Chung-Ang University VELAB Team Successfully Demonstrates 'Physical AI Architecture' That Judges and Moves on Its Own

In a significant leap forward for architectural technology, researchers at Chung-Ang University's VELAB (Virtual and Augmented Reality Laboratory) have successfully demonstrated "Physical AI Architecture." This groundbreaking technology moves beyond the current 3D Building Information Modeling (BIM) era by enabling buildings to possess an "AI brain" that perceives its environment and controls the structure in real-time, much like a living organism.

The research team, led by Professor Chae Young-ho, developed a "Kinetic Facade Predictive Cluster Control Framework," a next-generation building exterior technology. The key innovation lies in integrating Graph Neural Networks (GNN) and Reinforcement Learning (RL) with kinetic facades – exterior elements that move and adapt. This integration creates an "AI cluster intelligence" where thousands of exterior modules communicate and coordinate in real-time. This approach effectively solves the computational bottlenecks that previously hindered the application of such technologies to large-scale buildings.

To validate their framework, the team deployed the learned AI policies onto a 1:30 scale hardware prototype. The results were impressive: the system reduced solar heat gain in spectator areas by 10.3% while simultaneously decreasing unnecessary motor operations by 25.4%. Crucially, the AI was able to autonomously balance conflicting objectives – minimizing heat radiation to spectator areas while maximizing it for the stadium's turf – a feat that has long been a challenge for dynamic facade systems, thereby ensuring mechanical stability.

This pioneering research has earned international recognition, with its publication confirmed in a special AI issue of the 'JCDE (Journal of Computational Design and Engineering),' a globally top-tier journal. The VELAB, designated as a 'SW Star Lab' by the Ministry of Science and ICT, plans to expand this technology into high-value sectors such as future smart cities and the design of mega-scale landmark buildings. Professor Chae envisions this as the dawn of an era of "Autonomous Architectural Organisms," where AI seamlessly integrates with physical spaces to create truly responsive and intelligent buildings.

“

This achievement is a signal flare announcing the arrival of the era of true 'Autonomous Architectural Organisms,' where AI autonomously learns and solves complex real-world dilemmas, achieving perfect integration with physical space.