Personalized ambience: An integration of learning model and intelligent lighting control

The number of households and offices adopting automation system is on the rise. Although devices and actuators can be controlled through wireless transmission, they are mostly static with preset schedules, or at different times it requires human intervention. This paper presents a smart ambience system that analyzes the user's lighting habits, taking into account different environmental context variables and user needs in order to automatically learn about the user's preferences and automate the room ambience dynamically. Context information is obtained from Yahoo Weather and environmental data pertaining to the room is collected via Cubesensors to study the user's lighting habits. We employs a learning model known as the Reduced Error Prune Tree (REPTree) to analyze the users' preferences, and subsequently predicts the preferred lighting condition to be actuated in real time through Philips Hue. The system is able to ensure the user's comfort at all time by performing a closed feedback control loop which checks and maintains a suitable lighting ambience at optimal level.