Intelligent user profiling to provide assisted living facilities

Abstract

Ambient Intelligence is a new area of artificial intelligence and ubiquitous computing. In the home environment it is known as Smart Home, whereby the environment acts behind the scenes to provide services for the inhabitants. A Smart Home is any house designed to accommodate the needs of the occupants by means of automation. The main goal of this project is to simulate an adaptable smart house, achieved by implementing Intelligent User Profiling. User profiling is all about customizing and adapting systems to the user's specific needs, by means of a user profile. The simulation tool developed shows the physical boundaries of a house, together with a set of "sensors" which observe any changes in human behaviour or the environment. Basic appliances were included for the user to interact with. The program logs the events that occur and then, using Sequence Mining and Machine Learning techniques, a user profile is created. This profile is a collection of logical rules that describe the behavioural patterns of a particular user. The reasoning mechanism of the smart home is implemented by means of a Rule-Based Expert System. This Expert System imitates the decision-making abilities of a human being and provides the necessary adaptations and automations to facilitate the life of the house's inhabitants. From the evaluation performed it was found that the system can detect behavioural patterns efficiently and that together with the Expert System, the environment adapted to the user and human-interventions were reduced.