Department(s):

Marketing Communications
Preview

Shichao Liu

Shichao Liu

Augmented reality (AR) benefits architectural designers by enabling 3D design presentations while maintaining physical interaction with collaborators and stakeholders. Yet AR has not become an integral part of design in industry and academia because current AR tools do not support rapid prototyping and idea exploration.

Generative artificial intelligence (AI) could change that. With a $399,978 grant from the National Science Foundation (NSF), Architectural Engineering Assistant Professor Shichao Liu is leading a three-year project to develop and test AI-empowered models within AR for effectiveness in architectural design studios at WPI and Drexel University.

“Architectural engineers require extensive collaboration among architects, consultants, contractors, and other stakeholders, said Liu. “AI-empowered AR has the potential to be a powerful tool for visual ideation and representation during collaborative conceptual design and prototyping.”

As principal investigator (PI), Liu will be joined on the project by two WPI co-PIs: Gillian Smith, associate professor in the Department of Computer Science and director of the WPI Interactive Media and Game Development program, and Soroush Farzin, assistant teaching professor in the Department of Civil, Environmental, and Architectural Engineering. Kimberly LeChasseur,  senior research & evaluation associate in the Morgan Teaching & Learning Center, will evaluate the effectiveness of the developed tools.

Currently, AR is primarily used in demonstration settings, often serving as the final step to showcase architectural designs to clients, Liu said. A key barrier to AR is that AR tools lack the capability to facilitate rapid prototyping and idea exploration. Unlike computers or sketches, which allow for quick iterations, AR systems are complex to operate and often slower in graphic response. While AR provides an immersive display of models, the creation process leading up to this stage remains challenging.

Preview

An AI-generated image shows people collaborating with an AR design.

AI-generated image

The advent of generative AI, particularly through diffusion models, could revolutionize this process by enabling content creation in AR through verbal prompts, Liu said. This would mean that users, whether students or professional designers, no longer need to create models on computers and transfer them to AR for brainstorming or collaborative design. Instead, prototyping and conceptual design could take place directly within the AR environment, simplifying the creative process and enhancing collaborative potential.

Liu said that if the WPI researchers are successful in developing AI-empowered AR, the technology could be tested in architectural firms, which could lead to innovative tools for industry professionals that enhance ideation, collaboration, and impact. 

Keep Up With WPI Research News

Want to learn more about Research at WPI? Subscribe to the monthly newsletter.