CM/Collaborative Research: Simulation-based Software Tools for Automated Knitting

越来越多的纺织产品，特别是服装，是在复杂的自动化针织机上生产的，这种机器可以在一步内生产出完整的产品，并完全控制织物表面的形状、外观和机械性能的变化。这些机器承诺将使服装和其他针织产品的设计大众化：任何可以写成编织机程序的设计都可以像大规模生产产品一样轻松地进行任何数量的生产，并且每个实例都可以不同，在尺寸、合身和功能方面都有不同的变化。具有讽刺意味的是，很少有人能真正在这些机器上设计出用于自动化生产的定制针织服装，因为开发新图案是一个反复试验的过程，需要深奥的知识和对生产设备的了解。该项目旨在通过开发模拟技术来改变这种情况，使基于网络的交互设计工具能够准确预测最终结果，消除反复试验，让用户在第一时间直接制造他们的设计，对好的结果充满信心。对基于网络的交互式设计工具的重视将为扩大获取和实际使用提供一条切实可行的途径。该项目的技术目标包括：(1)通过扩展纱线级布料建模的先前工作，实现纱线力学的预测模拟，以提供与实验数据的校准匹配。(2)开发多尺度仿真模型，为交互式设计任务提供实时反馈，包括与设计相关的编织结构编辑。(3)构建基于模拟的设计工具，任何人都可以使用它来轻松地设计和模拟自己的定制针织产品，就像现在可以3D打印机械部件一样。(4)研究基于高级优化的设计工具，这些工具可以在不求助于低级针迹规范的情况下实现“智能”编辑。我们将使用一系列用户研究来评估我们的模拟工具，以衡量工具的有效性，以提高用户匹配设计目标的能力。

A growing fraction of textile products, particularly clothing, are made on sophisticated automated knitting machines that can produce complete products in a single step, with full control over variation in shape, appearance, and mechanical properties across the surface of the fabric. These machines promise to democratize the design of garments and other knitted products: any design that can be written down as a knitting machine program can be made in any quantity as easily as a mass-produced product, and every instance can be different, with variations in size, fit, and function. Ironically, very few people can actually design custom knitted garments for automated manufacture on these machines, because developing new patterns is a trial-and-error process requiring esoteric knowledge and access to the production equipment. This project aims to change this situation by developing simulation technology that enables web-based interactive design tools that accurately predict the end result, eliminating trial-and-error and letting users directly manufacture their designs with confidence of good results the first time. An emphasis on interactive web-based design tools will enable a practical pathway for expanding access and real-world usage. The project's technical aims include: (1) Achieve predictive simulation of yarn mechanics by extending prior work on yarn-level cloth modeling to provide a calibrated match to experimental data. (2) Develop multi-scale simulation models that provide real-time feedback for interactive design tasks involving design-related edits to knit structures. (3) Build simulation-based design tools that anyone can use to easily design and simulate their own custom knitted products in the same way one can now 3D print mechanical parts. (4) Investigate high-level optimization-based design tools that enable "smart" edits without resorting to low-level stitch specifications. We will evaluate our simulation tools using a sequence of user studies that measure the effectiveness of the tools to improve users' ability to match design targets.