Skip navigation
Please use this identifier to cite or link to this item: http://arks.princeton.edu/ark:/88435/dsp01vt150n137
Title: ROBOTIC ASSEMBLY: A GENERATIVE ARCHITECTURAL DESIGN STRATEGY THROUGH COMPONENT ARRANGEMENTS IN HIGHLY-CONSTRAINED DESIGN SPACES
Authors: Wu, Kaicong
Advisors: Kilian, Axel
Meggers, Forrest
Contributors: Architecture Department
Keywords: Architectural Design
Component Arrangements
Computer Vision
Deep Learning
Generative Design
Robotic Assembly
Subjects: Architecture
Robotics
Artificial intelligence
Issue Date: 2019
Publisher: Princeton, NJ : Princeton University
Abstract: Architectural assembly has been a neglected research topic within the field of computational design. Instead, manual assembly and passive fabrication processes directed by top-down controlled geometric models have attracted more attentions. Such design-to-production model is problematic because it is difficult for designers to overcome the constraints of their empirical knowledge. Moreover, an extensive amount of resources can be wasted when architectural components are manually assembled. However, what has yet to be determined is whether the applications of advanced assembly machines, especially architectural robots, can reduce resource use and create new design principles. How can robotic assembly become a generative strategy to design architectural forms through component arrangements? Three design models were developed to study the sequence, fitting, and configuration of robotic assembly. In relation to the model of Robotic Equilibrium Assembly, scaffold-free constructions were examined through tooling innovations to recreate the design of compression-only arch structures. For the model of Material Outline Assembly, a scanning procedure was carried out to fit foam fragments into a shell by flexibly approximating a human-designed surface geometry. For the model of Stochastic Assembly, deep learning was applied to autonomously achieve higher assembly goals of natural wood log structures. These models were implemented using various computational and robotic approaches and tested in small-scale experiments. This research contributes to architectural design by redefining the role of assembly machines as generators that can free design space from the constraints of existing knowledge regarding geometries, fabrications, and structures. The experiment results indicate that, even with unfamiliar design problems, potential solutions can be identified using robotic assembly to arrange architectural components. The design control shifts from top-down, human-centered geometric modeling to bottom-up, machine-centered component assembly. This can stimulate human designers to recognize and overcome their cognitive barriers, challenge existing architectural design criteria, and discover unknown design principles. Future work can be done to develop autonomous assembly of raw materials, connect the learning processes of virtual and physical assembly machines, and apply large-scale robotic assembly in built environments.
URI: http://arks.princeton.edu/ark:/88435/dsp01vt150n137
Alternate format: The Mudd Manuscript Library retains one bound copy of each dissertation. Search for these copies in the library's main catalog: catalog.princeton.edu
Type of Material: Academic dissertations (Ph.D.)
Language: en
Appears in Collections:Architecture

Files in This Item:
This content is embargoed until 2021-06-10. For more information contact the Mudd Manuscript Library.


Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.