System and method for asymptotically creating industrial software through human-computer interaction
Through the progressive human-computer interaction system, combined with low code modeling and intelligent interaction assistant, the problems of complexity and high failure rate of traditional industrial software development are solved, and rapid and reliable industrial software development is achieved to adapt to changes in complex demands.
Patent Information
- Application Number
- CN202311412958.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-27
- Publication Date
- 2024-11-15
AI Technical Summary
Traditional industrial software development process is complex, time-consuming and has a high failure rate. Especially when multi-system interaction is complex, it is difficult to respond quickly to demand changes, resulting in long development cycles and high risk of failure.
Adopt a human-computer interaction progressive system, combining a low-code modeling platform and an intelligent interaction assistant, receive the needs of operators through voice or text, interpret and feedback information, and call the low-code modeling platform to automatically generate industrial software, supporting industrial business engines, Establishment and update of data models and business models.
It greatly reduces the risk of errors in requirements and project failures, improves development efficiency and software quality, and allows non-developers to quickly build complex industrial software systems to adapt to changes in demand.
Smart Images

Figure CN118966989A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an application of artificial intelligence automatic programming, in particular to a system for creating industrial application software through deep learning using a human-computer interaction mode, and also to a method for creating industrial application software using such a system. Background Art
[0002] There are a large number of application demands in industrial fields such as manufacturing, logistics and warehousing. These industrial application demands have the characteristics of many customized requirements, difficult to clearly describe the requirements in advance, and numerous and changeable processes. The traditional industrial software production method still uses the programming method of professionals, which has defects such as long programming process, high failure rate, and poor adaptability. For example, for a typical automated logistics and warehousing application software in a manufacturing plant, traditional WMS software manufacturers generally need a development cycle of 3-6 months. If the internal systems interact complexly, such as the need to connect ERP software, MES manufacturing management software, material management software, process management software, stackers, automated warehouses, automatic guided vehicles (AGV / AMR), robots and other hardware systems, the complexity will soar rapidly. As a result, the development cycle is often measured in years, and the development failure rate is greatly increased, and the risk of unfinished projects is very high.
[0003] The above industrial application software can be built using low-code application technology, providing a landing platform for software development in the industrial application field. In recent years, low-code technology has developed rapidly. After professional training and a lot of practical exercises, developers can quickly build professional prototype systems. Compared with traditional development methods, it not only saves a lot of time, but also makes the software quality more reliable and has strong response capabilities when requirements change. However, low-code technology also requires operators to have professional system architecture and modeling experience, which is difficult for non-development professionals to master.
[0004] With the development and progress of artificial intelligence technology, the method of using robot artificial learning has been initially implemented in many fields. It is a good idea to use artificial intelligence technology in combination with traditional industrial application software models to build an intelligent programming platform to realize on-site use of industrial application software, on-site programming, and improve production efficiency.
[0005] Therefore, the present invention develops a method for developing industrial application software using a progressive artificial intelligence system with human-computer interaction. Summary of the invention
[0006] The purpose of the present invention is to provide a system for creating industrial software in a progressive manner through human-computer interaction, and at the same time to provide a method for creating industrial software through the above system.
[0007] The system of the present invention for incrementally creating industrial software through human-computer interaction includes a low-code modeling platform, an intelligent interaction assistant and a human-computer interaction module; the human-computer interaction module is used to receive instructions from an operator and receive information feedback from the intelligent interaction assistant, and the intelligent interaction assistant interprets the operator's instructions and the instruction information; the interpreted instruction information is sent to the low-code modeling platform for identification and fed back to the human-computer interaction module, and after the modeling platform identifies the instruction information, it calls the low-code modeling platform to automatically generate industrial software.
[0008] In the above-mentioned system for incrementally creating industrial software through human-computer interaction, the low-code modeling platform sets up an industrial business engine, an industrial domain model, a data model and a business model, and a corresponding relationship is established between the industrial business engine and the industrial domain model, the data model and the business model; the low-code modeling platform sets up an external interface to connect with the intelligent interaction assistant and the human-computer interaction module.
[0009] In the above-mentioned system for incrementally creating industrial software through human-computer interaction, the industrial business engine includes an industrial data acquisition engine, a custom process engine, a custom report engine, image detection training modeling for small customized scenarios, and a model application and iteration engine.
[0010] In the above-mentioned system for incrementally creating industrial software through human-computer interaction, the industrial domain model includes a general domain model and a special domain model.
[0011] In the above-mentioned system for incrementally creating industrial software through human-computer interaction, the data model includes domain model internal tables, domain model external tables, menu items, fields, sub-tables, related tables, statistical reports and other tables, and establishes mutual association relationships.
[0012] In the above-mentioned system for incrementally creating industrial software through human-computer interaction, the business model includes a creation and approval business model and an image detection business model.
[0013] The industrial software creation method of the human-computer interaction incremental creation industrial software system of the present invention comprises the following steps:
[0014] A: The operator inputs the scene requirement information and transmits it to the intelligent interactive assistant. The intelligent interactive assistant interprets the requirement information as instruction information and sends it to the low-code modeling platform for recognition and feedback to the human-computer interaction module.
[0015] B: After the low-code modeling platform recognizes the modeling instruction information, it calls the relevant business engine and creates the industrial application model software based on the instruction information. The low-code modeling platform does not recognize the instruction information and returns to perform operation A;
[0016] C: The operator modifies or supplements the scene requirement information based on the intelligent interactive assistant’s interpretation and feedback information. The intelligent interactive assistant interprets the information feedback again, and the operator and the intelligent interactive assistant determine the modeling areas and requirement information.
[0017] D: Confirm the demand information and send it to the low-code modeling platform. The low-code modeling platform calls the relevant business engine to connect to the relevant data model according to the scenario demand information and automatically builds the industrial application software;
[0018] E: During the use of industrial application software, we will continue to improve the industrial software and corresponding business engines and data models based on scenario changes and the interpretation of intelligent interactive assistants.
[0019] The specific process of the intelligent interactive assistant recognizing and interpreting natural language is as follows:
[0020] a: word segmentation and topic matching;
[0021] b: Is it a new topic? If it is an old topic, load the previously saved context information;
[0022] c: Match a certain interface command, confirm with the user, and establish the context information of a certain interface command;
[0023] d: Check whether the required operation parameters of the current interface command are complete. If any are missing, ask the user questions;
[0024] e: The interface commands and parameters are complete, and the user is asked for final confirmation.
[0025] The technical solution of the present invention is mainly to build a business system prototype within an industrial enterprise through human-computer interaction. The operator describes the requirements with the system intelligent interactive assistant through voice or text. The intelligent interactive assistant analyzes and interprets the voice or text and gives interpretation information. The interpretation information is fed back to the human-computer interaction interface, and the business engine of the modeling platform is called at the same time. If it is recognized by the business engine, the relevant model is called to establish industrial software. If it is not recognized by the business engine, the relevant information is returned. The operator further describes and explains it again through voice or text based on the feedback information, and then the intelligent interactive assistant interprets and recognizes it. This is repeated until it is recognized by the business engine, and the relevant business model is established to complete the industrial software production process.
[0026] The present invention provides business personnel with a descriptive method for building highly customized professional industrial software system prototypes suitable for small scenarios, allowing complex applications to be quickly verified, reducing the problem of rapid decay of effective demand understanding leading to implementation errors in the traditional chain from the owner's business manager to the implementer's business consultant to the implementer's development consultant to the implementer's product R&D personnel, thereby greatly reducing the risks of rework due to demand errors and project failure due to inability to adapt. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 It is a structural diagram of the human-computer interaction progressive industrial software creation system of the present invention;
[0028] Figure 2 It is a flow chart of the method for creating industrial software by human-computer interaction progressively of the present invention;
[0029] Figure 3 is a flow chart of the intelligent interactive assistant of the present invention interpreting natural language;
[0030] Figure 4 It is the operating UI of the human-computer interaction progressive industrial software creation system of the present invention. DETAILED DESCRIPTION
[0031] The following is a description of the human-computer interactive progressive industrial software creation system and method of the present invention in conjunction with the accompanying drawings and specific embodiments. The specific drawings and embodiments are only used for illustration and are not intended to limit the scope of protection.
[0032] The technical solution of the present invention is mainly to build a business system prototype within an industrial enterprise through human-computer interaction. The operator (which can be an implementation engineer or the user himself) can describe the requirements with the system intelligent interactive assistant through voice or text. After understanding and confirming the user's needs, the intelligent interactive assistant automatically calls the interface provided by the underlying industrial application low-code modeling platform to generate a business system prototype. In principle, the main difficulty of industrial software is the difficulty in defining and understanding requirements in professional fields and the difficulty in communication caused by the large number of participants. The present invention provides business personnel with a descriptive way to build a highly customized professional industrial software system prototype suitable for small scenarios, allowing complex applications to be quickly verified, reducing the problem of rapid attenuation of effective demand understanding leading to implementation errors in the traditional chain from the owner's business manager to the implementer's business consultant to the implementer's development consultant to the implementer's product R&D personnel, greatly reducing the risks of rework due to demand errors and avoiding project failures due to inability to adapt. Technically, the underlying application of low-code application building technology, assisted by artificial intelligence technologies such as voice and semantic analysis, provides a landing method for users who are familiar with business processes and rules but unfamiliar with industrial software design and development to quickly build professional system prototypes and conduct business verification, which can empower business consultants on the implementation side, business managers on the owner side and other personnel. Compared with traditional development methods, it not only saves a lot of time, but also makes the software quality more reliable and easier to deal with changes in requirements. Through the guidance of the intelligent interactive assistant, business personnel who do not have professional development experience can not only build a traditional simple system based on a single business form, but also "chat" with the intelligent interactive assistant to create a professional industrial software system that is associated with many data models, can adapt to the complex processes of real application scenarios, and integrates automated data interfaces and customized intelligent interactive assistant image recognition.
[0033] Typical application scenarios of this method are the following industrial software systems: production and manufacturing related systems such as manufacturing execution system (MES), logistics and warehousing related systems such as warehouse management system (WMS) or logistics execution system (LES), equipment monitoring related systems such as production safety monitoring, andon system, predictive maintenance system (PDM), etc.
[0034] In specific facts, such as Figure 1 As shown, first of all, it is necessary to establish an industrial software low-code data modeling platform, and the platform should at least support functions such as table creation, establishment of table associations, and customization of function module menus. The associations between tables include adding virtual association fields or calculated fields to tables, adding sub-tables or associated tables to tables, and supporting a base table to derive multiple concrete tables (that is, supporting object inheritance relationships). Tables are fixedly associated through integer type IDs. Generally, it is also necessary to provide front-end and back-end customization logic when operating the main and sub-tables. For example, when the check-in interface for the production process is opened, the (front-end) automatically fills in the current time and operator. After adding the check-in record, the (back-end) automatically detects whether the current process is completed and may then trigger the product work order production completion operation.
[0035] Secondly, it is necessary to expand the modeling platform with business engines related to industrial applications commonly used in various industrial fields, such as industrial data acquisition engines, custom process engines, custom reporting engines, image detection training modeling for small customized scenarios, and model application and iteration.
[0036] Then, the above basic data modeling capabilities and various business engines are provided with formal interfaces upward or externally, so that these capabilities can be further deployed and combined in addition to operating on the professional platform UI.
[0037] Finally, the intelligent interactive assistant is connected to support the process of chatting with users. Through semantic analysis (if it is a voice chat, it is converted to text first and then semantic analysis), it can deduce what domain model the user needs to build, or what data model to create or modify in the domain model, or what business model to create based on which business engine. After the user confirms, the data model or business model is created in real time. The user can continue to modify the generated application, still through the way of "chatting" with the intelligent interactive assistant.
[0038] Of course, the system of the present invention must also include a human-computer interaction module, which is equipped with keyboard, screen, voice and other devices as needed to issue various natural language commands and receive feedback information to achieve the function of human-computer interaction.
[0039] like Figure 2As shown in the figure, during the implementation process, the user first describes the demand in natural language. After receiving the demand, the intelligent interactive assistant performs semantic analysis and converts the user's natural language demand into the creation demand of the corresponding domain model, or data model, or business model. If the user's demand description is not enough to infer which specific operation is to be performed, or the user needs to further clarify the parameters of the operation, the intelligent interactive assistant should continue to interact and inquire with the user until the intelligent interactive assistant understands the user's upcoming operation and all its operation parameters. Therefore, the intelligent interactive assistant should support context. For example, if the user mentions the MES system and then mentions the working time demand, the intelligent interactive assistant should understand it as the working time sheet (current data model) in the MES system (this is a previously known domain model), without the user having to reiterate which domain model it is. When the intelligent interactive assistant obtains clear operation information, it should display the operation to the user in natural language, allowing the user to make the final confirmation or further modification.
[0040] When the user confirms the operation, the intelligent interactive assistant generates a call to an interface in a formalized manner (such as organizing it into a JSON object) and passes it to the industrial software low-code platform. After receiving the interface call, the platform creates or modifies the model in real time and presents it to the user, so what you see is what you get.
[0041] After that, after seeing the presented application prototype, the user continues to chat with the intelligent interactive assistant, gradually refining and adapting it to the scenario required by the user until the construction of the industrial software application prototype is completed.
[0042] like Figure 3 As shown in the figure, the process of the intelligent interactive assistant analyzing and interpreting the user's natural language is as follows:
[0043] First, semantic analysis is performed on the words sent by the user to determine which topic is being discussed, that is, which modeling platform interface (hereinafter referred to as the "platform interface"). The modeling platform should provide an interface (and parameters) for creating an application system. The interface is described in detail below. The method of semantic analysis is to segment the sentence, and important keywords (or synonyms) and each platform interface are weighted. Finally, the degree of match with each platform interface can be calculated, and the one with the highest match is determined as the new topic, and context information is created, such as the interface name, the degree of completion of each required parameter, etc.
[0044] If no new platform interface is matched, it is determined that it is not a new topic. At this time, the previously established context information can be loaded to check the required parameters. If it is not mentioned in the user's needs, it is necessary to ask the user. For example, when creating a table outside the domain model, all fields of the table must be clearly specified. At this time, the user should be asked which fields the table requires.
[0045] When the interface call information is complete, the upcoming operation will be presented to the user in natural language for final confirmation, and the optional parameters can be prompted.
[0046] The data model related interfaces are required, and the business model related interfaces are provided according to the business engine provided and encapsulated by the platform:
[0047] Create a domain model, parameter: domain model name. The software platform can contain several domain models, which are generally divided into general domain models and industry-specific domain models, such as general warehouse management domain model, general manufacturing management domain model, new energy vehicle manufacturing industry warehouse management domain model, electronic instrument manufacturing industry manufacturing management domain model, etc. Each domain model includes a series of data models and business models, and each model has an "Is it a basic model" flag. When creating a domain model, only models with the "basic model" flag are created by default. Other models can be introduced by users as needed.
[0048] Data model operation - add a table in the domain model, parameters: domain model name, table name. If the table required by the user already exists in the current domain model, it can be created directly after confirmation by the user. The user can then add, delete, and modify the table fields as needed. This interface is used for both table creation and table modification. After creation, the table can be directly opened to perform operations such as adding, deleting, modifying, and querying data.
[0049] Data model operation - add a table outside the domain model, parameters: table name, field list. If the user needs to create a table by himself, he must specify the table name and field names before creating the table.
[0050] Data model operation - add menu item, parameters: menu item, table name. Add the previously created table to the menu item to form an operation entry. Click the menu item to open the table for operation.
[0051] Data model operation - add / modify fields, parameters: table name, field name, field attributes (such as display type is list, list drop-down options). Modify the fields of an existing table.
[0052] Data model operation - add sub-table, parameters: main table name, main table associated fields, sub-table name. Associate the two tables, and when the main table is opened, the sub-table will automatically follow.
[0053] Data model operation - add associated table, parameters: main table name, main table associated field, associated table name. Associate two tables. When opening the main table, you can link to the associated table.
[0054] Data model operation - statistical report, parameters: menu item, report name, table name, statistical item, summary item. Perform summary statistics based on existing tables.
[0055] Business Engine - Create an approval business model: table name, approval field, approval role. This item customizes the approval process based on a table. There are many types of business engines, and the approval engine is one of the commonly used business engines.
[0056] Business Engine - Image Detection Business Model, parameters: table name, source image field, result image field, result field, model name. When adding an image to a specified table or image field, the corresponding model is automatically called to detect the target in the image (such as defect recognition in quality inspection), and the result is saved in the specified result field.
[0057] like Figure 4 The following is an example of interaction, taking the general MES domain model as an example:
[0058] The operator is referred to as "person"; the intelligent interactive assistant is referred to as "machine"; the low-code modeling platform is referred to as "platform".
[0059] Someone sent: I want to build an MES system to mainly manage work orders and related BOM materials;
[0060] Machine feedback: Match to the "general MES model" and let the user select or confirm to generate a prototype. Do you want to create "work order management", including work orders, workpieces and related materials?
[0061] The platform calls the general MES domain model and displays Figure 4 user interface.
[0062] People send: Yes
[0063] Machine feedback: Creation completed.
[0064] The modeling platform calls the existing tables in the general MES domain model interface.
[0065] People issue: Each product needs to go through several steps during the production process. The system needs to record the production process of each product so that the production process can be traced back afterwards.
[0066] Machine feedback: You can add production process management, technology and procedure management. Do you want to add them now?
[0067] Person sends: OK.
[0068] Machine feedback: Creation completed.
[0069] Platform operation: Add new tables that are not in the domain model.
[0070] Someone sent out: I need to increase after-sales management so that when the product is returned for repair, I can find the original production work order and process.
[0071] Machine feedback: What fields are needed in after-sales management and how are they related to existing tables?
[0072] Issued by: After-sales work order table, with work order status, after-sales work order number, and workpiece number fields, where the workpiece number is associated with the workpiece table.
[0073] Machine feedback: After-sales work order will be created, with the fields: work order status, after-sales work order number, workpiece number, and associated to the workpiece table. Confirm?
[0074] Person sent: Confirmed.
[0075] Machine feedback: Creation completed. Need to add to the menu?
[0076] Platform operation: Create an after-sales work order, add the following fields: work order status, after-sales work order number, workpiece number, and establish a link to the workpiece table.
[0077] People sent out: Add after-sales management menu item, and add after-sales work order menu.
[0078] Machine feedback: Menu item After-sales Management-After-sales Work Order, after clicking it, the after-sales work order form is displayed, is that right?
[0079] Man said: Yes.
[0080] Machine feedback: Creation completed.
[0081] Platform operation: Create after-sales management work orders.
[0082] Issued by: The work order status field needs to be changed to a drop-down list: Unprocessed, Processed, Not Processed; and an attachment field is added to upload files.
[0083] Machine feedback: Change the work order status field to a drop-down list, including Unprocessed, Processed, and Not Processed.
[0084] Platform operation: Change the work order status field to a drop-down list, divided into unprocessed, processed, and not processed.
[0085] Sent by: Add a notes table as a subtable of after-sales work order, including creation time, creator, and content fields.
[0086] Machine feedback: A notes table will be created with the following fields: creation time, creator, content field, and after-sales work order number.
[0087] Platform operation: Add a sub-table "Notes" to the after-sales work order table.
[0088] Issued by person: In the workpiece table, the association with the after-sales work order table is displayed.
[0089] Machine feedback: Add a virtual field "After-sales work order quantity" to the workpiece table. Click the quantity to open the after-sales work order associated with the workpiece. Confirm?
[0090] Platform operation: Add a virtual field "after-sales work order quantity" to the workpiece table.
[0091] Issued by: Add after-sales work order report to count the quantity in each status.
[0092] Machine Feedback: Please specify menu item.
[0093] Sent by a person: In after-sales management, under the after-sales work order report menu.
[0094] Machine feedback: Menu item After-sales Management - After-sales Work Order Report, add statistical report, the statistical item is the total number, and the summary item is the work order status. Confirm?
[0095] Person sent: OK.
[0096] Machine feedback: Completed.
[0097] Platform operation: Create statistical tables and establish associations.
[0098] Posted by: After-sales work order adds approval function
[0099] Machine feedback: Which role can approve?
[0100] People send out: use the after-sales review role.
[0101] Machine feedback: After-sales work orders will add approval status and approver fields. Add an after-sales review role, only this role can approve such documents. Please confirm the operation.
[0102] Sent by: Confirm
[0103] Machine feedback: Creation completed.
[0104] Sent by person: Add a device picture field and product type. When uploading the picture, the product type is automatically identified and filled in the product type field.
[0105] Machine feedback: In the after-sales work order table, add the fields device picture and product type, perform image detection on the device picture, output the result image to the field device picture detection result image, and output the detection result to the product type field. Confirm?
[0106] Person sent: Confirmed.
[0107] Machine feedback: Done!
[0108] Platform operation: The general MES system prototype has been created.
[0109] The internal operation process of the platform is:
[0110] Open the management application in development mode, activate the development menu, open the menu, System Settings - Development - Intelligent Interaction Assistant, and a chatbot window will appear. Through the dialogue, the following main functions can be achieved:
[0111] 1. Data model class
[0112] Add a new table (when adding a new table, if the object already exists, it will be overwritten, the same below)
[0113] Add menu
[0114] Modify Fields
[0115] Adding Domain Model
[0116] Add new association
[0117] Add subtable association
[0118] 2. Business Engine
[0119] Image Detection
[0120] Statistical Reports
[0121] Approval Process
[0122] For details, please see the examples in the Smart Interaction Assistant dialog box.
[0123] The functional interface of the intelligent interactive assistant, the data model class relies on the platform's secondary development engine, and the business engine class is provided by the platform's built-in functions (such as statistical reports) or various business plug-ins (such as approval processes, image detection, etc.).
[0124] If the table already exists, it will be overwritten and updated (if you want to make a patch update, you can use the Modify Field command) to associate tables, fields, and other tables. The UI displays a drop-down data table.
[0125] Drop-down options: The standard format is a list of "option 1; option 2" or a mapping list of "CR: create; RE: complete". It is also compatible with "option 1 option 2" or "CR = create RE = complete" (the first layer can be separated by ";" or Chinese or English spaces, and the second layer can be separated by ":" or "="). The UI will display as a drop-down option. Supports the special value "YesNoMap" which is equivalent to "0: No; 1: Yes"
[0126] The user can add, delete or modify fields in an existing table. If the specified field does not exist in the table, it will be added; if it does exist, it will be overwritten and updated.
[0127] For two tables in a master-child relationship, open the detail dialog box of the master table and display the child table below the dialog box. The table should be created in advance. In the associated table, there must be a field associated with the master table (that is, the field has specified the "associated table" and "associated field" parameters), otherwise an error will be reported during execution. For example, in the "Equipment Inspection Record" table, there is a "Device Id" field associated with the "Equipment" table. Fields associated with other tables must end with Id. The AI assistant will check if it does not end with Id and automatically add Id.
[0128] For two tables with an association relationship, a column showing the association table link is added to the main table list page. Clicking the link will jump to the associated table record corresponding to the row record. Adding a new association and adding a sub-table association are essentially the same, except that the UI display is different. The sub-table is displayed in the details dialog box of the main table, and the associated table is opened when clicking the link in the main table; you can even use associations and sub-table associations at the same time.
[0129] The essence of the domain model is addon, which can contain a series of data tables (DiMeta), pages (UiMeta), menu items, front-end and back-end custom logic, etc. The automatic loading path is addon / {domain model name}.xml.
[0130] For an image attachment field, automatically perform image detection when uploading an image, and record the output image with the annotation results to another output image field. The model is based on yolov8n and can be trained by yourself.
[0131] Image detection uses the class / AiDetect.php class, which relies on the functions provided by the yb-detect project. Currently, a fixed URL is used to access the interface: "http: / / localhost:8000 / ". yb-detect provides detection services implemented in c++ / php / c# in a production environment, using onnx format models (actually exported after training by the aidetect project); for installation and compilation, please refer to the yb-detect manual. You must install the opencv library and onnxruntime library, etc.
[0132] Multiple row statistics fields or column statistics fields can be specified, separated by spaces. You can specify a time statistics field, for example, "inspection time-year-month" means that "inspection time" is used as the time statistics field, and statistics are performed by "year-month". Here, "year-month" can also be changed to "year-month-day", "year-month-day-hour", "year-quarter", "year-week". The generated statistical report can be further improved using the custom report tool.
[0133] Generate an approval process on the specified table. The "Approve" button will be displayed on the list page, and you can view the approval record. Here, the simplest approval example is generated. Advanced configurations such as conditional approval, multi-level approval, and custom operations after approval are completed can all be operated in "Configure Approval Flow". When initiating an approval, there must be an employee with a specified approval role, otherwise an error will be reported if the approver cannot be found.
[0134] This is a special interface used to execute multiple functional interface commands in batches, such as creating multiple tables and establishing multiple associations at one time.
[0135] Implementation Mechanism
[0136] Front-end dialog box: page / dlgRobot.js
[0137] Interface function implementation: class / AC_Robot.php
[0138] Intelligent dialogue application: class / RobotImp.php
[0139] When the front end initiates a chat, the AC_Robot class implements the parsing logic:
[0140] If the parsed command is to execute a certain function, it will be executed directly; if parameters are missing or the execution is abnormal, an error will be reported and the chat will be returned.
[0141] When the analysis fails, the RobotImp class is used to extend the dialogue, realize the intelligent chat function, support context-related chat, and confirm before executing commands. It performs semantic analysis on the user's intention. If the command is matched but the parameters are incomplete, it should continue to ask the user to complete the parameters until the parameters meet the requirements of the functional command; before executing the functional command, the user should confirm. If the functional interface cannot be matched, it may be an unknown requirement raised by the user. The third-party intelligent platform will be extended to obtain the tables required to implement the user's requirements, and then returned to the user for confirmation and creation.
[0142] In summary, the present invention is described based on preferred embodiments, but the protection scope of the present invention should not be limited to the above embodiments, and improvements made on the basis of the concept of the present invention that do not deviate from the basic concept of the present invention should all fall within the protection scope of the present invention.
Claims
1. A system for incrementally creating industrial software through human-computer interaction, characterized in that: It includes a low-code modeling platform, an intelligent interactive assistant and a human-computer interaction module; the human-computer interaction module is used to receive the operator's instructions and receive information feedback interpreted by the intelligent interactive assistant, and the intelligent interactive assistant interprets the operator's instructions and interprets the instruction information; the interpreted instruction information is sent to the low-code modeling platform for identification and fed back to the human-computer interaction module, and after the modeling platform identifies the instruction information, it calls the low-code modeling platform to automatically generate industrial software.
2. The system for incrementally creating industrial software through human-computer interaction according to claim 1, characterized in that: The low-code modeling platform sets up an industrial business engine, an industrial domain model, a data model and a business model, and the industrial business engine establishes a corresponding relationship with the industrial domain model, the data model and the business model; The low-code modeling platform sets up an external interface to connect with the intelligent interaction assistant and the human-computer interaction module.
3. The system for incrementally creating industrial software through human-computer interaction according to claim 2, characterized in that: The industrial business engine includes an industrial data acquisition engine, a custom process engine, a custom report engine, an image detection training modeling for small customized scenarios, and a model application and iteration engine.
4. The system for incrementally creating industrial software through human-computer interaction according to claim 2, characterized in that: The industrial domain model includes a general domain model and a special domain model.
5. The system for creating industrial software by human-computer interaction progressively according to claim 2, 3 or 4, characterized in that: The data model includes domain model internal tables, domain model external tables, menu items, fields, sub-tables, related tables, statistical reports and other tables and establishes mutual association relationships.
6. The system for creating industrial software by human-computer interaction progressively according to claim 2, 3 or 4, characterized in that: The business model includes a creation and approval business model and an image detection business model.
7. An industrial software creation method for gradually creating an industrial software system by human-computer interaction as described in claims 1-6, characterized in that The following steps are involved: A: The operator inputs the scene requirement information and transmits it to the intelligent interactive assistant. The intelligent interactive assistant interprets the requirement information as instruction information and sends it to the low-code modeling platform for recognition and feedback to the human-computer interaction module. B: After the low-code modeling platform recognizes the modeling instruction information, it calls the relevant business engine and creates the industrial application model software based on the instruction information. The low-code modeling platform does not recognize the instruction information and returns to perform operation A; C: The operator modifies or supplements the scene requirement information based on the intelligent interactive assistant’s interpretation and feedback information. The intelligent interactive assistant interprets the information feedback again, and the operator and the intelligent interactive assistant determine the areas and requirement information that need to be modeled. D: Confirm the demand information and send it to the low-code modeling platform. The low-code modeling platform calls the relevant business engine to connect to the relevant data model according to the scenario demand information and automatically builds the industrial application software; E: During the use of industrial application software, we will continue to improve the industrial software and corresponding business engines and data models based on scenario changes and the interpretation of intelligent interactive assistants.
8. The industrial software creation method of claim 7, wherein: The process of the intelligent interactive assistant interpreting the operator's demand information includes the following steps: a: word segmentation and topic matching; b: Determine whether it is a new topic? If it is an old topic, load the previously saved context information; c: Match a certain interface command, confirm with the user, and establish the context information of a certain interface command; d: Check whether the required operation parameters of the current interface command are complete. If any are missing, ask the user questions; e: The interface commands and parameters are complete, and the user is asked for final confirmation.
Citation Information
Cited By
Method, system and equipment for realizing application generation and storage medium
CN119829026A
