Carnegie Mellon University
Browse

Beyond Automation: Supporting Human-Computer Collaboration in Designing with Active Materials

Download (17.19 MB)
thesis
posted on 2024-10-23, 20:41 authored by Humphrey YangHumphrey Yang

 Active materials design arises as a new concept in physical design and fabrication. These materials  have dynamic properties that can be activated by external stimuli, enriching physical devices'  functional versatility and interactivity upon integration. Their application spans various domains,  from engineering to design and art. In human-computer interaction, active materials allow  computational affordances to seamlessly blend into everyday life, augmenting physical interfaces'  interactions and versatility. However, as we begin to leverage active materials in physical design  practices, the need for a design infrastructure also surfaces. While plentiful digital fabrication tools  help us materialize ideas, designing active material systems is inherently challenging due to their  novelty and dynamism. Their exotic behaviors are unfamiliar to common designers; their  dynamism also requires spatiotemporal reasoning. The challenge is further exacerbated when  applying active materials in real-world design problems, where the designer must simultaneously  navigate constraints, opportunities, and objectives that arise from the design context. This thesis  acknowledges these challenges and asks: How can computational tools support designers to work  with active materials when addressing real-world, contextualized design problems?  

This work synthesizes a set of computational toolmaking motifs for active materials design. The  motifs administer a computational design tool's interaction with the user, empowering their ability  to manipulate and reason about active materials. This dissertation advocates a shift from  developing "tools that solve design problems" to "tools that help designers find solutions." These  tools collaborate closely with the designer instead of automating the design process. By allowing  both the computational design agent and the human designer to co-steer the course of design, both  parties may do what they do best in finding satisficing solutions for complex design tasks. In this  thesis, each chapter contextualizes the toolmaking motifs in different active material systems to  develop proof-of-concept computational tools. The work presented here discusses the toolmaking  techniques and user interactions that respond to different classes of active material design problems,  as well as their implications in supporting the user's design thinking and workflow. Each project  then uses the tools to develop demonstrative artifacts to validate their usefulness in helping  designers address contextualized design problems. The artifacts also highlight the novel design  opportunities enabled by the emergent media. 

Funding

Supporting Designers in Learning to Co-create with AI for Complex Computational Design Tasks

Directorate for Computer & Information Science & Engineering

Find out more...

History

Date

2024-09-01

Degree Type

  • Dissertation

Department

  • Human-Computer Interaction Institute

Degree Name

  • Doctor of Philosophy (PhD)

Advisor(s)

Lining Yao Nikolas Martelaro

Usage metrics

    Licence

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC