data overwrite save button 77 is selected (S360: Yes), the current design data (or the modified design data if modifications have been made by processing in S357 and / or S359) is overwritten and saved to the design data file read in processing S353 (S361).

[0195] Furthermore, after processing in S361, or as a result of the judgment in S360, if the design data overwrite save button 77 is not selected (S360: No), the support design AI unit 32 then determines whether the drawing button 78 is selected (S362). If the drawing button 78 is selected (S362: Yes), the AI ​​unit causes the CAD system to draw the product based on the current design data and displays the CAD drawing on the product design system 30's display device (not shown) (S363). This allows the designer to check for any problems by reviewing the current CAD drawing.

[0196] If, after processing in S363, or as a result of the determination in S362, the drawing button 78 is not selected (S362: No), the support design AI unit 32 then determines whether the exit button 79 is selected (S364). If the exit button 79 is not selected (S364: No), the process returns to S352. The processes from S352 to S364 are then repeatedly executed until the exit button 79 is selected. On the other hand, if, as a result of the determination in S364, it is determined that the exit button 79 is selected (S364: Yes), the support design AI process ends and the process returns to the design agent AI process (see Figure 8(a)). At this time, the selection screen 40 is displayed on the display device (not shown) of the product design system 30.

[0197] As described above, the AI ​​support design unit 32 can autonomously assist in product design while interacting with the designer via the AI ​​support design dialogue screen 70. Specifically, the AI ​​support design unit 32 determines that any parts of the product under design that deviate from its internal checklist and / or design rules require correction, and that this is an item requiring instruction from the designer. When the correction of such a part is accepted as an instruction from the designer, the design of that part is corrected, and the design process continues. In this way, the AI ​​support design can assist in product design while interacting with the designer, ensuring that the checklist and / or design rules are satisfied.

[0198] <3.4. Autonomous Design AI Processing> Next, the autonomous design AI processing (S37) will be explained with reference to Figures 17 to 19. First, Figure 17 is a flowchart showing the flow of the autonomous design AI processing (S37). As mentioned above, the autonomous design AI processing (S37) is a process executed by the autonomous design AI unit 33 when the autonomous design button 43 is selected by the designer on the selection screen 40.

[0199] When the execution of the autonomous design AI processing (S37) begins, the autonomous design AI unit 33 first displays the autonomous design AI dialogue screen 85 on a display device (not shown) of the product design system 30 (S371).

[0200] Here, the autonomous design AI dialogue screen 85 will be explained with reference to Figures 18 and 19. Figure 18(a) is a schematic diagram showing the initial screen of the autonomous design AI dialogue screen 85 displayed on the display device of the product design system 30 as a result of the processing in S371, and Figure 18(b) is a schematic diagram showing the autonomous design AI dialogue screen 85 displayed on the display device of the product design system 30 when design data has been read and the missing knowledge has been determined. Furthermore, Figure 19(a) is a schematic diagram showing the autonomous design AI dialogue screen 85 displayed on the display device of the product design system 30 when the missing knowledge has been selected by the designer and the designer has chosen to perform cooperation with an external device in order to acquire that knowledge as external knowledge, and Figure 19(b) is a schematic diagram showing the autonomous design AI dialogue screen 85 displayed on the display device of the product design system 30 when the cooperation with the external device has been completed.

[0201] The autonomous design AI dialogue screen 85 includes a design data loading button 86, a knowledge judgment display area 87, a knowledge content display area 88, an external linkage content display area 89, an execute button 90, a pass button 91, a design data overwrite save button 92, a drawing button 93, and an exit button 94.

[0202] The design data loading button 86 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, in order to load design data. The design data is data stored in the component design AI unit 31 and the support design AI unit 32, and as described above, it is the design data of products whose component design was inferred by the component design AI unit 31, or products whose design was assisted by the support design AI unit 32 based on its internal knowledge. Design data stored in the autonomous design AI unit 33 may be loaded again. Furthermore, the design data of products whose component design was inferred by the component design AI unit 31 includes not only the data stored in the component design AI unit 31 itself, but also data stored after design assistance was provided by the support design AI unit 32 to the data stored in the component design AI unit 31, and data stored after missing knowledge was supplemented by the autonomous design AI unit 33 through cooperation with external devices.

[0203] The autonomous design AI unit 33, based on the design data it has read, autonomously determines whether or not the design of the product has been carried out using the knowledge necessary for the design of the product, and displays the result of this determination in the knowledge determination display area 87 for each piece of knowledge necessary for the design. A specific example of the display in the knowledge determination display area 87 is shown in Figure 18(b). In the knowledge determination display area 87, the knowledge necessary for the design of the product "long umbrella" is listed, including "vertical strength analysis" for the component "rib" and "bending strength analysis" for the component "center shaft". Then, corresponding to each piece of knowledge, the "status" column indicates whether or not the design of the product "long umbrella" has been carried out using that knowledge.

[0204] For example, in the design of the product "long umbrella," if a "vertical strength analysis" has been performed on the "rib" component, "OK" will be displayed in the "Status" column, as shown in Figure 18(b), indicating that the product design is being carried out using that knowledge. On the other hand, if a "bending strength analysis" has not been performed on the "center shaft" component in the design of the product "long umbrella," "NON" will be displayed in the "Status" column, as shown in Figure 18(b), indicating that the product design is not being carried out using that knowledge. In this way, the autonomous design AI unit 33 can autonomously determine the knowledge lacking for product design and display that lack of knowledge to the designer in the knowledge judgment display area 87.

[0205] In the knowledge assessment display area 87, the items of knowledge lacking in the product design (in the example shown in Figure 18(b), "bending strength analysis" for the component "center rod") are configured to be selectable by the designer using a known input device (not shown), such as a mouse or stylus. If there are multiple areas of knowledge lacking, each item can be selected individually.

[0206] When the designer selects one of the missing pieces of knowledge from the knowledge determination display area 87, the specific content of that knowledge is displayed in the knowledge content display area 88, as shown in Figure 19(a), and the details of the external device that will cooperate to acquire that missing knowledge as external knowledge are displayed in the external cooperation content display area 89.

[0207] For example, in the example shown in Figure 19(a), the knowledge content display area 88 indicates that the missing knowledge is the bending strength analysis of the central rod, and the external linkage content display area 89 indicates that the analysis will be performed in cooperation with CAE (Computer-Aided Engineering) manufactured by Company A as external knowledge in order to acquire the knowledge of the bending strength analysis of the central rod. Based on the specific content of the missing knowledge displayed in the knowledge content display area 88 and the external device to be linked displayed in the external linkage content display area 89, the designer can decide whether to proceed with acquiring external knowledge by linking with an external device.

[0208] The execution button 90 is a button that can be selected by the designer from known input devices (not shown), such as a mouse or stylus, in order to perform coordination with an external device displayed in the external coordination content display area 89. When the execution button 90 is selected by the designer, the autonomous design AI unit 33 coordinates with the external device displayed in the external coordination content display area 89 and acquires any missing knowledge as external knowledge. This execution button 90 corresponds to the "input receiving means" of the present invention.

[0209] At this time, the autonomous design AI unit 33 displays a screen 95 provided by an external device (in the example shown in Figure 19(a), the analysis screen of a CAE manufactured by Company A) together with the autonomous design AI dialogue screen 85. This screen 95 displays an OK button 95a and an NG button 95b, which are configured to be selectable by the designer using a known input device (not shown), such as a mouse or stylus.

[0210] When the designer determines from the display content on screen 95 that there are no problems with the knowledge acquired from the external device and selects the OK button 95a, the autonomous design AI unit 33 reflects the knowledge acquired from the external device as external knowledge into the product design data. In addition, as shown in Figure 19(b), the autonomous design AI unit 33 changes the "Status" of the knowledge judgment display area 87 corresponding to the knowledge acquired from the external device (in the example in Figure 19(b), "bending strength analysis" of "center bar") from "NON" to "OK". Furthermore, the autonomous design AI unit 33 clears the display in the external linkage content display area 89 and displays the message "Analysis is not necessary" in the knowledge content display area 88. This shows the designer that the knowledge was obtained from an external device.

[0211] On the other hand, if the designer determines from the content displayed on screen 95 that there is a problem with the knowledge acquired from the external device and selects the NG button 95b, the autonomous design AI unit 33 discards the knowledge acquired from the external device and maintains the display of the knowledge judgment display area 87, the knowledge content display area 88, and the external linkage content display area 89.

[0212] The pass button 91 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, in order to deactivate the interaction with the external device displayed in the external interaction content display area 89. When the pass button 91 is selected by the designer, the autonomous design AI unit 33 proceeds with processing without interacting with the external device displayed in the external interaction content display area 89.

[0213] Furthermore, if the NG button 95b or the Pass button 91 is selected, the autonomous design AI unit 33 may, if it can suggest to the designer a collaboration with an external device other than the one displayed in the external collaboration content display area 89, display the collaboration with that other external device in the external collaboration content display area 89 and recommend it to the designer. If the designer wishes to collaborate with the external device newly displayed in the external collaboration content display area 89, they can select the Execute button 90 to acquire the missing knowledge from that external device.

[0214] The design data overwrite save button 92 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, to overwrite the design data of the product, which reflects external knowledge acquired through cooperation with an external device, with the design data loaded by the design data load button 86.

[0215] The drawing button 93 is a button that can be selected by the designer using a known input device (not shown), such as a mouse or stylus, to have the CAD system draw a product based on the design data. When the drawing button 93 is selected, the CAD drawing of the product based on the design data that incorporates external knowledge can be displayed on the display device (not shown) of the product design system 30. This allows the designer to repeatedly check from the CAD drawing whether there are any problems with the product design data that incorporates external knowledge.

[0216] The exit button 94 is a button that can be selected by the designer from a known input device (not shown), such as a mouse or stylus, in order to terminate the autonomous design AI processing (S37).

[0217] Returning to Figure 17, let's continue the explanation of the autonomous design AI processing (S37). After the processing in S371, the autonomous design AI unit 33 determines whether the design data loading button 86 on the autonomous design AI dialogue screen 85 has been selected (S372). If the design data loading button 86 has been selected (S372: Yes), the autonomous design AI unit 33 prompts the designer to select the design data file to be loaded, and then loads the design data from the selected file (S373). Next, the autonomous design AI unit 33 performs an analysis based on the loaded design data to autonomously determine whether the design of the product has been carried out using the knowledge necessary for the design of the product to be designed (S374). Then, as shown in Figure 18(b), the autonomous design AI unit 33 displays the analysis result in the knowledge judgment display area 87 of the autonomous design AI dialogue screen 85 (S375). This display in the knowledge judgment display area 87 allows the designer to understand whether the design of the product has been carried out using the knowledge necessary for the design of the product to be designed based on the loaded design data. Furthermore, the processing in S374 constitutes the "instruction awaiting item determination means" of the present invention.

[0218] After processing in S375, or as a result of the judgment in S372, if the design data loading button 86 is not selected (S372: No), the autonomous design AI unit 33 then checks the knowledge determination display area 87 and determines whether the execute button 90 or the pass button 91 is selected, with one of the knowledge items (items with "Status" set to "NON") that represent a lack of knowledge in the product design selected by the designer, the specific content of that knowledge displayed in the knowledge content display area 88, and the details of the external device that cooperates to acquire the missing knowledge as external knowledge displayed in the external cooperation content display area 89.

[0219] If, as a result of this decision, the execute button 90 is selected (S376: "Execute"), the autonomous design AI unit 33 cooperates with the external device displayed in the external cooperation content display area 89 to acquire the missing knowledge displayed in the knowledge content display area 88 as external knowledge (S377), and as a result of the cooperation, displays the screen 95 provided by the external device together with the autonomous design AI dialogue screen 85 (S378). This screen 95 allows the designer to confirm the external knowledge acquired from the external device. The process in S377 constitutes the "design continuation means" of the present invention.

[0220] Next, the autonomous design AI unit 33 determines whether the designer selected the OK button 95a or the NG button 95b, which are displayed together with the screen 95 (S379). If the OK button 95a was selected (S379: "OK"), the autonomous design AI unit 33 saves the knowledge acquired from the external device as external knowledge to the product design data, and then changes the "Status" in the knowledge judgment display area 87 for the knowledge acquired from the external device from "NON" to "OK" (S380). In addition, in the process of S380, the autonomous design AI unit 33 clears the display in the external linkage content display area 89 and displays the message "No analysis is required" in the knowledge content display area 88. Then, the autonomous design AI unit 33 proceeds to the process of S381.

[0221] On the other hand, if the autonomous design AI unit 33 determines in S379 that the NG button 95b has been selected (S379: "NG"), it skips the process in S380 and proceeds to the process in S381. Also, if the autonomous design AI unit 33 determines in S376 that the Pass button 91 has been selected (S376: "Pass"), it skips the processes in S377 to S380 and proceeds to the process in S381.

[0222] In the process of S381, the autonomous design AI unit 33 determines whether the design data overwrite save button 92 has been selected. If the design data overwrite save button 92 has been selected (S381: Yes), the current design data (design data with external knowledge reflected through the process of S378, etc.) is overwritten and saved to the design data file read in the process of S373 (S382).

[0223] Furthermore, after processing in S382, or as a result of the judgment in S381, if the design data overwrite save button 92 is not selected (S381: No), the autonomous design AI unit 33 then determines whether the drawing button 93 is selected (S383). If the drawing button 93 is selected (S383: Yes), the AI ​​unit has the CAD system create a drawing of the product based on the current design data and displays the CAD drawing on the display device (not shown) of the product design system 30 (S384). This allows the designer to check for any problems by reviewing the current CAD drawing.

[0224] If, after processing in S384, or as a result of the determination in S383, the drawing button 93 is not selected (S383: No), the autonomous design AI unit 33 then determines whether the exit button 94 is selected (S385). If the exit button 94 is not selected (S385: No), the process returns to S372. The processes from S372 to S385 are then repeatedly executed until the exit button 94 is selected. On the other hand, if, as a result of the determination in S385, it is determined that the exit button 94 is selected (S385: Yes), the autonomous design AI process ends and the process returns to the design agent AI process (see Figure 8(a)). At this time, the selection screen 40 is displayed on the display device (not shown) of the product design system 30.

[0225] As described above, the autonomous design AI unit 33 can autonomously assist in product design while interacting with the designer via the autonomous design AI dialogue screen 85. Specifically, the autonomous design AI unit 33 determines that acquiring knowledge that is lacking in the product design from an external device is a matter for which it should seek instructions from the designer. When the acquisition of the lacking knowledge from an external device is accepted as an instruction from the designer by operating the execution button 90, the design continues by coordinating with the external device that possesses the necessary knowledge to acquire it. In this way, the autonomous design AI unit 33 can autonomously acquire knowledge that is lacking in the product design from an external device while interacting with the designer.

[0226] <4. Designing a new product> Next, referring to Figures 20 to 22, the process for designing a product with a new structure for which the component design AI unit 31 does not have structural knowledge will be explained. Figure 20 shows the process for giving design instructions for a product with a new structure to the component design AI unit 31 from the knowledge registration button 54, where (a) is a schematic diagram showing the environmental condition input screen 50 before giving the design instructions, (b) is a schematic diagram showing the environmental condition input screen 50 after giving the design instructions, and (c) is a schematic diagram showing the contents of the CSV file for giving the design instructions.

[0227] Furthermore, Figures 21 and 22 illustrate the flow of the autonomous design AI unit when a design instruction for a product with a new structure is given to the component design AI unit 31, and the AI ​​unit acquires structural analysis of the product from external knowledge as missing knowledge. Figure 21(a) is a schematic diagram showing the autonomous design AI dialogue screen 85 at the stage when the missing knowledge is determined, Figure 21(b) is a schematic diagram showing the autonomous design AI dialogue screen 85 at the stage when structural analysis of the new product is selected as missing knowledge, and Figure 22 is a schematic diagram showing the autonomous design AI dialogue screen 85 at the stage when structural analysis of the new product is acquired as external knowledge.

[0228] When designing a product with a new structure in the product design system 30, first, the component design AI processing (S33) is executed from the selection screen 40, and the instruction to design the product with the new structure is registered as knowledge as an "autonomous will" to the component design AI unit 31 that performs component design inference. This knowledge registration is performed in the same way as when registering knowledge about a new component, by the designer selecting the knowledge registration button 54 on the environmental condition input screen 50 shown in Figure 20(a) and loading the CSV file shown in Figure 20(c).

[0229] For example, a CSV file for registering design instructions for a folding umbrella as a product with a new structure, as shown in Figure 20(c), contains the conceptual value "autonomous will" for the representative concept "#" and the conceptual value "folding umbrella" for the subordinate concept "umbrella". By loading this CSV file, the design of the folding umbrella is displayed as an autonomous will in the registered parts display area 56, as shown in Figure 20(b). The parts design AI unit 31 then performs structural design inference or structural design of the folding umbrella based on the environmental conditions entered (or pre-specified) in the environmental conditions input area 51 of the environmental conditions input screen 50, and then performs parts design inference for the folding umbrella.

[0230] Alternatively, an input area for providing design instructions for a new product may be provided on the environmental conditions input screen 50, allowing instructions for a new product to be given from that input area.

[0231] Currently, the component design AI unit 31 does not possess knowledge about the structure of the new product. Knowledge about the structure of the new product can be registered using the knowledge registration button 54 on the environmental conditions input screen 50. Furthermore, the autonomous design AI unit 33 can collaborate with an external device to acquire knowledge about the structure of the new product from external knowledge sources.

[0232] When the autonomous design AI unit 33 acquires knowledge about the structure of a new product from external knowledge, the designer first registers the instruction to design a product with the new structure as an "autonomous will" in the component design AI unit 31, then selects the structural calculation button 52 on the environmental conditions input screen 50 (however, the structural calculation is not performed because there is no knowledge about the structure), and then selects the save button 66 on the component design AI dialogue screen 60 that is displayed thereafter to save the design data.

[0233] The designer then selects the exit button 68 on the component design AI dialogue screen 60 to return to the selection screen 40, and then selects the autonomous design button 43 from the selection screen 40 to execute the autonomous design AI process (S37). After that, the designer selects the design data loading button 86 on the displayed autonomous design AI dialogue screen 85 to load the design data that was previously saved in the component design AI process (S33). The autonomous design AI unit 33 analyzes this design data and determines that it lacks the knowledge necessary to perform structural analysis of a new product, and displays a message to that effect ("Status" to "NON") in the knowledge judgment display area 87, as shown in Figure 21(a).

[0234] Furthermore, the design agent AI unit 34 may autonomously coordinate the component design AI unit 31 and the autonomous design AI unit 33 without using design data. For example, if the component design AI unit 31 registers a design instruction for a product with a new structure as an "autonomous intention," it may output to the design agent AI unit 34 that it lacks knowledge about the structure of that product. Based on this output, the design agent AI unit 34 may instruct the autonomous design AI unit 33 to acquire knowledge about the structure of that product from external knowledge sources in cooperation with an external device.

[0235] When the designer selects the knowledge for structural analysis displayed in the knowledge judgment display area 87, the specific content of that knowledge (in the example in Figure 21(b), "Design the structure of a folding umbrella") is displayed in the knowledge content display area 88, as shown in Figure 21(b). Additionally, details of the external device to which the designer will collaborate to acquire the missing knowledge as external knowledge (in the example in Figure 21(b), "Perform structural design at an umbrella specialist website") are displayed in the external collaboration content display area 89. After confirming the contents of the knowledge content display area 88 and the external collaboration content display area 89, the designer selects the execute button 90, which launches the structural analysis tool of the external device, and its screen 95 is displayed. The designer confirms the contents of screen 95, and if there are no problems with the structural design of the new product displayed on screen 95, selects the OK button 95a to acquire knowledge about the structure of the new product (knowledge of the structure of the newly designed product) as external knowledge. Then, the designer selects the design data overwrite save button 92, which overwrites and saves the design data, including the knowledge of the structure of the newly designed product.

[0236] The design of the new product can then be advanced by the designer launching the component design AI processing (S33) and performing structural calculations and component design inference from the design data, which includes knowledge of the structure of the new product.

[0237] As described above, if the product design system 30 does not possess knowledge about the structure of the product to be designed, the autonomous design AI unit 33 acquires knowledge about the structure of the product to be designed from an external device, and the specifications of the product's structure are inferred. Thus, if the product design system 30 does not possess knowledge about the structure of the product to be designed, it can autonomously acquire that knowledge from an external device and perform structural design of the product.

[0238] <5. Summary> As described above, according to the product design system 30 of this embodiment, the component design AI unit 31 infers the components of a product based on the environmental conditions under which the product will be used, utilizing the component design AI unit 31's internal knowledge of components. Furthermore, the support design AI unit 32 provides support for product design by utilizing the support design AI unit 32's internal knowledge of design. In addition, the autonomous design AI unit 33 autonomously determines the knowledge lacking in product design and collaborates with external devices possessing that knowledge. Finally, the design agent AI unit 34 coordinates the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33, thereby performing structural and component design of the product while making autonomous decisions. This enables autonomous product design.

[0239] Furthermore, according to the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33 of this embodiment, when designing a product, the system autonomously determines, based on the current design status, which items require instructions from the designer. The system then accepts input of instructions from the designer regarding the determined items. Based on these instructions, the product design continues. This allows the system to autonomously design a product while interacting with the designer.

[0240] <6. Variation> Although the present invention has been described above based on embodiments, it is easy to infer that the present invention is not limited in any way to the above embodiments, and that various improvements and modifications are possible without departing from the spirit of the present invention. For example, each embodiment, including the modified examples described above, may be constructed by modifying the embodiment by adding or replacing some or more parts of the configuration of other embodiments with parts or more parts of the configuration of other embodiments. Furthermore, the numerical values ​​given in the above embodiments are merely examples, and it is naturally possible to use other numerical values.

[0241] In the above embodiment, the case in which the design agent AI unit 34 coordinates the component design AI unit 31, the support design AI unit 32, and the autonomous design AI unit 33 via design data was described. However, the design agent AI unit 34 may also coordinate each of the design AI units 31 to 33 by operating the other design AI units 31 to 33 based on information output from each of the design AI units 31 to 33, without using design data.

[0242] For example, when the component design AI unit 31 has performed component design inference, it may output information to the design agent AI unit 34 indicating that it has done so, and the design agent AI unit 34 may then use that information to instruct the support design AI unit 32 to determine whether the combination of components inferred by component design inference satisfies the checklist and / or design rules.

[0243] Furthermore, if the component design AI unit 31 and / or the support design AI unit 32 have insufficient knowledge in structural design reasoning or structural design, component design reasoning, or support design, they may output information indicating the lack of knowledge to the design agent AI unit 34. Based on this information, the design agent AI unit 34 may have the autonomous design AI unit 33 acquire the missing knowledge as external knowledge in cooperation with an external device. This allows the design agent AI unit 34 to coordinate each of the design AI units 31 to 33 more autonomously. [Explanation of symbols]

[0244] 30 Product Design Systems 31. Component Design AI Unit 32. AI Unit for Assistive Design 33 Autonomous Design AI Unit 34 Design Agent AI Unit 40 Selection screen 41 Component Design Button 42. Support Design Button 43 Autonomous Design Button 50. Environmental Conditions Input Screen 51. Environmental Condition Input Area 52 Structural Calculation Button 53 Structural condition display area 54 Knowledge Registration Button 55 Inference Button 56 Registered parts display area 57. Part Modification Button 60 AI Dialogue Screen for Part Design 61 Structural condition display area 62 Re-inference button 64 Inference result display area 65 Selected Product Display Area 66 Save button 67 Drawing Buttons 70 AI Dialogue Screen for Assistive Design 71 Design Data Load Button 72 Analysis result display area 73 Correction instruction display area 74 AI Correction Button 75 CAD Modification Button 76 Modified content display area 77 Design data overwrite save button 78 Drawing Buttons 80 CAD drawings 81 CAD drawings 85 Autonomous Design AI Dialogue Screen 86 Design Data Load Button 87 Knowledge judgment display area 88 Knowledge content display area 89 External Linkage Content Display Area 90 Execute button 92 Design data overwrite save button 93 Drawing button 95 screens 95a OK button

Claims

1. A component design AI unit that infers the components of the product by utilizing the knowledge of components that the artificial intelligence possesses, based on the environmental conditions under which the product is used. An AI support design unit that utilizes the design knowledge internally possessed by artificial intelligence to assist in the design of the aforementioned product, An autonomous design AI unit that, based on its own will, determines the knowledge lacking in the design of the aforementioned product and cooperates with an external device possessing that knowledge, A product design system characterized by comprising: a component design AI unit, an assistance design AI unit, and a design agent AI unit that designs the product by coordinating the autonomous design AI unit.

2. The product design system according to claim 1, characterized in that the component design AI unit includes structural design reasoning means for inferring the structural specifications of the product based on the characteristics and / or features of the product to be designed.

3. The product design system according to claim 2, characterized in that, when the product design system does not possess knowledge regarding the structure of the product to be designed, the autonomous design AI unit acquires knowledge regarding the structure of the product to be designed from the external device and infers the specifications of the structure of the product.

4. The product design system according to claim 1, characterized in that the component design AI unit infers the components of the product based on structural conditions, which are conditions that the structure of the product should have, including the environmental conditions.

5. The product design system according to claim 4, characterized in that the component design AI unit includes a condition change acceptance means for accepting changes to the structural conditions from a designer if the inferred component of the product does not satisfy the structural conditions.

6. The product design system according to claim 1, characterized in that the component design AI unit generates a drawing of the product design data based on the component parts of the product that it has inferred, by linking it to a CAD system.

7. The product design system according to claim 1, characterized in that the AI ​​support design unit internally possesses design checklists and / or design rules as design knowledge.

8. The product design system according to claim 7, characterized in that the support design AI unit includes design checking means for identifying parts that deviate from the checklist and / or design rules in the design of the product using the parts of the product inferred by the parts design AI unit.

9. The product design system according to claim 8, wherein the support design AI unit includes a component modification means that modifies the component parts of the product to satisfy the checklist and / or the design rules when the design checking means identifies a part that deviates from the checklist and / or the design rules in the design of the product using the component parts of the product inferred by the component design AI unit.

10. The product design system according to claim 1, characterized in that the autonomous design AI unit can be linked with at least one of the following external devices: a design support tool providing system that provides design support tools for supporting the design of the product; an analysis tool providing system that provides analysis tools for analyzing the designed product; an estimation system for estimating the manufacturing and / or processing costs of the designed product; and a knowledge database for storing technical knowledge.

Images