AI-Driven Cooling Collaboration Awarded Japan–Singapore Joint Grant

Our project AI-Driven Climate Resilient Cooling: Robust Reinforcement Learning for Mixed-Mode Ventilation, co-led by Adrian Chong (IDEASLab, NUS) and Dr. Shohei Miyata (The University of Tokyo), has been awarded funding under the Japan Science and Technology Agency (JST) – Agency for Science, Technology and Research (A*STAR) Joint Call on Artificial Intelligence.

This three-year research effort will bring together researchers from Singapore and Japan to develop AI-based control strategies for mixed-mode ventilation (MMV) that cool only when necessary, prioritizing natural ventilation, ceiling fans, and other energy-efficient alternatives, with air-conditioning used only as needed. Achieving optimal MMV control is challenging due to constantly changing outdoor conditions (e.g., wind, temperature, rain), variable occupancy patterns, and the complex interaction between multiple subsystems. The project will integrate physics-informed digital twins, robust reinforcement learning controls, and explainable AI to address these challenges, enabling MMV systems that are robust, reliable, and build user trust in automated building operations.

The Singapore team will work closely with industry and regional partners to translate research outcomes into practical mixed-mode ventilation solutions that can be deployed in diverse building contexts. By bridging advanced AI control development with real-world application, the project aims to accelerate the adoption of sustainable cooling strategies across Southeast Asia.

Read more about the JST–A*STAR Joint Call on AI here