Intelligent building design method, system, electronic device and medium based on AI model

Abstract

The application provides an intelligent building design method and system based on an AI model, an electronic device and a medium. The method comprises: acquiring a standard database of building design; the standard database comprises a standard library meeting building design specifications; utilizing an AI model to perform data matching according to user design requirements and the standard database to obtain basic data meeting the user design requirements; the AI model performs scheme design based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme; the AI model optimizes the preliminary design scheme according to user feedback data to obtain an optimized design scheme; and the user feedback data is feedback data of the user on the preliminary design scheme. The application can significantly improve design efficiency and design scheme quality, reduce design cost, and ensure the quality of the design scheme through the intelligent optimization function of the AI model.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology and relates to an intelligent building design method, system, electronic device and medium based on AI model. Background Technology

[0002] Architectural design is a complex decision-making process involving numerous stages. It encompasses a wide range of professional knowledge, including architecture, structure, plumbing, and electrical systems. Furthermore, architectural designs have extremely high safety requirements, subject to numerous regulatory constraints. Traditional architectural design relies heavily on human designers, demanding strong expertise and meticulousness. However, human design is often inefficient and produces lower-quality solutions, requiring redesigns when user needs are not met, resulting in significant labor costs and hindering large-scale customization. Summary of the Invention

[0003] This application provides an AI-based intelligent building design method, system, electronic device, and medium to address the problems of low design efficiency and high labor costs in the design process of traditional building design methods.

[0004] In a first aspect, this application provides an intelligent building design method based on an AI model. The method includes: acquiring a standard database for building design; the standard database includes a standard library that meets building design specifications; using an AI model to match data with user design requirements and the standard database to obtain basic data that meets the user design requirements; the standard database includes a standard component library, a standard unit type library, a standard core tube library, and / or a standard unit library; the basic data includes successfully matched standard component models, standard unit type models, standard core tube models, and / or standard unit models; the AI ​​model performs scheme design based on preset design rules and the basic data to obtain corresponding three-dimensional design drawings as preliminary design schemes; the AI ​​model optimizes the preliminary design schemes based on user feedback data to obtain optimized design schemes; the user feedback data refers to user feedback data on the preliminary design schemes.

[0005] In one implementation of the first aspect, the method of matching user design requirements with the standard database based on an AI model to obtain basic data that conforms to the user design requirements includes: performing semantic parsing processing on the user design requirements based on a large language model to obtain parsed user design requirement features; and matching the parsed user design requirement features with the standard database to obtain basic data that conforms to the user design requirements.

[0006] In one implementation of the first aspect, the AI ​​model designs a scheme based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme, including: the AI ​​model designs a scheme based on the basic data that meets the user's design requirements according to the trained neural network model and preset design rules to obtain corresponding label data; the AI ​​model translates and transforms the label data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme that meets the user's design requirements.

[0007] In one implementation of the first aspect, the training method for the neural network model includes: collecting basic architectural design data; the basic architectural design data includes an architectural design language dictionary, architectural design codes, architectural design standards, architectural design parameters, and / or architectural design cases; preprocessing the collected basic architectural design data to obtain preprocessed data; training the neural network model based on the preprocessed data to obtain a trained neural network model; and testing and evaluating the trained neural network model.

[0008] In one implementation of the first aspect, the AI ​​model optimizes the preliminary design scheme based on user feedback data to obtain an optimized design scheme, including: receiving user feedback data on the preliminary design scheme; the user feedback data including but not limited to cost, size, materials, aesthetics, and / or suggestions for scheme modification; and the AI ​​model optimizing the preliminary design scheme based on a genetic algorithm and the user feedback data to obtain an optimized design scheme.

[0009] In one implementation of the first aspect, the method further includes: receiving the user design requirements and feeding the user design requirements back to the AI ​​model, so that the AI ​​model can perform data matching with the standard database based on the user design requirements; storing the user feedback in a user database of the standard database, so that the AI ​​model can learn and optimize the design.

[0010] Secondly, this application provides an intelligent building design system based on an AI model. The system includes: a standard database; the standard database includes a standard library that meets building design specifications; and an AI model, the AI ​​model including: a data matching unit configured to match data with user design requirements and the standard database to obtain basic data that meets the user design requirements; the standard database includes a standard component library, a standard unit library, a standard core tube library, and / or a standard unit library; the basic data includes successfully matched standard component models, standard unit models, standard core tube models, and / or standard unit models; a design generation unit configured to design a scheme based on preset design rules and the basic data to obtain corresponding 3D design drawings as preliminary design schemes; and a design optimization unit configured to optimize the preliminary design scheme based on user feedback data to obtain optimized design schemes; the user feedback data refers to user feedback on the preliminary design scheme.

[0011] In one implementation of the second aspect, the AI ​​model further includes: a user interaction unit configured to receive the user design requirements and feed them back to the data matching unit, so that the data matching unit can perform data matching with the standard database based on the user design requirements; the user interaction unit is further configured to store the user feedback in a user database within the standard database, so that the AI ​​model can learn and optimize the design.

[0012] Thirdly, this application provides an electronic device, which includes: a memory storing a computer program; and a processor communicatively connected to the memory, which executes the AI ​​model-based intelligent building design method described above when the computer program is invoked.

[0013] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by an electronic device, implements the AI ​​model-based intelligent building design method described above.

[0014] As described above, the intelligent building design method, system, electronic device, and medium based on an AI model described in this application have the following beneficial effects:

[0015] This application designs a standard database for architectural design. The standard database includes a standard library that meets architectural design codes. An AI model is used to match user design requirements with the standard database to obtain basic data that meets those requirements. The standard database includes a standard component library, a standard unit type library, a standard core tube library, and / or a standard unit library. The basic data includes successfully matched standard component models, standard unit type models, standard core tube models, and / or standard unit models. The AI ​​model designs schemes based on preset design rules and the basic data to obtain corresponding 3D design drawings as preliminary design schemes. The AI ​​model optimizes these preliminary design schemes based on user feedback data to obtain optimized design schemes. The user feedback data refers to user feedback on the preliminary design schemes, which can significantly improve design efficiency and design quality, reduce design costs, and ensure the quality of the design schemes through the intelligent optimization function of the AI ​​model.

[0016] This application integrates a standard database of architectural design with AI models to generate optimized design solutions that meet user design needs. It can assist designers in improving design efficiency and innovation in various design fields (such as architectural design, industrial design, graphic design, etc.), promote design innovation, and enhance the artistry and practicality of design works.

[0017] This application integrates a standard database of architectural design with an AI model for design, and provides a user interaction interface through the AI ​​model. It realizes intelligent automatic generation and optimization of design schemes based on user-input design parameters and constraints, which is based on integrated machine learning algorithms, natural language processing technology and 3D modeling technology. This improves design efficiency and design quality, and also supports users to interact with the architectural design system through a graphical interface or natural language. It has important industry significance. Attached Figure Description

[0018] Figure 1A The diagram shown is a schematic representation of a hardware application scenario for intelligent building design based on an AI model, as described in an embodiment of this application.

[0019] Figure 1B The diagram shown is a hardware application scenario diagram of an AI-based smart building design according to another embodiment of this application.

[0020] Figure 1C The diagram shown is a structural schematic of the standard component library described in an embodiment of this application.

[0021] Figure 1D The diagram shown is a structural schematic of the standard apartment type library described in the embodiments of this application.

[0022] Figure 1EThe diagram shown is a structural schematic of the standard core cylinder as described in the embodiments of this application.

[0023] Figure 2 The diagram shown is a flowchart of the AI ​​model-based intelligent building design method described in the embodiments of this application.

[0024] Figure 3 The diagram shown is a schematic representation of the training process of the neural network model described in this application embodiment.

[0025] Figure 4 The diagram shown is a flowchart of an AI model-based intelligent building design method according to another embodiment of this application.

[0026] Figure 5 The diagram shown is a flowchart illustrating an AI model-based intelligent building design method according to another embodiment of this application.

[0027] Figure 6 The diagram shown is a structural schematic of the AI-based intelligent building design system described in this application embodiment.

[0028] Figure 7 The diagram shown is a structural schematic of the electronic device described in an embodiment of this application.

[0029] Component designation explanation

[0030] 1 User Design Requirements Platform

[0031] 2. AI Model-Based Intelligent Building Design Platform

[0032] 21 Standard Database

[0033] 22 AI models

[0034] 3. AI Model-Based Intelligent Building Design System

[0035] 31 Standard Database

[0036] 32 AI models

[0037] 321 Data Matching Unit

[0038] 322 Design Generation Unit

[0039] 323 Design Optimization Unit

[0040] 324 User Interaction Unit

[0041] 4 Electronic devices

[0042] 41 Processor / Processing Unit

[0043] 42 Memory

[0044] 421 RAM

[0045] 422 Cache Memory

[0046] 423 Storage System

[0047] 424 Programs / Utilities

[0048] 4241 Program Module

[0049] 43 bus

[0050] 44 Input / Output Interfaces

[0051] 45 Network Adapter

[0052] Steps S1 to S6

[0053] Steps S311~S314 Detailed Implementation

[0054] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0055] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0056] With the development of artificial intelligence technology, traditional design methods can no longer fully meet the design needs of modern society. While existing design software offers a degree of automation, these tools still require significant manual operation and lack intelligent functionality. Therefore, developing an intelligent design system capable of automatically generating and optimizing design solutions based on specific user needs is of great importance.

[0057] The following embodiments of this application provide an AI model-based intelligent building design method, system, electronic device, and medium, which can effectively improve design efficiency and quality and reduce labor costs in the design process.

[0058] The following embodiments of this application provide an AI model-based intelligent building design method, system, electronic device, and medium, including but not limited to a user design requirement platform and an AI model-based intelligent building design platform. The following description will take the user design requirement platform and the AI ​​model-based intelligent building design platform as examples.

[0059] like Figure 1A As shown, this embodiment provides a hardware application scenario for intelligent building design based on an AI model, specifically including: a user design requirement platform 1 and an AI model-based intelligent building design platform 2. The AI ​​model-based intelligent building design platform 2 includes a standard database 21 and an AI model 22. Users send their design requirements to the AI ​​model-based intelligent building design platform 2 through the user design requirement platform 1. The AI ​​model-based intelligent building design platform 2 matches standard components that meet the user's design requirements from the standard database 21 and generates an optimized design scheme through the AI ​​model 22, which is then fed back to the user.

[0060] The user design requirement platform 1, in its specific implementation, is not limited to any particular form; it can be a visual human-computer interaction interface (see [reference]). Figure 1B As shown, the interaction can be in the form of voice, non-contact action, or file interaction such as text, images, audio, and video. The user design requirements platform supports graphical user interfaces and natural language processing technology, allowing users to interact with the platform in natural language.

[0061] like Figure 1B-1E As shown, the standard database 3 includes a standard resource library that conforms to design specifications for architectural design; the standard resource library includes standard component libraries, standard unit type libraries, standard core tube libraries, and standard unit libraries, etc., at various levels.

[0062] The AI-based intelligent building design platform 2 includes an AI model, which is a model trained through algorithms and data that can simulate human intelligent behavior. The AI ​​model may include at least one deep learning model, which can each play a role after training, such as Transformer model, large language model, Convolutional Neural Network (CNN) model, etc.

[0063] The user design requirement platform 1 inputs the requirement parameter data representing the user design requirements into the AI ​​model-based intelligent building design platform 2 in a specific input form. The AI ​​model-based intelligent building design platform 2 uses the trained AI model to automatically match existing standard resources at all levels from the standard database based on the requirement parameter data to generate project building schemes.

[0064] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0065] like Figure 2 As shown, this embodiment provides an intelligent building design method based on an AI model, the method including the following steps S1 to S6.

[0066] Step S1: Construct a standard database for architectural design; the standard database includes a standard library that conforms to design specifications; the standard library includes a standard component library, a standard unit type library, a standard core tube library, and a standard unit library, etc.

[0067] In one implementation, such as Figure 1C As shown, the standard component library is a database integrating various standard components (such as prefabricated components or parts) for use in architectural design, manufacturing, and engineering, aiming to improve design efficiency, reduce costs, and ensure quality. Standard components refer to pre-designed and manufactured parts or assemblies widely used in architectural design, such as floor slabs, doors and windows, stairs, and prefabricated wall panels. These components typically conform to certain industry standards and specifications, facilitating mass production and installation. Using parametric design software, various standard component models are generated, and these models are categorized and summarized into the standard component library. Each standard component model has its own independent attributes, including type, material, measurement, pricing, resource consumption, and process code, serving as the data foundation for subsequent construction and operation / maintenance phases. The types of standard component models include walls, doors and windows, roofs, stairs, balconies, and railings, etc.

[0068] Specifically, the standard wall component models include different types of standard wall component models such as interior wall models, exterior wall models, and partition wall models, as well as related structural node models and connector models. Standardizing these component models helps improve the efficiency and quality of wall construction, ensuring the stability and functionality of the walls.

[0069] Specifically, the standard component models for doors and windows include door and window models of various materials, sizes, and functions. The door models include wooden door models, aluminum alloy door models, fire door models, etc.; the window models include wooden window models, aluminum alloy window models, etc. Standardized door and window component models facilitate installation and replacement while meeting both aesthetic and practical requirements of architecture.

[0070] Specifically, standard roof component models include roof panel models, purlin models, roof truss models, and other roof component models used to form the top structure of a building. Standardized roof component models can ensure the waterproofing, insulation, and load-bearing performance of the roof, and facilitate construction and maintenance.

[0071] Specifically, standard component models such as staircases, balconies, and railings play a role in connection and safety protection in architectural design. Standardized model design can improve safety and reliability, and also facilitate mass production and installation.

[0072] In one implementation, such as Figure 1D As shown, the standard apartment layout library is an important resource in residential building design, containing a series of carefully designed and optimized residential apartment layout schemes. The standard apartment layout library consists of a collection of standard apartment layout models, and the attributes of each standard apartment layout model include apartment type, apartment area, orientation, building form, and economic standards.

[0073] The standard apartment layout model includes combinations of bedrooms, living rooms, and bathrooms, such as one-bedroom, one-living room, one-bathroom, two-bedroom, one-living room, one-bathroom, three-bedroom, two-living room, one-bathroom, and / or three-bedroom, two-living room, two-bathroom layouts. One-bedroom, one-living room, one-bathroom is the basic residential layout, suitable for singles or newlyweds. Two-bedroom, one-living room, one-bathroom is a common layout, suitable for small families. Three-bedroom, two-living room, one-bathroom or three-bedroom, two-living room, two-bathroom layouts provide additional living room space, suitable for large families or families needing more public areas. The apartment layout types in the standard apartment layout model play a defining role in the final presentation of the standard apartment layout model.

[0074] The standard apartment layout model includes the internal area, living room area, bedroom area, and bathroom area. The apartment layout area in the standard apartment layout model has an optimizing and adjusting effect on the final presentation of the model.

[0075] The standard apartment layout models primarily include single-level, split-level, staggered-level, and duplex designs. A single-level apartment has all rooms on the same floor, facilitating living and movement. A split-level apartment occupies two floors connected by an internal staircase, offering greater space utilization possibilities. A staggered-level apartment has rooms on different levels, increasing the sense of space. A duplex is conceptually one floor but has two levels of space, with a higher overall ceiling height, and can incorporate a mezzanine. The architectural forms of the standard apartment layout models optimize and adjust the final presentation of the standard apartment layout model.

[0076] The economic standards of the standard apartment layout models mainly include compact, mid-range, and comfortable types. Compact apartments have a smaller area and a compact functional layout, suitable for homebuyers with limited budgets. Mid-range apartments have a moderate area and a reasonable layout, meeting the needs of middle-income families. Comfortable apartments have a larger area and complete functions, providing a higher quality of life. These economic standards of the standard apartment layout models have an optimizing and adjusting effect on the final presentation of the model.

[0077] In one implementation, such as Figure 1E As shown, the standard core tube library is an important resource in the field of architectural design, containing various types of building core tube design cases and code requirements. The standard core tube library is a database integrating multiple standard core tube models. A standard core tube is the central area of ​​a high-rise building, including vertical transportation facilities (such as elevators and staircases), equipment shafts, etc. This part is part of the building's shared area. Each standard core tube model has its own independent attributes, including core tube type, core tube area, core tube shape, etc. The collection of all standard core tube models constitutes the standard core tube library.

[0078] The core tube types in the standard core tube model mainly include elevator shafts, staircases, ventilation shafts, cable shafts, public restrooms, and / or some equipment rooms. Elevator shafts provide vertical transportation within the building, facilitating the efficient movement of people and materials. Staircases serve as emergency evacuation routes, ensuring building safety. Ventilation shafts ensure air circulation, providing fresh air and expelling stale gases. Cable shafts can accommodate cables and other wiring to meet power and communication needs. Public restrooms provide convenient hygiene facilities. Equipment rooms store and manage various building equipment, such as air conditioning systems and water supply and drainage systems. The standard core tube is an indispensable part of modern high-rise building design, providing necessary functional spaces and playing a crucial role in the overall stability and seismic performance of the building.

[0079] In one implementation, the standard unit library consists of a collection of standard unit models, which are formed by combining a varying number of standard apartment layout models with standard core tube models. The combination of standard apartment layout models and standard core tube models can be in various combinations conforming to building standards, such as one elevator serving two households, one elevator serving three households, or two elevators serving four households.

[0080] This application uses AI models to filter and calculate data from the standard unit type library and standard core tube library based on building requirement parameters (such as building area, land area, plot ratio, shared area ratio, and permissible building height). This results in multiple combination patterns that meet user design needs, generating corresponding standard unit models. The collection of these standard unit models constitutes the standard unit library. The attributes of each standard unit model correspond to the building requirement parameters; for example, the attributes of each standard unit model include building area, shared area ratio, and / or permissible building height.

[0081] In one implementation, when the designer inputs building requirement parameters (such as land area, allowable building height, plot ratio, building area, shared area ratio, number of buildings, etc.), the AI ​​model calculates the preset number of floors for each building based on the building requirement parameters.

[0082] For example: The total building area is determined based on the land area and plot ratio; where land area × plot ratio = total building area. The building area of ​​each building is determined based on the total building area and the number of buildings. The preset number of floors N for each building is determined based on the allowed construction height, where N is less than or equal to the quotient of the allowed construction height H divided by the floor height h. The building area of ​​each floor is determined based on the building area of ​​the building and the preset number of floors for the corresponding building. The internal area of ​​the corresponding floor is determined based on the building area of ​​the floor and the shared area ratio.

[0083] The AI ​​model calculates the building area of ​​each floor based on the preset number of floors and the number of buildings.

[0084] For example: If the number of buildings is m, and the number of floors in each building is preset to N1, N2, N3, ..., Nm, then the building area of ​​each floor = total building area / (N1 + N2 + ... + Nm).

[0085] The AI ​​model filters and calculates standard unit models in the standard unit library based on the building area of ​​the floor, and outputs at least one standard unit model that meets preset floor constraints. Alternatively, the AI ​​model filters and calculates models in the standard apartment type library and the standard core tube library based on the building area of ​​the floor, and outputs a standard unit model composed of at least one standard apartment type model and a standard core tube model that meets preset floor constraints.

[0086] The AI ​​model sorts the at least one standard unit model by indicators, and optimizes and adjusts the constraint parameters in the floor constraints based on the sorting results, so that the standard unit model continues to be optimized and calculated until the obtained standard unit model meets the user's design requirements. For example, the indicators for sorting the standard unit model include the number of rooms and living rooms, the internal area, the room orientation and / or the number of balconies, etc. The constraint parameters in the floor constraints include the sum of the internal area, the sum of the shared area, the sum of the building area, and / or the elevator-to-unit ratio of all units in the standard unit and standard core tube arrangement.

[0087] The AI ​​model obtains building models based on the standard unit model and the preset number of floors for each building. Using these models as a standard, the architectural plan for the entire project is ultimately generated.

[0088] For example, when the AI ​​model uses the Transformer model, its input parameters include input sequence, vocabulary size, embedding dimension, positional encoding, attention mask, label sequence, batch size, learning rate, optimizer, and loss function.

[0089] In the architectural design scenario described in this embodiment, the input sequence of the AI ​​model includes architectural requirement parameters (such as land area, permissible building height, plot ratio, building area, shared area ratio, number of buildings, etc.) and architectural design code text. The input sequence is the model's input data, typically a sequence of words or tokens. For text data, the input sequence can be words, characters, or subwords.

[0090] In the architectural design scenario described in this embodiment, the vocabulary size of the AI ​​model is defined as the index in a fixed-size vocabulary to which each token in the input sequence is mapped. For each different token appearing in the input sequence, a unique index is assigned to each token.

[0091] In the architectural design scenario described in this embodiment, the embedding dimension of the AI ​​model is the dimension that converts the labels in the input sequence into embedding vectors, and is usually a hyperparameter; it is used to convert each label in the input sequence into an embedding vector of a fixed dimension.

[0092] In the architectural design scenario described in this embodiment, the positional encoding of the AI ​​model is used to provide positional information of the markers in the sequence.

[0093] In the architectural design scenario described in this embodiment, the attention mask of the AI ​​model is used to indicate which positions in the sequence are valid, especially when processing sequences of unequal length, to prevent the model from paying attention to padding marks.

[0094] In the architectural design scenario described in this embodiment, the label sequence of the AI ​​model is a training target required by the model for certain tasks. For example, the training target may be to match a standard unit model in the standard unit library, or / and to match a combination of a standard apartment model in the standard apartment library and a standard core tube model in the standard core tube library.

[0095] In the architectural design scenario described in this embodiment, the batch size of the AI ​​model is used to determine the number of samples processed simultaneously during model training.

[0096] In the architectural design scenario described in this embodiment, the learning rate of the AI ​​model is used as a parameter in the optimization algorithm to adjust the weight update step size.

[0097] In the architectural design scenario described in this embodiment, the AI ​​model's optimizer is used to update the model weights using algorithms such as Adam and SGD.

[0098] In the architectural design scenario described in this embodiment, the loss function of the AI ​​model is used to evaluate the difference between the model output and the true label, such as cross-entropy loss, mean squared error, etc.

[0099] When the AI ​​model uses the Transformer model, its output parameters include the output sequence, probability distribution, loss value, gradient, updated weights, accuracy, other performance metrics, and predicted labels.

[0100] In the architectural design scenario described in this embodiment, the output sequence of the AI ​​model represents the model's prediction result, which is usually a labeled sequence.

[0101] In the architectural design scenario described in this embodiment, the probability distribution of the AI ​​model is used to output a probability distribution for each output label, representing the probability of each possible label.

[0102] In the architectural design scenario described in this embodiment, the loss value of the AI ​​model is used to guide weight updates during training.

[0103] In the architectural design scenario described in this embodiment, the gradient of the AI ​​model is the derivative of the loss function with respect to the model parameters, which is used to calculate the weight update.

[0104] In the architectural design scenario described in this embodiment, the updated weights of the AI ​​model are the updated values ​​of the model parameters after the backpropagation and optimization steps.

[0105] In the architectural design scenario described in this embodiment, the accuracy of the AI ​​model is the proportion of correct predictions made by the model, and is used to evaluate the model's performance.

[0106] In the architectural design scenario described in this embodiment, other performance metrics of the AI ​​model include the F1 score (used for classification tasks), etc.

[0107] In the architectural design scenario described in this embodiment, the predicted labels of the AI ​​model are the predicted labels that the model ultimately outputs for the classification task.

[0108] In the architectural design scenario described in this embodiment, the application task of the AI ​​model is to match the model in the standard library according to the architectural requirement parameters. Therefore, its input sequence is architectural requirement parameters, architectural requirement text and architectural design specification text, and the output sequence is the label corresponding to the model in the standard library.

[0109] In the architectural design scenario described in this embodiment, the labels corresponding to the model output by the Transformer model can be translated and converted to obtain the three-dimensional design drawings of the model corresponding to the labels. That is, the final output is the three-dimensional design drawings (or three-dimensional model drawings) of the architectural scheme of the entire project matched by the AI ​​model.

[0110] Since the original elements of the standard unit library are basic component libraries, which already contain the quantities of work, labor, and other resource consumption and prices required for pricing, the project quantities and costs can be directly calculated, providing a basis for project decision-making. Specifically, the quantities are calculated as follows: Quantity = Component 1 quantity + Component 2 quantity + ... + Component N quantity; Project cost = Quantity × Unit price.

[0111] Since the original elements of the standard unit library are basic component libraries, which already contain process codes, when importing a standard schedule containing process codes into the system, BIM components can be associated with the schedule, triggering corresponding tasks at the corresponding nodes to automate task dispatch.

[0112] In one implementation, this application imports a schedule into the NAVISWORKS software. For example, the code for a certain operation in the schedule is WS001, and the attribute value of the BIM component named "Schedule" is also WS001. The operation code is associated with the component with the same attribute value. When the schedule is implemented to a certain operation, the software dispatches the task related to the operation.

[0113] It should be noted that NAVISWORKS software is a project review software designed specifically for architecture, engineering, and construction professionals to review integrated architectural, structural, and MEP models. NAVISWORKS software supports Building Information Modeling (BIM) for buildings or infrastructure, providing a comprehensive platform for all stakeholders in building construction to integrate, coordinate, and analyze building-related data.

[0114] Step S2: Use the AI ​​model to match data with the user's design requirements and the standard database to obtain basic data that meets the user's design requirements. The standard database includes a standard component library, a standard house type library, a standard core tube library, and / or a standard unit library; the basic data includes successfully matched standard component models, standard house type models, standard core tube models, and / or standard unit models.

[0115] In one embodiment of this application, the use of an AI model to match data with the user's design requirements and the standard database to obtain basic data that meets the user's design requirements includes:

[0116] Step S21: Perform speech parsing processing on the user design requirements based on the large language model to obtain the parsed user design requirement features.

[0117] Step S22: Match the parsed user design requirements features with the standard database to obtain basic data that meets the user design requirements.

[0118] Specifically, this application uses an AI model to call a large language model to perform semantic parsing on the user design requirements, thereby obtaining the parsed user design requirement features. Then, it matches these parsed user design requirement features with a standard database to obtain basic data that conforms to the user design requirements. The user design requirements can be voice and / or text information sent by the user regarding design needs.

[0119] More specifically, the overall process of this application for speech parsing and processing the user design requirements using a large language model includes the following steps S71 to S73.

[0120] Step S71: Preprocessing of speech and / or text information, including filtering redundant information and handling special characters. Speech recognition is performed on the user-input speech information, and the resulting text data is imported into a pre-trained preprocessing model. Common meaningless words, recurring phrases, and redundant information are removed. Special characters in the text (such as URLs, email addresses, phone numbers, etc.) are then identified and retained, transformed, or removed as needed. This ensures the standardization and consistency of the text data for convenient subsequent processing and analysis.

[0121] Step S72, text correction, includes text context correction and iterative optimization based on user feedback. The preprocessed text data is imported into a fine-tuned large language model for text correction. By combining the text's context with speech recognition, errors and ambiguities in the speech recognition process are corrected, improving the readability and accuracy of the text and providing a reliable data foundation for subsequent analysis. Furthermore, through continuous user-self-corrected text information, the correction model learns and optimizes, achieving self-improvement and enhancing user satisfaction.

[0122] Step S73: Extract design requirement information that meets user design needs. Using the standard database mentioned above, perform topic identification and classification on the corrected text. Extract keywords related to design requirement information from the corrected text to quickly understand the core content of the user's voice and / or text information, providing a basis for subsequent real-time generation of basic data that meets the user's design needs.

[0123] It's important to note that Large Language Models (LLMs) are a type of natural language artificial intelligence model with powerful capabilities for understanding and generating human language. Trained on massive amounts of text and image data, they can perform functions including text analysis, semantic understanding, and drawing completion. Their defining characteristic is the sheer scale of their data and parameters, often reaching billions or tens of billions of parameters.

[0124] In one implementation, the AI ​​model matches corresponding basic data from the standard database based on user design requirements and the trained neural network model, and generates a preliminary design scheme based on the user design requirements and the matched basic data. The basic data includes successfully matched standard component models, standard unit type models, standard core tube models, and / or standard unit models.

[0125] Step S3: The AI ​​model designs a scheme based on preset design rules and the basic data to obtain the corresponding three-dimensional design drawing as a preliminary design scheme.

[0126] In one embodiment of this application, the AI ​​model designs a scheme based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme, specifically including the following steps S31 to S32.

[0127] Step S31: The AI ​​model designs a scheme based on the trained neural network model and preset design rules, using the basic data that meets the user's design requirements, to obtain corresponding label data. The preset design rules are design rules that meet building design codes.

[0128] Step S32: The AI ​​model translates and transforms the tag data to obtain the corresponding 3D design drawing, which serves as a preliminary design scheme that meets the user's design requirements.

[0129] In one implementation, the AI ​​model automatically matches existing standard libraries at various levels with the design parameter data input by the user through the interactive interface. It then calls a background AI algorithm to intelligently combine and match the optimal standard components to generate a design scheme and model. The design parameter data includes information such as apartment type, area, and / or floor level. For example, a user inputs a requirement to design a house. The input design parameter data includes 20 floors, two apartment types: a three-bedroom apartment with an area of ​​90 square meters and a two-bedroom apartment with an area of ​​80 square meters, with 60 units of each type. This represents the user's specific needs. Based on this design parameter data, the trained neural network model calculates and matches standard combinations such as standard apartment types and standard units that meet the user's requirements in the standard database. A preliminary design scheme is generated based on the matched standard combinations. In other words, the preliminary design scheme is a scheme generated by calculating standard combinations such as standard apartment types and standard units that meet the user's specific needs from the enterprise standard library. The preliminary design scheme must include at least one design scheme that meets the user's requirements.

[0130] like Figure 3 As shown, in one embodiment of this application, the training process of the neural network model includes the following steps S311 to S314.

[0131] Step S311: Collect basic architectural design data; the basic architectural design data includes architectural design language dictionary, architectural design code, architectural design standard, architectural design parameters and / or architectural design cases.

[0132] Specifically, this application obtains design case data from corporate databases or the internet. For example, this application obtains the user's design data from a user database.

[0133] In one implementation, the design data collected in this application is divided into two parts: a training dataset for training the model and a test dataset for testing and evaluating the trained model.

[0134] Step S312: Preprocess the collected basic architectural design data to obtain preprocessed data.

[0135] Specifically, this application involves cleaning the collected data, extracting key features from the cleaned data, and preparing a training set based on the extracted key features. The key features include the drawing standards for each component, and the extraction of key features must be based on national architectural drawing standards.

[0136] Step S313: Train the neural network model based on the preprocessed data to obtain the trained neural network model.

[0137] Specifically, this application uses a deep learning model (such as a convolutional neural network CNN) to train the training set and generate a preliminary design scheme.

[0138] In one implementation, this application uses a CNN model to train the model to learn design styles and element combinations. For example, this application trains the model based on past design cases from different users. The trained model becomes familiar with different user preferences. When these users submit design requests, the trained model can recommend matching design styles and element combinations, thus accelerating the generation of design solutions and improving design efficiency. Here, "style" refers to a specific design effect preferred by users of different cultures, regions, and types, such as Nordic style or Chinese style. Furthermore, different element combinations create different style expressions; for example, Roman columns, fireplaces, and central halls are different element combinations, each belonging to a different style expression.

[0139] Step S314: Test and evaluate the trained neural network model.

[0140] Specifically, this application uses a test dataset to test and evaluate the trained model to ensure the accuracy of the evaluation results. If the design scheme generated by the trained model meets the user's design requirements, the evaluation result is accurate; if the design scheme generated by the trained model does not meet the user's design requirements, the evaluation result is inaccurate, and the model needs to be retrained.

[0141] In one embodiment of this application, the application uses a Generative Adversarial Network (GAN) model to generate candidate design schemes through the AI ​​model, and improves the design schemes through multiple rounds of iteration to better meet the specific needs of the user, thereby obtaining a preliminary design scheme that meets the specific needs of the user. The specific steps for obtaining a preliminary design scheme that meets the specific needs of the user by using a Generative Adversarial Network (GAN) model include the following steps S81 to S85.

[0142] Step S81: Data Acquisition and Processing. Users provide design requirements through the input module of the AI ​​model, including but not limited to functional requirements, structural constraints, and appearance preferences. Design requirements may also include basic data, such as historical design schemes, industry standards, user preferences, and environmental limitations.

[0143] The AI ​​model first preprocesses the user's input design requirements, transforming them into feature vectors suitable for the AI ​​model's processing. Preprocessing includes text parsing, numerical feature standardization, and categorical feature encoding.

[0144] Step S82: Generate candidate design schemes using a GAN model. The GAN model consists of a generator and a discriminator. The generator is used to generate candidate design schemes, and the discriminator is used to evaluate whether the generated schemes meet the design requirements and their similarity to real design samples. The specific structure is as follows:

[0145] Generator: The generator is a deep neural network that takes the feature vector of the user's design requirements as input. The generator generates a representation of the preliminary design scheme through a multi-layer neural network, and the output is the parameter set of the design scheme.

[0146] Discriminator: The discriminator is also a deep neural network. Its input consists of the design scheme generated by the generator and the actual design scheme. The goal of the discriminator is to determine whether the input design scheme is a high-quality and feasible solution, and to output a quality score.

[0147] Step S83: Multi-round iterative process of generation and evaluation. The Generative Adversarial Network (GAN) employs a multi-round training and generation process to gradually optimize the generator's output to meet the user's specific needs. The specific iterative process is as follows:

[0148] Initial Generation: In the first round, the generator produces several candidate design schemes based on the user's input design requirements. Each design scheme is represented by design parameters, such as apartment type, area, number of floors, materials, color, and other attributes.

[0149] Discriminant and Feedback: The discriminator evaluates each candidate design and outputs a matching score (e.g., a value between 0 and 1), which represents the quality of the design and the degree to which it meets the user's design requirements. The discriminator's feedback is used to calculate the generator's loss function.

[0150] Generator optimization: The generator's weights are adjusted based on feedback from the discriminator, ensuring that the generated design schemes in the next round better meet the user's design requirements. This process is achieved through the backpropagation algorithm.

[0151] User feedback participation: After several rounds of generation, the system will provide users with visual representations of some candidate design schemes (e.g., 3D models, 2D sketches, etc.). Users can provide feedback on the design schemes, such as adjusting certain parameters or selecting specific styles. User feedback will be used as additional input for the generator's next round of training to make the generated schemes more tailored to individual needs.

[0152] Step S84: Design Improvement. Through multiple rounds of generation and feedback, the generator gradually optimizes its output, improving the generated design in terms of functionality, appearance, and user preferences. Each generated design must not only meet basic design requirements but also continuously adapt to personalized user feedback. For example, users may request that the design have higher environmental friendliness in terms of materials or better match specific style preferences in terms of appearance. The AI ​​model achieves these improvements by iteratively adjusting the generator's weights over multiple rounds.

[0153] Step S85: Confirmation of the final design scheme. When the generator and discriminator reach a balance, that is, when the discriminator can no longer easily distinguish between the generated scheme and the actual design scheme, the system will determine the final candidate design scheme and provide it to the user. The final scheme will be further confirmed and fine-tuned by the user to form a complete preliminary design scheme.

[0154] In one implementation, a GAN model is used for architectural design generation. For example, in architectural design, the user provides input via an input module regarding building type (e.g., floor plan, area, material selection), functional requirements (e.g., load-bearing capacity), and aesthetic style (e.g., modern, traditional). The generator produces several candidate design schemes, outputting the geometric parameters of the floor plan, material selection, area, etc. A discriminator evaluates these schemes, determining, for example, whether the design meets building safety standards and satisfies the user's style preferences. User feedback can be provided through a graphical interface; the user may request adjustments to the floor plan or changes in material types, and the AI ​​model will meet these needs through multiple rounds of generation. Finally, after several iterations, the AI ​​model provides a personalized architectural design scheme that meets the user's design requirements—a preliminary design scheme—including a visual representation (e.g., a 3D model) and detailed design parameters.

[0155] It should be noted that the preliminary design scheme generated by the GAN model in this application includes at least one design scheme, so as to facilitate further optimization of the preliminary design scheme to obtain an optimized design scheme.

[0156] Generative Adversarial Networks (GANs) are deep learning models that learn data distributions by having two neural networks compete against each other. A GAN consists of two main parts: a generator and a discriminator. These two parts compete against each other in a zero-sum game framework, jointly driving the training and optimization of the GAN. The generator's role is to create synthetic data samples, while the discriminator's task is to distinguish between real and fake data samples.

[0157] Step S4: The AI ​​model optimizes the preliminary design scheme based on user feedback data to obtain an optimized design scheme; the user feedback data refers to user feedback on the preliminary design scheme.

[0158] Specifically, this application uses an AI model to optimize the generated preliminary design scheme based on user feedback data in order to select an optimized design scheme.

[0159] In one embodiment of this application, the AI ​​model optimizes the preliminary design scheme based on user feedback data to obtain an optimized design scheme, including steps S41 to S42.

[0160] Step S41: Receive user feedback data on the preliminary design scheme. The user feedback data includes, but is not limited to, suggestions for cost, dimensions, materials, aesthetics, and / or modifications to the scheme.

[0161] Step S42: The AI ​​model optimizes the preliminary design scheme based on the genetic algorithm and the user feedback data to obtain an optimized design scheme.

[0162] In one implementation, the AI ​​model generates a preliminary design scheme based on user feedback data and matched basic data. This preliminary scheme includes multiple design options, all of which meet the user's design requirements. Therefore, it is necessary to select the optimal scheme from these options. The user feedback data refers to the set optimization goals, including but not limited to cost, size, materials, aesthetics, and / or suggestions for scheme modifications.

[0163] In this implementation, the application utilizes a genetic algorithm and sets optimization objectives to optimize the preliminary design scheme to obtain an optimized design scheme. The optimization objectives include, but are not limited to, cost, size ratio, material selection, aesthetics, and / or proposed modifications.

[0164] In one implementation, the optimization objective set in this application includes an optimization objective, which includes, but is not limited to, cost, size ratio, material selection, and aesthetics. For example, if a user's design requirement is a 20-story building with a three-bedroom, one-living room apartment of 90 square meters, and two apartments per floor, the preliminary design schemes that meet these requirements include 20 design schemes. The design scheme with the lowest cost among these 20 schemes is then selected as the optimized design scheme. If the optimized design scheme also includes multiple schemes, these schemes can be presented to the user for selection, and further optimization objectives can be set, such as size ratio, to select the optimal size ratio as the optimized design scheme.

[0165] In another implementation, the optimization objectives set in this application include multiple objectives, including but not limited to cost, size ratio, material selection, and aesthetics. For example, if a user's design requirement is a 20-story building with a three-bedroom, one-living room apartment of 90 square meters, and two apartments per floor, the initial design schemes that meet these requirements include 20 schemes. The optimal design scheme is then selected from these 20 schemes based on both cost and aesthetics. If the optimal design scheme also includes multiple schemes, further optimization objectives can be set. On the one hand, these multiple schemes can be presented to the user for selection; on the other hand, further optimization objectives can be set, such as size ratio, selecting the optimal size ratio as the optimal design scheme.

[0166] It should be noted that genetic algorithms are based on natural selection and genetic mechanisms, and search for optimal or near-optimal solutions by simulating phenomena such as selection, crossover (hybridization) and mutation in the process of biological evolution.

[0167] like Figure 4 As shown, in one embodiment of this application, the AI ​​model-based intelligent building design method further includes the following steps S5 to S6.

[0168] Step S5: Receive the user design requirements and feed them back to the AI ​​model so that the AI ​​model can match the user design requirements with the standard database.

[0169] Specifically, this application receives the user design requirements and feeds them back to the AI ​​model, so that the AI ​​model can match the user design requirements with the standard database to obtain basic data that meets the user design requirements.

[0170] In one implementation, this application further provides feedback on the optimized design scheme to the user and receives user feedback, so that the AI ​​model can optimize the optimized design scheme based on the user feedback.

[0171] Specifically, this application uses an AI model to feed back the generated optimized design scheme to the user, and at the same time receives user feedback, so that the AI ​​model can optimize the optimized design scheme based on the user feedback.

[0172] In this implementation, the AI ​​model will feed back the generated optimized design scheme to the user. If the optimized design scheme includes multiple design schemes, the user can choose the most satisfactory design scheme from these multiple design schemes according to their needs or personal preferences.

[0173] In one implementation, users adjust or confirm the design scheme through the user interaction module of the AI ​​model.

[0174] In one implementation, the AI ​​model of this application provides a graphical user interface and a natural language interface, enabling users to intuitively operate and modify the design scheme.

[0175] In this implementation, the application provides a user-friendly graphical interface, allowing users to interact with the AI ​​model through dragging, clicking, and other methods.

[0176] In one implementation, this application provides a natural language interface, such as a dialog box, through which users can input text or voice to input their design requirements into the AI ​​model, so that the AI ​​model can provide corresponding optimized design solutions based on the user's design requirements. This implements natural language processing technology, enabling users to communicate with the AI ​​model through voice or text.

[0177] Step S6: Store the user feedback in the user database of the standard database to facilitate the AI ​​model's learning and optimization design.

[0178] In one implementation, this application stores the optimized design schemes generated by the AI ​​model and user feedback in a user database, and uses the user database as a resource for acquiring training sets. This allows the AI ​​model to learn and optimize designs, thus becoming familiar with different user preferences. When these users raise design requirements again, the trained model can recommend matching design styles and element combinations to the users, thereby accelerating the generation of design schemes and improving design efficiency.

[0179] Figure 5 The diagram shown is a flowchart illustrating an AI model-based intelligent building design method according to an embodiment of this application. Figure 5 As shown, this application obtains a standard component library through standard component models such as type, material, resource consumption, and WBS coding. Based on the standard component library and basic unit information such as usable floor area, living room area, and unit type, a standard unit library is obtained. Simultaneously, a standard unit library is obtained based on the standard unit library, standard core tube library, and user design requirements (such as building area and shared area ratio). A REVIT standard template is then obtained from the standard unit library. Finally, AI design is performed based on user design requirements and the obtained standard databases, including the standard component library, standard unit library, standard core tube library, standard unit library, and REVIT standard template, to obtain an optimized design solution that meets user design needs. This effectively improves design efficiency and quality, reduces manpower costs in the design process, and provides a better user experience.

[0180] The scope of protection of the AI ​​model-based intelligent building design method described in this application is not limited to the execution order of the steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principles of this application is included within the scope of protection of this application.

[0181] This application also provides an AI model-based intelligent building design system, which can implement the AI ​​model-based intelligent building design method described in this application. However, the implementation device of the AI ​​model-based intelligent building design method described in this application includes, but is not limited to, the structure of the AI ​​model-based intelligent building design system listed in this embodiment. All structural modifications and substitutions of the prior art made in accordance with the principles of this application are included within the protection scope of this application.

[0182] like Figure 6 As shown, this embodiment provides an intelligent building design system based on an AI model. The system 3 includes a standard database 31, an AI model 32, and a user interaction unit 33.

[0183] The standard database 31 includes a standard library that meets building design specifications; the standard library includes basic data such as a standard component library, a standard unit type library, a standard core tube library, and a standard unit library.

[0184] The AI ​​model 32 includes a data matching unit 321, a design generation unit 322, and a design optimization unit 323.

[0185] The data matching unit 321 is configured to perform data matching based on user design requirements and the standard database to obtain basic data that meets the user design requirements; the standard database includes a standard component library, a standard house type library, a standard core tube library and / or a standard unit library; the basic data includes successfully matched standard component models, standard house type models, standard core tube models and / or standard unit models.

[0186] The design generation unit 322 is configured to design a scheme based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme.

[0187] The design optimization unit 323 is configured to optimize the preliminary design scheme based on user feedback data to obtain an optimized design scheme; the user feedback data is the user's feedback data on the preliminary design scheme.

[0188] The user interaction unit 33 is configured to receive the user design requirements and feed them back to the data matching unit, so that the data matching unit can perform data matching with the standard database according to the user design requirements.

[0189] The user interaction unit 33 is also configured to store the user feedback in the user database of the standard database, so as to facilitate the AI ​​model to learn and optimize its design.

[0190] It should be noted that the functions or operations of the standard database 31, AI model 32, data matching unit 321, design generation unit 322, design optimization unit 323, and user interaction unit 33 described in this embodiment correspond one-to-one with the steps in the AI ​​model-based intelligent building design method described above, and therefore will not be repeated here.

[0191] This application also provides an electronic device, comprising: a memory and a processor; wherein the memory is used to store a computer program; the memory includes various media capable of storing program code, such as ROM, RAM, magnetic disk, USB flash drive, memory card, or optical disk. The processor is used to execute the computer program stored in the memory, so that the electronic device performs the dynamic surround view stitching method based on tire pressure monitoring as described above.

[0192] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0193] like Figure 7 As shown, the electronic device 4 of this application is embodied in the form of a general-purpose computing device. The components of the control terminal may include, but are not limited to: one or more processors or processing units 41, a memory 42, and a bus 43 connecting different system components (including the memory 42 and the processing unit 41).

[0194] Bus 43 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0195] Control terminals typically include various computer system-readable media. These media can be any available media that can be accessed by the control terminal, including volatile and non-volatile media, and removable and non-removable media.

[0196] Memory 42 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 421 and / or cache memory 422. The control terminal may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, memory 42 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 7 Not shown; usually referred to as a "hard drive"). Although Figure 7 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 43 via one or more data media interfaces. Memory 42 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this disclosure.

[0197] A program / utility 424 having a set (at least one) of program modules 4241 may be stored, for example, in memory 42. Such program modules 4241 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 4241 typically perform the functions and / or methods described in the embodiments of this disclosure.

[0198] This application also provides a computer-readable storage medium. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state drive (SSD)).

[0199] This application embodiment may also provide a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application embodiment are generated. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0200] When the computer program product is executed by a computer, the computer performs the method described in the foregoing method embodiments. The computer program product can be a software installation package; when the foregoing method is required, the computer program product can be downloaded and executed on the computer.

[0201] In summary, the AI-based intelligent building design method, system, electronic device, and medium provided in the above embodiments of this application have the following beneficial effects:

[0202] This application designs a standard database for architectural design. The standard database includes a standard library that meets architectural design codes. An AI model is used to match user design requirements with the standard database to obtain basic data that meets those requirements. The standard database includes a standard component library, a standard unit type library, a standard core tube library, and / or a standard unit library. The basic data includes successfully matched standard component models, standard unit type models, standard core tube models, and / or standard unit models. The AI ​​model designs schemes based on preset design rules and the basic data to obtain corresponding 3D design drawings as preliminary design schemes. The AI ​​model optimizes these preliminary design schemes based on user feedback data to obtain optimized design schemes. The user feedback data refers to user feedback on the preliminary design schemes, which can significantly improve design efficiency and design quality, reduce design costs, and ensure the quality of the design schemes through the intelligent optimization function of the AI ​​model.

[0203] This application integrates a standard database of architectural design with AI models to generate optimized design solutions that meet user design needs, thereby promoting design innovation and enhancing the artistry and practicality of design works.

[0204] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0205] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A smart building design method based on an AI model, characterized in that, The method includes: Obtain a standard database for architectural design; the standard database includes a standard library that meets architectural design specifications. The AI ​​model is used to match user design requirements with the standard database to obtain basic data that meets the user design requirements; the standard database includes a standard component library, a standard house type library, a standard core tube library and / or a standard unit library; the basic data includes successfully matched standard component models, standard house type models, standard core tube models and / or standard unit models. The AI ​​model designs a scheme based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme. The AI ​​model optimizes the preliminary design based on user feedback data to obtain an optimized design; the user feedback data refers to user feedback on the preliminary design.

2. The intelligent building design method based on an AI model according to claim 1, characterized in that, Using an AI model, data is matched between the user's design requirements and the standard database to obtain basic data that meets the user's design requirements, including: The user design requirements are semantically parsed based on a large language model to obtain the parsed user design requirement features. Data matching is performed based on the parsed user design requirements characteristics and the standard database to obtain basic data that meets the user design requirements.

3. The intelligent building design method based on an AI model according to claim 1, characterized in that, The AI ​​model designs a scheme based on preset design rules and the basic data to obtain a corresponding 3D design drawing, which serves as a preliminary design scheme, including: The AI ​​model designs a scheme based on the basic data that meets the user's design requirements, according to the trained neural network model and preset design rules, in order to obtain the corresponding label data. The AI ​​model translates and transforms the tag data to obtain corresponding 3D design drawings, which serve as preliminary design schemes that meet the user's design requirements.

4. The intelligent building design method based on an AI model according to claim 3, characterized in that, The training methods for the neural network model include: Collect basic architectural design data; the basic architectural design data includes architectural design language dictionary, architectural design code, architectural design standard, architectural design parameters and / or architectural design cases; The collected basic architectural design data is preprocessed to obtain preprocessed data. The neural network model is trained based on the preprocessed data to obtain the trained neural network model. The trained neural network model is then tested and evaluated.

5. The intelligent building design method based on an AI model according to claim 1, characterized in that, The AI ​​model optimizes the preliminary design based on user feedback data to obtain an optimized design, including: Receive the user feedback data; the user feedback data includes, but is not limited to, suggestions for cost, size, materials, aesthetics, and / or modification of the solution; The AI ​​model optimizes the preliminary design based on genetic algorithms and user feedback data to obtain an optimized design.

6. The intelligent building design method based on an AI model according to claim 1, characterized in that, The method further includes: The system receives the user design requirements and feeds them back to the AI ​​model, so that the AI ​​model can match the user design requirements with the standard database. The user feedback is stored in the user database of the standard database to facilitate the learning and optimization design of the AI ​​model.

7. An intelligent building design system based on an AI model, characterized in that, The system includes: A standard database, comprising a standard library that meets building design codes; AI model, the AI ​​model includes: The data matching unit is configured to perform data matching based on user design requirements and the standard database to obtain basic data that meets the user design requirements; the standard database includes a standard component library, a standard house type library, a standard core tube library, and / or a standard unit library; the basic data includes successfully matched standard component models, standard house type models, standard core tube models, and / or standard unit models. The design generation unit is configured to design a scheme based on preset design rules and the basic data to obtain a corresponding three-dimensional design drawing as a preliminary design scheme. The design optimization unit is configured to optimize the preliminary design scheme based on user feedback data to obtain an optimized design scheme; the user feedback data refers to user feedback on the preliminary design scheme.

8. The intelligent building design system based on an AI model according to claim 7, characterized in that, The system also includes: The user interaction module is configured to receive user design requirements and feed them back to the data matching unit, so that the data matching unit can perform data matching with the standard database based on the user design requirements. The user interaction unit is also configured to store the user feedback in a user database within the standard database, so that the AI ​​model can learn and optimize its design.

9. An electronic device, characterized in that, The electronic device includes: A memory that stores a computer program; The processor, which is communicatively connected to the memory, executes the AI-based intelligent building design method according to any one of claims 1 to 6 when calling the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by an electronic device, the program implements the AI-based intelligent building design method as described in any one of claims 1 to 6.

Images