Investigation Of Robotic Sewing Method Based On Process Modeling

Sewing is an important production process in the clothing manufacturing industry.Despite the rapid development of automation in sewing equipment such as sewing machines,sewing is still a labor-intensive production method.In recent years,industrial robots have developed rapidly,and the trend of replacing humans with machines in the manufacturing industry is becoming increasingly evident.Robots replacing manual labor and cooperating with sewing equipment will greatly improve the efficiency and quality of clothing production,reduce labor demand,and have important significance for the development of automation in the clothing manufacturing industry.However,for fabric sewing,the application of robots still faces many challenges.Firstly,the fabric is prone to deformation and wrinkling,causing deviations in the stitching and affecting the sewing quality.Secondly,mechanisms such as feeding teeth can cause dynamic disturbances to the sewing process,making it difficult for robots and sewing machines to collaborate.In addition,the variety of clothing product models makes it difficult to plan the robot sewing trajectory.Therefore,this article focuses on the challenges of robot sewing operations and studies a robot sewing method based on process modeling.Predicting the tension of the fabric sewing process based on ductility,classifying and identifying the sewing state,and implementing linear sewing process using impedance control.Using the grating scanning method to detect fabric edges,identify fabric curved edge traces,establish a robot geometric sewing model,and achieve the curved edge sewing process.Finally,a physics experimental platform was established and the method was validated by sewing different types of pockets.This article focuses on the two key issues of straight and curve sewing in the robot sewing process,and conducts the following research content:(1)Aiming at the problem of complex fabric performance parameters and difficulty in modeling the expected tension during the sewing process,a stretchability based robot fabric sewing tension prediction method is proposed.Using expert qualitative language values to represent fabric ductility,using SVM models to predict the ductility of unknown fabrics,using fuzzy logic to determine the nonlinear relationship between fabric characteristics and expected tension,and verifying the sewing effect of expected tension,laying the foundation for robot linear sewing based on expected tension.(2)A method based on wrinkle state recognition is proposed to address the issue of fabric wrinkles during robot linear sewing.Firstly,the sewing state is divided into three categories:collaborative state,pre needle wrinkle state,and post needle wrinkle state.SVM is used to achieve recognition of different sewing states.To address the issues of fabric wrinkles and uneven stitching caused by the lack of collaboration between robots and sewing machines,a robot based impedance control method is adopted,which adjusts the initial trajectory through feedback force signals to adapt to the speed changes of the sewing machine.The experimental results show that the robot can adjust the sewing state in a timely manner and maintain the desired tension,achieving a linear sewing process.(3)In the process of curved edge sewing,a robot curve sewing method based on geometric modeling is proposed to address the problem of difficulty in planning robot trajectories based on fabric shape.After image preprocessing,the method of image etching and raster scanning is used to realize the contour detection of cloth.The line trace is extracted by polynomial fitting the contour curve,and the geometric modeling of robot sewing curved cloth is established to realize the trajectory planning of the robot.The experimental results show that this method can realize the sewing of curved cloth with different curvature.(4)Construction and experimental verification of robot sewing platform.A sewing platform based on UR5e robotic arm was built.On the basis of determining the expected tension of the fabric,linear and curve sewing process methods were combined to conduct sewing experiments on different shapes of linen polyester pockets.Multiple straight edges and complex curved edges were successfully sewn.Verified the effectiveness of the robot sewing method based on process modeling.Finally,summarize the work and achievements of this article.The robot sewing method based on process modeling studied in this article has important research significance for achieving intelligent production in the clothing industry.