CHS: Medium: Collaborative Research: Computer-Aided Design and Fabrication for General-Purpose Knit Manufacturing

点击翻译按钮获取中文摘要

General-purpose fabrication of complex 3D shapes is possible today with industrial knitting machines, which provide a powerful alternative to conventional 3D printing, given their ability to use a wide range of feedstock to produce strong and durable artifacts that can withstand extreme deformations and stretch. Machine knitting is already used in applications ranging from composite preforms for aircraft parts, soft robotics, and medicine to functional clothing like body-armor, cut-proof gloves, and everyday apparel. However, controlling industrial knitting machines is a difficult task, as it involves determining an efficient sequence of low-level knitting machine instructions that produce a desired 3D form. The goal of this project is to develop a complete framework, from computer-aided design to fabrication, for general-purpose knit manufacturing that will allow users to start with an arbitrary 3D shape, convert the shape to a virtual knit structure, customize the knit structure, and simulate the behavior of the result. Project outcomes will enable widespread use of customized fabrication via machine knitting, with the potential to alter manufacturing practices and enable new products and applications in numerous industries, including robotics, furniture, automotive, space exploration, medicine, defense, and clothing. A key contribution of this project will be simulating at the meso-scale, using a large database of common patterns to preview behavior at the many-stitch level. Additional broad impacts will derive from seeding interdisciplinary cross-department and cross-college collaborations in both institutions.The key steps in the new CAD/CAM workflow, which involve converting arbitrary shapes to knit structures, will build on prior work by members of the team. The project will involve: customization of the knit structure using both high- and low-level operations; live simulation of the knit structure by exploiting a novel data-driven patch-based approach that draws on calibration data captured from photographs; and low-level instruction generation and verification to include a hint-based scheduler that allows (optional) user control of the scheduling process. Major contributions of this work will be the measurement-based simulation and the optimization steps of the pipeline, as well as the novel image-based yarn measurement technology to support them. The integration of a hinting system, and the use of a topological verifier, are innovations that will impart to the work the flexibility and guarantee of correctness required to enable the proposed pipeline.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.