A method for making personalized mattresses by combining body shape measurement data

Abstract

The present application provides a method for manufacturing a personalized mattress in combination with body shape measurement data, comprising: obtaining body shape measurement data of different customer body parts according to a pre-established personalized mattress manufacturing database, selecting a material combination scheme matching hardness, resilience and air permeability parameters from a material database according to the curve characteristics of each part, and forming a preliminary manufacturing scheme; for each revised manufacturing scheme, obtaining the overall comfort data of the mattress by simulating the pressure distribution of the human body under different sleeping postures, obtaining a comprehensive performance score of each scheme according to a preset comfort evaluation index, and selecting the scheme with the best comfort as a recommended scheme.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for manufacturing a personalized mattress by combining body shape measurement data. Background Art

[0002] In the process of making personalized mattresses, appropriate materials are selected for stacking and combination according to the body shape measurement data of different parts of the human body to achieve the purpose of adapting to the human body curve to the greatest extent. However, the hardness, resilience and air permeability of the material will change with the change of stacking order and thickness, which brings a technical problem: how to meet personalized needs while taking into account the performance of the material and the overall comfort of the mattress. Specifically, different materials will affect each other during the stacking process, and the performance of the upper material will be restricted by the lower material, and vice versa. For example, placing a material with higher hardness in the lower layer will affect the resilience of the upper soft material; while placing a material with poor air permeability in the upper layer will affect the air permeability of the entire mattress. At the same time, changes in the thickness of the stacking will also bring differences in material performance. Too much thickness may make the mattress too hard, while too little thickness may affect the support and durability. Therefore, the production of personalized mattresses requires careful weighing and optimization in terms of material selection, stacking order and thickness design to achieve the best balance of various performance indicators. Summary of the invention

[0003] The present invention provides a method for manufacturing a personalized mattress by combining body shape measurement data, which mainly includes:

[0004] According to the pre-established personalized mattress production database, we obtain the body shape measurement data of different customer parts, and select the material combination scheme that matches the hardness, resilience and air permeability parameters from the material database according to the curve characteristics of each part to form a preliminary production plan;

[0005] The finite element analysis method is used to simulate the mechanical properties of the material stacking sequence and thickness in the preliminary production plan, obtain the support force distribution and pressure distribution data under different stacking combinations, and judge whether the material performance parameters or stacking thickness settings meet the threshold requirements. If the material performance parameters or stacking thickness settings meet the threshold requirements, the preliminary production plan is added to the candidate plan set;

[0006] Analyze the candidate solution set, extract the material combination characteristics and layer thickness characteristics corresponding to different parts of the human body, establish a material selection and layer production model for personalized mattresses, and optimize the material layering sequence and thickness in the model based on the mutual influence and constraint relationship between material properties to obtain a set of revised production solutions;

[0007] For each revised production plan, the pressure distribution of the human body under different sleeping postures is simulated to obtain the overall comfort data of the mattress. According to the preset comfort evaluation index, the comprehensive performance score of each plan is obtained, and the plan with the best comfort is selected as the recommended plan;

[0008] Analyze the material property parameters in the recommended solution, select the most suitable material combination and stacking parameters based on the hardness, resilience and air permeability parameters of the stacked mattress, and combine the customer's body measurement data and personalized demand data. Use 3D modeling software to generate the structural model diagram and cross-section diagram of the personalized mattress as the basis for production and processing;

[0009] Carry out quality inspection on the finished personalized mattresses after production and processing to obtain the actual material performance parameters and human body pressure distribution data, and compare the detected data with the selected most matching material combination and stacking parameters to determine whether the detected data meets the comfort requirements. If the requirements are met, the mattress will be packaged and sold, otherwise it will be remade.

[0010] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0011] The present invention discloses a method for making a personalized mattress in combination with body shape measurement data. The method realizes the precise matching of mattress materials and structures by establishing a customer body shape database and a material database, combining finite element analysis and simulation. First, a preliminary material scheme is selected according to the customer's body shape data, and then the laminated structure is optimized by mechanical property simulation to form a set of candidate schemes. Then, a material selection and laminated manufacturing model is established, the scheme is optimized and a sleeping posture simulation is performed to select a recommended scheme with the best comfort. Finally, the most matching material combination and parameters are selected according to customer needs, and a three-dimensional structural model is generated to guide production. The finished product is quality tested to ensure that the comfort requirements are met. The present invention realizes the personalized customization of mattresses, improves sleeping comfort, and has important application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 The present invention is a flow chart of a method for making a personalized mattress by combining body shape measurement data. DETAILED DESCRIPTION

[0013] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0014] like Figure 1 In this embodiment, a method for making a personalized mattress by combining body shape measurement data may specifically include:

[0015] Step S101, based on the pre-established personalized mattress manufacturing database, obtain the body shape measurement data of different customer body parts, and select a material combination scheme that matches the hardness, resilience and air permeability parameters from the material database according to the curve characteristics of each part to form a preliminary manufacturing plan.

[0016] The whole body contour data of the customer is obtained, wherein the whole body contour data is collected by a three-dimensional scanner; the curve characteristics of key parts are extracted from the whole body contour data to obtain the pressure distribution parameters of the customer; the actual pressure distribution data of each part when the customer tries to lie down is obtained by using a pressure sensor array; according to the actual pressure distribution data, the material combination scheme established in advance is matched to determine the preliminary production scheme with the best support performance.

[0017] Exemplarily, the customer's body measurement data is obtained according to the pre-established personalized mattress production database, and the customer's whole body contour data is obtained using a three-dimensional scanner. The curve features of key parts such as shoulders, waist, and buttocks are extracted through image processing software to obtain pressure distribution parameters, and multiple sets of material combination solutions with matching hardness, resilience, and air permeability parameters are screened from the material database. The actual pressure distribution data of each part when the customer tries to lie down is obtained using a pressure sensor array, and the pressure distribution data is matched with the material combination solution through the least squares method to obtain a preliminary production plan with the best support performance. In the process of personalized mattress production, a high-precision three-dimensional laser scanner is first used to scan the customer's entire body to obtain point cloud data. The scanning resolution is set to 0.5 mm to ensure that subtle changes in body curves are captured. The point cloud data is imported into professional three-dimensional modeling software, such as Geomagic DesignX, and the customer's body model is reconstructed through the NURBS surface fitting algorithm. The curve features of key parts such as shoulders, waist, and buttocks are extracted from the reconstructed model, and parameters such as curvature and concavity are calculated. Based on these parameters, materials with hardness ranging from 20-40ILD, resilience ranging from 60% to 80%, and air permeability ranging from 100 to 300cm were selected from the preset material database. 3 / min·cm 2 A flexible pressure sensor array with a grid size of 64×128 and a sampling frequency of 100 Hz was used to record the changes in pressure distribution during the customer's 5-minute trial. The relationship between pressure data and material parameters was fitted using the least squares method, and the material combination with the smallest fitting error was selected as the preliminary solution.

[0018] Step S102, using the finite element analysis method, simulates the mechanical properties of the material stacking sequence and thickness in the preliminary manufacturing plan, obtains the support force distribution and pressure distribution data under different stacking combinations, and determines whether the material performance parameters or the stacking thickness settings meet the threshold requirements. If the material performance parameters or the stacking thickness settings meet the threshold requirements, the preliminary manufacturing plan is added to the candidate plan set.

[0019] The material stacking sequence and thickness information in the manufacturing scheme are obtained, and a three-dimensional geometric model is established in a finite element analysis software according to the material stacking sequence and thickness information; adaptive meshing is performed on the three-dimensional geometric model, and the adaptive meshing adopts a hexahedral unit type, and the initial mesh size is set to a preset value; a fixed constraint is applied to the bottom of the mattress of the three-dimensional geometric model, and a distributed pressure load is applied to the top of the mattress of the three-dimensional geometric model according to the human body pressure distribution diagram; the nonlinear equation group is solved by the Newton-Raphson iteration method to obtain support force distribution data and pressure distribution data; the displacement value and stress value of the key node are extracted from the support force distribution data and the pressure distribution data; it is determined whether the displacement value is less than a first preset threshold value and whether the stress value is less than a second preset threshold value; if the displacement value is less than the first preset threshold value and the stress value is less than the second preset threshold value, the manufacturing scheme is added to the candidate scheme set; if the displacement value is greater than or equal to the first preset threshold value or the stress value is greater than or equal to the second preset threshold value, the influence degree of each parameter on the displacement value and the stress value is calculated; and the parameters in the manufacturing scheme are adjusted according to the influence degree.

[0020] For example, according to the material stacking sequence and thickness information in the preliminary production plan, a three-dimensional geometric model is established in the finite element analysis software, and corresponding mechanical parameters are set for different material layers, including the elastic modulus of 0.01-0.1MPa and Poisson's ratio of 0.3-0.4 for memory foam, the elastic modulus of 0.5-1.5MPa and Poisson's ratio of 0.45-0.49 for latex, and the contact condition between the material layers is defined as friction contact, and the friction coefficient is set to 0.3. The established geometric model is adaptively meshed, the hexahedral unit type is selected, the initial mesh size is set to 5 mm, and it is automatically refined to 1 mm in areas with large stress gradients to ensure calculation accuracy. Apply boundary conditions and loads, apply fixed constraints at the bottom of the mattress, and apply distributed pressure loads on the top of the mattress according to the human body pressure distribution diagram, such as setting 1kPa in the head area, 2kPa in the shoulder area, 3kPa in the waist area, 2.5kPa in the hip area, and 1.5kPa in the leg area. Solve the nonlinear equations through the Newton-Raphson iteration method to calculate the support force distribution and pressure distribution data. The solution of the nonlinear equation is gradually approached through iteration. In each iteration, a linearized problem is formed, and the linear system is solved to update the displacement vector u. The linearized equation is expressed as:

[0021] K(u)△u=-R(u)

[0022] In the formula, K(u) represents the stiffness matrix, which depends on the displacement u, Δu represents the displacement increment, and R(u) represents the residual force vector. According to the physical model of the support force and pressure distribution, the mathematical model of the nonlinear equation group is established to determine the unknowns, known parameters and boundary conditions of the equation group. The Newton-Raphson iteration method is used to solve the nonlinear equation group, and the correction amount of the current iteration point is calculated according to the Jacobian matrix and residual vector of the equation group. In each iteration process, by updating the value of the unknown number, a new iteration point is obtained, and it is judged whether the convergence condition is met. If the condition is met, the iteration is terminated, otherwise the iterative calculation continues. For the calculation process of the Newton-Raphson iteration method, the trust region method or line search method can also be used to improve the convergence speed and calculation efficiency of the iteration. While iteratively solving, by monitoring the norm change of the residual vector, the iteration step size and convergence threshold are dynamically adjusted to balance the calculation accuracy and efficiency. After the iterative calculation is completed, the distribution data of the support force and pressure are obtained according to the converged solution. The displacement and stress values ​​of key nodes are extracted from the calculation results and compared with the preset thresholds. If the maximum displacement is less than 20 mm and the maximum stress is less than 0.1 MPa, the preliminary production plan is judged to meet the threshold requirements and added to the candidate plan set. Otherwise, the sensitivity analysis method is used to calculate the influence of each parameter on the displacement and stress, and the parameters are automatically adjusted according to the influence. At the same time, the Latin hypercube sampling method is used to generate multiple groups of stacking combinations for batch analysis until a solution that meets the requirements is found or the preset number of iterations is reached. In the finite element analysis software, a 2000mm×1800mm×300mm mattress 3D model was created according to the preliminary production plan, which contained 5 layers of materials: 20mm thick memory foam with an elastic modulus of 0.05MPa and a Poisson's ratio of 0.35, 50mm thick high-density foam with an elastic modulus of 0.8MPa and a Poisson's ratio of 0.4, 100mm thick spring layer with an equivalent elastic modulus of 2MPa and a Poisson's ratio of 0.3, 50mm thick latex with an elastic modulus of 1MPa and a Poisson's ratio of 0.47, and 80mm thick base foam with an elastic modulus of 0.3MPa and a Poisson's ratio of 0.38. The interlayer contact was set to friction contact with a friction coefficient of 0.3. Adaptive meshing was used, with an initial mesh size of 5mm, which was automatically refined to 1mm in the human body contact area, and the total number of units was about 5 million. A fixed constraint is applied to the bottom of the mattress, and a load is applied to the top according to the standard human body pressure distribution diagram, including 1kPa for the head, 2kPa for the shoulder, 3kPa for the waist, 2.5kPa for the hip, and 1.5kPa for the leg. The Newton-Raphson iteration method is used to solve the nonlinear equations, with the convergence accuracy set to 0.1% and the maximum number of iterations set to 100. The displacement and stress values ​​of 200 key nodes are extracted from the results, and the average and standard deviation are calculated. If the maximum displacement is less than 20mm and the maximum stress is less than 0.1MPa, the solution is added to the candidate set.Otherwise, the influence of each parameter on displacement and stress is calculated through sensitivity analysis. If it is found that the thickness of memory foam has the greatest influence on shoulder pressure, this parameter is adjusted first. The Latin hypercube sampling method is used to generate 50 different stacking combinations, each with a memory foam thickness of 15-25mm, a high-density foam thickness of 40-60mm, a latex thickness of 40-60mm, and a basic foam thickness of 70-90mm. These 50 groups of solutions are batch analyzed, and the solutions that meet the requirements are selected and added to the candidate set. If the candidate set is empty or the number of analyses reaches 200, the iteration is stopped and the optimal solution is output.

[0023] Step S103, analyze the candidate solution set, extract the material combination characteristics and layer thickness characteristics corresponding to different parts of the human body, and establish a material selection and layer production model for a personalized mattress. In the model, according to the mutual influence and constraint relationship between material properties, the material layering sequence and thickness are optimized to obtain a set of revised production solutions.

[0024] The material combination characteristics and stacking thickness characteristics of each solution in the candidate solution set are obtained, and a feature vector matrix is ​​established, wherein each row of the feature vector matrix represents a candidate solution. The feature vector matrix is ​​subjected to principal component analysis, and the principal components whose explained variance ratio exceeds a preset threshold are retained to obtain a simplified feature representation. According to the simplified feature representation, a material selection and stacking production model for a personalized mattress is constructed, and the input of the model is the pressure distribution value of human body parts and body characteristic parameters. If the hardness difference of adjacent layers of materials exceeds a preset threshold, the adjacent layers of materials are adjusted; if the total thickness is not within a preset range, the stacking thickness is readjusted. A genetic algorithm is used to optimize the stacking order and thickness. The encoding method of the genetic algorithm is real number encoding. The fitness function of the genetic algorithm is designed in combination with the prediction results and constraints of the model, and a revised production plan is obtained through evolution of a preset number of iterations.

[0025] Exemplarily, the material combination characteristics and stacking thickness characteristics of each solution are extracted from the candidate solution set, and a feature vector matrix is ​​established, in which each row represents a candidate solution, and the columns contain 25 features including material type, hardness, thickness, density, and air permeability of the five parts of the head, shoulders, waist, buttocks, and legs. The principal component analysis method is used to reduce the dimension of the feature vector matrix, retain the principal components with an explained variance ratio of more than 80%, and determine the optimal number of principal components through cross-validation to obtain a simplified feature representation. Based on the simplified feature representation, a support vector regression algorithm is used to construct a material selection and stacking production model for a personalized mattress, with inputs such as pressure distribution values ​​of five parts of the human body, height, weight, BMI index, etc., and outputs the corresponding material combination and stacking thickness. For the constructed model, material performance constraints are introduced, such as the hardness difference of adjacent layers of materials does not exceed 30%, the total thickness is controlled within the range of 200-300 mm, and the air permeability decreases with depth. The stacking order and thickness are optimized through genetic algorithms. The encoding method uses real number encoding. Each gene represents the type and thickness of a layer of material. The fitness function is designed based on the prediction results and constraints of the support vector regression model. The crossover operation uses arithmetic crossover, and the mutation operation uses Gaussian mutation. After 100 generations of evolution, a set of revised production plans are obtained. In the process of making personalized mattresses, features are first extracted from 100 candidate plans to construct a 25×100 feature matrix. Each plan contains material types for five parts: head, shoulders, waist, buttocks, and legs, such as 1 for memory foam, 2 for latex, 3 for spring, hardness 10-50kPa, thickness 10-100mm, density 20-80kg / m 3, breathability 1-10 and other features. Principal component analysis was applied to the feature matrix. By calculating the eigenvalues ​​and cumulative contribution rates, the first 7 principal components were selected, explaining 83.5% of the total variance. The optimal number of principal components was determined using 5-fold cross validation, and a simplified 7×100 feature matrix was obtained. The support vector regression algorithm was used to build the model. The input was the pressure distribution values ​​of five parts of the human body, such as head 1kPa, shoulder 2kPa, waist 3kPa, hip 2.5kPa, leg 1.5kPa, height 170cm, weight 65kg, BMI index 22.5 and other features. The output was a 25-dimensional material combination and stacking thickness vector. Material performance constraints were introduced, such as the hardness difference between adjacent layers was controlled within 30%, the total thickness was limited to 200-300mm, and the breathability showed a decreasing trend of 5% from the top layer to the bottom layer. The stacking structure was optimized using a genetic algorithm, with a population size of 50 and a coding length of 50. Each layer of material used 2 bits to represent the type, 8 bits to represent the thickness, a crossover probability of 0.8, and a mutation probability of 0.1. The fitness function was designed in combination with the predicted comfort score of 0-100 and the constraint violation degree of the support vector regression model. After 100 generations of evolution, the optimal stacking solution was obtained, which was a top layer of 20mm memory foam, a second layer of 40mm latex, a third layer of 100mm pocket springs, a fourth layer of 30mm high-density sponge, and a bottom layer of 60mm hard support, with a total thickness of 250mm and a predicted comfort score of 92.5.

[0026] Step S104: for each revised production plan, the overall comfort data of the mattress is obtained by simulating the pressure distribution of the human body under different sleeping postures. According to the preset comfort evaluation index, the comprehensive performance score of each plan is obtained, and the plan with the best comfort is selected as the recommended plan.

[0027] Construct a three-dimensional mattress model and use finite element analysis software to set the material parameters of the mattress model; wherein the material parameters include elastic modulus, Poisson's ratio and density. Obtain posture models of different body types from the standard human body model database and construct a human body simulation model; the human body simulation model includes bones, muscles and skin layers. Place the human body simulation model on the mattress model, apply gravity load, and obtain contact status and pressure distribution data under different postures. Combine the heat conduction equation and mass diffusion equation to obtain the temperature distribution and humidity distribution data of the mattress surface. Calculate the comprehensive performance score of each solution according to the comfort evaluation index; select the solution with the highest comprehensive performance score as the recommended solution; wherein the comfort evaluation index includes pressure uniformity, support, temperature regulation and breathability. Pressure uniformity is calculated by the following formula:

[0028]

[0029] Where σ represents pressure uniformity, P irepresents the pressure value of the i-th measuring point, P avg It represents the average pressure value of all measuring points, and n represents the total number of measuring points.

[0030] Exemplarily, according to the revised production plan, a three-dimensional mattress model is constructed in the finite element analysis software, and material parameters including elastic modulus, Poisson's ratio, and density are set. The adaptive meshing method is used, and the initial mesh size is set to 5 mm, which is automatically refined to 1 mm in the human body contact area. Supine, side-lying, and prone posture models of different body types are selected from the standard human body model database to construct a human body simulation model, including bones, muscles, and skin layers, and set the corresponding material properties and contact characteristics. The human body model is placed on the mattress model, and a gravity load is applied to simulate the contact state and pressure distribution under different postures. The contact analysis method combining Hertz contact theory and finite element method is used to calculate the pressure distribution of the contact surface between the human body and the mattress, extract the average pressure value and pressure peak value of key parts such as the head, shoulders, waist, buttocks, and legs, and obtain the deformation and stress distribution data of the mattress. Combined with the heat conduction equation and the mass diffusion equation, the temperature field and humidity field are simulated and analyzed to obtain the temperature distribution and humidity distribution data of the mattress surface. According to the preset comfort evaluation indicators, including pressure uniformity (expressed by standard deviation), support (expressed by the pressure ratio of key parts), temperature regulation (expressed by the surface temperature change rate) and air permeability (expressed by water vapor permeability), the comprehensive performance score of each solution is calculated, and the final score is obtained by weighted summation method. The solution with the highest score is selected as the recommended solution. In the process of personalized mattress production, ANSYS software is first used to build a 2000mm×1800mm×250mm three-dimensional mattress model, and the elastic modulus of the top memory foam is set to 0.05MPa, Poisson's ratio is 0.35, and the density is 50kg / m 3 . Adaptive meshing was used, with an initial mesh size of 5mm, and the contact area was refined to 1mm, with a total number of about 2 million units. A standard male model of 175cm / 70kg was selected from the human body model library, which contained 206 bones and more than 600 muscles. The skin elastic modulus was set to 1MPa and the Poisson's ratio was set to 0.48. Three postures were simulated: supine, left-side lying, and right-side lying, with a gravity load of 700N applied to each posture. The initial contact area was calculated using Hertz contact theory, and the nonlinear contact problem was solved by finite element iteration, with the convergence accuracy set to 0.1%. The pressure data of five key parts were extracted: head 10cm×10cm, shoulder 20cm×30cm, waist 30cm×20cm, buttocks 25cm×25cm, and legs 40cm×30cm. At the same time, thermal and moisture coupling analysis was performed, with the human body surface temperature set to 37℃, relative humidity 60%, ambient temperature 25℃, relative humidity 50%, and heat conduction equation and mass diffusion equation solved. The time step was 60s, and the total simulation time was 8 hours. Calculate comfort index, pressure uniformity

[0031]

[0032] Where σ represents pressure uniformity, P i represents the pressure value of the i-th measuring point, P avg It represents the average pressure value of all measuring points, n represents the total number of measuring points, and the support

[0033]

[0034] In the formula, S represents support, P waist Indicates waist pressure, P head Indicates head pressure, P shoulder Indicates shoulder pressure, P hip Indicates hip pressure, P leg Indicates leg pressure. The formula calculates the ratio of waist pressure to one-fourth of the average inverse of the pressure in the other four parts. Temperature regulation T = |dT / dt|max, T represents temperature, t represents time, dT / dt is the rate of change of temperature over time, that is, the derivative of temperature with respect to time, which represents the absolute value of the maximum rate of temperature change of the mattress during use. Breathability, where S represents support, P waist Indicates waist pressure, P head Indicates head pressure, P shoulder Indicates shoulder pressure, P hip Indicates hip pressure, P leg Indicates leg pressure. The formula calculates the ratio of waist pressure to one-fourth of the average inverse of the pressure in the other four parts. Temperature regulation T = |dT / dt|max, T represents temperature, t represents time, dT / dt represents the rate of change of temperature over time, that is, the derivative of temperature with respect to time, which represents the absolute value of the maximum rate of temperature change of the mattress during use. Breathability:

[0035]

[0036] In the formula, V represents the gas permeability of the material, m represents the mass of gas passing through area A in time t, and ΔP represents the pressure difference on both sides of the gas. The comprehensive score is calculated by weighted summation W=0.4σ+0.3S+0.2T+0.1V, and the solution with the highest score is selected as the final recommendation.

[0037] Step S105: Analyze the material attribute parameters in the recommended solution, and select the most suitable material combination and stacking parameters according to the hardness, resilience, and air permeability parameters of the stacked mattress, combined with the customer's body measurement data and personalized demand data. Generate a structural model diagram and cross-section diagram of the personalized mattress through 3D modeling software as a basis for production and processing.

[0038] Obtain the hardness value, rebound coefficient and air permeability of each layer of material in the recommended solution, and establish a material property parameter vector; receive the customer's height, weight, body shape measurement data, and generate personalized demand data based on the customer's preference questionnaire results for hardness, rebound resilience and air permeability; use a gradient boosting decision tree algorithm, with customer data as input and the material property parameters as output, to train a matching model; based on the prediction results of the matching model, select materials with matching hardness, rebound resilience and air permeability from the material library, determine the model and thickness of each layer of material, and generate a laminated structure solution for the mattress; if the laminated structure solution meets the compatibility requirements, create a three-dimensional model of the mattress through three-dimensional modeling software, including drawing a two-dimensional cross-sectional profile, generating a three-dimensional entity through a stretching operation, and assigning corresponding attributes and colors to each layer of material; generate an overall structural model diagram of the mattress and cross-sectional diagrams of each cross section, and add a labeling function to indicate the parameters and size information of each layer of material.

[0039] For example, material property parameters are extracted from the recommended solution, including the hardness value (unit: N), rebound coefficient (percentage), air permeability (cm 3 / cm 2 ·s), establish parameter vectors, and obtain customer's height, weight, body shape and other body measurement data at the same time. Through questionnaire survey, convert customer's preference for hardness, resilience and breathability into scores of 1-10 as personalized demand data. Perform feature engineering on customer data, calculate BMI index, segment age and generate new feature vectors. Use gradient boosting decision tree algorithm, take processed customer data as input and material property parameters as output, train matching model to predict the best material combination and stacking parameters. According to the prediction results of matching model, select the material with the closest hardness, resilience and breathability from material library, determine the specific model and thickness of each layer of material, and generate personalized mattress stacking structure scheme. In this process, perform compatibility check of material combination to ensure that the hardness difference of adjacent layers does not exceed 30%, and the total thickness is controlled within 200-300mm. Use Solidworks software to create a three-dimensional model of mattress according to stacking structure scheme. First, draw two-dimensional section profile, generate three-dimensional entity through stretching operation, and then assign corresponding attributes and colors to each layer of material. Generate the overall structure model diagram of the mattress and the cross-section diagram of each section, add the annotation function to indicate the parameters and size information of each layer of material, as the technical basis for subsequent production and processing. In the process of personalized mattress design, first extract the material attribute parameters from the recommended solution, such as the top layer memory foam hardness 25N, rebound coefficient 85%, air permeability 0.5cm 3 / cm 2·s. Obtain the customer's body shape data of 178cm in height, 75kg in weight, and 35 years old, and convert the customer's preferences for hardness, resilience, and breathability into 8, 7, and 6 points through an online questionnaire. The calculated BMI is 23.7, and the age range is 30-40 years old. Use the XGBoost algorithm to train the matching model, set the maximum tree depth to 6, the learning rate to 0.1, the number of iterations to 500, and use five-fold cross validation. The model output predicts that the best material combination includes a top layer of 20mm memory foam, a middle layer of 50mm latex, and a bottom layer of 180mm pocket springs, with hardnesses of 23N, 35N, and 60N, respectively. Check the material compatibility, the hardness difference between adjacent layers is 52% and 71%, exceeding the 30% threshold, and is adjusted to a top layer of 25mm memory foam (hardness 28N), a transition layer of 30mm high-density sponge (hardness 40N), a middle layer of 60mm latex (hardness 50N), and a bottom layer of 135mm pocket springs (hardness 65N), with a total thickness of 250mm. Use Solidworks software to model, first draw a 2000mm×1800mm rectangular bottom surface, stretch the thickness of each layer in turn, and set the material properties: memory foam density 50kg / m 3 , rebound coefficient 85%; high density sponge density 60kg / m 3 , rebound coefficient 75%; latex density 75kg / m 3 , rebound coefficient 90%; pocket spring density 100kg / m 3 , rebound coefficient 95%. Generate three-view and isometric drawings, and the cross-section view shows the four-layer structure, annotates the thickness and material parameters of each layer, and exports DWG format files as a basis for production.

[0040] Step S106, conduct quality inspection on the finished personalized mattress after production and processing, obtain actual material performance parameters and human body pressure distribution data, and compare the detected data with the selected most matching material combination and stacking parameters to determine whether the detected data meets the comfort requirements. If the requirements are met, the mattress will be packaged and sold, otherwise it will be remade.

[0041] Obtaining material performance parameters of a finished personalized mattress includes: testing multiple preset test points of the finished personalized mattress using a hardness tester, a rebound tester, and an air permeability tester to obtain the actual hardness value, rebound coefficient, and air permeability data of each layer of material. According to the material performance parameters, the finished personalized mattress is tested for human pressure distribution through a pressure sensor array, and a weight distribution simulator is used to simulate multiple preset postures to record the pressure values ​​of each body part. Judging whether the material performance parameters and human pressure distribution data meet the preset conditions includes: comparing the material performance parameters and human pressure distribution data with the most matching material combination and stacking parameters selected in the design stage, and calculating the deviation value of each parameter. If the material performance parameters and human pressure distribution data meet the preset conditions, the personalized mattress is evaluated for comfort using a random forest algorithm and a comfort score is output. Determining whether the finished personalized mattress is qualified according to the comfort score includes: if the comfort score exceeds a preset threshold, the personalized mattress is judged to be qualified; if the comfort score does not exceed the preset threshold, the detection data is fed back to the production system, and the process parameters are adjusted to re-produce the finished personalized mattress.

[0042] Exemplarily, a hardness tester, a rebound tester and an air permeability tester are used to test the material performance parameters of the finished personalized mattress. Three tests are performed at the four corners, the midpoints of the four sides and the center of the mattress, a total of 9 key points, and the average value is taken to obtain the actual hardness value, rebound coefficient and air permeability data of each layer of material. A high-precision pressure sensor array is used to test the human body pressure distribution of the mattress. A weight distribution simulator is used to simulate three postures of supine, side and prone respectively. Each posture lasts for 5 minutes, and the pressure values ​​of the head, shoulders, waist, buttocks and legs are recorded. The material performance parameters and human body pressure distribution data obtained by the test are compared and analyzed with the most matching material combination and stacking parameters selected in the design stage, and the root mean square error is used to calculate the deviation value of each parameter. According to the preset comfort threshold requirements, it is judged whether the test data meets the conditions, and the hardness deviation does not exceed ±10%, the rebound coefficient deviation does not exceed ±5%, the air permeability deviation does not exceed ±15%, and the pressure distribution uniformity deviation does not exceed ±20%. The random forest algorithm is used to comprehensively evaluate the comfort of the mattress, taking various parameters as input and outputting a comfort score of 0-100. If all parameter deviation values ​​are within the allowable range and the comfort score exceeds 80 points, the mattress will be packaged and sold. Otherwise, the test data will be fed back to the production system, and the process parameters will be adjusted and re-made.

[0043] In the process of personalized mattress quality inspection, the Shore A-type hardness tester was first used to test 9 key points (four corners, four sides and center) three times each, and the average hardness of the top memory foam was 28.5±0.3, the middle latex hardness was 49.8±0.5, and the bottom spring hardness was 64.7±0.7. The rebound coefficients measured by the rebound tester were 86.2%, 91.5% and 95.8% respectively. The air permeability tester measured the air permeability of 0.48, 0.35 and 0.22cm respectively. 3 / cm 2 ·s. Subsequently, a 64×128 grid high-precision pressure sensor array was used, and a 70kg weight distribution simulator was used to simulate three sleeping positions, each lasting 5 minutes, with a sampling frequency of 100Hz. The average pressure values ​​of the head, shoulders, waist, hips and legs in the supine position were recorded to be 0.9, 1.8, 2.7, 2.3 and 1.4kPa, respectively. The root mean square error was used to calculate the deviation of each parameter, with a hardness deviation of 1.8%, a rebound coefficient deviation of 1.5%, a permeability deviation of 4.2%, and a pressure distribution uniformity deviation of 6.5%, all within the preset threshold range. Finally, the random forest algorithm (number of trees 100, maximum depth 10) was used to comprehensively evaluate the comfort, with inputs including the hardness of each layer, rebound coefficient, permeability and pressure values ​​of five parts. After training, a comfort score of 92.5 was obtained. Since all parameters meet the requirements and the comfort score exceeds 80 points, the system automatically determines that the mattress is qualified and starts the packaging process for sale.

[0044] It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present application. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for making a personalized mattress by combining body shape measurement data, characterized in that: The method comprises: According to the pre-established personalized mattress production database, we obtain the body shape measurement data of different customer parts, and select the material combination scheme that matches the hardness, resilience and air permeability parameters from the material database according to the curve characteristics of each part to form a preliminary production plan; The finite element analysis method is used to simulate the mechanical properties of the material stacking sequence and thickness in the preliminary production plan, obtain the support force distribution and pressure distribution data under different stacking combinations, and judge whether the material performance parameters or stacking thickness settings meet the threshold requirements. If the material performance parameters or stacking thickness settings meet the threshold requirements, the preliminary production plan is added to the candidate plan set; Analyze the candidate solution set, extract the material combination characteristics and layer thickness characteristics corresponding to different parts of the human body, establish a material selection and layer production model for personalized mattresses, and optimize the material layering sequence and thickness in the model based on the mutual influence and constraint relationship between material properties to obtain a set of revised production solutions; For each revised production plan, the pressure distribution of the human body under different sleeping postures is simulated to obtain the overall comfort data of the mattress. According to the preset comfort evaluation index, the comprehensive performance score of each plan is obtained, and the plan with the best comfort is selected as the recommended plan; Analyze the material property parameters in the recommended solution, select the most suitable material combination and stacking parameters based on the hardness, resilience and air permeability parameters of the stacked mattress, and combine the customer's body measurement data and personalized demand data. Use 3D modeling software to generate the structural model diagram and cross-section diagram of the personalized mattress as the basis for production and processing; Carry out quality inspection on the finished personalized mattresses after production and processing to obtain the actual material performance parameters and human body pressure distribution data, and compare the detected data with the selected most matching material combination and stacking parameters to determine whether the detected data meets the comfort requirements. If the requirements are met, the mattress will be packaged and sold, otherwise it will be remade.

2. The method according to claim 1, characterized in that According to the pre-established personalized mattress production database, the body shape measurement data of different customer body parts are obtained, and according to the curve characteristics of each part, a material combination scheme matching hardness, resilience and air permeability parameters is selected from the material database to form a preliminary production scheme, including: Acquire the whole body contour data of the customer, wherein the whole body contour data is collected by a three-dimensional scanner; Extracting curve features of key parts from the whole body contour data to obtain pressure distribution parameters of the customer; Use pressure sensor array to obtain actual pressure distribution data of each part when customers try lying down; The actual pressure distribution data is matched with a pre-established material combination scheme to determine a preliminary manufacturing scheme with optimal support performance.

3. The method according to claim 1, characterized in that The finite element analysis method is used to simulate the mechanical properties of the material stacking sequence and thickness in the preliminary production plan, obtain the support force distribution and pressure distribution data under different stacking combinations, and judge whether the material performance parameters or stacking thickness settings meet the threshold requirements. If the material performance parameters or stacking thickness settings meet the threshold requirements, the preliminary production plan is added to the candidate plan set, including: Obtaining the material stacking sequence and thickness information in the manufacturing plan, and establishing a three-dimensional geometric model in a finite element analysis software according to the material stacking sequence and thickness information; Performing adaptive meshing on the three-dimensional geometric model, wherein the adaptive meshing adopts a hexahedral unit type, and an initial mesh size is set to a preset value; Applying a fixed constraint at the bottom of the mattress of the three-dimensional geometric model, and applying a distributed pressure load at the top of the mattress of the three-dimensional geometric model according to a human body pressure distribution diagram; The Newton-Raphson iteration method is used to solve the nonlinear equations to obtain the support force distribution data and pressure distribution data; Extracting displacement values ​​and stress values ​​of key nodes from the support force distribution data and the pressure distribution data; Determining whether the displacement value is less than a first preset threshold and whether the stress value is less than a second preset threshold; If the displacement value is less than a first preset threshold and the stress value is less than a second preset threshold, adding the manufacturing solution to a candidate solution set; If the displacement value is greater than or equal to a first preset threshold or the stress value is greater than or equal to a second preset threshold, then calculating the influence of each parameter on the displacement value and the stress value; The parameters in the production plan are adjusted according to the degree of influence.

4. The method according to claim 1, characterized in that The candidate solution set is analyzed to extract the material combination characteristics and layer thickness characteristics corresponding to different parts of the human body, and a material selection and layer production model for a personalized mattress is established. In the model, the material layering sequence and thickness are optimized according to the mutual influence and constraint relationship between the material properties, and a set of revised production solutions are obtained, including: Obtaining material combination characteristics and stacking thickness characteristics of each solution in the candidate solution set, and establishing a feature vector matrix, wherein each row of the feature vector matrix represents a candidate solution; Performing principal component analysis on the eigenvector matrix, retaining principal components whose explained variance ratio exceeds a preset threshold, and obtaining a simplified feature representation; According to the simplified characteristic representation, a material selection and layering manufacturing model of a personalized mattress is constructed, wherein the input of the model is the pressure distribution value of human body parts and body characteristic parameters; If the hardness difference between adjacent layers of materials exceeds a preset threshold, the adjacent layers of materials are adjusted; If the total thickness is not within the preset range, readjust the stacking thickness; A genetic algorithm is used to optimize the stacking sequence and thickness. The encoding method of the genetic algorithm is real number encoding. The fitness function of the genetic algorithm is designed in combination with the prediction results and constraint conditions of the model, and a revised production plan is obtained through evolution of a preset number of iterations.

5. The method according to claim 1, characterized in that: For each revised production plan, the overall comfort data of the mattress is obtained by simulating the pressure distribution of the human body under different sleeping postures, and the comprehensive performance score of each plan is obtained according to the preset comfort evaluation index, and the plan with the best comfort is selected as the recommended plan, including: Constructing a three-dimensional mattress model, and setting material parameters of the mattress model using finite element analysis software; Wherein, the material parameters include elastic modulus, Poisson's ratio and density; Obtain posture models of different body types from the standard human body model database and construct a human body simulation model; The human body simulation model includes bones, muscles and skin layers; Placing the human body simulation model on the mattress model, applying gravity load, and obtaining contact state and pressure distribution data under different postures; Combining the heat conduction equation and mass diffusion equation, the temperature distribution and humidity distribution data of the mattress surface are obtained; Calculate the comprehensive performance score of each solution based on the comfort evaluation index; The solution with the highest comprehensive performance score is selected as the recommended solution; The comfort evaluation indexes include pressure uniformity, support, temperature regulation and breathability; Pressure uniformity is calculated using the following formula: Where σ represents pressure uniformity, P i represents the pressure value of the i-th measuring point, P avg It represents the average pressure value of all measuring points, and n represents the total number of measuring points.

6. The method according to claim 1, characterized in that The material attribute parameters in the recommended solution are analyzed, and the most matching material combination and stacking parameters are selected according to the hardness, resilience, and air permeability parameters of the stacked mattress, combined with the customer's body measurement data and personalized demand data, and the structural model diagram and cross-sectional diagram of the personalized mattress are generated through 3D modeling software as the basis for production and processing, including: Obtain the hardness value, rebound coefficient and air permeability of each layer of material in the recommended solution, and establish a material property parameter vector; Receive the customer's height, weight and body shape measurement data, and generate personalized demand data based on the customer's preference questionnaire results on hardness, resilience and breathability; Adopting the gradient boosting decision tree algorithm, taking the customer data as input and the material property parameters as output, to train the matching model; According to the prediction results of the matching model, materials with matching hardness, resilience and air permeability are selected from the material library, the model and thickness of each layer of material are determined, and a stacking structure plan of the mattress is generated; If the stacked structure scheme meets the compatibility requirements, a three-dimensional model of the mattress is created through three-dimensional modeling software, including drawing a two-dimensional cross-sectional profile, generating a three-dimensional entity through a stretching operation, and assigning corresponding properties and colors to each layer of material; Generate the overall structural model diagram of the mattress and the cross-sectional diagrams of each section, and add the annotation function to indicate the parameters and size information of each layer of material.

7. The method according to claim 1, characterized in that The quality inspection of the finished personalized mattress after production and processing is carried out to obtain the actual material performance parameters and human body pressure distribution data, and the detected data is compared with the selected most matching material combination and stacking parameters to determine whether the detected data meets the comfort requirements. If the requirements are met, the mattress is packaged and sold, otherwise it is remade, including: Obtaining material performance parameters of the finished personalized mattress, including: using a hardness tester, a rebound tester, and an air permeability tester to test multiple preset test points of the finished personalized mattress to obtain actual hardness values, rebound coefficients, and air permeability data of each layer of material; According to the material performance parameters, the human body pressure distribution test is performed on the personalized mattress product through a pressure sensor array, a weight distribution simulator is used to simulate multiple preset postures, and the pressure value of each body part is recorded; Determining whether the material performance parameters and the human body pressure distribution data meet preset conditions includes: comparing the material performance parameters and the human body pressure distribution data with the most matching material combination and lamination parameters selected in the design stage, and calculating the deviation value of each parameter; If the material performance parameters and the human body pressure distribution data meet the preset conditions, a random forest algorithm is used to evaluate the comfort of the personalized mattress product and a comfort score is output; Determining whether the personalized mattress finished product is qualified according to the comfort score includes: if the comfort score exceeds a preset threshold, determining that the personalized mattress finished product is qualified; If the comfort score does not exceed the preset threshold, the detection data is fed back to the production system, and the process parameters are adjusted to re-produce the personalized mattress product.

Images