Chung-Ang University Develops Physical AI-Based Kinetic Facade Technology

Implementation of Core Technology Paving the Way for Autonomous Building Control

Chung-Ang University announced on April 29 that Professor Youngho Chae's research team at the Graduate School of Advanced Imaging's Virtual Environment Laboratory (VELAB) has developed a "large-scale kinetic facade predictive cluster control framework," a next-generation architectural exterior technology, and successfully implemented it as a full-scale mock-up.

(From left) Researcher Shin Su-cheol, Professor Lee Won-hyung, Professor Chae Young-ho. Chung-Ang University

The key aspect of this research is the integration of a "physical AI brain" into buildings, enabling them to make autonomous decisions and movements.

The research team combined graph neural networks (GNN) and reinforcement learning with kinetic facade technology, which allows exterior materials to autonomously move and regulate energy in response to weather changes. By establishing "AI swarm intelligence" that allows thousands of exterior modules to communicate in real time, the team overcame the bottleneck caused by a surge in computational volume and opened up the possibility of applying this technology to large-scale buildings.

Furthermore, when the trained AI policy was directly deployed on a 1:30 scale hardware prototype for verification experiments, the results showed a 10.3% reduction in solar gain for the seating area and a 25.4% decrease in unnecessary motor operation.

In particular, by allowing the AI to autonomously balance conflicting objectives—minimizing radiant heat for the seating area and maximizing radiant heat for the stadium lawn—the team simultaneously achieved mechanical stability, addressing a chronic issue of dynamic exterior materials.

Selected as a "SW Star Lab" by the Ministry of Science and ICT, Chung-Ang University's VELAB plans to expand this technology into high-value-added industries, such as future-oriented smart cities and the design of ultra-large landmarks.

Professor Chae commented, "This achievement signals the dawn of a true 'Autonomous Architectural Organism' era, where AI independently learns and resolves complex real-world dilemmas, becoming perfectly integrated with physical space."

Meanwhile, the results of this research, with Researcher Shinchul Shin as the first author, have been accepted for publication in the AI special issue of the international journal 'JCDE (Journal of Computational Design and Engineering)'.

This content was produced with the assistance of AI translation services.