Location prediction approach for context-aware energy management system

Given that over the past few decades energy demand has been continuously increasing and the context-aware energy management systems for renewable sources and electrical loads have become much more sophisticated, the paper presents the application of location prediction methods in such systems. CAEMSs are highly distributed, can manage large amounts of energy-related data and have to be able to react rapidly and intelligently when conditions change, leading to real-time challenges. We address one of the most valuable context prediction tasks — learning human habits and behavioral patterns for indoor location prediction. The motivation behind choosing this problem is that knowledge of the location, presumably entered by the user, can initiate several preparation routines to maximize comfort and minimize energy consumption. Automatic adjustment of light or temperature, by heating rooms prior to their occupation, can be stated as an example of such a routine.