Enhancing Building Energy Efficiency with a Modular AI-based Smart Lighting System

This study presents a modular smart lighting system that employs an open-loop, daylight-based control strategy supported by artificial intelligence (AI). Unlike conventional closed-loop systems that rely on multiple indoor sensors and complex wiring, the proposed system adopts a distributed architecture in which a Master LED Luminaire (MLL) coordinates Slave LED Luminaires (SLLs). The MLL incorporates a Support Vector Regression (SVR)-based daylight prediction model that estimates indoor daylight availability using fixed architectural parameters such as room dimensions, window geometry, and surface reflectance together with real-time outdoor irradiance data. The system was implemented and validated in a public building under real operating conditions. Comparative evaluation with traditional on–off control and sensor-based closed-loop control indicated that, while the closed-loop approach achieved the lowest artificial lighting demand (56.4%), the proposed system required only slightly more (66.4%) yet offered clear advantages in scalability, ease of installation, and reduced infrastructure costs by eliminating the need for multiple sensors. Integration with IoT-enabled communication further allows real-time parameter updates and adaptive dimming control. The findings demonstrate that the proposed method provides a scalable and cost-effective retrofit solution, addressing the installation, maintenance, and adaptability challenges of conventional lighting control systems.