A digital assisted design method integrating artificial intelligence and knitting process knowledge

Abstract

The present application provides a digital assisted design method integrating artificial intelligence and knitting process knowledge, including: establishing a knitted fabric physical property database according to the knitting process and wool fabric characteristics, the knitted fabric physical property data including the fabric's elastic coefficient, thickness, density, and friction coefficient parameters; obtaining a three-dimensional model of a human body surface by performing a three-dimensional scan of the human body, obtaining human body motion data, matching the knitted fabric physical property data with the three-dimensional model of the human body surface, and constructing a mapping relationship between the knitted fabric and the human body surface; evaluating the fit and motion adaptability of the knitted fabric, determining the optimal acquisition precision of the knitted fabric based on the evaluation results, and maximizing the freedom of movement while meeting the fit requirements.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a digital auxiliary design method integrating artificial intelligence and knitting process knowledge. Background Art

[0002] In the innovative practice of diffusion-assisted design of knitted surfaces, a core technical problem is particularly prominent in the process of integrating the advanced algorithms of artificial intelligence with deep knitting process knowledge: how to accurately grasp the degree of detail of parameter collection to ensure that the knitted fabric and the human body surface are in an ideal fit while avoiding the problems of inefficient data processing and extended design cycle caused by over-refinement. This problem not only concerns the delicate balance between the quality and quantity of collected data, but also deeply touches on the complexity of the inherent flexibility and elasticity of knitted fabrics in predicting mechanical behavior, further exacerbating the challenge of maintaining design accuracy under various deformations. In the design process, selecting the most critical design elements from a large set of parameters and building an efficient model of the interaction between these elements is a task that relies heavily on cross-domain knowledge integration and goes beyond the scope of traditional empirical rules. Although the intervention of artificial intelligence provides a new perspective for parameter optimization, ensuring that the algorithm can not only absorb the deep insights of domain expertise but also flexibly adapt to data-oriented iterative learning, this combination of intelligence and practical wisdom has become a key issue that needs to be solved in order to promote the progress of knitted surface design and even the transformation of the entire clothing design industry to a more intelligent, personalized and refined one. Summary of the invention

[0003] The present invention provides a digital auxiliary design method integrating artificial intelligence and knitting process knowledge, which mainly includes:

[0004] According to the knitting process and wool fabric characteristics, a knitted fabric physical property database is established. The knitted fabric physical property data includes the fabric's elastic coefficient, thickness, density, and friction coefficient parameters.

[0005] By performing 3D scanning on the human body to obtain a 3D model of the human body surface and human body movement data, the physical property data of the knitted fabric is matched with the 3D model of the human body surface to construct a mapping relationship between the knitted fabric and the human body surface;

[0006] Evaluate the fit and sports adaptability of knitted fabrics, and determine the optimal collection fineness of knitted fabrics based on the evaluation results to maximize freedom of movement while meeting the fit requirements;

[0007] The fuzzy comprehensive evaluation method is used to combine the fit, sports adaptability, comfort and emotional experience factors of wool knitting to build a comprehensive evaluation model for knitted fabric design parameters, including optimizing the knitted fabric design parameters through genetic algorithms to obtain the optimal parameter combination;

[0008] Optimize the design consistency between knitted fabrics and design parameters, input the optimized wool knitting design parameters into digital knitting equipment for manufacturing, and perform quality inspection on the manufactured wool knitting to determine the consistency between the knitted fabrics and design parameters. If the consistency does not meet the preset requirements, return to the comprehensive evaluation model for re-optimization;

[0009] A simulation test was conducted on the human body wearing wool knitwear. A fuzzy comprehensive evaluation model of knitted fabric performance was established based on the user's objective evaluation data and surface fit to evaluate the comfort and freedom of movement of knitted fabrics. The design parameters were optimized based on the evaluation results and re-input into the digital knitting equipment for manufacturing. The performance of knitted fabrics and user experience were improved through iterative design until the expected design goals were achieved.

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

[0011] The present invention discloses a digital assisted design method integrating artificial intelligence and knitting process knowledge, which ensures that the design parameters of knitted fabrics can maximize the fit and motion adaptability of the human body, and through a highly integrated design and evaluation process, ensures that the wool knitted fabrics can fully consider the user's comfort and motion needs during the design and production process, while improving the accuracy of fabric design and the consistency of production; the technical effects also include improving the user customization capability of knitted fabrics, optimizing the functionality of fabrics, and greatly improving the comfort and satisfaction of the wearer by precisely controlling the matching of fabric properties with human body dynamics. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 The present invention is a flow chart of a digital-aided design method integrating artificial intelligence and knitting process knowledge.

[0013] Figure 2 It is a schematic diagram of a digital-aided design method integrating artificial intelligence and knitting process knowledge according to the present invention.

[0014] Figure 3 This is another schematic diagram of a digital-aided design method that integrates artificial intelligence and knitting process knowledge according to the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0016] like Figure 1-3 In this embodiment, a digital auxiliary design method integrating artificial intelligence and knitting process knowledge may specifically include:

[0017] Step S101, establishing a knitted fabric physical property database according to the knitting process and wool fabric properties, the knitted fabric physical property data including the fabric's elastic coefficient, thickness, density, and friction coefficient parameters.

[0018] According to the knitting process and wool fabric characteristics, the physical parameter data of elastic coefficient, thickness, density and friction coefficient of different types of knitted fabrics are obtained, and a knitted fabric physical property database is established; the data in the knitted fabric physical property database are statistically analyzed to obtain the distribution law and correlation of the knitted fabric physical property parameters under different knitting processes and wool fabric characteristics, and determine the correlation between the knitting process and wool fabric characteristics and the knitted fabric physical properties; a prediction model between the knitting process and wool fabric characteristic parameters and the knitted fabric physical property parameters is established, and the knitted fabric elastic coefficient, thickness, density and friction coefficient physical property parameters are predicted by inputting the knitting process and wool fabric characteristic parameters; the physical property parameters of the knitted fabric are predicted for a new knitting process and wool fabric combination using the established prediction model, and the predicted elastic coefficient, thickness, density and friction coefficient parameter values ​​are obtained; the prediction model is optimized by comparing and analyzing the prediction results with the actual test results.

[0019] For example, according to the knitting process and wool fabric characteristics, the physical parameter data such as elastic coefficient, thickness, density and friction coefficient of different types of knitted fabrics are collected and sorted to establish a knitted fabric physical property database. The data in the database are statistically analyzed to obtain the distribution law and correlation of the physical property parameters of knitted fabrics under different knitting processes and wool fabric characteristics, and the correlation between the knitting process and wool fabric characteristics and the physical properties of knitted fabrics is determined. A support vector regression (SVR) model is used to establish a prediction model between the knitting process and wool fabric characteristic parameters and the knitted fabric physical property parameters. The input features of the SVR model include knitting process parameters (such as needle type, needle pitch, tightness, etc.) and wool fabric characteristic parameters (such as wool diameter, wool length, wool curl, etc.), and the output is the elastic coefficient, thickness, density and friction coefficient of the knitted fabric. Before establishing the prediction model, the data needs to be preprocessed, including data cleaning, feature scaling, feature selection and other steps. The hyperparameters of the prediction model, such as the penalty coefficient C, kernel function type and kernel function parameters, are optimized using grid search and cross-validation methods. The trained prediction model is applied to a new knitting process and wool fabric combination to predict the physical property parameters of the knitted fabric. New knitting process and wool fabric combinations can be generated through orthogonal experimental design or expert knowledge recommendation to cover a wider range of knitting process and fabric property spaces. By comparing and analyzing the prediction results with the actual test results, the prediction model is continuously optimized and improved to improve the accuracy and reliability of the prediction. A knowledge base and experience base for the prediction of knitted fabric physical properties can be established. The knowledge base contains the association rules and expert experience between knitting process and wool fabric properties and knitted fabric physical properties, and the experience base contains historical data and prediction results. When collecting and collating knitted fabric physical property data, 100 different knitted fabric samples can be selected, each sample is tested 10 times, and 1000 sets of data are obtained. The Pearson correlation coefficient is used to analyze the correlation between knitting process parameters and wool fabric property parameters and knitted fabric physical property parameters. The significance level is 0.05, and the parameter pairs with correlation coefficients greater than 0.6 are considered to have significant correlation. When building a prediction model based on the SVR algorithm, you can choose the Gaussian kernel function, use grid search and 5-fold cross validation to optimize the model hyperparameters, the search range of the penalty coefficient C is [0.1, 1, 10, 100], and the search range of the kernel function parameter γ is [0.01, 0.1, 1, 10]. The data set is randomly divided into a training set and a test set in a ratio of 8:2, and the SVR model is trained and evaluated. The root mean square error (RMSE) and determination coefficient (R^2) are used to evaluate the prediction performance of the prediction model, requiring RMSE to be less than 0.1 and R^2 to be greater than 0.9.When using the trained prediction model to predict the new knitting process and wool fabric combination, the orthogonal experimental design method can be used to select three knitting process parameters (such as needle type, needle pitch, and tightness) and three wool fabric characteristic parameters (such as wool diameter, wool length, and wool curl), and each parameter takes three levels, for a total of 27 combinations. Each combination is predicted 5 times, and the average value is taken as the final prediction result. The prediction results are compared and analyzed with the actual test results, and the relative error is calculated, requiring the relative error to be less than 5%. It is also possible to establish a knowledge base of knitted fabric physical properties, and use association rule mining algorithms, such as the Apriori algorithm, to set the minimum support to 0.1 and the minimum confidence to 0.8, to mine the association rules between knitting process parameters and wool fabric characteristic parameters and knitted fabric physical characteristic parameters, and form knowledge rules in the form of IF-THEN.

[0020] Step S102, by performing a three-dimensional scan on the human body to obtain a three-dimensional model of the human body surface, obtain human body movement data, match the physical property data of the knitted fabric with the three-dimensional model of the human body surface, and construct a mapping relationship between the knitted fabric and the human body surface.

[0021] The three-dimensional point cloud data of the human body is obtained, and the three-dimensional mesh model of the human body surface is obtained by performing point cloud filtering, denoising and fusion on the three-dimensional point cloud data; the three-dimensional spatial coordinates of the marker points are synchronously collected by using an infrared camera by attaching reflective marker points to key parts of the human body, and the three-dimensional coordinates of the marker points are converted into three-dimensional motion trajectory data of the human body joints; the physical property parameters of the knitted fabric are matched with the three-dimensional model of the human body surface, and the mapping relationship between the physical properties of the knitted fabric and the shape of the human body surface is established by mapping the elastic coefficient, thickness and density parameters of the knitted fabric to the corresponding areas of the three-dimensional model of the human body surface.

[0022] For example, a high-precision 3D scanner Artec Eva is used to perform an all-round scan of the human body to obtain the 3D point cloud data of the human body, and the point cloud is filtered, denoised and fused through the point cloud data processing software Artec Studio to obtain a 3D mesh model of the human body surface. The OptiTrack motion capture system is used to collect the 3D motion data of the human body in different actions. By attaching reflective markers to key parts of the human body, 12 infrared cameras are used to synchronously collect the 3D spatial coordinates of the markers, and the OptiTrack Motive software is used to convert the 3D coordinates of the markers into 3D motion trajectory data of the human body joints. The physical property parameters such as the elastic coefficient, thickness, and density of the knitted fabric are converted into a 3D texture map, and the 3D mesh model of the human body surface is divided into different areas such as the trunk, limbs, and head. The 3D texture map of the physical property parameters of the knitted fabric is mapped to the corresponding area of ​​the 3D model of the human body surface using the UV mapping algorithm, and the mapping relationship between the physical properties of the knitted fabric and the shape of the human body surface is established. The mapped physical property parameters of the knitted fabric are interpolated and smoothed by Gaussian filtering. According to human motion data and physical property parameters of knitted fabrics, the deformation and stress distribution of knitted fabrics during human motion were simulated by ANSYS explicit dynamics finite element analysis method. The stress and deformation characteristics of knitted fabrics in different human surface areas were visualized and analyzed by strain cloud diagrams and stress cloud diagrams. The quantitative relationship between the physical property parameters of knitted fabrics and the shape characteristics and motion characteristics of human surface was obtained by multivariate regression analysis and variance analysis of the force and deformation data of knitted fabrics, and a multivariate quadratic regression model for matching the physical properties of knitted fabrics with the human surface was constructed. The support vector regression (SVR) algorithm was used to take the physical property parameters of knitted fabrics and the shape characteristics of human surface as input and the comfort and fit scores of knitted fabrics as output to construct a prediction model for matching the physical properties of knitted fabrics with the human surface. The LIBSVM toolbox was used to train and optimize the SVR model. The optimal hyperparameters of the SVR model were selected by grid search and cross-validation methods to improve the generalization performance of the prediction model. The multivariate quadratic regression model for matching the physical properties of knitted fabrics with the human surface and the SVR prediction model were integrated into the Unity3D engine to build a three-dimensional scene and human-computer interaction interface for the virtual fitting system. Maya software is used to build a three-dimensional model of knitted clothing, and the physical property parameters of knitted fabrics such as elasticity coefficient and friction coefficient are assigned to the corresponding parts of the clothing model. Through human body measurement data drive and human body shape semantic matching technology, a virtual human body model that matches the real human body shape and size is generated.The PhysX physics engine is used to simulate the dynamic wearing effect of knitted clothing on the virtual human model, and the real-time rendering pipeline of Unity3D is used to render the virtual fitting scene with realistic light and shadow materials, allowing users to interact and operate the virtual human model and knitted clothing in real time through the mouse and keyboard, observe the virtual fitting effect of knitted clothing from multiple perspectives, and evaluate the comfort and fit of knitted clothing. When the ArtecEva 3D scanner is used to scan the human body, the scanning accuracy can reach 0.1mm, the scanning speed is 16fps, and the scanning distance is 0.4-1.0m. Through multi-perspective scanning of the human body, point cloud data containing 5 million 3D points can be obtained. When processing the point cloud data using ArtecStudio software, the spatial dispersion filtering algorithm can be used to remove noise points, and the point cloud can be converted into a triangular mesh model through the Poisson surface reconstruction algorithm. The number of faces of the mesh model can reach 300,000. When using the OptiTrack motion capture system to collect human motion data, 12 Prime13W infrared cameras can be used, each with a sampling frequency of 240Hz and a spatial resolution of 1.3MP. By attaching reflective markers with a diameter of 14 mm at the 22 key joints of the human body, the three-dimensional spatial coordinates of the markers are calculated using the triangulation principle, and the three-dimensional trajectory curve of human motion is fitted by the least squares method. When mapping the physical property parameters of knitted fabrics to the human surface model, a parametric mapping algorithm based on NURBS surfaces can be used. By solving the weight coefficients of the NURBS surface control points, the optimal match between physical parameters and geometric shapes can be achieved, and the mapping error can be controlled within 1%. When performing finite element simulation of knitted fabrics in the process of human motion, tetrahedral unit meshes can be used to divide the human body and knitted fabric models, and the density of the knitted fabric is set to 200 kg / m. 3, the elastic modulus is 50MPa, the Poisson's ratio is 0.4, and the friction coefficient is 0.3. The strain and stress distribution of knitted fabrics at different time steps are calculated by explicit dynamics solution algorithm. The simulation time step is 0.01s, the total simulation time is 10s, and the average error of the calculation results is less than 5%. When constructing a prediction model for matching the physical properties of knitted fabrics with the human body surface, the Gaussian kernel function can be selected. The hyperparameters of SVR are optimized by 5-fold cross validation and grid search algorithm. Finally, the optimal SVR model with penalty coefficient C=10 and kernel function parameter g=0.1 is obtained. The prediction accuracy on the test set can reach more than 90%. When building a virtual fitting system, the PhysX physical simulation plug-in of the Unity3D engine can be used. The spring coefficient of the fabric is set to 50N / m, the damping coefficient is 0.5N·s / m, and the friction coefficient is 0.4. By solving the motion equations of the fabric mesh nodes, the real-time dynamic simulation of knitted fabrics in the process of human body movement is realized. The rendering frame rate can reach 60fps, providing users with a smooth and realistic virtual fitting experience.

[0023] Step S103, evaluating the fit and sports adaptability of the knitted fabric, and determining the optimal collection fineness of the knitted fabric based on the evaluation results, so as to maximize the freedom of movement while meeting the fit requirements.

[0024] According to the physical property parameters of knitted fabrics and the three-dimensional model of the human body surface, the finite element simulation analysis method is used to simulate the strain and stress distribution of knitted fabrics under different human body movement postures, and the fitting degree of knitted fabrics in different parts and the quantitative indicators of deformation during movement are obtained; the principal component analysis method is used to reduce the dimension of the evaluation indicators of knitted fabric fitting degree and motion adaptability, extract the key factors that have the greatest impact on the performance of knitted fabrics, and establish a quantitative relationship model between knitted fabric performance and key physical parameters; according to the quantitative relationship model between knitted fabric performance and key physical parameters, a multi-objective optimization algorithm is used to search for the optimal combination of the physical parameters of the knitted fabric; the optimal value range of the physical parameters of the knitted fabric is used as a constraint condition to optimize the data acquisition parameters in the three-dimensional scanning and motion capture process; the three-dimensional model of the knitted fabric is combined with the human body movement data to construct a dynamic simulation model of the knitted fabric during human body movement, evaluate the fitting degree and motion adaptability of the knitted fabric under different acquisition fineness, and verify the reliability of the determined optimal acquisition fineness.

[0025] For example, according to the physical property parameters of knitted fabrics and the three-dimensional model of the human body surface, the finite element simulation analysis method is used to simulate the strain and stress distribution of knitted fabrics under different motion postures of the human body, and the fit and motion adaptability of knitted fabrics are intuitively evaluated through strain cloud diagrams and stress cloud diagrams, and the fit degree of knitted fabrics in different parts and the quantitative index of deformation during motion are obtained. The knitted fabric fit and motion adaptability evaluation index obtained by finite element simulation analysis are used as input data for principal component analysis, and the principal component analysis method is used to extract the key factors that have the greatest impact on the performance of knitted fabrics, such as the elastic modulus, density, friction coefficient and other physical parameters of knitted fabrics, and a quantitative relationship model between the performance of knitted fabrics and key physical parameters is established by multiple linear regression or nonlinear regression methods. The elastic modulus, density and friction coefficient of knitted fabrics are used as independent variables, and the fit evaluation index of knitted fabrics is used as the dependent variable. The coefficients of the regression equation are fitted by the least squares method, and the significance of the regression model and the significance of the regression coefficient are evaluated by the F test and the t test. The quantitative relationship model is optimized by stepwise regression or ridge regression methods to improve the goodness of fit and prediction accuracy of the quantitative relationship model. According to the established quantitative relationship model between knitted fabric performance and key physical parameters, the fit and freedom of movement of knitted fabrics are taken as optimization objective functions, and the value range of key physical parameters is taken as constraint conditions. Heuristic search algorithms such as genetic algorithm or particle swarm algorithm are used to evaluate the quality of the solution through the fitness function, and the population is updated through selection, crossover, mutation and other operations to search for the optimal combination of knitted fabric physical parameters. The convergence condition of the optimization algorithm is set as the rate of change of the objective function value is less than a certain threshold, or the number of iterations reaches the preset upper limit. The conflict between the fit and freedom of movement of knitted fabrics is balanced by the Pareto optimal solution set, and a series of optional optimization schemes are obtained. The optimal solution set of knitted fabric physical parameters is taken as the constraint conditions in the process of three-dimensional scanning and motion capture, and the data acquisition parameters such as scanning accuracy, scanning speed, camera resolution, and marker density are optimized by reverse engineering methods. Under the premise of ensuring the quality of three-dimensional reconstruction of knitted fabrics, the data acquisition cost and computing resource consumption are minimized, and the optimal acquisition fineness of three-dimensional reconstruction of knitted fabrics is determined. Using 3D clothing design software such as MarvelousDesigner, the optimized physical parameters of knitted fabrics and acquisition precision settings are imported into the clothing design process. The dynamic deformation and stress of knitted fabrics during human movement are simulated through the physical simulation engine, and a virtual prototype model of knitted fabrics and human movement is constructed. The fit and motion adaptability of knitted fabrics under different acquisition precisions are intuitively evaluated through visualization methods such as strain cloud maps and stress cloud maps, and the effectiveness and reliability of the determined optimal acquisition precision are verified.Based on the optimal acquisition precision, a large amount of high-precision three-dimensional scanning data and human motion capture data of knitted fabrics are obtained. Machine learning algorithms such as support vector machines or random forests are used. The physical parameters of knitted fabrics and human motion characteristics are used as input, and the fit and motion adaptability evaluation indicators of knitted fabrics are used as output to construct a knitted fabric performance prediction model. The hyperparameters of the model are optimized through methods such as cross-validation and grid search to improve the generalization ability and robustness of the prediction model. The trained knitted fabric performance prediction model is encapsulated as a service module, and data interaction and function integration are carried out with the virtual fitting system through communication protocols such as WebSocket to realize real-time evaluation and feedback of the fit and motion adaptability of knitted fabrics. After the user selects the knitted fabric in the virtual fitting system and enters his or her body parameters, the knitted fabric performance prediction model is automatically called to calculate the fit and motion adaptability scores of the knitted fabric, and the knitted fabric is recommended and selected according to the scores, thereby realizing personalized customization and intelligent design of knitted clothing. When using the finite element simulation analysis method to simulate the deformation of knitted fabrics during human movement, the Mechanical module in ANSYS software can be used. The three-dimensional solid model of knitted fabrics can be constructed using SolidWorks software and imported into ANSYS for meshing. The mesh unit type can be tetrahedral unit SOLID187, and the unit size is controlled at about 1mm. The material properties of knitted fabrics can be obtained from data manuals or material tests. For example, the elastic modulus of spandex knitted fabrics is 0.5MPa, the Poisson's ratio is 0.4, and the density is 200kg / m. 3 . Human body motion posture can be collected by MotionCapture system, such as OptiTrackPrime13W camera, with a sampling frequency of 240Hz, and the kinematic data can be imported into ANSYS as the boundary condition of knitted fabric. It can be seen from the strain cloud map that the strain of knitted fabric is large at joints such as knees and elbows, with the maximum strain value of 15%. It can be seen from the stress cloud map that the stress of knitted fabric is large at places with large curvature such as shoulders and waists, with the maximum stress value of 1MPa. The average strain value and stress value of different parts of knitted fabric are used as evaluation indicators of fit and motion adaptability. They are substituted into SPSS software for principal component analysis, and two principal components are extracted, with a cumulative contribution rate of more than 95%. Elastic modulus, density, and friction coefficient are selected as the key physical parameters of knitted fabrics. The multivariate linear regression method is used to establish a quantitative relationship model with the fit evaluation index. The regression coefficient is fitted by the least squares method, and the adjustment R of the regression equation is 2is 0.85, and the P value of the F test is less than 0.05, indicating that the model is significant. A sensitivity analysis of the key physical parameters showed that the standardized regression coefficient of the elastic modulus was the largest, which was 0.6, and had the most significant effect on the fit of knitted fabrics. Taking the established regression model as the objective function and the range of elastic modulus, density, and friction coefficient as constraints, a genetic algorithm was used to optimize the parameters. The population size was set to 50, the crossover probability was 0.7, the mutation probability was 0.1, and the number of iterations was 100. The optimal parameter combination was determined by the Pareto frontier multi-objective genetic algorithm to be an elastic modulus of 0.3MPa, a density of 180kg / m 3 , friction coefficient is 0.4. Under this parameter, the fit and freedom of movement of knitted fabrics are balanced, and the weighted sum of the two optimization objectives is minimized. According to the optimized physical parameters of knitted fabrics, the accuracy and speed of 3D scanning are reversely solved. It is determined that when the scanning accuracy is 0.5mm and the scanning speed is 10fps, the precision of 3D reconstruction of knitted fabrics is optimal, and the error with the actual fit evaluation result is less than 5%. At the same time, the time for data acquisition and processing is the shortest, not exceeding 1 minute. The 3D model of knitted fabrics obtained under the optimal acquisition precision is imported into MarvelousDesigner software, the physical parameters of the fabric and the human body motion sequence are set, and the dynamic deformation of knitted fabrics during human body movement is simulated by the cloth simulation engine. The stress-strain cloud map of knitted fabrics in different postures is rendered and generated, and compared with the finite element simulation results, the relative error of the fit evaluation index of the two is less than 10%, which verifies the effectiveness of the determined optimal acquisition precision. 100 subjects of different body shapes were selected, and three-dimensional scanning and motion capture were performed on them with the best precision. A total of 10,000 sets of matching data of knitted fabrics and human movements were obtained. 80% of them were randomly selected as training sets and 20% as test sets. The support vector machine regression model was constructed using the scikit-learn library in Python. The optimal hyperparameters were obtained through 5-fold cross-validation and grid search. The mean absolute error MAE of the model on the test set was 0.02, and the determination coefficient R 2 The prediction model is 0.95, indicating that the accuracy of the knitted fabric performance prediction model is high. The prediction model is encapsulated as a RESTful API using the Flask framework and deployed on the server of the virtual fitting system. The front-end page interacts with the back-end API through Ajax requests, realizing the real-time evaluation function of the knitted fabric fit and sports adaptability. The average response time is less than 1 second, and the user experience is smooth.

[0026] Step S104, using a fuzzy comprehensive evaluation method, combining the fit, sports adaptability, comfort and emotional experience factors of wool knitting, to construct a comprehensive evaluation model for knitted fabric design parameters, including optimizing the knitted fabric design parameters through a genetic algorithm to obtain the optimal parameter combination.

[0027] According to the evaluation indicators of knitted fabric fit, sports adaptability, comfort and emotional experience, the hierarchical analysis method is used to construct a comprehensive evaluation index system for knitted fabric design, the importance weight of each indicator is obtained, and the weighted average method is used to calculate the comprehensive score of each indicator, and the mapping relationship between knitted fabric design parameters and comprehensive evaluation scores is established; according to the value range of knitted fabric design parameters, the triangle membership function is used to fuzzy process each parameter, and each parameter value is converted into a membership value, and the fuzzy matrix of knitted fabric design parameters is constructed. The fuzzy comprehensive evaluation is performed on the fuzzy matrix to obtain the comprehensive evaluation membership of the knitted fabric design scheme; according to the relationship between the knitted fabric design parameters and the evaluation indicators, the knitted fabric design parameters are fuzzy processed, and the comprehensive evaluation membership of the knitted fabric design scheme is obtained. The nonlinear relationship between them is considered, and a BP neural network is used to construct a comprehensive evaluation model for the design parameters of knitted fabrics. The membership value obtained by fuzzy comprehensive judgment is used as the input of the BP neural network, and the comprehensive evaluation score of the knitted fabric is used as the output. The knitted fabric design parameters are taken as optimization variables, and the output value of the comprehensive evaluation model is taken as the fitness function. A genetic algorithm is used to optimize the design parameters of knitted fabrics and search for the optimal parameter combination to maximize the comprehensive evaluation score of the knitted fabric. This includes balancing the conflicts and constraints between multiple evaluation indicators such as fit, sports adaptability, comfort and emotional experience in the iterative process of the genetic algorithm, and obtaining the Pareto optimal solution set of the knitted fabric design parameters.

[0028] For example, based on multiple evaluation indicators such as the fit, sports adaptability, comfort and emotional experience of knitted fabrics, the hierarchical analysis method is used to construct a comprehensive evaluation index system for knitted fabric design. The comprehensive evaluation target of knitted fabric design is decomposed into the target layer, the criterion layer and the index layer to form a hierarchical structure. In each layer, the factors of the same layer are compared pairwise by expert scoring to construct a judgment matrix, and the 1-9 scaling method is used to determine the element values ​​of the judgment matrix. A consistency test is performed on each judgment matrix, and the maximum eigenvalue and consistency index CI of the matrix are calculated. When CI<0.1, the judgment matrix is ​​considered to have satisfactory consistency. The weight vector of the judgment matrix is ​​calculated using the sum-product method or the geometric mean method, and the weight vectors of each level are combined to obtain the weight value of each indicator in the comprehensive evaluation index system. According to the value range of knitted fabric design parameters, the triangle membership function is used for fuzzy processing. The triangle membership function is μ(x)=max(min((xa) / (ma),(bx) / (bm)),0), where m is the peak point of the triangle, and a and b are the two boundary values ​​of the triangle. By adjusting the values ​​of a, b, and m, different fuzzy concepts and language variables can be flexibly expressed. The maximum-minimum synthesis method is used to perform fuzzy comprehensive evaluation on the fuzzy matrix of knitted fabric design parameters to obtain the comprehensive evaluation membership of the design scheme. The calculation steps of the maximum-minimum synthesis method are to take the maximum value of the fuzzy matrix by row and the minimum value by column to obtain the fuzzy evaluation vector; take the maximum value of the fuzzy evaluation vector by row to obtain the comprehensive evaluation membership of the knitted fabric design scheme. According to the nonlinear and complex relationship between the knitted fabric design parameters and the evaluation indicators, a three-layer BP neural network is used to construct a comprehensive evaluation model. The membership value obtained by the fuzzy comprehensive evaluation is used as the input of the BP neural network, and the comprehensive evaluation score of the knitted fabric is used as the output. The number of nodes in the middle hidden layer is determined according to the empirical formula. The BP neural network is trained by gradient descent method. By setting appropriate learning rate, momentum factor and iteration number, the connection weights and thresholds of the network are continuously updated until the mean square error is less than the specified training target. The design parameters of knitted fabrics are taken as optimization variables, the output value of BP neural network model is taken as fitness function, and genetic algorithm is used to globally optimize the design parameters. A penalty term is introduced in the fitness function to penalize individuals that exceed the constraints. The general form of the penalty term is P(x)=∑λi·max(0,gi(x)), where λi is the penalty coefficient and gi(x) is the i-th constraint function. By reasonably setting the population size, crossover probability, mutation probability and other parameters of the genetic algorithm, and using the elite retention strategy and non-dominated sorting method to deal with multi-objective optimization problems, the Pareto optimal solution set of knitted fabric design parameters is searched.The optimal parameter combination obtained by genetic algorithm optimization is applied to the process design and production of knitted fabrics. By adjusting the process parameters of the knitting machine such as needle type, needle pitch, bed speed, etc., the variety, fineness, strength and other indicators of wool yarn are optimized, and the pre-treatment processes of wool yarn such as dyeing and oil processing are optimized to achieve the improvement and stable control of the comprehensive performance of knitted fabrics. When constructing a comprehensive evaluation index system for knitted fabric design, fit, sports adaptability, comfort and emotional experience can be used as criterion layer indicators, and their weights are determined by expert scoring, which are 0.3, 0.3, 0.2 and 0.2 respectively. Under the fit index, three secondary indicators such as fabric elasticity and recovery are set; under the sports adaptability index, four secondary indicators such as fabric stretching and flexion are set; under the comfort index, three secondary indicators such as fabric breathability and hygroscopicity are set; under the emotional experience index, two secondary indicators such as fabric feel and appearance are set. The 1-9 scale is used to compare the secondary indicators. For example, if the ratio of the importance of elasticity to recovery is 5:1, the element in the corresponding position of the judgment matrix is ​​5. The CI value of the judgment matrix is ​​0.05, which passes the consistency test. The weight vector is calculated using the geometric mean method, and the weight values ​​of the secondary indicators are 0.25, 0.15, ..., 0.08. According to the elastic modulus parameters of knitted fabrics, a triangular membership function is designed, with m taking the value of 80, a taking the value of 50, and b taking the value of 120, mapping the continuous elastic modulus value to the membership value in the interval [0,1]. For different knitted fabric design schemes, a fuzzy matrix consisting of 8 design parameters such as elastic modulus, thickness, and density is constructed. The maximum and minimum synthesis method is used for fuzzy comprehensive evaluation, and the comprehensive evaluation membership of each scheme is 0.85, 0.76, and 0.92, respectively. With the membership degree as the input of BP neural network and the comprehensive evaluation score as the output, a three-layer BP neural network with 8-10-1 structure was constructed, the learning rate was 0.05, the momentum factor was 0.8, the sigmoid function was used as the activation function, and the training was carried out through the back propagation algorithm. After 5000 iterations, the mean square error dropped below 0.001. The value range of knitted fabric design parameters was used as the search space of genetic algorithm, the population size was set to 50, the crossover probability was 0.6, the mutation probability was 0.05, the constraint penalty term was introduced, the penalty coefficient was 100, and after 500 generations of evolution, the optimal combination of design parameters was obtained: elastic modulus 75, thickness 0.8mm, density 220g / m2, ..., and the comprehensive evaluation score was 95 points. In knitting production, the optimal parameter combination corresponding to No. 14 needles, 30 mesh / inch needle type, 1.2 times draft ratio, etc. is applied to the production settings of the large circular knitting machine. 16.5μm Australian wool is selected as raw material, and after wool yarn pre-treatment processes such as dyeing and shrinkage prevention, wool knitted fabrics with excellent comprehensive performance are finally produced.

[0029] Step S105, optimize the design consistency between the knitted fabric and the design parameters, input the optimized wool knitting design parameters into the digital knitting equipment for manufacturing; at the same time, perform quality inspection on the manufactured wool knitting to determine the consistency between the knitted fabric and the design parameters; if the consistency does not meet the preset requirements, return to the comprehensive evaluation model for re-optimization.

[0030] According to the optimal design parameter combination of wool knitting obtained by genetic algorithm optimization, the microstructure parameters of the fabric structure, needle type and needle pitch, and coil height of the knitted fabric are modeled by computer-aided design software, and the feasibility and optimization effect of the optimal design parameter combination of wool knitting are verified by finite element simulation analysis, and the knitting design model and process parameter planning scheme are obtained; the knitting design parameters are imported into the control system of the CNC knitting machine, and the process parameters of the needle cylinder speed, draft ratio, and line tension are set; the manufactured wool knitted fabric is scanned in all directions by a three-dimensional laser scanner, the three-dimensional point cloud data of the fabric surface is obtained, and pre-processing operations such as noise filtering and smoothing compression are performed to obtain a three-dimensional digital model of the knitted fabric; the three-dimensional digital model of the knitted fabric is compared with the digital design model, and the degree of consistency between the knitted finished product and the design parameters is determined by calculating the geometric parameters of the shape deviation and size error between the two models, as well as the physical parameters of the thickness, density, and elastic modulus of the fabric; if the consistency error between the knitted finished product and the design parameters exceeds the preset threshold, the three-dimensional scanning data and physical property test data of the knitted fabric are fed back to the three-dimensional digital model of the knitted fabric for re-optimization, and the optimized digital design parameters are generated and re-imported into the CNC knitting machine for production and manufacturing.

[0031] For example, the optimal design parameter combination of wool knitting obtained by genetic algorithm optimization, such as double-sided rib, needle type needle pitch of 8mm, coil height of 2mm, etc., is modeled in three dimensions using computer-aided design software, and the feasibility and optimization effect of the design parameter combination are verified by finite element simulation analysis. The simulation results show that the design parameter combination can improve the air permeability of wool knitted fabric by 20%, the warmth retention by 15%, and the softness by 10%, thereby obtaining the optimal knitting design model and process parameter planning scheme. The knitting design parameters are imported into the control system of the CNC knitting machine, and the process parameters such as the needle cylinder speed of 800rpm, the draft ratio of 2, and the line tension of 10cN are set. A three-dimensional laser scanner is used to perform an all-round scan of the manufactured wool knitted fabric to obtain the three-dimensional point cloud data of the fabric surface, and the noise is filtered by the Gaussian filter algorithm, and the surface fitting algorithm is used for smoothing and compression to obtain a three-dimensional digital model of the knitted fabric. The model was compared with the digital design model. The shape deviation was less than 5mm by calculating the Hausdorff distance between the two models. The size error was less than 1% by calculating the key dimensions of the two models. The thickness of the knitted fabric was 1mm and the density was 210g / m 2 , the elastic modulus is 8MPa, and the consistency between the knitted product and the design parameters is judged to be more than 95%. If the consistency error exceeds the preset threshold of 5%, the 3D scanning data and physical property test data of the knitted fabric are fed back to the 3D digital model of the knitted fabric, and re-optimized through the reverse engineering algorithm driven by shape deviation and the structural optimization algorithm driven by physical property parameters, and the optimized digital design parameters are generated and re-imported into the CNC knitting machine for production and manufacturing until the consistency between the knitted product and the design parameters meets the requirements.

[0032] Step S106, conduct a simulation test on human body wearing wool knitwear, establish a fuzzy comprehensive evaluation model of knitted fabric performance based on the user's objective evaluation data and surface fit to evaluate the comfort and freedom of movement of the knitted fabric; optimize the design parameters based on the evaluation results, re-input the optimized design parameters into the digital knitting equipment for manufacturing, and improve the performance of the knitted fabric and user experience through iterative design until the expected design goals are achieved.

[0033] The three-dimensional data of the human body surface and the three-dimensional data of the knitted surface when the user wears wool knitted fabrics are obtained, and the fit evaluation index of the user surface and the knitted surface is obtained. The data of the change of the joint angle when the user wears the knitted fabrics for different actions is recorded, and the evaluation index of the freedom of movement of the knitted fabric is judged according to the size of the joint range of motion. The fit evaluation index, the thermal insulation performance evaluation index and the freedom of movement evaluation index are used as inputs, and the fuzzy comprehensive evaluation method is used to establish a fuzzy comprehensive evaluation model of the knitted fabric design parameters to evaluate the comprehensive performance of the knitted fabric. If the comprehensive evaluation result does not meet the expected design goal, the evaluation result is correlated with the design parameters, and the structural parameters and process parameters of the knitted fabric are adjusted and optimized by the optimization algorithm to obtain the optimized design parameters. The optimized design parameters are input into the control program of the digital knitting equipment, and the digital knitting is performed by controlling the needle selection, needle movement and needle removal actions of the knitting equipment. The physical properties of the optimized knitted fabric samples produced are tested to obtain objective evaluation data on the thickness, quality, elasticity and strength of the fabric to determine whether it meets the design requirements. If the physical property test results of the optimized knitted fabric meet the design requirements, the design parameters of the knitted fabric will be applied; if the test results do not meet the requirements, the design parameters will be adjusted and optimized until the requirements are met.

[0034] For example, in order to obtain the three-dimensional data of the human body surface and the three-dimensional data of the knitted surface when the user wears wool knitted fabrics, a three-dimensional scanner can be used to scan the human body and knitted fabrics, with a scanning accuracy of up to 5mm and a scanning time of 2 seconds. Then the acquired point cloud data is imported into the Geomagic Design X software, and the two sets of point cloud data are automatically registered by the registration command, and the registration error is controlled within 3mm. After registration, the deviation analysis command is used to generate a deviation cloud map of the knitted fabric and the human body surface. According to the area ratio of different color areas in the deviation cloud map, it is concluded that the fit between the knitted fabric and the human body surface is 85%. In terms of the evaluation of the thermal comfort performance of knitted fabrics, the FLIRT530 infrared thermal imager is used to continuously collect the surface temperature of the human skin wearing knitted fabrics under the conditions of an ambient temperature of 18°C ​​and a relative humidity of 60%, and the skin temperature distribution data of different parts are obtained. By analyzing the standard deviation and temperature gradient of the skin temperature distribution, it is found that the skin temperature fluctuations on the back and abdomen are small, and the temperature gradient is 5°C / 10cm, indicating that the knitted fabric has good thermal insulation performance. The Vicon motion capture system was used to collect the angles of the lower limb joints of users wearing knitwear and performing actions such as walking and squatting at a sampling frequency of 100Hz. By calculating the kinematic parameters of the hip, knee, ankle and other joints, it was found that the influence of knitwear on the range of motion of the lower limb joints was less than 10%, indicating that the knitwear had a high degree of freedom of movement. Based on the above three evaluation indicators, a fuzzy comprehensive evaluation model of knitted fabric performance was constructed. The weights of each indicator were determined by the AHP hierarchical analysis method, and the membership function was used to calculate the membership. Finally, the comprehensive performance score of the knitted fabric was 85, which achieved the expected design goal. In order to further improve the performance of knitted fabrics, genetic algorithms were used to optimize the structural parameters such as coil length and tightness of knitted fabrics and the process parameters such as dyeing temperature and time. The multi-objective optimization solution was performed by Matlab software, and the optimized parameter combination of coil length of 12mm, tightness of 15 mesh / 10cm, dyeing temperature of 80℃ and dyeing time of 60min was obtained. The optimized parameters were input into the control program of the StollCMS530E2 flat knitting machine to intelligently control the knitting process and produce optimized knitted fabric samples. The thickness, quality, elasticity and strength of knitted fabric samples were tested using instruments such as the YG141D digital fabric thickness meter and the FA2104N electronic balance, and all indicators met the design requirements. A database for knitted fabric design optimization can also be established to record the design parameters, comprehensive evaluation results and physical performance test data of each iterative optimization, and to explore the inherent relationship between design parameters and fabric performance. Specifically, a knitted fabric design optimization case library is established, and the inherent relationship between design parameters and fabric performance is analyzed using an association rule mining algorithm. A series of rules are obtained, such as the thickness of knitted fabrics is positively correlated with the coil length, and the strength is positively correlated with the tightness, etc., providing data support for the design optimization of knitted fabrics.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and supplements without departing from the principle of the present invention. These improvements and supplements should also be regarded as the scope of protection of the present invention.

Claims

1. A digital assisted design method integrating artificial intelligence and knitting process knowledge, characterized in that: The method comprises: According to the knitting process and wool fabric characteristics, a knitted fabric physical property database is established. The knitted fabric physical property data includes the fabric's elastic coefficient, thickness, density, and friction coefficient parameters. By performing 3D scanning on the human body to obtain a 3D model of the human body surface and human body movement data, the physical property data of the knitted fabric is matched with the 3D model of the human body surface to construct a mapping relationship between the knitted fabric and the human body surface; Evaluate the fit and sports adaptability of knitted fabrics, and determine the optimal collection fineness of knitted fabrics based on the evaluation results to maximize freedom of movement while meeting the fit requirements; The fuzzy comprehensive evaluation method is used to combine the fit, sports adaptability, comfort and emotional experience factors of wool knitting to build a comprehensive evaluation model for knitted fabric design parameters, including optimizing the knitted fabric design parameters through genetic algorithms to obtain the optimal parameter combination; Optimize the design consistency between knitted fabrics and design parameters, input the optimized wool knitting design parameters into digital knitting equipment for manufacturing, and perform quality inspection on the manufactured wool knitting to determine the consistency between the knitted fabrics and design parameters. If the consistency does not meet the preset requirements, return to the comprehensive evaluation model for re-optimization; Conduct simulation tests on human bodies wearing wool knitwear, establish a fuzzy comprehensive evaluation model of knitted fabric performance based on the user's objective evaluation data and surface fit to evaluate the comfort and freedom of movement of knitted fabrics, optimize design parameters based on the evaluation results, and re-input the optimized design parameters into digital knitting equipment for manufacturing. Improve the performance of knitted fabrics and user experience through iterative design until the expected design goals are achieved; The evaluation of the fit and sports adaptability of the knitted fabric is performed, and the optimal collection fineness of the knitted fabric is determined based on the evaluation results, so as to maximize the freedom of movement while meeting the fit requirements, including: According to the physical property parameters of knitted fabrics and the three-dimensional model of the human body surface, the finite element simulation analysis method is used to simulate the strain and stress distribution of knitted fabrics under different movement postures of the human body, and the fitting degree of knitted fabrics in different parts and the quantitative index of deformation during movement are obtained; The principal component analysis method is used to reduce the dimensionality of the knitted fabric fit and deformation quantitative indicators during movement, extract the key physical parameters that have the greatest impact on the performance of knitted fabrics, and establish a quantitative relationship model between the knitted fabric performance and the key physical parameters; According to the quantitative relationship model between the performance of knitted fabrics and key physical parameters, a multi-objective optimization algorithm is used to search for the optimal combination of the physical property data of the knitted fabrics; Taking the optimal value range of knitted fabric physical property data as a constraint condition, the data acquisition parameters in the 3D scanning and motion capture process are optimized; The three-dimensional model of knitted fabrics is combined with human motion data to construct a dynamic simulation model of knitted fabrics during human motion. The fit and motion adaptability of knitted fabrics under different acquisition fineness are evaluated, and the reliability of the determined optimal acquisition fineness is verified.

2. A digital assisted design method integrating artificial intelligence and knitting process knowledge according to claim 1, characterized in that: According to the knitting process and wool fabric characteristics, a knitted fabric physical property database is established, and the knitted fabric physical property data includes the elastic coefficient, thickness, density, and friction coefficient parameters of the fabric, including: According to the knitting process and wool fabric characteristics, the physical parameter data of elasticity coefficient, thickness, density and friction coefficient of different types of knitted fabrics are obtained, and a knitted fabric physical property database is established; Statistical analysis was performed on the data in the knitted fabric physical property database to obtain the distribution law and correlation of knitted fabric physical property parameters under different knitting processes and wool fabric properties, and to determine the correlation between knitting processes and wool fabric properties and knitted fabric physical properties; Establish a prediction model between knitting process and wool fabric characteristic parameters and knitted fabric physical characteristic parameters, and predict the elastic coefficient, thickness, density and friction coefficient physical characteristic parameters of knitted fabric by inputting knitting process and wool fabric characteristic parameters; Using the established prediction model, the physical properties of knitted fabrics are predicted for new knitting processes and wool fabric combinations to obtain predicted elastic coefficient, thickness, density and friction coefficient parameter values; The prediction model is optimized by comparing the prediction results with the actual test results.

3. The digital assisted design method integrating artificial intelligence and knitting process knowledge according to claim 1 is characterized in that: The method of performing a three-dimensional scan on a human body to obtain a three-dimensional model of a human body curved surface, obtaining human body movement data, matching the physical property data of the knitted fabric with the three-dimensional model of the human body curved surface, and constructing a mapping relationship between the knitted fabric and the human body curved surface includes: Acquire three-dimensional point cloud data of a human body, and obtain a three-dimensional mesh model of a human body surface by performing point cloud filtering, denoising and fusion on the three-dimensional point cloud data; By attaching reflective markers to key parts of the human body, using infrared cameras to synchronously collect the three-dimensional spatial coordinates of the markers, and converting the three-dimensional coordinates of the markers into three-dimensional motion trajectory data of human joints; The physical property parameters of the knitted fabric are matched with the three-dimensional model of the human body surface, and the mapping relationship between the physical properties of the knitted fabric and the human body surface shape is established by mapping the elastic coefficient, thickness and density parameters of the knitted fabric to the corresponding area of ​​the three-dimensional model of the human body surface.

4. The digital assisted design method integrating artificial intelligence and knitting process knowledge according to claim 1 is characterized in that: The fuzzy comprehensive evaluation method is used to combine the fit, sports adaptability, comfort and emotional experience factors of wool knitting to construct a comprehensive evaluation model for knitted fabric design parameters, including optimizing the knitted fabric design parameters through a genetic algorithm to obtain the optimal parameter combination, including: According to the evaluation indicators of knitted fabric fit, sports adaptability, comfort and emotional experience, the hierarchical analysis method is used to construct a comprehensive evaluation index system for knitted fabric design, obtain the importance weight of each indicator, and use the weighted average method to calculate the comprehensive score of each indicator, and establish the mapping relationship between knitted fabric design parameters and comprehensive evaluation scores. According to the value range of knitted fabric design parameters, the triangular membership function is used to fuzzy process each parameter, and each parameter value is converted into a membership value. The fuzzy matrix of knitted fabric design parameters is constructed, and the fuzzy comprehensive evaluation of the fuzzy matrix is ​​performed to obtain the comprehensive evaluation membership of the knitted fabric design scheme. According to the nonlinear relationship between knitted fabric design parameters and evaluation indexes, a comprehensive evaluation model of knitted fabric design parameters was constructed using BP neural network. The membership value obtained by fuzzy comprehensive evaluation was used as the input of BP neural network, and the comprehensive evaluation score of knitted fabric was used as the output. Taking the design parameters of knitted fabrics as optimization variables and the output values ​​of the comprehensive evaluation model as the fitness function, a genetic algorithm is used to optimize the design parameters of knitted fabrics and search for the optimal parameter combination to maximize the comprehensive evaluation score of the knitted fabrics. This includes balancing the conflicts and constraints among multiple evaluation indicators such as fit, motion adaptability, comfort and emotional experience during the iterative process of the genetic algorithm to obtain the Pareto optimal solution set of the design parameters of knitted fabrics.

5. The digital-aided design method integrating artificial intelligence and knitting process knowledge according to claim 1, characterized in that: The knitted fabric and the design parameters are optimized for design consistency, the optimized wool knitting design parameters are input into the digital knitting equipment for manufacturing, and the manufactured wool knitting is quality tested to determine the consistency of the knitted fabric and the design parameters. If the consistency does not meet the preset requirements, the comprehensive evaluation model is returned for re-optimization, including: According to the optimal design parameter combination of wool knitting obtained by genetic algorithm optimization, computer-aided design software is used to model the microstructure parameters of the knitted fabric, needle type and stitch length, and coil height. The feasibility and optimization effect of the optimal design parameter combination of wool knitting are verified through finite element simulation analysis, and the knitting design model and process parameter planning scheme are obtained. Import the knitting design parameters into the control system of the CNC knitting machine and set the process parameters of needle cylinder speed, draft ratio and thread tension; A 3D laser scanner is used to perform an all-round scan on the manufactured wool knitted fabric, and the 3D point cloud data of the fabric surface is obtained. The pre-processing operations of noise filtering and smoothing compression are performed to obtain a 3D digital model of the knitted fabric. The three-dimensional digital model of the knitted fabric is compared with the digital design model, and the degree of consistency between the knitted finished product and the design parameters is determined by calculating the geometric parameters of the shape deviation and the size error between the two models, as well as the physical parameters of the thickness, density and elastic modulus of the fabric; If the consistency error between the knitted product and the design parameters exceeds the preset threshold, the 3D scanning data and physical property test data of the knitted fabric will be fed back to the 3D digital model of the knitted fabric for re-optimization, and the optimized digital design parameters will be generated and re-imported into the CNC knitting machine for production.

6. The digital-aided design method integrating artificial intelligence and knitting process knowledge according to claim 1, characterized in that: The method involves conducting a simulation test on human body wearing wool knitwear, establishing a fuzzy comprehensive evaluation model of knitted fabric performance based on the user's objective evaluation data and surface fit to evaluate the comfort and freedom of movement of the knitted fabric, optimizing the design parameters based on the evaluation results, and re-inputting the optimized design parameters into the digital knitting equipment for manufacturing, and improving the performance and user experience of the knitted fabric through iterative design until the expected design goals are achieved, including: Acquire three-dimensional curved surface data of a human body and three-dimensional curved surface data of the knitted fabric when the user wears the wool knitted fabric, and obtain a fitting evaluation index of the user curved surface and the knitted curved surface; Record the change data of joint angles when users wear knitwear and perform different actions, and determine the knitwear's freedom of movement evaluation index based on the range of joint motion; Taking the fit evaluation index, thermal insulation performance evaluation index and freedom of movement evaluation index as input, a fuzzy comprehensive evaluation method is used to establish a fuzzy comprehensive evaluation model of knitted fabric design parameters, and the comprehensive performance of knitted fabrics is evaluated. If the comprehensive evaluation results do not meet the expected design goals, the evaluation results are correlated with the design parameters, and the structural parameters and process parameters of the knitted fabric are adjusted and optimized through the optimization algorithm to obtain the optimized design parameters; The optimized design parameters are input into the control program of the digital knitting equipment, and the digital knitting is performed by controlling the needle selection, needle movement and needle removal actions of the knitting equipment; Conduct physical property tests on the optimized knitted fabric samples to obtain objective evaluation data on the thickness, quality, elasticity and strength of the fabric to determine whether it meets the design requirements; If the physical property test results of the optimized knitted fabric meet the design requirements, the design parameters of the knitted fabric are applied; If the test results do not meet the requirements, adjust and optimize the design parameters until they meet the requirements.

Images