Spatial and Temporal aspects for pattern representation and discovery in Intelligent Environments

Core to the development of an intelligent environment is the understanding of where, when, and for how long, actions take place. Our work addresses the process of discovering patterns of behaviour which may reflect preferences of a user on the use of devices. A solution to this problem is essential for the automatization of environments like Smart Homes so that user preferences can be learnt automatically by the system. We present a language that allows us to represent in a clear and non ambiguous way useful patterns that can occur in intelligent environments like Smart Homes. The language is coupled with an algorithm to learn patterns, related to the proposed language, from data collected by sensors.