Towards Co-creation: Exploring Modalities of Vision-augmented Robotic Fabrication for Automation and Beyond
Restricted (Penn State Only)
- Author:
- Capunaman, Ozguc
- Graduate Program:
- Architecture
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 04, 2024
- Committee Members:
- Frank Jacobus, Program Head/Chair
Jose Pinto Duarte, Major Field Member
Benay Gursoy Toykoc, Chair & Dissertation Advisor
Chris McComb, Special Member
Loukas Kalisperis, Major Field Member
Onur Yuce Gun, Special Member
Katie Fitzsimons, Outside Unit & Field Member - Keywords:
- Computer Vision
Robotic Fabrication
Additive Manufacturing
Digital Twin
3D Scanning
Conformal 3D Printing - Abstract:
- The evolution of architectural design practice and research has been profoundly influenced by the advent and integration of digital technologies. However, despite the technological advances of computer-aided design (CAD) and manufacturing (CAM) that have reshaped architectural practices since the 1960s, the adoption of these digital tools within the context of unstructured physical environments and unpredictable material behaviors remains a pressing challenge. Traditional robotic systems, designed for predictable and controlled environments and predominantly governed by rigid and unidirectional information flow, consistently fall short of addressing the dynamic complexities of materialization. This often results in significant discrepancies between computational design intentions and their material realizations, exacerbated by environmental variability and material uncertainty. As architectural forms become increasingly complex through computational design workflows and as the field’s interest in novel material systems and sustainable construction practices surges, the need for systems that can effectively bridge the gap between digital precision and physical uncertainty becomes critical. In response to these challenges, this dissertation introduces the Vision-augmented Robotic Fabrication (V-aRF) framework to enhance the adaptability and responsiveness of robotic systems in architectural fabrication. The V-aRF framework augments traditional robotic fabrication workflows with advanced sensing capabilities, enabling them to adapt dynamically to varying material conditions and environmental contexts. Through a systematic exploration of three primary operational modalities enabled by V-aRF – namely (1) informing robotic systems of the state of the environment, (2) adapting fabrication operations, and (3) assessing materialization quality – this research demonstrates how enhanced robotic feedback loops can lead to increased materialization accuracy and enable novel manipulation capabilities to address evolving challenges of architectural design. Within the scope of this dissertation, these explorations were conducted across various scales and material systems, including additive manufacturing of clay, concrete, and mycelium-based composites, as well as hot knife carving, demonstrating the versatility and effectiveness of the V-aRF framework in diverse fabrication scenarios. This shift towards a more integrated, responsive, and adaptive approach not only promises to enhance the efficiency and effectiveness of robotic fabrication but also sets the stage for a future where architectural robotics transcends traditional automation. Envisioning a scenario beyond automation, this dissertation argues for a co-creative role of robotic systems, where they are not merely passive executors of predetermined tasks but active participants in ideation and materialization processes. The V-aRF framework, as a foundational step towards this goal, establishes the technological and methodological frameworks to facilitate the integration of robotic systems in open-ended explorations and enable a more reciprocal interaction between the user, the material, and the robot. This shift promises to redefine the boundaries of architectural fabrication, both the creative potential and operational efficiency of robotic systems in complex materialization scenarios.