CM/Collaborative Research: A Computational Approach to Customizing Design

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

Creating physical things,including shapes and structures, cyber-mechanical devices, and other functional artifacts and machines, is a laborious and often error-prone process, typically limited to domain experts and engineers. Often the manufacturing expert is distinct from the device designer, leading to a knowledge and ability gap that makes it difficult if not impossible for end users to create custom products. Nonetheless, there is a need for customization and personalization in product design across many application spaces, including, for example, healthcare, sports gear, furniture, and vehicles. This project will investigate algorithms, methods, and tools to allow casual end-users to leverage the knowledge of engineering experts to easily and effectively create validated and verifiable manufacturable designs for custom objects. The project will engage students in K-12 schools and community colleges, women, and minorities through the software and design artifacts, thus promoting and inspiring STEM education. This project will develop a complete computational pipeline to enable designing and manufacturing parameterized objects that can be customized and instantiated by non-experts. More specifically, given a parametric Computer Aided Design (CAD) model, the investigators will develop translation methods that convert it to a customizable digital model with a small number of intuitive parameters. The framework will also provide mechanisms to compose these digital models into new ones. Given desired parameter values, each model will be translated to a manufacturing process and its digital instructions. To this end, a number of fundamental questions will be addressed including: (1) general representations for customizable, manufacturable digital models, (2) translation processes for converting standard parametric CAD models to these new representations, (3) methods that enable composition of these digital models into new ones, and (4) computational backends that translate the new digital models into digital instructions that drive specific manufacturing machines.