Architecture doctoral candidate works to make robots more intelligent, adaptive

UNIVERSITY PARK, Pa. — Helping robots become more aware of their physical environment can allow designers to enable new materialization techniques that have not been possible up to this point, such as 3D printing on an uneven surface, according to Özgüç Bertuğ Çapunaman, a doctoral candidate in architecture with a concentration in computational design at the Penn State College of Arts and Architecture’s Stuckeman School. Çapunaman successfully defended his dissertation on June 4 and will graduate in August with his doctorate. 

“My research looks into the idea of using robots in architectural digital fabrication and beyond, and trying to understand how we can make these systems more intelligent using sensing technologies so that they can adapt to changes in their environment,” Çapunaman said.

Çapunaman, from Istanbul, Turkey, works with the Form and Matter (ForMat) Lab in the Stuckeman Center for Design Computing and collaborates with the Additive Construction Lab at Penn State to research how robots can become more aware of their surroundings and adapt accordingly to their environments.

“Robots do not have any inherent understanding of what’s going on around them. As far as these systems are concerned, they are moving through a series of points in space,” he said. “My idea is that we can start introducing novel computer vision and graphics methodologies to the domain of fabrication so that the robots can perceive their environment and intelligently adapt their operation, enabling novel materialization techniques that have not been possible in the past, like 3D printing on an uneven surface.”

In 2023, Çapunaman spent three months in San Francisco at Autodesk Research, which provides software products and services for the architecture, engineering, construction, manufacturing, media, education and entertainment industries, to explore his line of research with the company’s broader vision of autonomous robotics. He introduced his 3D-scanning work from Penn State and helped build a spatial robotic perception pipeline for robotic assembly tasks.

“I want to help architects and designers understand that [robots] are our tools as long as we claim them to be ours and as long as we put in the work to understand why they do what they do,” Çapunaman said. “Robotic fabrication in an architectural context often involves a plethora of unique challenges, from novel material systems to complex geometries. I’m interested in making it possible for robots to have more capacity to understand their world so that they can make things safer and faster for us and [do] more complex [tasks].”

Çapunaman said that he envisions a world where robots would collaborate with humans in the design and fabrication space.

“Especially with the recent surge in artificial intelligence research, I believe that we are at the cusp of a major paradigm shift toward more intelligent and autonomous fabrication modalities. Beyond autonomous digital fabrication, my work can open novel fabrication scenarios in creative processes, mimicking how humans engage with open-ended problems through a series of seeing, acting and seeing,” Çapunaman said. “My ultimate goal would be to make robots not necessarily executors of things but more like collaborators so that you work with them rather than making them work for you. The goal is to be co-creative [with robots] so that we both can bring design ideas to the table by taking turns in manipulating our physical environment.”

Çapunaman’s research publications detail the limitations of robotic arms and introduce “a vision-based sensing framework to digitally reconstruct and register work environments, enabling tooling on indefinite surfaces,” exploring robotic perception across different scales and material systems.

After his graduation this summer, he plans to return to Turkey and continue his research.

“Being a creative reasoner, Özgüç not only has high-level design computing skills but also a profound interest in theory, which greatly elevates the quality of our discussions and of his doctoral research,” said Benay Gürsoy, assistant professor of architecture, ForMat Lab director and Çapunaman’s thesis adviser.

The field of robotics and the surge of interest in artificial intelligence is a rapidly evolving topic, Çapunaman said, noting every study or project takes his research one step closer to a deeper understanding of robots, even if they do not yet know how their work will ultimately be applied.

“This line of research is still in its infancy and there is no single application [for our work], but that is how we learn and start figuring things out,” Çapunaman said.