Furniture design scheme customization method and system based on big data

Abstract

The invention relates to the technical field of intelligent furniture design and big data, and discloses a furniture design scheme customization method and system based on big data, and the method comprises the steps: collecting physiological data, environment data and multi-source behavior data of a user, carrying out the time-space alignment of the physiological data, environment data and behavior data through a heterogeneous data alignment engine, and carrying out the customization of a furniture design scheme. Generating a three-dimensional feature tensor; constructing a dynamic demand model fusing spatial-temporal characteristics and cognitive deviation compensation; generating candidate design schemes based on the generative adversarial network and the historical knowledge graph; carrying out material matching on the candidate schemes, and carrying out sustainability verification; and feeding back a verification result to the dynamic demand model for iterative optimization. According to the invention, furniture personalized customization is realized, and the user satisfaction is improved. The intelligent generation scheme ensures the design quality and effect, and enhances the market competitiveness.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent furniture design and big data, and particularly relates to a method and system for customizing furniture design solutions based on big data. Background Art

[0002] In recent years, with the deep integration of Internet of Things, artificial intelligence and big data technologies, the furniture design field has gradually evolved towards intelligence and personalization. The existing technologies mainly focus on the following aspects: First, a personalized recommendation system based on user portraits generates a preference prediction model by analyzing users' historical behavior data (such as e-commerce browsing records and smart home interaction logs), and combines CAD technology to achieve rapid rendering of furniture solutions; Second, ergonomic dynamic adaptation technology uses flexible sensors to collect human body pressure distribution and posture data, and combines biomechanical models to optimize furniture structure parameters; Third, environment perception and space layout optimization technology constructs a digital twin space with the help of BIM models and IoT sensors, and generates furniture layout solutions through reinforcement learning algorithms. In addition, under the promotion of the sustainable design concept, the Material Life Cycle Assessment (LCA) method has gradually been applied to the furniture industry to guide the selection of environmentally friendly materials through carbon emission calculation models. However, the existing technologies mostly focus on single-dimensional optimization, lacking cross-modal data fusion and dynamic coordination mechanisms, and are difficult to meet the customization requirements in complex scenarios.

[0003] The current technical system still has significant limitations: First, the ability to fuse heterogeneous data is insufficient. Physiological data (such as body pressure distribution), environmental data (such as light intensity) and behavior data (such as operation habits) are difficult to be effectively aligned due to differences in time and space benchmarks, resulting in deviations in user demand modeling; Second, the problem of static demand perception is prominent. Traditional models rely on historical data training and cannot correct users' cognitive biases in real time (such as the misalignment between subjective preferences and objective needs), resulting in a low matching degree between the generated solutions and real needs; Third, the process of generating solutions lacks industry knowledge constraints. Existing Generative Adversarial Networks (GANs) are prone to generating designs that do not conform to engineering specifications (such as incompatible material structures), and are not linked with the knowledge graph of historical solutions, resulting in an imbalance between design innovation and practicality; Finally, the dimension of sustainability verification is single. Existing LCA methods mostly focus on the environmental protection attributes of materials, ignoring the maintenance costs during the use stage and the hardware scalability, and are difficult to support full-life cycle optimization. The above defects seriously restrict the implementation and industrial promotion of intelligent furniture design systems. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for customizing furniture design solutions based on big data, which improves the user matching degree, engineering feasibility and environmental friendliness of the design solutions by constructing a spatio-temporal feature fusion model, a cognitive bias compensation mechanism and a closed-loop iterative optimization system.

[0006] To solve the above technical problems, the present invention provides the following technical solutions. A method for customizing furniture design solutions based on big data includes: collecting user physiological data, environmental data and multi-source behavior data, performing spatio-temporal alignment on the physiological data, environmental data and multi-source behavior data by using a heterogeneous data alignment engine to generate a three-dimensional feature tensor; constructing a dynamic demand model that fuses spatio-temporal features and cognitive bias compensation; generating candidate design solutions based on a generative adversarial network and a historical knowledge graph; performing material matching on the candidate design solutions and conducting sustainability verification; and feeding back the verification results to the dynamic demand model for iterative optimization.

[0007] As a preferred embodiment of the method for customizing furniture design solutions based on big data according to the present invention, the user physiological data includes spinal curvature distribution data and body pressure distribution data obtained through intelligent wearable devices.

[0008] The environmental data includes spatial temperature and humidity, light intensity and sound field characteristics collected through IoT sensors.

[0009] The multi-source behavior data includes e-commerce platform browsing logs, smart home operation records and CAD floor plan files.

[0010] As a preferred embodiment of the method for customizing furniture design solutions based on big data according to the present invention, the dynamic demand model includes a dual-channel feature extraction network. The first channel uses a gated recurrent unit to process time-series behavior data and extract user operation habit cycle features. The second channel uses a three-dimensional convolutional network to analyze spatial environment data and extract functional area division features. The structure of cognitive bias compensation includes a demand difference detection layer and a weight adjustment layer, which are respectively used to compare the feature distance between the user's subjective selection solution and the system-recommended solution and dynamically adjust the dual-channel feature fusion weight coefficient according to the degree of difference.

[0011] As a preferred embodiment of the method for customizing furniture design solutions based on big data according to the present invention, the cognitive bias compensation includes establishing a user cognitive feature space, mapping historical design solutions into a feature vector set, calculating the offset between the current user selection solution and the centroid of the feature space to generate a compensation vector, and correcting the output of the initial dynamic demand model through a vector field interpolation algorithm.

[0012] As a preferred solution of a furniture design plan customization method based on big data according to the present invention, wherein: the generation of candidate design plans includes constructing an industry knowledge graph, and the graph nodes include material type nodes, structural style nodes, and style element nodes; the edge relationships include material-structure compatibility edges, style-color association edges, and plan-user evaluation edges;

[0013] The graph attention generator performs multi-hop reasoning in the knowledge graph according to the demand vector to generate a style constraint rule set, and performs a three-hop neighborhood search. The first hop retrieves the nodes directly associated with the demand vector; the second hop extends to the nodes related to function compatibility; the third hop associates with the aesthetic evaluation nodes; encodes the constraint rules as the prior conditions of the generative adversarial network to control the direction of plan generation.

[0014] As a preferred solution of a furniture design plan customization method based on big data according to the present invention, wherein: the sustainability verification includes establishing a dynamic database of self-repairing material parameters, generating a hierarchical optimization strategy. The first layer uses the filtering method to eliminate the plans that do not meet the hard constraints; the second layer applies the improved NSGA-Ⅲ algorithm to search for the optimal solution on the Pareto front;

[0015] Reserve physical interfaces when generating the material configuration plan to allow the replacement of functional components in the later stage; add hidden expansion slots to adapt to future intelligent hardware upgrades;

[0016] Calculate the whole life cycle of the candidate plan, predict the durability of the surface coating by combining historical temperature and humidity data; simulate the wear trend of the joint structure according to the user behavior data; calculate the time cost index of disassembly and replacement of each component.

[0017] As a preferred solution of a furniture design plan customization method based on big data according to the present invention, wherein: the iterative optimization includes extracting the material carbon emission intensity, structural safety redundancy, and user preference matching index from the sustainability verification results to form a structured feedback data set, including the current iteration cycle and the data of the previous three iterations;

[0018] Monitor the fluctuation of the user preference matching index, enhance the intensity of deviation compensation intervention, and trigger the emergency update mode of the knowledge graph; when the change of the matching index decreases continuously for three times, freeze the compensation intensity parameter, and enable the dual-channel feature weight rebalancing and the incremental fine-tuning function of the dynamic demand model;

[0019] Dynamically adjust the fusion weight of the spatial function partition feature and the temporal behavior feature according to the change direction of the structural safety redundancy index;

[0020] Set a dual - condition iteration termination rule. When the standard is not met, trigger cross - module collaborative optimization, adjust the generator style constraint conditions, and enable the alternative material combination scheme library; when the standard is met, generate the final scheme, save the current model parameter snapshot to the version library, and automatically create a knowledge graph backtracking path.

[0021] As a preferred scheme of a furniture design scheme customization system based on big data according to the present invention, it includes a data acquisition module, a data processing module, a candidate design module, a verification and optimization module, and an iterative optimization module;

[0022] The data acquisition module collects real - time physiological data, environmental data, and multi - source behavior data of users;

[0023] The data processing module uses a heterogeneous data alignment engine to perform spatio - temporal alignment on the collected data, and is also responsible for the deviation compensation function to effectively compare and adjust the weights between user preferences and system - recommended schemes;

[0024] The candidate design module generates candidate design schemes based on a generative adversarial network and a historical knowledge graph;

[0025] The verification and optimization module conducts sustainability verification on the candidate design schemes and proposes optimization strategies according to the verification results;

[0026] The iterative optimization module converts the sustainability verification results into a feedback data set, monitors the change of the user preference fit index, and performs iterative optimization of the model according to the feedback.

[0027] A computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of a furniture design scheme customization method based on big data.

[0028] A computer - readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of a furniture design scheme customization method based on big data.

[0029] The beneficial effects of the present invention: Realize cross - modal association of physiological, environmental, and behavior data through a spatio - temporal encoder to generate a high - precision three - dimensional feature tensor, solving the data island problem of traditional methods; adopt a dual - channel network and a deviation compensation controller to correct users' cognitive deviations in real time, with the demand matching degree increased by more than 30%; combine the industry knowledge graph to constrain the generative adversarial network, making the proportion of generated schemes meeting engineering specifications increase to more than 92%; integrate material performance prediction, usage intensity simulation, and maintainability assessment to reduce the carbon emissions in the whole life cycle of furniture; drive the co - evolution of model parameters and the knowledge graph through multi - index feedback, improving the efficiency of scheme iteration. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0031] Figure 1 Schematic flowchart of a method for customizing furniture design solutions based on big data provided by an embodiment of the present invention.

[0032] Figure 2 Schematic diagram of the working modules of a system for customizing furniture design solutions based on big data provided by an embodiment of the present invention. Detailed implementation manners

[0033] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific implementation manners of the present invention in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0034] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0035] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0036] The present invention is described in detail in conjunction with schematic diagrams. When detailing the embodiments of the present invention, for the convenience of explanation, the cross-sectional views showing the device structure will be enlarged locally out of the general scale, and the schematic diagrams are only examples and should not limit the protection scope of the present invention here. In addition, in actual production, three-dimensional spatial dimensions including length, width, and depth should be included.

[0037] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper, lower, inner, and outer" is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first, second, or third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0038] Unless otherwise clearly defined and limited in the present invention, the terms "installation, connection, and coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can also be a mechanical connection, an electrical connection, or a direct connection, and can also be indirectly connected through an intermediate medium, or can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0039] Example 1, referring to Figure 1 , which is the first embodiment of the present invention. This embodiment provides a method for customizing furniture design solutions based on big data, including:

[0040] S1: Collect user physiological data, environmental data, and multi-source behavior data, and use a heterogeneous data alignment engine to perform spatio-temporal alignment on the physiological data, environmental data, and behavior data to generate a three-dimensional feature tensor.

[0041] Furthermore, the user physiological data includes the spinal curvature distribution data and body pressure distribution data obtained through smart wearable devices, and the collected sitting body pressure distribution matrix, the sensor node density of which is not less than 100 per dm 2 ; the environmental data is the spatial temperature and humidity, light intensity, and sound field characteristics collected by IoT sensors;

[0042] The multi-source behavior data includes e-commerce platform browsing logs, smart home operation records, and CAD floor plan files.

[0043] It should be noted that a dynamic time warping algorithm is applied to the behavior data stream to generate a time series feature vector with a unified time reference. The environmental parameters and physiological data are projected onto a three-dimensional space grid, and a spatial feature tensor is generated through bilinear interpolation; a cross-attention matrix A of behavior-environment-physiological data is established ij , and the calculation formula is:

[0044]

[0045] Among them, Q i is the behavior feature vector, K j is the environmental feature vector, d is the dimension scaling factor, T is the matrix transpose, Pi , P j are the occurrence positions of the i-th and j-th behavioral events respectively, and σ is the spatial decay coefficient.

[0046] S2: Construct a dynamic demand model that integrates spatio-temporal features and cognitive bias compensation.

[0047] Furthermore, construct a two-channel feature extraction network. The gated recurrent unit of the first channel adopts a double-layer GRU structure with a hidden layer dimension of 128. The input sequence is the composite behavior sequence after time alignment, and the output is the feature vector of the user operation cycle. The three-dimensional convolutional network of the second channel contains 4 convolutional blocks, each block contains a 3×3×3 convolutional kernel with a stride of 2, and downsamples the spatial feature grid to output the feature vector of the spatial functional partition.

[0048] The structure of cognitive bias compensation includes a demand difference detection layer and a weight adjustment layer, which are used to compare the feature distances between the user's subjective selection scheme and the recommended scheme and dynamically adjust the two-channel feature fusion weight coefficient according to the degree of difference. Real-time compare the feature vectors of the user's selected scheme and the system's recommended scheme, and calculate the difference degree indexes in three dimensions of color distribution, functional partition, and material combination. When the color difference degree > 0.3, increase the weight of the temporal behavior channel (increment Δw = 0.1×difference degree); when the functional partition difference degree > 0.25, enhance the weight of the spatial environment channel (increment Δw = 0.15×difference degree), and set the upper limit of the weight adjustment amplitude to ±40% to prevent excessive deviation from the dynamic demand model.

[0049] It should be noted that cognitive bias compensation maps historical design schemes to a set of feature vectors by establishing a user cognitive feature space, calculates the offset of the current user's selected scheme from the centroid of the feature space to generate a compensation vector, generates 128-dimensional feature vectors from the 3D model files of historical design schemes and user evaluation data, including geometric contour analysis, material texture classification, and color space distribution, performs dimensionality reduction on the set of feature vectors, projects them onto a two-dimensional plane to form a feature distribution map; uses the density clustering algorithm to identify the high-density regions in the feature distribution map, calculates the weighted average of each cluster center, and the weight is determined by the user score and usage duration of the corresponding scheme, and connects the main cluster centers to form the topological structure of the feature space. Correct the output of the initial dynamic demand model through the vector field interpolation algorithm, and its correction formula is:

[0050]

[0051] where, V corr is the corrected demand vector, V init is the initial demand vector, α is the compensation intensity factor, w k is the similarity weight, C kis the clustering center of the feature space, n is the total number of vectors, and k is the index.

[0052] S3: Generate candidate design solutions based on the generative adversarial network and the historical knowledge graph.

[0053] Furthermore, construct an industry knowledge graph. The graph nodes include material type nodes, structural style nodes, and style element nodes; the edge relationships include material-structure compatibility edges, style-color association edges, and scheme-user evaluation edges.

[0054] It should be noted that the node creation rules include that the material nodes contain 12 physical properties such as density, elastic modulus, and thermal conductivity; the structural nodes record the connector type, load-bearing level, and assembly complexity parameters; the style nodes store the era feature labels, color matching rules, and decorative element correlation degrees.

[0055] The edge relationship is defined as follows: the weight of the material-structure edge is determined by the frequency of the combination of the two in historical solutions; the style-color edge uses the HSV color space similarity to calculate the association strength.

[0056] Through the graph attention generator, perform multi-hop reasoning in the knowledge graph according to the demand vector, generate a style constraint rule set, and perform a three-hop neighborhood search. The first hop retrieves the nodes directly associated with the demand vector; the second hop extends to the nodes related to functional compatibility; the third hop associates with the aesthetic evaluation nodes; encode the constraint rules as the prior conditions of the generative adversarial network to control the direction of scheme generation.

[0057] Furthermore, based on the generative adversarial network, the overall layout scale uses the VGG19 network to extract spatial division features and outputs a 4096-dimensional feature vector; the local detail scale identifies the connection structure features through a segmentation network; the material texture scale uses the GLCM algorithm to calculate the surface texture contrast and homogeneity index.

[0058] S4: Perform material matching on the candidate solutions and conduct sustainability verification.

[0059] Furthermore, establish a dynamic database of self-healing material parameters and generate a hierarchical optimization strategy. The first layer uses the filtering method to eliminate the solutions that do not meet the hard constraints; the second layer applies the improved NSGA-Ⅲ algorithm to search for the optimal solution on the Pareto front.

[0060] It should be noted that the hard constraint filtering conditions of the first layer include structural safety: load-bearing capacity ≥ 1.5 times the design load; budget constraint: total material cost ≤ 95% of the user budget; size adaptability: product external dimension < available space dimension - 50mm (for each dimension).

[0061] The specific implementation steps of the improved NSGA-Ⅲ algorithm in the second layer are as follows:

[0062] (1) Reference point generation: Uniformly arrange m reference points in the three-dimensional target space (cost, performance, environmental protection);

[0063] (2) Population initialization: Randomly select individuals from the solutions passing through the filtering layer;

[0064] (3) Adaptive crossover and mutation: The crossover probability is The mutation probability is where t is the number or generation of the current iteration, and T max is the maximum number of iterations;

[0065] (4) Elite retention strategy: Retain the top 30% of individuals in terms of the HV index in the non-dominated solution set in each generation.

[0066] Reserve physical interfaces when generating the material configuration plan to allow for later replacement of functional components; Add hidden expansion slots to adapt to future intelligent hardware upgrades;

[0067] Calculate the full life cycle of the candidate solutions, predict the durability of the surface coating in combination with historical temperature and humidity data; Simulate the wear trend of the joint structure based on the user behavior data; Calculate the time cost index for disassembly and replacement of each component.

[0068] It should be noted that for the prediction of the surface coating durability, after collecting the historical temperature and humidity data and performing smoothing processing through a sliding window, it is input into a bidirectional LSTM network to predict the performance decay over a five-year period. The network outputs the annual decrease in glossiness, the crack propagation rate, and the color difference value, and generates coating maintenance cycle suggestions in combination with material properties. Focus on monitoring the waterproof performance decay in a continuously high-humidity environment and trigger a warning for material replacement;

[0069] Simulate the wear trend of the joint structure. Based on the standard wear coefficient table in the mechanical engineering handbook, calculate the life of key components in combination with user operation frequency data (such as the number of daily adjustments, load-bearing). Use the simplified version of the ISO 3408 ball screw wear rate formula: wear amount = basic wear rate × number of operations × load factor, and preset the basic wear rate of the plastic / metal joint surface with reference to the industry average;

[0070] Calculate the time cost index for disassembly and replacement of each component. Decompose the disassembly process into basic action units such as arm extension, grasping, and moving, and assign standard time values to each action (such as grasping small parts = 3.5 TMU). Select the standard disassembly template according to the component connection type (bolt / clasp / welding), and increase the processing time by 30% for rusty components. Use an anti-static operation process for precision components.

[0071] S5: Feed the verification result back to the dynamic demand model for iterative optimization.

[0072] Furthermore, extract the material carbon emission intensity, structural safety redundancy, and user preference fit index from the sustainability verification results to form a structured feedback dataset, including the current iteration cycle and the data of the previous three iterations. The material carbon emission intensity is the carbon dioxide equivalent emission of the recorded unit material during its life cycle; the structural safety redundancy is the ratio of the measured load-bearing capacity to the design requirements (rounded to two decimal places); the user preference fit index is quantified by the cosine value of the spatial angle between the solution feature vector and the user demand vector.

[0073] Monitor the fluctuation of the user preference fit index. When the fluctuation of the user preference fit index exceeds 15%, increase the intervention intensity of the deviation compensation controller, with the maximum adjustment amplitude not exceeding 50% of the initial value, trigger the emergency update mode of the knowledge graph, and preferentially load the data of recently high-scoring solutions; when the change of the fit index is less than 2% for three consecutive iterations, freeze the compensation intensity parameter and only enable the incremental fine-tuning function of the dynamic demand model.

[0074] According to the change direction of the structural safety redundancy index, when the safety factor decreases, increase the fusion weight of the spatial function partition feature (the increase amplitude is positively correlated with the decrease rate); when the safety factor increases, enhance the fusion weight of the temporal behavior feature (the increase amplitude is negatively correlated with the increase amplitude).

[0075] The knowledge graph evolves progressively. For solutions with a carbon emission intensity lower than 20% of the industry average, automatically create a strong association edge with the environmental protection material library; for solutions in the top 10% of the user preference fit index, generate derivative nodes and associate them with high-frequency design elements; when the user preference fit index is higher than 0.85, shrink the graph attention search radius to the two-hop neighborhood; when a new material type is detected to be stored in the library, temporarily expand the search radius to the four-hop neighborhood for exploration.

[0076] Set a dual-condition iteration termination rule. The primary condition is that the fluctuation amplitude of the user preference fit index for three consecutive iterations is less than 2%; the secondary condition is that the material carbon emission intensity reaches the preset environmental protection standard value.

[0077] When the standard is not met, trigger cross-module collaborative optimization, adjust the style constraint conditions of the generator, and enable the alternative material combination solution library; after reaching the standard, generate the final solution, save the snapshot of the current model parameters to the version library, and automatically create a knowledge graph backtracking path.

[0078] Example 2, the second example of the present invention, which is different from the previous example in that:

[0079] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs, Read-Only Memories), random access memories (RAMs, Random Access Memories), magnetic disks, or optical discs.

[0080] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus, or device.

[0081] More specific examples (a non-exhaustive list) of computer-readable media include the following: electrical connection parts with one or more wirings (electronic devices), portable computer disk cartridges (magnetic devices), random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), optical fiber devices, and portable compact disc read-only memories (CDROMs). Additionally, a computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, then editing, interpreting, or processing it in other suitable ways as necessary, and then storing it in a computer memory.

[0082] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits with logic gate circuits for implementing logical functions on data signals, application specific integrated circuits with appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0083] Embodiment 3, referring to Figure 2 , which is an embodiment of the present invention, provides a furniture design solution customization system based on big data, including a data acquisition module, a data processing module, a candidate design module, a verification and optimization module, and an iterative optimization module;

[0084] The data acquisition module collects the physiological data, environmental data, and multi-source behavior data of the user in real time;

[0085] The data processing module uses a heterogeneous data alignment engine to perform spatio-temporal alignment on the collected data, and is also responsible for the deviation compensation function to effectively compare and adjust the weights between the user preferences and the system recommended solutions;

[0086] The candidate design module generates candidate design solutions based on a generative adversarial network and a historical knowledge graph;

[0087] The verification and optimization module conducts sustainability verification on the candidate design solutions and proposes optimization strategies according to the verification results;

[0088] The iterative optimization module converts the sustainability verification results into a feedback data set, monitors the change of the user preference fit index, and performs iterative optimization of the model according to the feedback.

[0089] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A furniture design solution customization method based on big data, characterized in that: Including Collecting user physiological data, environmental data and multi-source behavior data, and using a heterogeneous data alignment engine to perform spatio-temporal alignment on the physiological data, environmental data and multi-source behavior data to generate a three-dimensional feature tensor; Constructing a dynamic demand model that integrates spatio-temporal features and cognitive bias compensation; Generating candidate design solutions based on a generative adversarial network and a historical knowledge graph; Performing material matching on the candidate design solutions and conducting sustainability verification; Feeding back the verification results to the dynamic demand model for iterative optimization.

2. The method for customizing a furniture design solution based on big data according to claim 1, characterized in that: The user physiological data includes spinal curvature distribution data and body pressure distribution data obtained through intelligent wearable devices; The environmental data includes spatial temperature and humidity, light intensity and sound field characteristics collected through IoT sensors; The multi-source behavior data includes e-commerce platform browsing logs, smart home operation records and CAD floor plan files.

3. The method for customizing a furniture design solution based on big data according to claim 2, characterized in that: The dynamic demand model includes a dual-channel feature extraction network. The first channel uses a gated recurrent unit to process time-series behavior data and extract user operation habit cycle features. The second channel uses a three-dimensional convolutional network to analyze environmental data and extract functional area division features. The structure of cognitive bias compensation includes a demand difference detection layer and a weight adjustment layer, which are respectively used to compare the feature distances between the user's subjective selection solution and the recommended solution and dynamically adjust the dual-channel feature fusion weight coefficients according to the degree of difference.

4. The method for customizing a furniture design solution based on big data according to claim 3, characterized in that: The cognitive bias compensation includes establishing a user cognitive feature space, mapping historical design solutions into a set of feature vectors, calculating the offset between the current user selection solution and the centroid of the feature space to generate a compensation vector, and correcting the output of the initial dynamic demand model through a vector field interpolation algorithm.

5. The method for customizing a furniture design solution based on big data according to claim 4, characterized in that: The generation of candidate design solutions includes constructing an industry knowledge graph, where the graph nodes include material type nodes, structural style nodes and style element nodes; The edge relationships include material-structure compatibility edges, style-color association edges and solution-user evaluation edges; Performing multi-hop reasoning in the knowledge graph according to the demand vector through a graph attention generator to generate a style constraint rule set, and performing a three-hop neighborhood search. The first hop retrieves nodes directly associated with the demand vector. The second hop extends to nodes related to functional compatibility. The third hop associates with aesthetic evaluation nodes. Encoding the constraint rules as prior conditions of the generative adversarial network to control the solution generation direction.

6. The method for customizing a furniture design solution based on big data according to claim 5, characterized in that: The sustainability verification includes establishing a dynamic database of self-healing material parameters, generating a hierarchical optimization strategy. The first layer uses a filtering method to eliminate solutions that do not meet the hard constraints. The second layer applies an improved NSGA-Ⅲ algorithm to search for the optimal solution on the Pareto front; Reserving physical interfaces when generating a material configuration solution to allow for later replacement of functional components; adding hidden expansion slots to adapt to future smart hardware upgrades; Calculating the full life cycle of the candidate solution, predicting the durability of the surface coating in combination with historical temperature and humidity data; simulating the wear trend of joint structures based on user behavior data; Calculating the time cost index for disassembly and replacement of each component.

7. The method for customizing a furniture design solution based on big data according to claim 6, characterized in that: The iterative optimization includes extracting the material carbon emission intensity, structural safety redundancy, and user preference fit index from the sustainability verification results to form a structured feedback data set, including the current iteration cycle and the data of the previous three iterations; Monitor the fluctuation of the user preference fit index, enhance the intensity of deviation compensation intervention, and trigger the emergency update mode of the knowledge graph; When the fit index change decreases continuously for three times, freeze the compensation intensity parameter, and enable the dual-channel feature weight rebalancing and incremental fine-tuning function of the dynamic demand model; Dynamically adjust the fusion weight of the spatial function partition feature and the temporal behavior feature according to the change direction of the structural safety redundancy index; Set the dual-condition iterative termination rule. When the standard is not met, trigger cross-module collaborative optimization, adjust the generator style constraint conditions, and enable the alternative material combination scheme library; after meeting the standard, generate the final scheme, save the snapshot of the current dynamic demand model parameters to the version library, and automatically create a knowledge graph backtracking path.

8. A system adopting a method for customizing a furniture design solution based on big data as described in any one of claims 1 to 7, characterized in that: Including a data acquisition module, a data processing module, a candidate design module, a verification and optimization module, and an iterative optimization module; The data acquisition module collects the physiological data, environmental data, and multi-source behavior data of users in real time; The data processing module uses a heterogeneous data alignment engine to perform spatio-temporal alignment on the collected data, and is also responsible for the deviation compensation function to effectively compare and adjust the weights between user preferences and system recommended solutions; The candidate design module generates candidate design solutions based on the generative adversarial network and the historical knowledge graph; The verification and optimization module conducts sustainability verification on the candidate design solutions and proposes optimization strategies according to the verification results; The iterative optimization module converts the sustainability verification results into a feedback data set, monitors the change of the user preference fit index, and performs iterative optimization of the model according to the feedback.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images